Browse the corpus

Introduction Cardiovascular disease is a major cause of morbidity and mortality in patients with diabetes mellitus.1 Type 2 diabetes has long been recognized as a major risk factor for the development of coronary artery disease (CAD),2 but diabetes and insulin resistance are now also recognized as independent risk factors for the development of heart failure.3,4 In one observational study, 33% of men and 45% of women with diabetes developed heart failure over 5.5-year follow-up.5 The risk of developing heart failure was independent of age, gender, CAD, or hypertension. Once established, heart failure in patients with diabetes is associated with worse clinical outcomes, independent of CAD.6 Observational studies have demonstrated that patients with systolic heart failure and diabetes have a mortality that is almost doubled compared with their normoglycaemic counterparts.7,8 Several mechanisms have been proposed for the association between diabetes and heart failure including endothelial dysfunction, abnormal calcium handling, myocardial fibrosis, and inappropriate activation of the renin–angiotensin–aldosterone system.9,10 Observational research has shown that diabetes is associated with an increase in left-ventricular (LV) mass11 and impaired diastolic function.12 Over time, these structural and functional changes lead to impaired systolic function and the clinical syndrome of heart failure.13

e renin–angiotensin–aldosterone system.9,10 Observational research has shown that diabetes is associated with an increase in left-ventricular (LV) mass11 and impaired diastolic function.12 Over time, these structural and functional changes lead to impaired systolic function and the clinical syndrome of heart failure.13 Cardiovascular magnetic resonance (CMR) offers a unique opportunity to assess these changes in cardiac structure and function as well as changes in myocardial perfusion in a single examination. In this study, we undertook a comprehensive assessment of cardiac structure, function, and perfusion by CMR in patients with diabetes, prediabetes, and normal insulin sensitivity, in whom significant CAD was excluded by coronary angiography. We hypothesized that patients with diabetes have increased LV mass, abnormal strain patterns, and impaired myocardial perfusion and that patients with prediabetes would have similar, albeit less severe, changes.

, prediabetes, and normal insulin sensitivity, in whom significant CAD was excluded by coronary angiography. We hypothesized that patients with diabetes have increased LV mass, abnormal strain patterns, and impaired myocardial perfusion and that patients with prediabetes would have similar, albeit less severe, changes. Methods Selection of patients and recruitment We prospectively screened 399 consecutive patients with non-obstructive (no stenosis >30% luminal narrowing by visual analysis) CAD on routine coronary angiography typically clinically indicated for the investigation of chest pain at our tertiary cardiology centre. Patients with a history of previous myocardial infarction, coronary revascularization or other significant heart disease, contraindications to CMR or adenosine, and known claustrophobia were excluded. Of the remaining patients, 72 agreed to participate in the study. The study was approved by the local Ethics Committee. All patients gave fully informed written consent.

arction, coronary revascularization or other significant heart disease, contraindications to CMR or adenosine, and known claustrophobia were excluded. Of the remaining patients, 72 agreed to participate in the study. The study was approved by the local Ethics Committee. All patients gave fully informed written consent. Patient classification Smoking history, clinic non-invasive blood pressure, lipid profile, and drug history were recorded. In line with current American Diabetes Association (ADA) guidelines,14 diabetes mellitus was defined as fasting glucose ≥7 mmol/L, HbA1c ≥6.5%, or a past history of diabetes. Duration of diabetes was reported by the patients. Those that did not meet this definition were defined as non-diabetes. Prediabetes was defined as fasting glucose 5.6–6.0 mmol/L or HbA1c 5.7–6.4%. Those that did not meet the definition of diabetes or prediabetes were defined as controls. Hypertension was defined as clinic SBP >140 mmHg (the level at which treatment is recommended in current ADA guidelines). Hypercholesterolaemia was defined as current use of HMG-CoA reductase inhibitors (statins) or low-density lipoprotein (LDL) >100 mg/dL (the level at which statins are recommended in ADA guidelines). All subjects had their height, weight, hip circumference, and waist circumference measured. The homeostatic model assessment of insulin resistance (HOMA-IR) was calculated using the following formula: serum fasting glucose (µU/mL) × serum fasting insulin (mg/dL)/405.15 In diabetic patients, this was only calculated for those not taking exogenous insulin. Waist–hip ratio (WHR), body mass index (BMI), and Mosteller body surface area (BSA) were calculated.

esistance (HOMA-IR) was calculated using the following formula: serum fasting glucose (µU/mL) × serum fasting insulin (mg/dL)/405.15 In diabetic patients, this was only calculated for those not taking exogenous insulin. Waist–hip ratio (WHR), body mass index (BMI), and Mosteller body surface area (BSA) were calculated. Blood and urine sampling Urine was tested for albumin–creatinine ratio. Blood samples were obtained from participants after 8 h of fasting and tested for glucose, insulin, total cholesterol, LDL cholesterol, high-density lipoprotein cholesterol, serum urea and electrolytes, and glycosylated haemoglobin (HbA1c). CMR protocol CMR was performed on a 1.5 T whole-body magnetic resonance scanner (Intera, Philips Medical Systems, Best, The Netherlands) using vectorcardiographic gating and a 5-element cardiac phased-array receiver coil. Data were acquired during breath holding at end expiration. From scout CMR images, the LV long and short axes were determined. Then, tagged CMR images were acquired at the apex, mid-ventricle, and base [complementary spatial modulation of magnetization method (CSPAMM) using multishot echo planar imaging, flip angle sweep applied to the radiofrequency excitation pulses of subsequent cardiac phases, two orthogonal line tags acquired per slice in a 14 s breath hold, typical FOV 300 mm, matrix 128 × 128, slice thickness 10 mm, tag separation 8 mm, 18 phases, temporal resolution 30 ms, and typical repetition time (TR)/echo time (TE)/flip angle 30 ms/6 ms/25°]. Slices were positioned using the highly reproducible ‘3 of 5 technique’.16

ne tags acquired per slice in a 14 s breath hold, typical FOV 300 mm, matrix 128 × 128, slice thickness 10 mm, tag separation 8 mm, 18 phases, temporal resolution 30 ms, and typical repetition time (TR)/echo time (TE)/flip angle 30 ms/6 ms/25°]. Slices were positioned using the highly reproducible ‘3 of 5 technique’.16 Next, myocardial perfusion CMR was planned in the mid-LV short axis orientation with a saturation recovery fast gradient echo method accelerated by two-fold sensitivity encoding (SENSE), TR/TE/flip angle 2.7/1.0/15°, typical field of view 380 × 380 mm, image matrix 160 × 160, in plane spatial resolution 2.4 × 2.4 mm, slice thickness 10 mm, 60 dynamics, preparation pulse delay (to middle of k-space) 150 ms, and shot duration 130 ms.17 For perfusion acquisition, a contrast dose of 0.05 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer-Schering Pharma, Berlin, Germany) was administered at a rate of 5 mL/s followed by a 20 mL saline flush. Breath holding was carried out during the first pass of contrast agent. The same mid-myocardial section was imaged twice–once in mid-systole and end-diastole by choosing two appropriate trigger delays from the cine images. A stress perfusion scan was performed during maximal vasodilatation, stimulated by intravenous infusion of adenosine at a dose of 140 µg/min/kg for 4 min.

ontrast agent. The same mid-myocardial section was imaged twice–once in mid-systole and end-diastole by choosing two appropriate trigger delays from the cine images. A stress perfusion scan was performed during maximal vasodilatation, stimulated by intravenous infusion of adenosine at a dose of 140 µg/min/kg for 4 min. Then, a retrospectively triggered, balanced steady-state free precession (SSFP) short-axis cine stack covering the entire LV was acquired (TR/TE/flip angle 3.3/1.64/60°, typical FOV 380 × 380 mm, matrix 192 × 256, slice thickness 10 mm, slice gap: 0–1 mm, 12 slices, temporal resolution 40 ms, 20 phases). After a 15 min delay from stress perfusion imaging, a rest perfusion scan was performed in identical locations and using the same approach as the stress acquisition followed by an additional top-up bolus of 0.1 mmol/kg Gd-DTPA. LV volumes and mass quantification Using dedicated image analysis software (Q Mass 6.1.6, Medis, Leiden University, Leiden, The Netherlands), the epicardial and endocardial borders were traced off line on the LV cine stack.18 End-diastolic, end-systolic (ES) LV volumes, stroke volume (SV), ejection fraction (EF), and LV mass including papillary muscles were calculated. Relative wall mass (RWM) was calculated by dividing LV mass by LV end-diastolic volume (EDV).19 Left atrial volume was calculated from the SSFP 4- and 2-chamber cine at end-systole using the following formula:20 LAvolume=8×4chamberarea×2chamberarea3×π×lengthoftheleftatria.

s including papillary muscles were calculated. Relative wall mass (RWM) was calculated by dividing LV mass by LV end-diastolic volume (EDV).19 Left atrial volume was calculated from the SSFP 4- and 2-chamber cine at end-systole using the following formula:20 LAvolume=8×4chamberarea×2chamberarea3×π×lengthoftheleftatria. Quantitative myocardial blood flow estimation Analysis was performed with the same dedicated software package. Endocardial and epicardial contours were outlined on one representative dynamic perfusion frame with the optimum blood-to-myocardial contrast, with papillary muscles excluded, and copied to all other dynamic frames (Figure 1). The position of individual contours was then manually corrected to account for any respiratory motion. In order to obtain an arterial input function (AIF), another region of interest was drawn inside the LV blood pool. This method has previously been shown to be highly reproducible.21 Myocardial blood flow (MBF) was measured in the systolic phase where the myocardium is thicker and easier to contour and has previously been shown to be as accurate as diastolic MBF measurement.22 AIF was measured in the diastolic phase when there was the largest area of blood pool to contour. Signal intensity vs. time profiles were then generated for the mid-LV myocardial slice as a whole without dividing into segments and for the LV blood pool. Signal intensity vs. time data generated in MASS were analysed with MATLAB7 R2009b, (The MathWorks Inc., Natick, USA). Fermi-constrained deconvolution was used to generate estimates of absolute MBF in mL/g/min.23 The myocardial perfusion reserve (MPR) was calculated by dividing hyperaemic (stress) by baseline (rest) MBF. Figure 1: Mid-myocardial stress perfusion contours (top left) and blood flow after Fermi-constrained deconvolution (bottom left) with MBF (green) and AIF (red). Mid-myocardial CSPAMM tagging showing contours (top right) and Lagrangian circumferential strain over time (bottom right) with each colour representing a different layer: epicardium (blue), mid-myocardium (cyan), and endocardium (green)

after Fermi-constrained deconvolution (bottom left) with MBF (green) and AIF (red). Mid-myocardial CSPAMM tagging showing contours (top right) and Lagrangian circumferential strain over time (bottom right) with each colour representing a different layer: epicardium (blue), mid-myocardium (cyan), and endocardium (green) CSPAMM analysis Analysis was performed for the entire myocardial slice from each acquisition with a dedicated tagging analysis software package using HARP analysis (TagTrack, GyroTools, Zurich, CH, Switzerland). Endocardial and epicardial contours were drawn by a semi-automated process for each slice, and a mid-myocardial contour was automatically calculated between the endocardial and epicardial contours (Figure 1). The software then tracked the contours throughout all phases of the cardiac cycle. Circumferential Lagrangian strain and strain rate were measured at the mid-ventricular level, which has previously been shown to be the most reproducible.16 Early diastolic strain rate was defined as peak rate in the first four phases after end-systole. LV twist was calculated by subtracting the basal rotation from the apical rotation. LV torsion was calculated by the following formula:24 Torsion shear angle=(Apical rotation−Basal rotation)×(Apical radius+Basal radius)Length×2.

diastolic strain rate was defined as peak rate in the first four phases after end-systole. LV twist was calculated by subtracting the basal rotation from the apical rotation. LV torsion was calculated by the following formula:24 Torsion shear angle=(Apical rotation−Basal rotation)×(Apical radius+Basal radius)Length×2. Statistical analysis Statistical analysis was performed using IBM SPSS® Statistics 19.0. Continuous variables were expressed as means ± standard deviation (SD). Means of the diabetes and non-diabetes groups were compared using an unpaired two-tailed t-test with assumed equal variance. To compare the diabetes, prediabetes, and control groups, analysis of variance with post hoc Tukey correction was used. A P-value of <0.05 was considered statistically significant. Kolmogorov–Smirnov test was used to establish that MPR was not normally distributed. Correlations between MPR and strain measurements were assessed by Spearman’s test. Univariable analyses were performed to identify predictors of increased LV mass index (LVMI), RWM, and LV torsion and reduced MPR. Variables with a probability value of <0.1 in the univariable analysis were included in a stepwise multivariable analysis, based on a linear regression model. Error bars for mean values denote 95% confidence interval.

rformed to identify predictors of increased LV mass index (LVMI), RWM, and LV torsion and reduced MPR. Variables with a probability value of <0.1 in the univariable analysis were included in a stepwise multivariable analysis, based on a linear regression model. Error bars for mean values denote 95% confidence interval. Results Of the 72 recruited patients, three could not complete the scan because of claustrophobia. Two patients in the diabetes group had tagging data of poor quality that could not be interpreted. In two patients, one of the perfusion data sets was of insufficient quality for quantitative analysis (excessive through plane motion). These four patients were excluded from the final analysis. Of the 65 patients with completed scans, 19 were classed as diabetic and 46 non-diabetics (of whom 30 were defined as prediabetic). The mean duration of diabetes was 7.3 ± 9.0 years.

sufficient quality for quantitative analysis (excessive through plane motion). These four patients were excluded from the final analysis. Of the 65 patients with completed scans, 19 were classed as diabetic and 46 non-diabetics (of whom 30 were defined as prediabetic). The mean duration of diabetes was 7.3 ± 9.0 years. Patient characteristics of the study groups are shown in Table 1. The proportion of males (68 vs. 41%, P = 0.048) was higher in the diabetics than non-diabetics. Serum creatinine (93.7 ± 8.7 vs. 85.4 ± 6.4 µmol/L, P = 0.01) and WHR (0.95 ± 0.11 vs. 0.90 ± 0.09, P = 0.046) were higher in diabetics than non-diabetics. There was no significant difference between diabetics and non-diabetics for age, BMI, hypertension, hypercholesterolaemia, albumin–creatinine ratio, or smoking history. HOMA-IR was significantly higher in diabetics (not taking exogenous insulin) than in non-diabetics (10.0 ± 9.1 vs. 2.5 ± 1.5, P < 0.001), but there was no significant difference between prediabetics and controls. Table 1 Patient characteristics with total numbers in each group

umin–creatinine ratio, or smoking history. HOMA-IR was significantly higher in diabetics (not taking exogenous insulin) than in non-diabetics (10.0 ± 9.1 vs. 2.5 ± 1.5, P < 0.001), but there was no significant difference between prediabetics and controls. Table 1 Patient characteristics with total numbers in each group Diabetes Non-diabetes Prediabetes Control Number 19 46 30 16 Agea 59 ± 6 57 ± 7 57 ± 8 57 ± 7 Male 13 (68)† 19 (41) 13 (43) 6 (38) Hypertension 7 (37) 16 (35) 11 (37) 5 (31) Hypercholesterolaemia 16 (84) 38 (82) 26(87) 12 (75) Current smoking 2 (9) 6 (13) 4 (13) 2 (13) BMI (kg/m2)a 30.81 ± 4.6 29 ± 4.9 30.1 ± 5.0 27.7 ± 4.5 WHRa 0.95 ± 0.11† 0.90 ± 0.09 0.91 ± 0.09 0.88 ± 0.09 Clinic SBP (mmHg)a 134.3 ± 14.43 132.8 ± 15.3 131.6 ± 15.0 135.1 ± 15.9 Clinic DBP (mmHg)a 76.4 ± 9.7 77.2 ± 9.5 76.9 ± 8.7 77.8 ± 11.1 LDL cholesterol (mmol/L) 75.7 ± 16.6† 116.2 ± 18.9 115.8 ± 42.1 116.6 ± 34.0 HbA1c (%)a 7.7 ± 1.6† 5.77 ± 0.29 5.9 ± 0.2 5.5 ± 0.2 Serum creatinine (µmol/L)a 93.7 ± 8.7† 85.4 ± 6.4 86.6 ± 14.2 83.1 ± 10.6 Urine ACR (mg/mmol)a 3.8 ± 8.7 1.2 ± 1.83 1.2 ± 1.9 1.1 ± 1.6 HOMA-IRa 10.0 ± 9.1† 2.5 ± 1.5 2.7 ± 1.4 2.2 ± 1.7 Medication ACE-i/ARB 14 (74)† 14 (30) 10 (33) 4 (25) Beta-blocker 11(58) 17 (37) 14 (47) 3 (19) Calcium channel blocker 4(21) 11 (24) 7 (23) 4 (25) Thiazide diuretic 4 (21) 3 (7) 3 (10) 0 (0) HMG-CoA reductase inhibitor (statin) 14 (61)† 17 (37) 10 (33) 7 (44) Metformin 9 (47)† 0 (0) 0 (0) 0 (0) Insulin 5 (26)† 0 (0) 0 (0) 0 (0) Percentage of nominal values in parentheses. aData are mean ± SD. †P < 0.05 when compared with non-diabetes.

Diabetes Non-diabetes Prediabetes Control Number 19 46 30 16 Agea 59 ± 6 57 ± 7 57 ± 8 57 ± 7 Male 13 (68)† 19 (41) 13 (43) 6 (38) Hypertension 7 (37) 16 (35) 11 (37) 5 (31) Hypercholesterolaemia 16 (84) 38 (82) 26(87) 12 (75) Current smoking 2 (9) 6 (13) 4 (13) 2 (13) BMI (kg/m2)a 30.81 ± 4.6 29 ± 4.9 30.1 ± 5.0 27.7 ± 4.5 WHRa 0.95 ± 0.11† 0.90 ± 0.09 0.91 ± 0.09 0.88 ± 0.09 Clinic SBP (mmHg)a 134.3 ± 14.43 132.8 ± 15.3 131.6 ± 15.0 135.1 ± 15.9 Clinic DBP (mmHg)a 76.4 ± 9.7 77.2 ± 9.5 76.9 ± 8.7 77.8 ± 11.1 LDL cholesterol (mmol/L) 75.7 ± 16.6† 116.2 ± 18.9 115.8 ± 42.1 116.6 ± 34.0 HbA1c (%)a 7.7 ± 1.6† 5.77 ± 0.29 5.9 ± 0.2 5.5 ± 0.2 Serum creatinine (µmol/L)a 93.7 ± 8.7† 85.4 ± 6.4 86.6 ± 14.2 83.1 ± 10.6 Urine ACR (mg/mmol)a 3.8 ± 8.7 1.2 ± 1.83 1.2 ± 1.9 1.1 ± 1.6 HOMA-IRa 10.0 ± 9.1† 2.5 ± 1.5 2.7 ± 1.4 2.2 ± 1.7 Medication ACE-i/ARB 14 (74)† 14 (30) 10 (33) 4 (25) Beta-blocker 11(58) 17 (37) 14 (47) 3 (19) Calcium channel blocker 4(21) 11 (24) 7 (23) 4 (25) Thiazide diuretic 4 (21) 3 (7) 3 (10) 0 (0) HMG-CoA reductase inhibitor (statin) 14 (61)† 17 (37) 10 (33) 7 (44) Metformin 9 (47)† 0 (0) 0 (0) 0 (0) Insulin 5 (26)† 0 (0) 0 (0) 0 (0) Percentage of nominal values in parentheses. aData are mean ± SD. †P < 0.05 when compared with non-diabetes. When diabetics, prediabetics, and controls were compared, serum creatinine was higher in diabetics than controls (93.7 ± 8.7 vs. 83.1 ± 10.6 µmol/L, P = 0.03). There were no other differences in patient characteristics.

Diabetes Non-diabetes Prediabetes Control Number 19 46 30 16 Agea 59 ± 6 57 ± 7 57 ± 8 57 ± 7 Male 13 (68)† 19 (41) 13 (43) 6 (38) Hypertension 7 (37) 16 (35) 11 (37) 5 (31) Hypercholesterolaemia 16 (84) 38 (82) 26(87) 12 (75) Current smoking 2 (9) 6 (13) 4 (13) 2 (13) BMI (kg/m2)a 30.81 ± 4.6 29 ± 4.9 30.1 ± 5.0 27.7 ± 4.5 WHRa 0.95 ± 0.11† 0.90 ± 0.09 0.91 ± 0.09 0.88 ± 0.09 Clinic SBP (mmHg)a 134.3 ± 14.43 132.8 ± 15.3 131.6 ± 15.0 135.1 ± 15.9 Clinic DBP (mmHg)a 76.4 ± 9.7 77.2 ± 9.5 76.9 ± 8.7 77.8 ± 11.1 LDL cholesterol (mmol/L) 75.7 ± 16.6† 116.2 ± 18.9 115.8 ± 42.1 116.6 ± 34.0 HbA1c (%)a 7.7 ± 1.6† 5.77 ± 0.29 5.9 ± 0.2 5.5 ± 0.2 Serum creatinine (µmol/L)a 93.7 ± 8.7† 85.4 ± 6.4 86.6 ± 14.2 83.1 ± 10.6 Urine ACR (mg/mmol)a 3.8 ± 8.7 1.2 ± 1.83 1.2 ± 1.9 1.1 ± 1.6 HOMA-IRa 10.0 ± 9.1† 2.5 ± 1.5 2.7 ± 1.4 2.2 ± 1.7 Medication ACE-i/ARB 14 (74)† 14 (30) 10 (33) 4 (25) Beta-blocker 11(58) 17 (37) 14 (47) 3 (19) Calcium channel blocker 4(21) 11 (24) 7 (23) 4 (25) Thiazide diuretic 4 (21) 3 (7) 3 (10) 0 (0) HMG-CoA reductase inhibitor (statin) 14 (61)† 17 (37) 10 (33) 7 (44) Metformin 9 (47)† 0 (0) 0 (0) 0 (0) Insulin 5 (26)† 0 (0) 0 (0) 0 (0) Percentage of nominal values in parentheses. aData are mean ± SD. †P < 0.05 when compared with non-diabetes. When diabetics, prediabetics, and controls were compared, serum creatinine was higher in diabetics than controls (93.7 ± 8.7 vs. 83.1 ± 10.6 µmol/L, P = 0.03). There were no other differences in patient characteristics. LV structure LV mass (112.8 ± 39.7 vs. 91.5 ± 21.3 g, P = 0.01), EDV (171.0 ± 43.0 vs. 151.1 ± 26.2 mL P = 0.03), and SV (95.9 ± 25.7 vs. 86.6 ± 12.2 mL P = 0.05) were higher in diabetics than non-diabetics. There were no significant differences in LVMI, RWM, LA area, or LA area indexed to BSA (Table 2). Table 2 CMR measured mean and SD LV structure, function, and perfusion

= 0.01), EDV (171.0 ± 43.0 vs. 151.1 ± 26.2 mL P = 0.03), and SV (95.9 ± 25.7 vs. 86.6 ± 12.2 mL P = 0.05) were higher in diabetics than non-diabetics. There were no significant differences in LVMI, RWM, LA area, or LA area indexed to BSA (Table 2). Table 2 CMR measured mean and SD LV structure, function, and perfusion Diabetes Non-diabetes Prediabetes Control Mean SD Mean SD Mean SD Mean SD LV mass (g) 112.8† 39.7 91.5 21.3 90.3 18.7 93.7 26.0 LVMI Mosteller (g/m2) 52.5 12.8 47.7 8.9 46.8 7.0 49.4 11.7 EDV (mL) 171.0† 43.0 151.1 26.2 152.0 25.2 149.4 29.0 ESV volume (mL) 73.5 22.9 64.5 17.8 64.3 16.6 64.8 20.4 SV 95.9* 25.7 86.6 12.2 87.7 12.7 84.6 11.3 RWM (g/mL) 0.65 0.14 0.61 0.10 0.60 0.09 0.63 0.11 LA volumes (mL) 100.0 27.3 91.3 17.2 89.6 19.0 94.3 13.4 LA volume index to BSA (mL/m2) 47.2 10.0 47.7 7.7 46.4 8.7 50.0 5.2 EF (%) 57.5 5.5 57.8 5.2 58.0 5.1 57.4 5.6 Peak circumferential strain −0.18 0.03 −0.18 0.03 −0.18 0.03 −0.18 0.03 Peak systolic strain rate (S−1) −0.90 0.22 −0.91 0.17 −0.93 0.19 −0.88 0.13 Peak early diastolic strain rate (S−1) 0.52* 0.25 0.40 0.20 0.41 0.17 0.40 0.24 LV twist (degrees) 10.88† 2.61 9.37 0.41 9.60 2.43 8.93 1.70 LV torsion (degrees) 9.65* 1.90 8.59 1.91 8.69 2.11 8.41 1.51 Stress MBF (mL/g/min) 3.39* 0.85 4.05 1.35 4.10 1.42 3.96 0.31 Rest MBF (mL/g/min) 1.81 0.80 1.54 0.52 1.54 0.57 1.54 0.11 MPR (mL/g/min) 2.10† 0.76 2.84 1.25 2.88 1.34 2.76 0.27 †P < 0.05 when compared with non-diabetes. *P = 0.05 when compared with non-diabetes.

Diabetes Non-diabetes Prediabetes Control Mean SD Mean SD Mean SD Mean SD LV mass (g) 112.8† 39.7 91.5 21.3 90.3 18.7 93.7 26.0 LVMI Mosteller (g/m2) 52.5 12.8 47.7 8.9 46.8 7.0 49.4 11.7 EDV (mL) 171.0† 43.0 151.1 26.2 152.0 25.2 149.4 29.0 ESV volume (mL) 73.5 22.9 64.5 17.8 64.3 16.6 64.8 20.4 SV 95.9* 25.7 86.6 12.2 87.7 12.7 84.6 11.3 RWM (g/mL) 0.65 0.14 0.61 0.10 0.60 0.09 0.63 0.11 LA volumes (mL) 100.0 27.3 91.3 17.2 89.6 19.0 94.3 13.4 LA volume index to BSA (mL/m2) 47.2 10.0 47.7 7.7 46.4 8.7 50.0 5.2 EF (%) 57.5 5.5 57.8 5.2 58.0 5.1 57.4 5.6 Peak circumferential strain −0.18 0.03 −0.18 0.03 −0.18 0.03 −0.18 0.03 Peak systolic strain rate (S−1) −0.90 0.22 −0.91 0.17 −0.93 0.19 −0.88 0.13 Peak early diastolic strain rate (S−1) 0.52* 0.25 0.40 0.20 0.41 0.17 0.40 0.24 LV twist (degrees) 10.88† 2.61 9.37 0.41 9.60 2.43 8.93 1.70 LV torsion (degrees) 9.65* 1.90 8.59 1.91 8.69 2.11 8.41 1.51 Stress MBF (mL/g/min) 3.39* 0.85 4.05 1.35 4.10 1.42 3.96 0.31 Rest MBF (mL/g/min) 1.81 0.80 1.54 0.52 1.54 0.57 1.54 0.11 MPR (mL/g/min) 2.10† 0.76 2.84 1.25 2.88 1.34 2.76 0.27 †P < 0.05 when compared with non-diabetes. *P = 0.05 when compared with non-diabetes. On comparing diabetics, prediabetics, and controls, LV mass was higher in diabetics than prediabetics (112.8 ± 39.7 vs. 90.3 ± 18.7 g, P = 0.02).

Diabetes Non-diabetes Prediabetes Control Mean SD Mean SD Mean SD Mean SD LV mass (g) 112.8† 39.7 91.5 21.3 90.3 18.7 93.7 26.0 LVMI Mosteller (g/m2) 52.5 12.8 47.7 8.9 46.8 7.0 49.4 11.7 EDV (mL) 171.0† 43.0 151.1 26.2 152.0 25.2 149.4 29.0 ESV volume (mL) 73.5 22.9 64.5 17.8 64.3 16.6 64.8 20.4 SV 95.9* 25.7 86.6 12.2 87.7 12.7 84.6 11.3 RWM (g/mL) 0.65 0.14 0.61 0.10 0.60 0.09 0.63 0.11 LA volumes (mL) 100.0 27.3 91.3 17.2 89.6 19.0 94.3 13.4 LA volume index to BSA (mL/m2) 47.2 10.0 47.7 7.7 46.4 8.7 50.0 5.2 EF (%) 57.5 5.5 57.8 5.2 58.0 5.1 57.4 5.6 Peak circumferential strain −0.18 0.03 −0.18 0.03 −0.18 0.03 −0.18 0.03 Peak systolic strain rate (S−1) −0.90 0.22 −0.91 0.17 −0.93 0.19 −0.88 0.13 Peak early diastolic strain rate (S−1) 0.52* 0.25 0.40 0.20 0.41 0.17 0.40 0.24 LV twist (degrees) 10.88† 2.61 9.37 0.41 9.60 2.43 8.93 1.70 LV torsion (degrees) 9.65* 1.90 8.59 1.91 8.69 2.11 8.41 1.51 Stress MBF (mL/g/min) 3.39* 0.85 4.05 1.35 4.10 1.42 3.96 0.31 Rest MBF (mL/g/min) 1.81 0.80 1.54 0.52 1.54 0.57 1.54 0.11 MPR (mL/g/min) 2.10† 0.76 2.84 1.25 2.88 1.34 2.76 0.27 †P < 0.05 when compared with non-diabetes. *P = 0.05 when compared with non-diabetes. On comparing diabetics, prediabetics, and controls, LV mass was higher in diabetics than prediabetics (112.8 ± 39.7 vs. 90.3 ± 18.7 g, P = 0.02). LV function There was no significant difference in EF, peak Lagrangian circumferential strain, or systolic strain rate between diabetics and non-diabetics (Table 2). LV twist (10.88 ± 2.6 vs. 9.37 ± 0.41°, P = 0.02) and shear torsion angle (9.65 ± 1.90 vs. 8.59 ± 1.91°, P = 0.047) were significantly higher in diabetics than non-diabetics. Early diastolic strain rate (0.52 ± 0.25 vs. 0.40 ± 0.20 S−1, P = 0.051) was higher in diabetics than non-diabetics bordering on significance.

). LV twist (10.88 ± 2.6 vs. 9.37 ± 0.41°, P = 0.02) and shear torsion angle (9.65 ± 1.90 vs. 8.59 ± 1.91°, P = 0.047) were significantly higher in diabetics than non-diabetics. Early diastolic strain rate (0.52 ± 0.25 vs. 0.40 ± 0.20 S−1, P = 0.051) was higher in diabetics than non-diabetics bordering on significance. Diabetics had LV twist that was significantly greater than prediabetics (10.88 ± 2.61 vs. 9.60 ± 2.43°, P = 0.04). Myocardial blood flow There was no significant difference in MBF at stress or at rest between diabetics and non-diabetics (Figure 2). However, MPR was significantly decreased (2.10 ± 0.76 vs. 2.84 ± 1.25 mL/g/min, P = 0.01) in diabetics compared with non-diabetics. Figure 2: Mean and 95% confidence intervals of LV mass, LVMI, and EF (top); LV twist, LV torsion, and circumferential strain (middle); and stress MBF, rest MBF, and MPR (below). Within each graph diabetes is on the left, prediabetes in the middle, and controls on the right MPR was decreased in diabetics compared with prediabetics of borderline significance (2.10 ± 0.76 vs. 2.88 ± 1.34 mL/g/min, P = 0.05). There was no difference in MPR between prediabetics and controls. Calculation of Spearman’s correlation coefficient revealed a modest significant correlation between MPR and early diastolic strain rate (r = −0.310, P = 0.01) and LV torsion (r = −0.306, P = 0.01).

MPR was decreased in diabetics compared with prediabetics of borderline significance (2.10 ± 0.76 vs. 2.88 ± 1.34 mL/g/min, P = 0.05). There was no difference in MPR between prediabetics and controls. Calculation of Spearman’s correlation coefficient revealed a modest significant correlation between MPR and early diastolic strain rate (r = −0.310, P = 0.01) and LV torsion (r = −0.306, P = 0.01). Multivariable linear regression analysis In the multivariable linear regression model of all patients (Table 3), only male sex was associated with increased LVMI (Beta = 0.57, P < 0.001). RWM had a significant association with WHR only (Beta = 0.44, P < 0.001). The only factor significantly associated with increased LV torsion shear angle was a history of diabetes (Beta = 0.25, P = 0.047). On univariable regression analysis, age, sex, and diabetes had a correlation with decreased MPR (P < 0.1) and were included in the multivariable analysis. In the multivariate analysis, diabetes (Beta = −0.27, P = 0.03) and age (Beta = −0.27, P = 0.02) had a significant association with decreased MPR. Table 3 Univariable and multivariable linear regression analysis of all patients

a correlation with decreased MPR (P < 0.1) and were included in the multivariable analysis. In the multivariate analysis, diabetes (Beta = −0.27, P = 0.03) and age (Beta = −0.27, P = 0.02) had a significant association with decreased MPR. Table 3 Univariable and multivariable linear regression analysis of all patients LVMI (g/m2) RWM (g/mL) LV torsion (degrees) MPR (mL/g/min) Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Age 0.44 0.51 0.08 0.19 0.01 0.02* Sex <0.001 <0.001* <0.001 0.052 0.39 0.099 0.24 Hypertension 0.44 0.64 0.13 0.30 Diabetes 0.09 0.47 0.11 0.05 0.047* 0.01 0.03* Prediabetes 0.09 0.15 0.10 0.43 0.13 Microalbuminuria 0.68 0.53 0.42 0.99 BMI 0.66 0.13 0.86 0.57 WHR 0.007 0.87 <0.001 <0.001* 0.57 0.87 Current smoking 0.14 0.73 0.19 0.93 Hypercholesterolaemia 0.58 0.93 0.82 0.30 Factors with P < 0.1 in the univariable analysis were included in the multivariable analysis. *Significant correlation P < 0.05. In the multivariable linear regression model of only non-diabetic patients (Table 4), only male sex was associated with increased LVMI (Beta = 0.50, P < 0.001). RWM had a significant association with WHR (Beta = 0.34, P = 0.02). None of the risk factors had a significant association with LV torsion. MPR only had a significant association with male sex (Beta = 0.49, P = 0.001). Table 4 Univariable and multivariable linear regression analysis in non-diabetic patients

P < 0.001). RWM had a significant association with WHR (Beta = 0.34, P = 0.02). None of the risk factors had a significant association with LV torsion. MPR only had a significant association with male sex (Beta = 0.49, P = 0.001). Table 4 Univariable and multivariable linear regression analysis in non-diabetic patients LVMI (g/m2) RWM (g/mL) LV Torsion (degrees) MPR (mL/g/min) Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Age 0.28 0.83 0.12 0.12 Sex <0.001 <0.001* 0.020 0.16 0.20 0.026 0.026* Hypertension 0.88 0.64 0.36 0.35 Microalbuminuria 0.39 0.35 0.22 0.86 BMI 0.33 0.86 0.82 0.94 WHR 0.09 0.68 0.023 0.023* 0.55 0.83 Current smoking 0.09 0.06 0.63 0.31 0.64 Hypercholesterolaemia 0.06 0.06 0.43 0.67 0.37 HOMA-IR 0.14 0.95 0.97 0.43 Factors with P < 0.1 in the univariable analysis were included in the multivariable analysis. *Significant correlation P < 0.05. Discussion We have demonstrated that Type 2 diabetes mellitus, when compared with both non-diabetes and prediabetes, is associated with increased LV mass and LV torsion and decreased MPR.

LVMI (g/m2) RWM (g/mL) LV Torsion (degrees) MPR (mL/g/min) Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Univariable P-value Multivariable P-value Age 0.28 0.83 0.12 0.12 Sex <0.001 <0.001* 0.020 0.16 0.20 0.026 0.026* Hypertension 0.88 0.64 0.36 0.35 Microalbuminuria 0.39 0.35 0.22 0.86 BMI 0.33 0.86 0.82 0.94 WHR 0.09 0.68 0.023 0.023* 0.55 0.83 Current smoking 0.09 0.06 0.63 0.31 0.64 Hypercholesterolaemia 0.06 0.06 0.43 0.67 0.37 HOMA-IR 0.14 0.95 0.97 0.43 Factors with P < 0.1 in the univariable analysis were included in the multivariable analysis. *Significant correlation P < 0.05. Discussion We have demonstrated that Type 2 diabetes mellitus, when compared with both non-diabetes and prediabetes, is associated with increased LV mass and LV torsion and decreased MPR. Data from the large observational Multi-Ethnic Study of Atherosclerosis study11 have previously demonstrated that diabetes is associated with an increase in LV mass. After regression analysis, diabetes was associated with a 3.5 g greater LV mass (95% confidence interval 1.2–5.8 g). The findings of the present study corroborate these findings and also demonstrate that diabetic subjects have an increase in LV mass compared with subjects with prediabetes.

with an increase in LV mass. After regression analysis, diabetes was associated with a 3.5 g greater LV mass (95% confidence interval 1.2–5.8 g). The findings of the present study corroborate these findings and also demonstrate that diabetic subjects have an increase in LV mass compared with subjects with prediabetes. Previous observational studies have demonstrated that insulin resistance and prediabetic states are associated with concentric LV remodelling19 and abnormal LV function.25,26 In our study, patients with prediabetics had LVMI and RWM similar to controls; however, in the multivariable linear regression analysis of all subjects and non-diabetics RWM (a marker of LV remodelling) had a significant correlation with WHR but not BMI or HOMA-IR. The role of visceral adipose tissue in cardiac remodelling and dysfunction is increasingly being recognized,27 and our finding of a correlation between WHR and LV remodelling independent of diabetes status or HOMA-IR adds further evidence to this hypothesis.

ad a significant correlation with WHR but not BMI or HOMA-IR. The role of visceral adipose tissue in cardiac remodelling and dysfunction is increasingly being recognized,27 and our finding of a correlation between WHR and LV remodelling independent of diabetes status or HOMA-IR adds further evidence to this hypothesis. Further, we have demonstrated that diabetics have increased LV twist and torsion despite unchanged circumferential strain. This finding has previously been reported in patients with Type 1 diabetes mellitus and was attributed to subendocardial myocardial dysfunction secondary to small vessel disease.28 Our results and our finding of a negative correlation between MPR and LV torsion suggest that this hypothesis is also applicable in Type 2 diabetes. It is plausible that abnormal myocardial perfusion in Type 2 diabetes contributes to the progressive changes in cardiac strain and function reported in this study and in previous studies.12,29 The hypothesis that impaired MPR leads to cardiac dysfunction raises the possibility of future therapeutic trials of pharmacological agents that increase coronary microvascular function to prevent heart failure. One recent study in which patients with diabetes were randomized to receive the selective phosphodiesterase inhibitor sildenafil or placebo30 reported that sildenafil reversed LV remodelling as well as decreasing LV torsion compared with placebo. The possible mechanistic link between microvascular disease and cardiac dysfunction does suggest a need for further trials into drugs that vasodilate the coronary microvasculature.

inhibitor sildenafil or placebo30 reported that sildenafil reversed LV remodelling as well as decreasing LV torsion compared with placebo. The possible mechanistic link between microvascular disease and cardiac dysfunction does suggest a need for further trials into drugs that vasodilate the coronary microvasculature. It has been reported in previous CMR-based research that diabetic patients when compared with controls have increased LV torsion12 and decreased MPR.31 In both of these studies, the diabetic patients had significantly higher blood pressure (BP) than the controls. It has previously been reported that hypertension leads to increased torsional shear angle, decreased circumferential strain,32 and decreased MPR.33,34 In our study, there was no significant difference in clinic BP between the study groups making the reported changes likely due to diabetic status of the patients. We had hypothesized that patients with prediabetes would also show abnormalities in LV structure, function, and perfusion. The techniques employed in this study did not demonstrate any abnormalities in prediabetes. Our findings have also confirmed MPR measurements by positron emission tomography35 where MPR was also reduced in diabetics but not prediabetics.36 Other CMR techniques such as T1 mapping37,38 or measuring aortic distensibility39,40 have been demonstrated to show changes in diabetes and may be more sensitive to detect subclinical cardiac changes in prediabetes.

ements by positron emission tomography35 where MPR was also reduced in diabetics but not prediabetics.36 Other CMR techniques such as T1 mapping37,38 or measuring aortic distensibility39,40 have been demonstrated to show changes in diabetes and may be more sensitive to detect subclinical cardiac changes in prediabetes. Limitations The three patient groups had similar characteristics, but there were more men and higher creatinine in the diabetic group than non-diabetic group. We attempted to control for this retrospectively by indexing results to BSA where appropriate and by conducting a multivariable linear regression analysis. The correlations reported are modest which reflects the fact that multiple factors influence ventricular structure, function, and perfusion. Diabetics also had significantly higher use of angiotensin-converting-enzyme (ACE) inhibitors and statins than non-diabetic patients. This may have influenced our results, but any effects would be expected to improve the studied parameters in patients taking ACE inhibitors or statins.

Limitations The three patient groups had similar characteristics, but there were more men and higher creatinine in the diabetic group than non-diabetic group. We attempted to control for this retrospectively by indexing results to BSA where appropriate and by conducting a multivariable linear regression analysis. The correlations reported are modest which reflects the fact that multiple factors influence ventricular structure, function, and perfusion. Diabetics also had significantly higher use of angiotensin-converting-enzyme (ACE) inhibitors and statins than non-diabetic patients. This may have influenced our results, but any effects would be expected to improve the studied parameters in patients taking ACE inhibitors or statins. We have only carried out CSPAMM and MPR analysis at the mid-ventricular short axis level because this has previously been demonstrated to be the most reproducible level for analysis of both of these techniques.16,21 We have not reported the segmental analysis of CSPAMM or MPR because we are investigating a diffuse process that affects the whole myocardium. With regard to MPR, our aim was to optimize image acquisition and our study aims to examine ubiquitous changes in the myocardium rather than to detect coronary disease (which was excluded at patient inclusion). For this purpose, single slice coverage was considered sufficient. Patients were recruited at the time of coronary angiography, typically for investigation of chest pain. It is possible that these patients by nature of their symptoms are not representative of all diabetic or non-diabetic patients.

We have only carried out CSPAMM and MPR analysis at the mid-ventricular short axis level because this has previously been demonstrated to be the most reproducible level for analysis of both of these techniques.16,21 We have not reported the segmental analysis of CSPAMM or MPR because we are investigating a diffuse process that affects the whole myocardium. With regard to MPR, our aim was to optimize image acquisition and our study aims to examine ubiquitous changes in the myocardium rather than to detect coronary disease (which was excluded at patient inclusion). For this purpose, single slice coverage was considered sufficient. Patients were recruited at the time of coronary angiography, typically for investigation of chest pain. It is possible that these patients by nature of their symptoms are not representative of all diabetic or non-diabetic patients. Conclusions Patients with diabetes, without CAD, have increased LV mass, LVMI, and LV torsion and decreased MPR when compared with non-diabetic patients. There is a significant association between decreased MPR and increased LV torsion suggesting a possible mechanistic link between microvascular disease and cardiac dysfunction in diabetes. Conflict of interest: None declared. Funding Libyan Ministry of Higher Education & Scientific Research, Tripoli, Libya to A.M.L. and British Heart Foundation to S.P. (FS/10/62/28409). S.P. and J.P.G. received unrestricted educational grants from Philips Healthcare. Funding to pay the Open Access publication charges for this article was provided by The University of Leeds.

Background Aortic stenosis (AS) is the most common form of valvular heart disease in the Western world affecting 5% of those aged over 75 years.1 The standard treatment once symptoms or LV dysfunction develops is aortic valve replacement (AVR).2 However, the perioperative risks of this procedure rise with increasing age and co-morbidity.3,4 Medical therapy to delay the onset of symptoms and progression of AS would be highly desirable, but to date, no medical therapy has been shown to be beneficial in patients with AS. Early retrospective studies of statins suggested that the progression of AS could be delayed,5,6 but subsequent larger randomized trials were negative,7–9 underlining the importance of prospective trials. The response of the myocardium is likely to be as important as the degree of valve stenosis,10 and both the aortic valve area11 and measures of the myocardial response [degree and pattern of left ventricular hypertrophy (LVH),12 presence of fibrosis,13 reductions in longitudinal strain14] have been shown to determine prognosis in these patients.

ium is likely to be as important as the degree of valve stenosis,10 and both the aortic valve area11 and measures of the myocardial response [degree and pattern of left ventricular hypertrophy (LVH),12 presence of fibrosis,13 reductions in longitudinal strain14] have been shown to determine prognosis in these patients. The renin–angiotensin system (RAS) has a major influence on myocardial physiology, and there is some evidence for this in AS: it regulates the degree of LVH,15 the extent of fibrosis in the myocardium,16 and may even play a role in aortic valve thickening.17 RAS inhibitors reduce LVM independent of blood pressure (BP) (suggesting a direct myocardial effect),18 can reduce the extent of myocardial fibrosis,19 and have been shown to improve clinical outcome through LV remodelling in other disease areas—e.g. post-myocardial infarction,20 heart failure,21 and hypertension-induced LVH.18 Inhibition of the RAS with angiotensin-converting enzyme inhibitors (ACEi) would therefore seem an attractive option to improve LV remodelling and myocardial physiology in AS, resulting in better tolerance to the valve obstruction, potentially delaying the onset of symptoms and reducing need for aortic valve surgery. ACEi have, however, been traditionally regarded as contraindicated in moderate or severe AS due to the theoretical danger of syncope caused by afterload reduction, and current guidelines still advise caution. There are however no clinical studies indicating harm, and in fact, the limited animal and human data that exist do not suggest harm, and even suggest benefit.22–24 Recent retrospective studies suggest a potential benefit in AS patients taking ACEi or ARBs25 as well as reductions in the progression of AS,26 but these studies are subject to significant selection and other biases, in a similar way to the early retrospective statin studies in AS, and a prospective clinical trial is required.

etrospective studies suggest a potential benefit in AS patients taking ACEi or ARBs25 as well as reductions in the progression of AS,26 but these studies are subject to significant selection and other biases, in a similar way to the early retrospective statin studies in AS, and a prospective clinical trial is required. With the Ramipril In Aortic Stenosis (RIAS) trial, we therefore sought to carry out the first prospective, randomized, placebo-controlled study of ramipril in AS. Given the perceived historical problems of ACEi in this population and the large scale required for a clinical outcome trial, we planned an intermediate (physiological) study to determine whether there were any positive physiological changes, and an absence of harm, before embarking on a large-scale clinical outcome study. The aims of this study were as follows: To examine changes in myocardial physiology, in particular the regression of left ventricular mass (LVM), as well as other LV physiological parameters (perfusion, LV strain, fibrosis) using multi-parametric cardiac magnetic resonance (CMR) in patients with moderate to severe AS. To assess the safety and tolerability of ramipril in these patients. To examine potential improvements in effort tolerance.

With the Ramipril In Aortic Stenosis (RIAS) trial, we therefore sought to carry out the first prospective, randomized, placebo-controlled study of ramipril in AS. Given the perceived historical problems of ACEi in this population and the large scale required for a clinical outcome trial, we planned an intermediate (physiological) study to determine whether there were any positive physiological changes, and an absence of harm, before embarking on a large-scale clinical outcome study. The aims of this study were as follows: To examine changes in myocardial physiology, in particular the regression of left ventricular mass (LVM), as well as other LV physiological parameters (perfusion, LV strain, fibrosis) using multi-parametric cardiac magnetic resonance (CMR) in patients with moderate to severe AS. To assess the safety and tolerability of ramipril in these patients. To examine potential improvements in effort tolerance. Methods The study protocol and detailed methods have been previously published,27 and a concise summary is provided here. The protocol was approved by the Oxfordshire research ethics committee C (reference 07/H606/139) and the Medicines and Healthcare products Regulatory Agency (clinical trial authorization 21584/0226/001-004). The trial was also registered with the European Community clinical trials database (EudraCT no 2007-005224-32) and the ISRCTN register (24 616 095). All patients gave their written informed consent.

/H606/139) and the Medicines and Healthcare products Regulatory Agency (clinical trial authorization 21584/0226/001-004). The trial was also registered with the European Community clinical trials database (EudraCT no 2007-005224-32) and the ISRCTN register (24 616 095). All patients gave their written informed consent. Study population Subjects were recruited from clinics at the John Radcliffe Hospital and surrounding institutions. All patients aged >18 years with moderate or severe AS by standard echocardiographic criteria [valve area <1.5 cm2, or peak velocity >3.0 m/s (peak valve gradient >36 mmHg)],2 who were asymptomatic as judged by patient-reported symptoms, and who did not have indications for valve replacement surgery were invited to participate. All had normal LV function (ejection fraction >50% by echocardiography) and no other significant (>mild) valvular heart disease, excess hypo- or hypertension (BP < 100/40 or >200/110 mmHg). Intolerance of ACEi or angiotensin receptor blockers (ARBs) or their prescription over the previous 3 months were also exclusion criteria.

normal LV function (ejection fraction >50% by echocardiography) and no other significant (>mild) valvular heart disease, excess hypo- or hypertension (BP < 100/40 or >200/110 mmHg). Intolerance of ACEi or angiotensin receptor blockers (ARBs) or their prescription over the previous 3 months were also exclusion criteria. Study design and drug titration schedule This was a randomized, double-blind, placebo-controlled trial to assess physiological changes in the myocardium with ramipril; it was not powered for clinical end points. Figure 1 summarizes the study design. After baseline visits, subjects were randomized to ramipril or placebo for 1 year. An initial pack of study medication (ramipril 2.5 mg daily or placebo) was provided for 2 weeks to ensure no adverse symptoms. The ramipril was then increased to 5 mg daily and again to 10 mg daily at 12 weeks. The occurrence of adverse events and any changes in laboratory parameters were noted throughout the study. Full assessments occurred at baseline, 6 months and 1 year, and researchers were blinded to the randomization until after data analysis by the statisticians. The primary outcome was the change in LVM from baseline to 12 months measured by CMR. Secondary end points included changes in left ventricular ejection fraction (LVEF), change in other myocardial functional parameters assessed by CMR (perfusion, T1 values, strain), and echocardiography (including diastolic parameters); change in B-type natriuretic peptide (BNP); and change in distance walked on ETT. Figure 1 Study design and flow chart.

es in left ventricular ejection fraction (LVEF), change in other myocardial functional parameters assessed by CMR (perfusion, T1 values, strain), and echocardiography (including diastolic parameters); change in B-type natriuretic peptide (BNP); and change in distance walked on ETT. Figure 1 Study design and flow chart. Cardiovascular magnetic resonance Patients were scanned using a 1.5-T Avanto CMR system (Siemens, Erlangen, Germany). Comprehensive CMR assessment was carried out at baseline and 1 year, while at 6 months only LV volumetric analysis was performed to determine the time course of any changes in the primary end point. LVM, LV volumes, and LV function were assessed using a stack of steady-state free precession short-axis cine images, in accordance with the Society for Cardiovascular Magnetic Resonance guidelines.28 The aortic valve was imaged using short-axis steady-state free precession (SSFP) cine sequences at the valve tips in mid-systole, and transvalvular velocity was measured using breath-hold through-plane phase-contrast velocity mapping just distal to the aortic valve (at the vena contracta).

netic Resonance guidelines.28 The aortic valve was imaged using short-axis steady-state free precession (SSFP) cine sequences at the valve tips in mid-systole, and transvalvular velocity was measured using breath-hold through-plane phase-contrast velocity mapping just distal to the aortic valve (at the vena contracta). Myocardial strain was assessed using a grid-based ‘tagging’ sequence29 in the horizontal long-axis (four chamber) and three short-axis views (basal, mid-ventricular, apical), each during a single breath hold. Diffuse myocardial interstitial fibrosis was assessed using non-contrast myocardial T1 mapping, as a surrogate marker for this in AS.30 A Short Modified Look Locker Inversion recovery (ShMOLLI) technique was used,31 in a single mid-ventricular slice, with the assumption that the degree of diffuse fibrosis was similar throughout the myocardium, and the average T1 value over this whole slice was calculated as previously described.30 Late gadolinium enhancement (LGE) imaging was also performed to assess more patchy, confluent areas of fibrosis 10 min after the injection of gadolinium for perfusion (see below) using a standard inversion recovery technique.32 This was repeated with the phase-encoding direction swapped to exclude artefact.

escribed.30 Late gadolinium enhancement (LGE) imaging was also performed to assess more patchy, confluent areas of fibrosis 10 min after the injection of gadolinium for perfusion (see below) using a standard inversion recovery technique.32 This was repeated with the phase-encoding direction swapped to exclude artefact. Myocardial perfusion reserve was assessed according to guidelines,33 following administration of adenosine at a rate of 140 µg/kg/min for 3 min. Gadolinium-based contrast (Omniscan 0.03 mmol/kg at 6 mL/s; Nycomed Amersham, Little Chalfont, UK) was administered intravenously, to maintain a linear relationship between signal intensity and perfusion. Perfusion imaging was performed in three short-axis sections during the first pass of the contrast bolus34 and repeated at rest at least 20 min later. Myocardial perfusion reserve index was calculated for all 16 segments, and the average value was used (MPRI: the ratio of stress to rest normalized myocardial perfusion upslopes). Echocardiography Transthoracic echocardiography was carried out, particularly to assess diastolic function, using a Philips iE33 advanced echo system (Philips Medical Systems, Best, Netherlands). A full echo study was performed according to guidelines for assessment and classification of AS and chamber quantification, including Tissue Doppler measurements of systolic tissue deformation.2,35 LV diastolic function was assessed using tissue Doppler measurements of medial and lateral mitral annular velocities in early diastole and mitral valve inflow.35,36

to guidelines for assessment and classification of AS and chamber quantification, including Tissue Doppler measurements of systolic tissue deformation.2,35 LV diastolic function was assessed using tissue Doppler measurements of medial and lateral mitral annular velocities in early diastole and mitral valve inflow.35,36 Exercise treadmill testing Exercise testing was carried out to assess the maximum walking distance as a continuous variable in the study population (6 min walk tests are less useful in active patients). The Naughton protocol was chosen37 as it contains multiple small increments in speed and incline, to provide a smooth increase in workload more applicable to a continuous variable than fewer large increases in workload, which occur in other protocols. The test was carried out under medical supervision with continuous 12-lead electrocardiogram recording and regular BP monitoring. Clinical events These were not a primary aim of this study, which was too small to examine this robustly, but any major adverse cardiac events (death, AVR or hospital admission with cardiac symptoms) were recorded. Any decision to refer patients for AVR was taken by the patient's treating cardiologist, who was also blinded to the study medication.

t a primary aim of this study, which was too small to examine this robustly, but any major adverse cardiac events (death, AVR or hospital admission with cardiac symptoms) were recorded. Any decision to refer patients for AVR was taken by the patient's treating cardiologist, who was also blinded to the study medication. Statistical methods A full description of these is published with the trial protocol,27 and a brief summary is included here. Analyses were carried out by an independent statistician (SG) at the Centre for Statistics in Medicine, University of Oxford in accordance with the trial statistical analysis plan. The sample size was based on changes in LVM, using a baseline from a prior study of AS with CMR (142 ± 35 g/m2),38 and a 15% (21.3 g/m2) reduction in LVM—the mean change from a large meta-analysis of antihypertensive treatments.18 We used a one-sided test (only including reduction in LVM) with 85% power (β error) and 95% confidence (α error). The number needed in each study group with these calculations was 43, and a total of 50 patients per group was planned, allowing for a 15% drop-out rate. Primary and secondary analyses were conducted on the modified intention to treat (mITT) population, including all participants who received study medication and at least one follow-up measurement.

y group with these calculations was 43, and a total of 50 patients per group was planned, allowing for a 15% drop-out rate. Primary and secondary analyses were conducted on the modified intention to treat (mITT) population, including all participants who received study medication and at least one follow-up measurement. Results One hundred patients with moderate (n = 80) or severe (n = 20) AS were recruited between October 2008 and April 2011, and baseline patient characteristics are summarized in Table 1. Four patients withdrew after randomization, leaving 96 patients who received the trial medication (47 placebo; 49 ramipril). Nineteen patients withdrew during the trial (mostly within the first 6 months), leaving 77 patients who completed the 1-year assessment—only 2 patients (one from each group) withdrew between 6 and 12 months; see Figure 2 for CONSORT diagram. Treatment groups were balanced at baseline with respect to demographics, symptoms, CMR, echocardiography, and exercise testing data (Table 1). Any differences were slight and the numbers too small to draw any inferences. Table 1 Baseline patient characteristics

between 6 and 12 months; see Figure 2 for CONSORT diagram. Treatment groups were balanced at baseline with respect to demographics, symptoms, CMR, echocardiography, and exercise testing data (Table 1). Any differences were slight and the numbers too small to draw any inferences. Table 1 Baseline patient characteristics Characteristics Placebo group (n = 47) Ramipril group (n = 49) P-value (unpaired t-test) Male gender, n (%) 36 (75.0) 35 (71.4) 0.70 Age 70.0 (14.6) 67.2 (13.7) 0.34 BMI 28.0 (5.4) 29.2 (4.8) 0.27 Systolic BP (mmHg) 135 (18) 130 (16) 0.13 Diastolic BP (mmHg) 77 (9) 77 (6) 0.57 Smoker, n (%) 3 (6.1) 5 (10.0) 0.48 Ex-smoker, n (%) 8 (16.7) 2 (4.1) 0.04 Hypertension, n (%) 17 (35.4) 11 (22.4) 0.16 Myocardial infarction, n (%) 1 (2.1) 0 (0.0) 0.32 CABG, n (%) 2 (4.2) 2 (4.3) 0.98 Stents, n (%) 2 (4.2) 2 (4.1) 0.98 Diabetes, n (%) 2 (4.2) 1 (2.0) 0.55 Medications β-Blockers, n (%) 10 (20.8) 10 (20.8) 0.80 Statins, n (%) 27 (56.3) 20 (41.7) 0.16 Aspirin, n (%) 16 (33.3) 20 (41.7) 0.40 Diuretics, n (%) 10 (20.8) 5 (10.4) 0.04 Calcium channel blockers, n (%) 10 (20.8) 5 (10.4) 0.16 CMR parameters LVM (g) 154.8 (40.0) 158.4 (49.8) 0.80 LVM index (g/m2) 80.7 (19.5) 79.6 (20.5) 0.80 LVEF (%) 72.6 (8.2) 70.8 (8.0) 0.27 Peak velocity (m/s) 3.2 (0.6) 3.1 (0.7) 0.33 Aortic valve area (cm2) 1.2 (0.4) 1.2 (0.3) 0.59 Number with moderate AS (%) 37 (79%) 39 (80%) 0.85 T1 value (ms) 961 (39) 951 (24) 0.35 MPRI 1.20 (0.41) 1.37 (0.37) 0.13 Longitudinal strain (%) −9.7 (2.4) −9.7 (2.9) 0.17 Circumferential strain (%) −15.9 (4.5) −17.1 (2.2) 0.17 Echo parameters AV max (m/s) 3.5 (0.5) 3.3 (0.5) 0.12 AV mean (m/s) 2.4 (0.4) 2.3 (0.4) 0.09 Septal E/E' ratio 13.0 (5.7) 12.5 (4.8) 0.72 Lateral E/E' ratio 11.0 (5.3) 10.1 (5.7) 0.52 S-wave (cm/s) 6.3 (1.3) 6.2 (0.8) 0.85 Biomarkers BNP (pmol/L) 20.9 (35.5) 14.9 (21.5) 0.33 Exercise tolerance Exercise distance (m) 985 (360) 1030 (386) 0.57 Values are mean (standard deviation) unless indicated otherwise.

0.09 Septal E/E' ratio 13.0 (5.7) 12.5 (4.8) 0.72 Lateral E/E' ratio 11.0 (5.3) 10.1 (5.7) 0.52 S-wave (cm/s) 6.3 (1.3) 6.2 (0.8) 0.85 Biomarkers BNP (pmol/L) 20.9 (35.5) 14.9 (21.5) 0.33 Exercise tolerance Exercise distance (m) 985 (360) 1030 (386) 0.57 Values are mean (standard deviation) unless indicated otherwise. BMI, body mass index; LVM, left ventricular mass; LVEF, left ventricular ejection fraction; MPRI, myocardial perfusion reserve index; BNP, B-type natriuretic peptide. Figure 2 CONSORT flow chart—numbers of subjects randomized and allocated to each treatment. Primary end point: change in LVM There was a modest but significant difference between treatment groups at 1 year: the ramipril group showed a reduction in mean LVM of −3.9 g compared with an increase of +4.5 g in the placebo group, leading to an overall mean difference of 8.4 g (P = 0.0057, Figure 3). The change in LVM was also progressive, with a similar reduction in LVM with ramipril and increase with placebo at 6 months, though with half the degree of change that occurred at 1 year: mean difference between groups 4.0 g (P = 0.089). Figure 3 Changes in LV mass with ramipril 10 mg daily or placebo over a 12-month period.

ange in LVM was also progressive, with a similar reduction in LVM with ramipril and increase with placebo at 6 months, though with half the degree of change that occurred at 1 year: mean difference between groups 4.0 g (P = 0.089). Figure 3 Changes in LV mass with ramipril 10 mg daily or placebo over a 12-month period. Secondary end points Other parameters of LV physiology Most assessments of myocardial physiology using CMR and echocardiography did not show significant differences between treatment groups (Table 2). This included LVEF, myocardial T1 values (as a surrogate marker of interstitial fibrosis), perfusion indices, strain, and diastolic function. Systolic myocardial velocity (measured using tissue Doppler S-wave) worsened slightly in the placebo group compared with ramipril (−0.5 vs. 0.0 cm/s, P = 0.04). However, these group differences are small and baseline S-wave measurements were lower than normally expected, so this may reflect measurement variability. Table 2 Primary and secondary end points

d using tissue Doppler S-wave) worsened slightly in the placebo group compared with ramipril (−0.5 vs. 0.0 cm/s, P = 0.04). However, these group differences are small and baseline S-wave measurements were lower than normally expected, so this may reflect measurement variability. Table 2 Primary and secondary end points Placeboa Ramiprila Differenceb P-value n = 41 n = 36 Primary end point Change in LVM (g) 6 months +2.0 ± 1.6 −2.0 ± 1.7 −4.0 ± 2.3 0.089 12 months +4.5 ± 2.1 −3.9 ± 2.2 −8.4 ± 3.0 0.006 Change in LVM index (g/m2) 6 months +1.3 ± 0.9 −1.0 ± 1.0 −2.3 ± 1.3 0.078 12 months +3.5 ± 1.5 −1.0 ± 1.6 −4.4 ± 2.1 0.036 Secondary end points Change in CMR parameters LVEF (%) −0.3 ± 1.0 +1.0 ± 1.0 1.3 ± 1.3 0.328 AVmax (m/s) +0.1 ± 0.1 0.0 ± 0.1 −0.1 ± 0.1 0.277 Aortic valve area (cm2) −0.2 ± 0.05 0.0 ± 0.1 −0.2 ± 0.1 0.067 T1 values (ms) −2 ± 6 +4 ± 7 6 ± 9 0.530 MPRIc +0.1 ± 0.1 0.0 ± 0.1 −0.1 ± 0.2 0.518 Longitudinal strain (%) +1.0 ± 0.7 +0.4 ± 0.7 −0.6 ± 1.0 0.550 Circumferential strain (%) +0.1 ± 0.8 +0.7 ± 0.8 +0.6 ± 1.1 0.584 Change in echo parameters AVmax (m/s) +0.03 ± 0.49 +0.05 ± 0.30 −0.025 ± 0.10 0.801 AV mean (m/s) +0.04 ± 0.30 +0.05 ± 0.25 −0.012 ± 0.06 0.841 Septal E/E' +0.7 ± 1.1 +1.6 ± 1.1 +0.9 ± 1.6 0.530 Lateral E/E' +0.4 ± 1.1 +1.2 ± 1.4 +0.8 ± 1.8 0.632 S-wave (cm/s) −0.5 ± 0.2 0.0 ± 0.2 +0.5 ± 0.2 0.040 Change in other parameters Systolic BP (mmHg) −2.9 ± 2.1 −5.5 ± 2.2 −2.7 ± 3.0 0.374 Diastolic BP (mmHg) −1.4 ± 1.1 −3.6 ± 1.8 −2.2 ± 1.6 0.160 BNP (pmol/L) +8.2 ± 3.4 −0.5 ± 3.7 −8.6 ± 5.1 0.086 Exercise distance (m) +29 ± 25 −20 ± 26 −49 ± 36 0.176 P-values in bold indicate P < 0.05. Abbreviations as in Table 1.

.040 Change in other parameters Systolic BP (mmHg) −2.9 ± 2.1 −5.5 ± 2.2 −2.7 ± 3.0 0.374 Diastolic BP (mmHg) −1.4 ± 1.1 −3.6 ± 1.8 −2.2 ± 1.6 0.160 BNP (pmol/L) +8.2 ± 3.4 −0.5 ± 3.7 −8.6 ± 5.1 0.086 Exercise distance (m) +29 ± 25 −20 ± 26 −49 ± 36 0.176 P-values in bold indicate P < 0.05. Abbreviations as in Table 1. aMean change from baseline, adjusted for baseline. bResult of linear regression assessing change from baseline, adjusted for baseline. cn = 30 for this parameter. Aortic valve area (by CMR direct planimetry) At 1 year, there was a trend towards a slower rate of progression of AS in the ramipril group, with a static aortic valve area (0.0 cm2) compared with a reduction in the placebo group (−0.2 cm2), P = 0.067. This did not, however, impact on the peak velocity across the aortic valve which did not differ significantly between the ramipril and placebo groups: change at 1 year +0.03 and +0.12 m/s, respectively, P = 0.28. Changes in BNP There was a trend towards stabilization in BNP in the ramipril group (−0.50 pmol/L) and increase in the placebo group (+8.2 pmol/L), but the difference between groups was not statistically significant (P = 0.086), and these changes were very small, which, coupled with the low mean baseline values (15–21 pmol/L), limits any interpretation. Changes in exercise tolerance There was no significant difference in the mean change in distance walked on the treadmill at 12 months compared with baseline (−20.1 m in the ramipril group vs. +28.7 m in the placebo group; P = 0.18).

Changes in BNP There was a trend towards stabilization in BNP in the ramipril group (−0.50 pmol/L) and increase in the placebo group (+8.2 pmol/L), but the difference between groups was not statistically significant (P = 0.086), and these changes were very small, which, coupled with the low mean baseline values (15–21 pmol/L), limits any interpretation. Changes in exercise tolerance There was no significant difference in the mean change in distance walked on the treadmill at 12 months compared with baseline (−20.1 m in the ramipril group vs. +28.7 m in the placebo group; P = 0.18). Blood pressure BP reduced slightly in both groups at 12 months, with systolic pressure falling by −5.5 vs. −2.9 mmHg for ramipril and placebo, respectively, though differences between groups were not statistically significant (P = 0.37). Adverse events Ramipril was well tolerated. There was one serious adverse event in a patient who developed neutropenia, leading to discontinuation of the trial medication (which was placebo). The trial was not powered for clinical events, and there were the same number of major adverse cardiac events (5) in each group, and a similar number of AVRs: ramipril 4 vs. placebo 2; P = 0.52. Discussion In the first prospective, randomized, placebo-controlled trial of ACEi in patients with moderate and severe asymptomatic AS, we have shown that ramipril reduces the hypertrophic response of the myocardium, may have additional benefits, and is well tolerated.

Adverse events Ramipril was well tolerated. There was one serious adverse event in a patient who developed neutropenia, leading to discontinuation of the trial medication (which was placebo). The trial was not powered for clinical events, and there were the same number of major adverse cardiac events (5) in each group, and a similar number of AVRs: ramipril 4 vs. placebo 2; P = 0.52. Discussion In the first prospective, randomized, placebo-controlled trial of ACEi in patients with moderate and severe asymptomatic AS, we have shown that ramipril reduces the hypertrophic response of the myocardium, may have additional benefits, and is well tolerated. LV response to ramipril in AS We found a modest reduction in LVM with ramipril, and although the observed differences are small, these represent group mean differences, suggesting an overall shift in LVM within the cohort. There was also a progressive change over the year (Figure 3), which may have continued if ramipril had been given for longer, and these changes occurred despite the fixed outflow tract obstruction from the AS. There were small differences in BP between the groups at 12 months, but the differences were not statistically significant, suggesting that this reduction in mass was driven by direct effects of ramipril on the myocardium, similar to the benefits described in the hypertensive population,39 though a contribution from the BP is difficult to rule out given the moderate group sizes. Baseline LVM index was also significantly smaller than reference group used for power calculations (80 vs. 142 g/m2, respectively), but the previous study38 examined patients just prior to AVR, and it is likely that these were at the more severe end of the spectrum, which may explain the difference (our study included many patients with moderate AS). Patients with severe LVH from non-valvular causes have an adverse prognosis,12 and reductions in LVM have been associated with improvements in prognosis.40 Reduced LVH in response to pressure overload in AS might therefore lead to improvements in prognosis, and some studies suggest a stronger prognostic association with LVM than the severity of AS per se.12 It would thus be reasonable to hypothesize that ramipril may lead to improved prognosis in AS by reducing LVH, and this is supported by findings from a retrospective study that showed reduced mortality in patients taking ACEi in AS.25 The current study was not powered for clinical outcomes however, and any prognostic benefit from RAS inhibition in AS requires a larger clinical outcome trial to address this question.

ducing LVH, and this is supported by findings from a retrospective study that showed reduced mortality in patients taking ACEi in AS.25 The current study was not powered for clinical outcomes however, and any prognostic benefit from RAS inhibition in AS requires a larger clinical outcome trial to address this question. Other CMR measures of LV physiology—perfusion, strain, T1 values (an indirect CMR measure of interstitial cardiac fibrosis30)—or presence of LGE did not differ between groups for the most part. There were small improvements in the echocardiographic tissue Doppler S-wave (a measure of longitudinal contraction) with ramipril, though baseline values were lower than normal—this may reflect reduced long-axis contraction in established AS and would be in keeping with the reduced longitudinal strain values in both groups, but it could also be partly due to mitral annular calcification restricting systolic mitral annular velocities (this is relatively common in AS and was not assessed in our study). There were also no differences in exercise capacity between groups, either at baseline or any changes during the trial, in keeping with small, if any, differences in myocardial physiological parameters. All patients had normal ejection fractions at baseline, however, which may partly explain the lack of a consistent beneficial effect on myocardial physiology with ramipril (it may have been difficult to demonstrate improvements in normal function), though the study may also be under-powered to show changes in these parameters.

ents had normal ejection fractions at baseline, however, which may partly explain the lack of a consistent beneficial effect on myocardial physiology with ramipril (it may have been difficult to demonstrate improvements in normal function), though the study may also be under-powered to show changes in these parameters. We hypothesized at the outset that ramipril might improve interstitial cardiac fibrosis by lowering circulating angiotensin II, a promoter of interstitial cardiac fibrosis,41 and losartan has previously been shown to cause regression of interstitial cardiac fibrosis and improvements in diastolic function in hypertensive patients.19 Our inability to demonstrate changes in T1 values (a surrogate marker of interstitial fibrosis30) may have been due to the imaging technique being insufficiently sensitive to pick up small changes in interstitial fibrosis. Alternatively, the reduction in LVM may be due to a reduction in myocyte size, as recently shown in patients post-AVR,42 and there may have been little, if any, regression of fibrosis. Further, in other studies, the regression in cardiac fibrosis in hypertensive patients (associated with improvements in diastolic parameters) was only seen in those with severe interstitial fibrosis—no changes in either fibrosis or diastolic function were seen in those with mild or moderate fibrosis.19 As the majority of patients in our cohort had moderate AS (in whom we have previously shown T1 values did not differ from normal controls30), the fibrosis burden may have been too light for significant changes with medical treatment to be demonstrated.

tolic function were seen in those with mild or moderate fibrosis.19 As the majority of patients in our cohort had moderate AS (in whom we have previously shown T1 values did not differ from normal controls30), the fibrosis burden may have been too light for significant changes with medical treatment to be demonstrated. Our study did not show any significant improvement in perfusion indices, despite myocardial perfusion previously being associated with LVM in patients with severe AS.43 The number of patients who completed perfusion imaging at both time points in our study was small however (n = 30), and the reduction in LVM may have been too small and the relationship between LVM and perfusion too weak to show any changes. This relationship between perfusion and increased LVM may also differ in moderate AS (which represented the majority of our cohort).

at both time points in our study was small however (n = 30), and the reduction in LVM may have been too small and the relationship between LVM and perfusion too weak to show any changes. This relationship between perfusion and increased LVM may also differ in moderate AS (which represented the majority of our cohort). Potential effect on the aortic valve The ramipril group showed a trend towards reduced progression of the AS (change in valve area 0 vs. −0.2 cm2 in the ramipril and placebo groups respectively, though the statistical strength of the difference was weak (P = 0.067). There is nonetheless some evidence to support the involvement of the RAS in aortic valve calcification. Activation of the local RAS in aortic valves has been seen in AS;44,45 angiotensin II is implicated in aortic valve thickening;17 and aortic valve weights (extracted at the time of AVR) are significantly lower in AS patients taking ARBs.46 Retrospective clinical studies show conflicting results however: O'Brien et al.47 reported significant reductions in aortic valve calcification in AS treated with ACEi, while Rosenhek et al.5 found no change in the progression of aortic valve disease or calcification. The ongoing ROCK-AS trial (NCT00699452) aims to analyse the degree of valvular inflammation, calcification, lipid accumulation, and fibrosis from histology of aortic valves removed at surgery from AS patients taking candesartan compared with placebo and may help to answer these questions.

alve disease or calcification. The ongoing ROCK-AS trial (NCT00699452) aims to analyse the degree of valvular inflammation, calcification, lipid accumulation, and fibrosis from histology of aortic valves removed at surgery from AS patients taking candesartan compared with placebo and may help to answer these questions. Clinical events and safety There were no differences in the progression to AVR or major adverse clinical events between the two groups. However, the trial was not powered for clinical events, consequently the number of events was small and the length of time too short for any meaningful conclusions. Encouragingly, there was no increase in adverse events for patients taking ramipril, which supports the hypothesis that ACEi are safe in AS, as suggested by a number of small retrospective studies.23,24 Our group size was modest though, and a larger study with longer follow-up would be required to evaluate these aspects more robustly. Limitations The data in this study are encouraging but were based on a relatively small sample size, with limited follow-up of 1 year. It assessed detailed physiological changes, which are feasible in this sample size with CMR, but the study was significantly under-powered for any assessment of clinical outcome. A further clinical outcome study would be required to assess this, and preliminary power calculations suggest that such a study would require around 1400 subjects over 4 years to determine any significant effect on clinical events.

R, but the study was significantly under-powered for any assessment of clinical outcome. A further clinical outcome study would be required to assess this, and preliminary power calculations suggest that such a study would require around 1400 subjects over 4 years to determine any significant effect on clinical events. Conclusion In this first prospective trial of ACEi in patients with severe and moderate AS, we demonstrated that they are likely to be well tolerated and may reduce LVH. The study size was modest however, and a larger trial powered for clinical outcomes is required to confirm these physiological changes and determine whether they translate into improved clinical outcomes. If this is shown, ACEi could potentially be of benefit to significant numbers of patients with asymptomatic AS and would be the first medical treatment for this condition. Funding The work was supported by a Heart Research UK project grant RG2512 and the Oxford Comprehensive Biomedical Research Centre, funded by the UK National Institute for Health Research (A.P.B., B.D.P., S.N., S.G.M.). S.B. was supported by a British Heart Foundation Clinical Research Training Fellowship FS/10/015/28104. Funding to pay the Open Access publication charges for this article was provided by the University of Oxford.

l Research Centre, funded by the UK National Institute for Health Research (A.P.B., B.D.P., S.N., S.G.M.). S.B. was supported by a British Heart Foundation Clinical Research Training Fellowship FS/10/015/28104. Funding to pay the Open Access publication charges for this article was provided by the University of Oxford. Acknowledgements We acknowledge all the cardiology consultants who referred patients for the study from the John Radcliffe and surrounding hospitals (including the Royal Berkshire, Northampton General, Milton Keynes, Horton, and Wycombe Hospitals). We would also like to thank Merryn Voysey from the Centre for Statistics in Medicine for senior statistical support. Conflict of interest: none declared.

Introduction Cardiac magnetic resonance (CMR) has become an established method for myocardial perfusion imaging.1 CMR offers superior spatial resolution in comparison to other perfusion imaging modalities. Additionally, an elevated temporal resolution enables the dynamic visualization of the first-pass wash-in of contrast agent in rest and stress conditions. In normal hearts, the myocardium is perfused relatively homogeneously across all myocardial segments. This results in a homogeneous display of CMR perfusion signals both in the amplitude and in the temporal direction, with the peak myocardial signal intensity occurring nearly simultaneously in all segments a few beats after the peak arterial input signal. In contrast, in ischaemic hearts, abnormal segments display a peak myocardial signal that is both reduced in amplitude and delayed resulting in lower peak signal intensity and temporal dyssynchrony across the ventricle. There have been various approaches to quantification of first-pass perfusion signal intensities; however, to our knowledge, the temporal dyssynchrony of perfusion signals has not yet been exploited directly by any diagnostic algorithm. In this study, we hypothesized that important diagnostic information can be derived by analysing first-pass perfusion signals in the temporal direction.

erfusion signal intensities; however, to our knowledge, the temporal dyssynchrony of perfusion signals has not yet been exploited directly by any diagnostic algorithm. In this study, we hypothesized that important diagnostic information can be derived by analysing first-pass perfusion signals in the temporal direction. We sought to describe perfusion dyssynchrony analysis, a novel approach to the analysis of perfusion CMR data. This is specifically designed to isolate and measure the temporal dyssynchrony of myocardial perfusion independently from absolute myocardial blood flow (MBF; Figure 1). Specifically, we tested the potential of perfusion dyssynchrony analysis as a tool for the detection of haemodynamically significant coronary artery disease (CAD) assessed by fractional flow reserve (FFR) and to differentiate between patients with single- and multi-vessel CAD. Figure 1 Areas of abnormal myocardial perfusion are characterized by reduced and delayed wash-in of contrast agent. These features are the basis for visual analysis. Quantitative analysis detects and measures absolute differences of perfusion (vertical arrow). To achieve this, myocardial signal intensity curves require temporal realignment before deconvolution with the arterial input function. This is particularly important when high-resolution voxel-wise perfusion quantification is performed. Dyssynchrony analysis instead does not take into account changes in the amplitude of signal intensity but rather isolates and measures the temporal dyssynchrony of the wash-in curves (horizontal arrow). SI, signal intensity.

s particularly important when high-resolution voxel-wise perfusion quantification is performed. Dyssynchrony analysis instead does not take into account changes in the amplitude of signal intensity but rather isolates and measures the temporal dyssynchrony of the wash-in curves (horizontal arrow). SI, signal intensity. Methods Study population Patients referred for stress perfusion CMR were retrospectively included. All patients had undergone invasive coronary angiography and FFR assessment in all vessels with visually >50% severity stenosis within 3 months of the CMR scan. FFR <0.8 was considered haemodynamically significant. Patients were assigned to the Normal, Single-vessel, and Multi-vessel groups according to the results of the invasive assessment. Patients with previous coronary artery bypass grafting, hypertrophic cardiomyopathy, aortic stenosis, or other primary myopathic or valvular disease were excluded. This study was performed in accordance with the principles set by the Declaration of Helsinki and was conducted in accordance with local ethical standards. All participants gave written informed consent.

rafting, hypertrophic cardiomyopathy, aortic stenosis, or other primary myopathic or valvular disease were excluded. This study was performed in accordance with the principles set by the Declaration of Helsinki and was conducted in accordance with local ethical standards. All participants gave written informed consent. CMR acquisition The CMR scans, including adenosine stress and rest perfusion, functional and scar imaging, were carried out at 3.0T (Philips Achieva-TX, Philips Medical Systems) using standard acquisition protocols.2 A k-t SENSE gradient echo method was used, and typical sequence parameters were repetition time/echo time 3.0/1.0 ms, flip angle 15°, 90° saturation prepulse, 120 ms prepulse delay, spatial resolution 1.2 × 1.2 × 10 mm3. Perfusion data were acquired in three left ventricular (LV) short-axis views covering 16 standard myocardial segments during adenosine-induced hyperaemia over 3 min (140 μg/kg/min) and 15 min later at rest using 0.075 mmol/kg gadobutrol (Gadovist, Schering, Berlin, Germany) at 4 mL/s followed by a 20 mL saline flush. A dual-bolus contrast agent scheme was used as previously described.3 Functional data were acquired with steady-state free precession cine sequences prescribed in short axis and long axis of the LV.4 Right and LV volumes and function and LV mass were measured according to standard analysis criteria.5 Late gadolinium enhancement (LGE) images were acquired 15 min after injection of a top up bolus of contrast agent performed after rest perfusion imaging to a total dose of gadolinium of 0.2 mEq/kg of body weight.4

LV.4 Right and LV volumes and function and LV mass were measured according to standard analysis criteria.5 Late gadolinium enhancement (LGE) images were acquired 15 min after injection of a top up bolus of contrast agent performed after rest perfusion imaging to a total dose of gadolinium of 0.2 mEq/kg of body weight.4 Visual CMR analysis The scans were visually assessed by consensus of at least two expert readers (level of accreditation III according to the guidelines of the Society for Cardiovascular Magnetic Resonance—SCMR) as part of routine clinical assessment.6,7 Rest and stress images were reviewed in conjunction with LGE images.8 Perfusion defects were defined based on standardized criteria set by the SCMR.5 Each cardiac segment was assigned to the appropriate perfusion territory, with segment 15 assigned to the dominant coronary artery (defined by the observer analysing the angiogram).9 A visual score was given for image quality of each dataset using a 4-point scale: 1—poor, 2—fair, 3—good, and 4— excellent. The severity of respiratory and dark rim artefacts was also scored on a 4-point and 3-point scale, respectively. For respiratory artefacts: 1—non-diagnostic; 2—severe artefacts but diagnostic; 3—mild artefacts; 4—no artefacts. For dark rim artefacts: 1—circumferential; 2—segmental; 3—absent.

, 3—good, and 4— excellent. The severity of respiratory and dark rim artefacts was also scored on a 4-point and 3-point scale, respectively. For respiratory artefacts: 1—non-diagnostic; 2—severe artefacts but diagnostic; 3—mild artefacts; 4—no artefacts. For dark rim artefacts: 1—circumferential; 2—segmental; 3—absent. Perfusion dyssynchrony analysis After automated respiratory motion correction and image segmentation,10 a grid of 60 angular positions located on chords perpendicular to the myocardial centerline was generated.11 Transmural contrast agent wash-in signal intensity curves were then extracted for each angular position and filtered in the spatial and temporal domain using a binomial filter.12,13 For each patient, perfusion dyssynchrony analysis was performed on a total of 180 radial segments (60 segments/slice) and on both stress and rest perfusion datasets. The temporal dyssynchrony of LV perfusion was measured as four perfusion dyssynchrony indices; the variance and the coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve (TTMU), and the variance and coefficient of variation of the time to peak myocardial signal intensity (TTP; Figure 2). Variances (V-TTMU and V-TTP) are expressed in square seconds (s2). Coefficients of variation (C-TTMU and C-TTP) are represented as percentages. Figure 2 Schematic representation of perfusion dyssynchrony analysis. First-pass myocardial signal intensity curves are shown for two myocardial segments (Segment A and Segment B) and for the AIF measured from the LV cavity. A1 and B1 indicate the point of maximum signal intensity upslope in each segment. A2 and B2 indicate the peak of signal intensity in each segment. In this schematic example, the calculated perfusion dyssynchrony indices are coefficient of variation of time to maximum upslope (C-TTMU) 22%; variance of time to maximum upslope (V-TTMU) 2.3 s2; coefficient of variation of time to peak signal (C-TTP) 8.8%; variance of time to peak signal (V-TTP) 1 s2.

in each segment. In this schematic example, the calculated perfusion dyssynchrony indices are coefficient of variation of time to maximum upslope (C-TTMU) 22%; variance of time to maximum upslope (V-TTMU) 2.3 s2; coefficient of variation of time to peak signal (C-TTP) 8.8%; variance of time to peak signal (V-TTP) 1 s2. All temporal intervals were calculated starting from the initial upslope of the LV arterial input function (AIF) visually identified by the operator performing the analysis from the basal LV slice. The range of data analysed was set by default to 20 beats from the AIF upslope but could be manually reduced by the operator to avoid respiratory motion occurring towards the end of the acquisition. Perfusion dyssynchrony analysis was repeated twice by the same operator and by a different blinded operator to measure the intra- and inter-observer variabilities. Quantitative perfusion analysis Quantitative perfusion analysis was performed using Fermi deconvolution according to the methods described by Wilke14 and Jerosch-Herold15 using in-house software previously validated against positron emission tomography,16 FFR,17 microspheres,18 and hardware perfusion phantom data.19,20 Myocardial perfusion reserve (MPR) was defined as the ratio between stress and rest perfusion estimates obtained in each coronary perfusion territory.

nd Jerosch-Herold15 using in-house software previously validated against positron emission tomography,16 FFR,17 microspheres,18 and hardware perfusion phantom data.19,20 Myocardial perfusion reserve (MPR) was defined as the ratio between stress and rest perfusion estimates obtained in each coronary perfusion territory. Catheter laboratory protocol Invasive coronary angiography was performed with standard methods.21 FFR was measured in all vessels that showed visually a >50% diameter stenosis in two orthogonal views during intracoronary adenosine-induced hyperaemia (140 μg/kg/min) with a 0.014-inch coronary pressure sensor–tipped wire (Volcano Therapeutics, San Diego, CA or St Jude Medical, St Paul, MN, USA).22

standard methods.21 FFR was measured in all vessels that showed visually a >50% diameter stenosis in two orthogonal views during intracoronary adenosine-induced hyperaemia (140 μg/kg/min) with a 0.014-inch coronary pressure sensor–tipped wire (Volcano Therapeutics, San Diego, CA or St Jude Medical, St Paul, MN, USA).22 Statistical analysis The Medcalc software (Medcalc, Belgium) and Analyse-it software (Analyse-it Software Limited, United Kingdom) were used. Data are presented as mean ± standard deviation. Intra- and inter-observer reproducibilities were determined by Bland–Altman plots and regression analysis. The correlation between perfusion dyssynchrony indices and other quantitative or semi-quantitative CMR parameters was assessed using regression analysis. ANOVA and t-tests were used for comparison of results as appropriate. Receiver operating characteristic (ROC) analysis determined the accuracy of each perfusion dyssynchrony index in the diagnosis of CAD and in differentiating between patients with single-vessel and multi-vessel CAD. Optimal diagnostic cut-offs were determined by the best sum of sensitivity and specificity. Mann–Whitney and χ2 tests were used to test the qualitative measurements for statistical significance. No formal power analysis was carried out.23

and in differentiating between patients with single-vessel and multi-vessel CAD. Optimal diagnostic cut-offs were determined by the best sum of sensitivity and specificity. Mann–Whitney and χ2 tests were used to test the qualitative measurements for statistical significance. No formal power analysis was carried out.23 Results A total of 98 subjects were included in the analysis: 35 subjects in the Normal Group, 32 patients in the Single-vessel Group, and 31 patients in the Multi-vessel Group. Baseline data and demographics are shown in Table 1. The FFR results are shown in Table 2. Table 1 Demographics and risk factors for CAD All vessels (n = 98) Normal group (n = 35) Single-vessel group (n = 32) Multi-vessel group (n = 31) Male gender 73 (74%) 21 (60%) 25 (78%) 27 (87%) Age 60 ± 9 59 ± 10 60 ± 7 62 ± 7 Hypertension 51 (52%) 16 (46%) 19 (59%) 16 (52%) Dyslipidaemia 65 (66%) 17 (49%) 24 (75%) 24 (77%) Diabetes 18 (18%) 4 (11%) 9 (28%) 5 (16%) Current Smoker 15 (15%) 2 (6%) 10 (31%) 3 (10%) Previous PCI 12 (12%) 0 (0%) 9 (28%) 3 (10%) Family history of CAD 28 (29%) 4 (11%) 13 (41%) 11 (35%) CAD, coronary artery disease; PCI, percutaneous coronary intervention. Table 2 FFR results

All vessels (n = 98) Normal group (n = 35) Single-vessel group (n = 32) Multi-vessel group (n = 31) Male gender 73 (74%) 21 (60%) 25 (78%) 27 (87%) Age 60 ± 9 59 ± 10 60 ± 7 62 ± 7 Hypertension 51 (52%) 16 (46%) 19 (59%) 16 (52%) Dyslipidaemia 65 (66%) 17 (49%) 24 (75%) 24 (77%) Diabetes 18 (18%) 4 (11%) 9 (28%) 5 (16%) Current Smoker 15 (15%) 2 (6%) 10 (31%) 3 (10%) Previous PCI 12 (12%) 0 (0%) 9 (28%) 3 (10%) Family history of CAD 28 (29%) 4 (11%) 13 (41%) 11 (35%) CAD, coronary artery disease; PCI, percutaneous coronary intervention. Table 2 FFR results All vessels Normal group Single-vessel group Multi-vessel group Vessels FFR measured 119/294 (40%) 6/105 (6%) 47/96 (49%) 66/93 (71%) Vessels with FFR >0.8 25/294 (9%) 6/105 (6%) 17/96 (18%) 2/93 (2%) Vessels with FFR <0.8 94/294 (32%) 0/105 (0%) 30/96 (31%) 64/93 (69%) FFR negative vessels 0.89 ± 0.06 0.91 ± 0.05 0.88 ± 0.07 0.92 ± 0.09 FFR positive vessels 0.60 ± 0.14 — 0.63 ± 0.14 0.59 ± 0.16 FFR, fractional flow reserve. Perfusion CMR Detailed functional CMR and LGE findings as well as the detailed results of visual assessment and quantitative perfusion analysis are shown in Table 3. A total of 294 perfusion territories were included in the analysis. The number of segments visually positive for stress-induced abnormalities was 2.3 ± 2.5 in the Single-vessel and 8.1 ± 3.6 in the Multi-vessel group (P < 0.001). There was a significant difference between MPR values in FFR positive and negative perfusion territories (P < 0.0001 for all comparisons). Table 3 CMR findings

he number of segments visually positive for stress-induced abnormalities was 2.3 ± 2.5 in the Single-vessel and 8.1 ± 3.6 in the Multi-vessel group (P < 0.001). There was a significant difference between MPR values in FFR positive and negative perfusion territories (P < 0.0001 for all comparisons). Table 3 CMR findings All subjects (n = 98) Normal group (n = 35) Single-vessel group (n = 32) Multi-vessel group (n = 31) LV EF (%) 59 ± 6 60 ± 5 59 ± 5 56 ± 7 LV EDV (mL/m2) 80 ± 11 74 ± 4 76 ± 6 87 ± 15 LV ESV (mL/m2) 35 ± 12 31 ± 5 30 ± 6 38 ± 12 RV EF (%) 55 ± 5 56 ± 5 54 ± 6 56 ± 7 RV EDV (mL/m2) 86 ± 10 80 ± 3 90 ± 7 88 ± 16 RV ESV (mL/m2) 36 ± 8 34 ± 7 41 ± 8 36 ± 9 LA (cm2) 25 ± 5 22 ± 2 24 ± 3 26 ± 4 RA (cm2) 22 ± 4 20 ± 3 23 ± 3 24 ± 5 Visual perfusion positive segments (average ± SD per patient) 5.0 ± 4.9 — 2.3 ± 2.5 8.1 ± 3.6 LGE positive segments (average ± SD per patient) 0.5 ± 1.4 — 0.8 ± 1.6 0.9 ± 1.7 MPR (all territories) 2.62 ± 1.1 2.9 ± 1.1 2.8 ± 1.1 2.2 ± 0.8 MPR (territories with FFR <0.8) 1.9 ± 0.6 — 1.9 ± 0.7 1.9 ± 0.6 MPR (territories with FFR >0.8) 2.9 ± 1.1 2.9 ± 1.1 3.0 ± 1.0 2.8 ± 0.8 CAD, coronary artery disease; CMD, coronary microvascular disease; LV, left ventricle; EF ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume; MPR, myocardial perfusion reserve; RV, right ventricle; LA, left atrium; RA, right atrium; LGE, late gadolinium enhancement; SD, standard deviation; FFR, fractional flow reserve.

disease; CMD, coronary microvascular disease; LV, left ventricle; EF ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume; MPR, myocardial perfusion reserve; RV, right ventricle; LA, left atrium; RA, right atrium; LGE, late gadolinium enhancement; SD, standard deviation; FFR, fractional flow reserve. Perfusion dyssynchrony analysis Detailed results of perfusion dyssynchrony analysis are shown in Table 4. All tested perfusion dyssynchrony indices increased significantly during stress in comparison with rest values, with the exception of the Normal group where perfusion dyssynchrony during stress did not differ from rest values. Moreover, perfusion dyssynchrony values increased proportionally to the extent of haemodynamically significant CAD, with more severe dyssynchrony being induced by adenosine stress in patients with multi-vessel disease. These results were particularly significant when C-TTMU or TTP-derived indices were used. In contrast to perfusion dyssynchrony indices, however, the average TTMU and TTP did not differ between groups and between stress and rest (Table 4). Pearson's analysis showed no correlation between average TTMU and TTMU-derived perfusion dyssynchrony indices (R = 0.19; R2 = 0.038 vs. V-TTMU; R = 0.31; R2 = 0.11 vs. C-TTMU). Similarly, no correlation was found between average TTP values and V-TTP or C-TTP (R = 0.02; R2 = 0.001 vs. V-TTP; R = 0.33; R2 = 0.1 vs. C-TTP). Table 4 Comparison between stress and rest values of TTMU, TTP, and perfusion dyssynchrony indices

indices (R = 0.19; R2 = 0.038 vs. V-TTMU; R = 0.31; R2 = 0.11 vs. C-TTMU). Similarly, no correlation was found between average TTP values and V-TTP or C-TTP (R = 0.02; R2 = 0.001 vs. V-TTP; R = 0.33; R2 = 0.1 vs. C-TTP). Table 4 Comparison between stress and rest values of TTMU, TTP, and perfusion dyssynchrony indices Stress Rest P (stress vs. rest) V-TTMU Normal group 2.7 ± 2.4 s2 3.2 ± 3.1 s2 0.71 Single-vessel group 3.7 ± 4.0 s2 2.8 ± 1.7 s2 0.30 Multi-vessel group 4.8 ± 2.8 s2 2.5 ± 1.7 s2 <0.001 P (between groups) 0.03 0.56 C-TTMU Normal group 22 ± 10.1% 24.2 ± 12.7% 0.15 Single-vessel group 27.7 ± 11.2% 22.3 ± 7.3% 0.02 Multi-vessel group 34.1 ± 12.8% 23.4 ± 15.6% 0.003 P (between groups) 0.0002 0.89 V-TTP Normal group 1.5 ± 1.1 s2 2.2 ± 1.7 s2 0.17 Single-vessel group 3.7 ± 3.3 s2 2.1 ± 1.1 s2 0.01 Multi-vessel group 8.8 ± 6.9 s2 2.4 ± 1.3 s2 <0.0001 P (between groups) <0.0001 0.07 C-TTP Normal group 8.1 ± 2.9% 8.5 ± 4.3% 0.36 Single-vessel group 14.6 ± 5.2% 10 ± 3.6% <0.0001 Multi-vessel group 21.7 ± 12.6% 9.6 ± 3.7% <0.0001 P (between groups) <0.0001 0.40 Average TTMU Normal group 5.6 ± 1.9 s 7.1 ± 3.4 s 0.54 Single-vessel group 6.4 ± 2.7 s 6.5 ± 2.1 s 0.70 Multi-vessel group 6.2 ± 1.9 s 6.9 ± 1.7 s 0.13 P (between groups) 0.24 0.67 Average TTP Normal group 14.4 ± 5.1 s 15.2 ± 4.3 s 0.92 Single-vessel group 12.5 ± 3.4 s 12.5 ± 2.4 s 0.95 Multi-vessel group 13.4 ± 3.3 s 14.7 ± 2.3 s 0.06 P (between groups) 0.17 0.67 C-TTMU, coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity.

f variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity. On ROC analysis, perfusion dyssynchrony allowed discrimination between normal subjects and patients with CAD (Table 5). TTP-derived indices performed better than TTMU-derived indices. The most accurate parameter for the diagnosis of CAD was C-TTP, with a sensitivity of 0.89, specificity of 0.86, and area under the ROC curve of 0.94. The best C-TTP cut-off for the diagnosis of CAD was >10%. V-TTMU, C-TTMU, and C-TTP were more accurate than visual assessment for the diagnosis of CAD (P = 0.0002, P = 0.017, and P = 0.049, respectively). V-TTMU and C-TTP were more accurate than quantitative analysis for the diagnosis of CAD (P = 0.004 and P = 0.04, respectively). Table 5 ROC analysis for the prediction of CAD

-TTMU, C-TTMU, and C-TTP were more accurate than visual assessment for the diagnosis of CAD (P = 0.0002, P = 0.017, and P = 0.049, respectively). V-TTMU and C-TTP were more accurate than quantitative analysis for the diagnosis of CAD (P = 0.004 and P = 0.04, respectively). Table 5 ROC analysis for the prediction of CAD Area under ROC curve 95% CI SE Z P Best cut-off Sensitivity 95% CI Specificity 95% CI TTMU 0.39 0.27–0.51 0.060 −1.8 0.9639 — — — — — V-TTMU 0.63 0.52–0.75 0.059 2.27 0.0117 1.9 s2 0.726 0.598–0.831 0.486 0.314–0.660 C-TTMU 0.72 0.61–0.83 0.056 3.96 <0.0001 23% 0.730 0.603–0.834 0.629 0.449–0.785 TTP 0.4 0.28–0.52 0.063 −1.56 0.9410 — — — — — V-TTP 0.88 0.81–0.94 0.034 11.15 <0.0001 2.2 s2 0.825 0.709–0.909 0.857 0.697–0.952 C-TTP 0.94 0.9–0.98 0.022 20.02 <0.0001 10% 0.889 0.784–0.954 0.857 0.697–0.952 Visual assessment 0.87 0.82–0.93 0.028 13.50 <0.0001 ≥1 positive perfusion territory 0.746 0.621–0.847 1 0.900–1.000 Quantitative analysis 0.84 0.76–0.93 0.044 7.77 <0.0001 1.8 0.84 0.727–0.921 0.714 0.537–0.854 C-TTMU, coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity.

f variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity. All perfusion dyssynchrony indices allowed identification of multi-vessel disease (Table 6). The most accurate parameter was V-TTP, with a sensitivity of 0.74, specificity of 0.79, and area under the ROC curve of 0.84. The best V-TTP diagnostic cut-off for multi-vessel CAD was >3.3 s2. V-TTMU and C-TTMU were more accurate than visual assessment for the diagnosis of multi-vessel disease (P = 0.03 for both). Table 6 ROC analysis for prediction of multi-vessel CAD

ivity of 0.74, specificity of 0.79, and area under the ROC curve of 0.84. The best V-TTP diagnostic cut-off for multi-vessel CAD was >3.3 s2. V-TTMU and C-TTMU were more accurate than visual assessment for the diagnosis of multi-vessel disease (P = 0.03 for both). Table 6 ROC analysis for prediction of multi-vessel CAD Area under ROC curve 95% CI SE Z P Best cut-off Sensitivity 95% CI Specificity 95% CI TTMU 0.48 0.36–0.61 0.063 −0.26 0.6044 — — — — — V-TTMU 0.68 0.57–0.80 0.058 3.15 0.0008 2.8 s2 0.774 0.589–0.904 0.621 0.493–0.738 C-TTMU 0.69 0.58–0.80 0.056 3.43 0.0003 33% 0.516 0.331–0.698 0.776 0.658–0.869 TTP 0.49 0.37–0.61 0.061 −0.23 0.5914 — — — — — V-TTP 0.84 0.77–0.92 0.039 8.75 <0.0001 3.3 s2 0.742 0.554–0.881 0.791 0.674–0.881 C-TTP 0.81 0.72–0.89 0.043 7.09 <0.0001 12% 0.806 0.625–0.925 0.657 0.531–0.768 Visual assessment 0.84 0.76–0.92 0.043 7.95 <0.0001 ≥2 positive perfusion territories 0.710 0.520–0.858 0.970 0.896–0.966 Quantitative analysis 0.79 0.70–0.87 0.045 6.40 <0.0001 1.6 0.903 0.742–0.980 0.657 0.531–0.768 C-TTMU, coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity. Results of average TTMU and TTP were not significant in the prediction of CAD or multi-vessel CAD.

Area under ROC curve 95% CI SE Z P Best cut-off Sensitivity 95% CI Specificity 95% CI TTMU 0.48 0.36–0.61 0.063 −0.26 0.6044 — — — — — V-TTMU 0.68 0.57–0.80 0.058 3.15 0.0008 2.8 s2 0.774 0.589–0.904 0.621 0.493–0.738 C-TTMU 0.69 0.58–0.80 0.056 3.43 0.0003 33% 0.516 0.331–0.698 0.776 0.658–0.869 TTP 0.49 0.37–0.61 0.061 −0.23 0.5914 — — — — — V-TTP 0.84 0.77–0.92 0.039 8.75 <0.0001 3.3 s2 0.742 0.554–0.881 0.791 0.674–0.881 C-TTP 0.81 0.72–0.89 0.043 7.09 <0.0001 12% 0.806 0.625–0.925 0.657 0.531–0.768 Visual assessment 0.84 0.76–0.92 0.043 7.95 <0.0001 ≥2 positive perfusion territories 0.710 0.520–0.858 0.970 0.896–0.966 Quantitative analysis 0.79 0.70–0.87 0.045 6.40 <0.0001 1.6 0.903 0.742–0.980 0.657 0.531–0.768 C-TTMU, coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity. Results of average TTMU and TTP were not significant in the prediction of CAD or multi-vessel CAD. The results of correlation analysis between perfusion dyssynchrony indices, FFR values, visually positive segments, and severity of ischaemia measured as MPR values are shown in Table 7. The correlation between perfusion dyssynchrony results and severity of ischaemia was weak, with TTP-derived indices again performing better. Table 7 Results of Pearson's correlation analysis between perfusion dyssynchrony indices and FFR values, number of visually positive segments and MPR

values are shown in Table 7. The correlation between perfusion dyssynchrony results and severity of ischaemia was weak, with TTP-derived indices again performing better. Table 7 Results of Pearson's correlation analysis between perfusion dyssynchrony indices and FFR values, number of visually positive segments and MPR FFR Visual analysisa MPR V-TTMU R2 = 0.02 P = 0.282 R2 = 0.04 P = 0.047 R2 = 0.04 P = 0.038 C-TTMU R2 = 0.04 P = 0.093 R2 = 0.05 P = 0.017 R2 = 0.06 P = 0.012 V-TTP R2 = 0.12 P = 0.004 R2 = 0.13 P < 0.001 R2 = 0.09 P = 0.003 C-TTP R2 = 0.10 P = 0.008 R2 = 0.13 P < 0.001 R2 = 0.05 P = 0.029 C-TTMU, coefficient of variation of the time to maximum upslope of the myocardial signal intensity curve; C-TTP, coefficient of variation of the time to peak myocardial signal intensity; V-TTMU, variance of the time to maximum upslope of the myocardial signal intensity curve; V-TTP, variance of the time to peak myocardial signal intensity. aNumber of positive segments. Bland–Altman graphs and Pearson's r analysis for inter- and intra-observer variabilities are shown in Figures 3 and 4. C-TTP and C-TTMU were the most reproducible perfusion dyssynchrony indices for intra-operator and inter-operator variabilities. Figure 3 Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Bland–Altman graphs for each parameter. Figure 4 Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Pearson's r analysis for each parameter.

Bland–Altman graphs and Pearson's r analysis for inter- and intra-observer variabilities are shown in Figures 3 and 4. C-TTP and C-TTMU were the most reproducible perfusion dyssynchrony indices for intra-operator and inter-operator variabilities. Figure 3 Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Bland–Altman graphs for each parameter. Figure 4 Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Pearson's r analysis for each parameter. Results of overall image qualitative assessment, respiratory artefacts, and dark rim artefacts are presented in Figure 5. No significant differences were observed between groups. The average angular extent of dark rim artefact was 35° (range 8–45°). Figure 5 Results of image quality analysis. No significant differences were observed among groups for image quality (A), respiratory motion (B), and dark rim artefacts (C). Discussion Perfusion dyssynchrony analysis introduces a novel pathophysiological concept of temporal heterogeneity for the analysis of perfusion CMR data. In this study, we tested the potential of perfusion dyssynchrony analysis to be used as a tool for the detection of CAD and to identify patients with multi-vessel disease.

sion Perfusion dyssynchrony analysis introduces a novel pathophysiological concept of temporal heterogeneity for the analysis of perfusion CMR data. In this study, we tested the potential of perfusion dyssynchrony analysis to be used as a tool for the detection of CAD and to identify patients with multi-vessel disease. The main findings of this study are as follows: (i) CAD is associated with temporal dyssynchrony of first-pass perfusion signals, which is proportional to the number of vessels with haemodynamically relevant stenosis as assessed by FFR; (ii) perfusion dyssynchrony is induced by stress; (iii) indices of perfusion dyssynchrony can reliably detect the presence of CAD, and (iv) multi-vessel CAD. Visual assessment of stress and rest perfusion scans is based on the identification of areas of reduced and delayed wash-in of contrast agent. Quantitative analysis instead measures the absolute reduction myocardial signal intensity without generally accounting for temporal delay, which can potentially be a source of errors,24 particularly when voxel-wise techniques are used. We have previously described and validated automated algorithms for correction of the temporal delay for voxel-wise quantification.25 However, the observed association between heterogeneous temporal delay and myocardial ischaemia leads us to hypothesize that specific indices of temporal dyssynchrony could be developed and used to detect areas of abnormal perfusion.25 To our knowledge, the assessment of myocardial perfusion in term of its temporal component has not yet been explored.

between heterogeneous temporal delay and myocardial ischaemia leads us to hypothesize that specific indices of temporal dyssynchrony could be developed and used to detect areas of abnormal perfusion.25 To our knowledge, the assessment of myocardial perfusion in term of its temporal component has not yet been explored. Diastolic blood flow in unobstructed epicardial coronary arteries is very fast, with the dead volume in the epicardial coronaries being replenished several times every heartbeat with fresh blood inflowing from the aorta.26 Normal myocardium has preserved vasodilatory reserve and shows a temporally homogeneous perfusion (the myocardium is perfused with uniform amount of blood and wash-in happens at approximately the same time in all segments).25 The presence of flow-limiting CAD can deeply influence the propagation of the contrast agent through the coronary circulation, with reduced amplitude and a temporal spread of the signal intensity curves in different coronary territories. Our results demonstrate for the first time that the degree of temporal dyssynchrony correlates with the extent of haemodynamically significant CAD and is maximal in patients with multi-vessel disease. We hypothesized that perfusion dyssynchrony is generated as a result of the interaction between several factors, including the site and severity of the epicardial lesions and the relationship between coronary resistance and down-stream coronary capacitance (Figure 6). Figure 6 Proposed pathophysiology of myocardial perfusion dyssynchrony. In normal hearts (left), myocardial perfusion shows a temporally homogeneous perfusion. The presence of flow-limiting CAD influences the propagation of the contrast agent through the coronary circulation. Temporal dyssynchrony is caused by the interaction between flow-limiting stenoses and down-stream coronary capacitance. This effect is proportional to the number of flow-limiting coronary lesions. LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; LM, left main coronary artery; RCA, right coronary artery.

y is caused by the interaction between flow-limiting stenoses and down-stream coronary capacitance. This effect is proportional to the number of flow-limiting coronary lesions. LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; LM, left main coronary artery; RCA, right coronary artery. Four indices of perfusion dyssynchrony were evaluated in this study, including the variance of the time to maximum myocardial signal upslope (V-TTMU), the coefficient of variation of the time to maximum myocardial signal upslope (C-TTMU), the variance of the time to peak myocardial signal (V-TTP), and the coefficient of variation of the time to peak myocardial signal (C-TTP). Our results show a progressive increase of perfusion dyssynchrony proportionally to the number of diseased vessels. Importantly, while perfusion dyssynchrony values showed this correlation, the average values of TTMU and TTP did not correlate with the severity of the disease. Moreover, perfusion dyssynchrony indices showed weak correlation with other invasive and non-invasive measurements of the severity of ischaemia, including FFR values (considered as a continuous rather than dichotomous variable), visual ischaemic burden, and MPR values. All dyssynchrony indices were capable of detecting single- and multi-vessel CAD. However, TTP-derived indices (V-TTP and C-TTP) performed better than TTMU-derived indices. This might reflect the relatively higher signal-to-noise ratio of TTP measures. V-TTP and C-TTP were also the most reproducible indices on both inter- and intra-observer variabilities.

e of detecting single- and multi-vessel CAD. However, TTP-derived indices (V-TTP and C-TTP) performed better than TTMU-derived indices. This might reflect the relatively higher signal-to-noise ratio of TTP measures. V-TTP and C-TTP were also the most reproducible indices on both inter- and intra-observer variabilities. We have previously described and validated transmural perfusion gradient (TPG) analysis against FFR, an another tool to detect stress-induced perfusion abnormalities on high-resolution CMR scans.13,27 TPG analysis was designed to take advantage of the high-spatial resolution of CMR, exploiting the differences in first-pass perfusion observed between the inner (sub-endocardial) layers of the LV wall and the outer (sub-epicardial) layers. TPG identifies areas with inducible perfusion abnormalities based on the transmural redistribution of signal during first pass. In principle, TPG analysis is very similar to visual assessment, taking into consideration a combination of amplitude and temporal delay of signal intensity curves at segmental level. Conversely, perfusion dyssynchrony analysis differs from TPG analysis as it does not take into account the amplitude of the signal intensity curves and was rather designed to isolate and measure the temporal dyssynchrony of the wash-in curves to produce one value for the entire left ventricle.

urves at segmental level. Conversely, perfusion dyssynchrony analysis differs from TPG analysis as it does not take into account the amplitude of the signal intensity curves and was rather designed to isolate and measure the temporal dyssynchrony of the wash-in curves to produce one value for the entire left ventricle. Perfusion dyssynchrony analysis has several potential advantages over other post-processing techniques. It is based on regional differences in the TTMU and TTP rather than on absolute signal intensity values, making it very robust to signal inhomogeneities as well as different data acquisition schemes. In addition, unlike quantitative analysis, perfusion dyssynchrony analysis is not influenced by the complex relationship between contrast agent concentration and signal intensity. This is reflected by low inter- and intra-observer variabilities. We could find in the literature only very limited attempts to discuss the variability of stress and rest perfusion values in normal subjects by using deconvolution-based quantification algorithms. These results explored the variability of absolute perfusion values rather than the temporal dyssynchrony of the signals.28

We could find in the literature only very limited attempts to discuss the variability of stress and rest perfusion values in normal subjects by using deconvolution-based quantification algorithms. These results explored the variability of absolute perfusion values rather than the temporal dyssynchrony of the signals.28 Limitations This study included highly selected populations of patients with suspected CAD. Thus, the data on diagnostic accuracy reflect the accuracy to discriminate between these specific groups rather than a general population. Perfusion dyssynchrony analysis will need to be tested on different scanners, field strength, and with different acquisition protocols. Moreover, patients were included in the study retrospectively, and this constitutes an additional limitation. The presence of coronary microvascular disease or of collateral coronary vessels could in theory affect perfusion dyssynchrony measurements. Conclusions In conclusion, perfusion dyssynchrony analysis is a novel method to measure temporal differences of myocardial perfusion and appears to be highly accurate in identifying patients with haemodynamically significant CAD. In particular, TTP-derived indices showed high accuracy and excellent reproducibility and might represent an additional useful tool in the non-invasive assessment of myocardial ischaemia. Funding Funding to pay the Open Access publication charges for this article was provided by the Engineering and Physical Sciences Research Council (EPSCR).

Conclusions In conclusion, perfusion dyssynchrony analysis is a novel method to measure temporal differences of myocardial perfusion and appears to be highly accurate in identifying patients with haemodynamically significant CAD. In particular, TTP-derived indices showed high accuracy and excellent reproducibility and might represent an additional useful tool in the non-invasive assessment of myocardial ischaemia. Funding Funding to pay the Open Access publication charges for this article was provided by the Engineering and Physical Sciences Research Council (EPSCR). Acknowledgements The authors acknowledge financial support from the Department of Health via the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre award to Guy's & St Thomas’ NHS Foundation Trust in partnership with King's College London and King's College Hospital NHS Foundation Trust. The Centre of Excellence in Medical Engineering funded by the Wellcome Trust and EPSRC under grant number WT 088641/Z/09/Z. King's College London and UCL Comprehensive Cancer Imaging Centre. Funded by CRUK and EPSRC in association with MRC and DoH (England). Funded by the British Heart Foundation award RE/08/003. Conflict of interest: M.B. is an employee of Philips Healthcare.

A 52-year-old woman with atopic asthma and a long smoking habit underwent CT pulmonary angiography (CTPA) for investigation of atypical, pleuritic chest pain. There was no history of heart disease or cardiovascular risk factors. As per usual practice for non-cardiac imaging, CTPA was performed without ECG gating. While there was no pulmonary embolism, CT showed an apparent large left atrial filling defect (arrows), demonstrated in axial (Panel A) and sagittal (Panel B) views. This CT abnormality raised the suspicion of left atrial thrombus or tumour, prompting further cardiac imaging. Transoesophageal echocardiography, shown in short-axis (Panel C; see Supplementary data online, Movie S1) and four-chamber (Panel D; see Supplementary data online, Movie S2) views, showed the left atria to be of normal size and devoid of any visible mass. There was, however, an atrial septum aneurysm, which exhibited 2 cm shift from midline (arrows). Bi-ventricular size and function were normal, and there was no intra-atrial shunt seen on colour Doppler. No medical or surgical intervention was required for this incidental radiographic finding.

evoid of any visible mass. There was, however, an atrial septum aneurysm, which exhibited 2 cm shift from midline (arrows). Bi-ventricular size and function were normal, and there was no intra-atrial shunt seen on colour Doppler. No medical or surgical intervention was required for this incidental radiographic finding. This report highlights a potential hazard of imaging the heart with non-ECG-gated CT, which is common practice when non-cardiac chest pain is suspected. Owing to aneurysmal septal motion, an unusual ‘ghost’ artifact appeared on this patient's CTPA, mimicking a left atrial mass. Diagnostic uncertainly led to unnecessary invasive testing. This case provokes the question: should ECG gating be added routinely when CTPA is performed for undifferentiated chest pain? (Ao, aorta; LA, left atrium; LV, left ventricle; PA, pulmonary artery; RA, right atrium). J.M.T. is supported by a Wellcome Trust research training fellowship (104492/Z/14/Z). J.H.F.R. is part supported by HEFCE. M.R.B. and J.H.F.R are supported by the NIHR Cambridge Biomedical Research Centre and the British Heart Foundation. Supplementary data are available at European Heart Journal— Cardiovascular Imaging online.

Introduction Unrecognized myocardial infarcts (UMIs) have been described in 19–44% of the general population, with prevalence increasing by 10% per decade and their presence associated with a similar or worse prognosis than recognized myocardial infarcts (RMIs).1,2 The majority of population-based studies have focused on the use of electrocardiogram (ECG) for the detection of UMI; however, not all infarcts result in pathological Q waves.3 Late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR) imaging has become the clinical gold standard for the detection of myocardial scarring, with a significantly higher detection rate of UMIs than using ECG alone.4 These UMIs detected on CMR have significant clinical implications, with those with evidence of myocardial scarring in the absence of clinically apparent prior infarct more likely to have chest pain and poorer left ventricular (LV) function and suffer from a greater number of major adverse cardiovascular events.4,5

ne.4 These UMIs detected on CMR have significant clinical implications, with those with evidence of myocardial scarring in the absence of clinically apparent prior infarct more likely to have chest pain and poorer left ventricular (LV) function and suffer from a greater number of major adverse cardiovascular events.4,5 Earlier studies have described the prevalence of UMIs in elderly populations at high risk of cardiovascular disease (CVD). These studies, by including those with known CVD, have thereby conflated the prevalence of UMIs in the general population.4,6 The prevalence of UMIs in a population without prior cardiovascular events and who are not at high risk has not been previously undertaken. Furthermore, a prior study of 75-year-olds has suggested that UMIs may not be associated with traditional CVD risk factors.7 Thus, the identification of a cohort considered as low or intermediate risk that have suffered from UMIs may provide insight into novel predisposing aetiological factors. The aim of this study was to investigate the prevalence of UMIs in a large, non-high-risk, asymptomatic cohort, assessed with magnetic resonance imaging (MRI), and the association between UMIs and risk markers for CVD.

Earlier studies have described the prevalence of UMIs in elderly populations at high risk of cardiovascular disease (CVD). These studies, by including those with known CVD, have thereby conflated the prevalence of UMIs in the general population.4,6 The prevalence of UMIs in a population without prior cardiovascular events and who are not at high risk has not been previously undertaken. Furthermore, a prior study of 75-year-olds has suggested that UMIs may not be associated with traditional CVD risk factors.7 Thus, the identification of a cohort considered as low or intermediate risk that have suffered from UMIs may provide insight into novel predisposing aetiological factors. The aim of this study was to investigate the prevalence of UMIs in a large, non-high-risk, asymptomatic cohort, assessed with magnetic resonance imaging (MRI), and the association between UMIs and risk markers for CVD. Methods Study population Following local ethics committee approval, a cohort of n = 1651 volunteers were recruited to the imaging arm of the Tayside Screening for Cardiovascular Events (TASCFORCE) study. Volunteers were recruited to the study via general practitioner (GP) surgeries, advertising on radio and via leaflets distributed at local public events and via large local employers. They were enrolled into TASCFORCE if they met the inclusion criteria that they (i) were above the age of 40 years, (ii) were free from CVD or other indication for statin therapy as recommended by the Scottish Intercollegiate Guidelines Network (SIGN) report 97 (www.sign.ac.uk) for ‘Risk Estimation and the Prevention of Cardiovascular Disease’ published in February 2007, and (iii) had a 10-year risk of coronary heart disease <20% as predicted by the Adult Treatment Panel III (ATP-III) algorithm.8 Exclusion criteria included the following: (i) pregnancy; (ii) known primary muscle disease; (iii) known atherosclerotic disease—including angina, previous myocardial infarction, peripheral arterial disease, amputation, previous revascularization surgery, hypertension, heart failure, or cerebrovascular event; (iv) known diabetes; (v) active liver disease; (vi) other known illness or contraindication to MRI; (vii) participation in a clinical trial; (viii) inability to give informed consent; (ix) known alcohol abuse; and (x) a blood pressure of >145/95mmHg. Participants who had a serum B-type natriuretic peptide (BNP) level greater than their gender-specific median (7.5 pg/mL in men, 15.3 pg/mL in women), indicating possible non-specific stress on the cardiovascular system, were invited for whole-body MRI angiography. Except for 14 men and 2 women, all participants had a BNP level of <100 pg/mL, which is the threshold used for the diagnosis of heart failure, and therefore the vast majority of participants had a BNP in a ‘normal’ range. Of 1651 volunteers, 122 were excluded because of claustrophobia or MR safety concerns, with 1529 (91.4%; 931 females and 577 males, mean age 54 years, range 40–83 years) completing their imaging.

hold used for the diagnosis of heart failure, and therefore the vast majority of participants had a BNP in a ‘normal’ range. Of 1651 volunteers, 122 were excluded because of claustrophobia or MR safety concerns, with 1529 (91.4%; 931 females and 577 males, mean age 54 years, range 40–83 years) completing their imaging. Image acquisition The MRI protocol has been described in detail elsewhere,9 but in brief, imaging was performed using a 3 T Magnetom Trio Scanner (Siemens, Erlangen, Germany). Whole-body MR angiography was performed using a dual bolus injection technique with the CMR CINEs performed before the first contrast injection and the LGE sequences performed between the first and second contrast bolus injections. For CMR, a body matrix radiofrequency coil (6 elements) was used in combination with a spine array (up to 24 elements).

erformed using a dual bolus injection technique with the CMR CINEs performed before the first contrast injection and the LGE sequences performed between the first and second contrast bolus injections. For CMR, a body matrix radiofrequency coil (6 elements) was used in combination with a spine array (up to 24 elements). ECG-gated, segmented, breath-hold, cinematic (CINE) TrueFISP images were acquired from the atrioventricular ring to the LV apex using 2D ECG-gated, breath-hold, segmented (CINE) TrueFISP sequence with retrospective gating. Retrospective ECG gating was used, with 25 cardiac phases reconstructed (25 lines per segment) and 2 image slices acquired per breath-hold. Parallel imaging was also implemented [integrated parallel acquisition technique (iPAT) ×2], resulting in a scan time of <15 s per breath-hold. An ECG-gated, segmented, breath-hold, CINE 2D turbo Inversion Recovery ‘TI-scout’ sequence was implemented (in a central short-axis orientation) 8–10 min after the injection of 10 mL of 0.5 mmol/mL gadoteric acid. At a median of 11 min post-contrast (range 9–16 min), a short-axis stack of ECG-gated, segmented, 2D phase-sensitive inversion recovery images were acquired.10 Image analysis LV mass and volume and whole-body atheroma burden quantification were performed as previously described.9

ECG-gated, segmented, breath-hold, cinematic (CINE) TrueFISP images were acquired from the atrioventricular ring to the LV apex using 2D ECG-gated, breath-hold, segmented (CINE) TrueFISP sequence with retrospective gating. Retrospective ECG gating was used, with 25 cardiac phases reconstructed (25 lines per segment) and 2 image slices acquired per breath-hold. Parallel imaging was also implemented [integrated parallel acquisition technique (iPAT) ×2], resulting in a scan time of <15 s per breath-hold. An ECG-gated, segmented, breath-hold, CINE 2D turbo Inversion Recovery ‘TI-scout’ sequence was implemented (in a central short-axis orientation) 8–10 min after the injection of 10 mL of 0.5 mmol/mL gadoteric acid. At a median of 11 min post-contrast (range 9–16 min), a short-axis stack of ECG-gated, segmented, 2D phase-sensitive inversion recovery images were acquired.10 Image analysis LV mass and volume and whole-body atheroma burden quantification were performed as previously described.9 Analysis of the LGE sequences was performed offline on a diagnostic PACS radiological workstation (Kodak Carestream PACS Client Suite Version 10.1 sp1, Rochester, NY, USA) by one of the two experienced observers [one with 5 years (J.R.W.M.) and one with 15 years of experience (J.G.H.)] independently, with each scan being recorded as either scar-positive or scar-negative. All positive scans were then classified by consensus opinion with LGE classified as being sub-endocardial, mid-myocardial, epicardial, or other. Thickness was scored as <50%, >50%, or transmural. Sub-endocardial and transmural scarring were classed as UMI. The location and extent were scored using the American Heart Association (AHA) 17 segment model.11

fied by consensus opinion with LGE classified as being sub-endocardial, mid-myocardial, epicardial, or other. Thickness was scored as <50%, >50%, or transmural. Sub-endocardial and transmural scarring were classed as UMI. The location and extent were scored using the American Heart Association (AHA) 17 segment model.11 The whole-body MRA analysis was performed as previously described.9 In brief, the arterial tree was divided into 31 segments with each scored according to the degree of narrowing of the lumen diameter, with stenosis graded at the narrowest part of the vessel. Each vessel was scored from 0 to 4, where 0, segment with no stenosis; 1, <50% stenosis; 2, 51–70% stenosis; 3, 71–99% stenosis; and 4, vessel occlusion. The ‘standardized atheroma score’ (SAS) was calculated by summing each individual segment's stenosis score and divided by the number of diagnostic segments (n) before dividing by 4 that is the maximum potential score [Eq. (1)]: (1) SAS=[ΣMRA score/n4]×100

2, 51–70% stenosis; 3, 71–99% stenosis; and 4, vessel occlusion. The ‘standardized atheroma score’ (SAS) was calculated by summing each individual segment's stenosis score and divided by the number of diagnostic segments (n) before dividing by 4 that is the maximum potential score [Eq. (1)]: (1) SAS=[ΣMRA score/n4]×100 Statistical analysis Data are expressed as mean ± standard deviation (SD) for normally distributed continuous variables, median (range) for non-normally distributed continuous and ordinal variables, and n (%) for nominal variables. To test the null hypothesis that samples originate from the same distribution, an independent sample Kruskal–Wallis test was used, with post hoc Mann–Whitney U test to further evaluate between group differences when a significant difference was observed from the Kruskal–Wallis test. Fisher's exact test was used to compare nominal data. All data were analysed using SPSS statistical package (version 21.0, IBM SPSS, Chicago, IL, USA). Significance was assumed when P < 0.05.

ey U test to further evaluate between group differences when a significant difference was observed from the Kruskal–Wallis test. Fisher's exact test was used to compare nominal data. All data were analysed using SPSS statistical package (version 21.0, IBM SPSS, Chicago, IL, USA). Significance was assumed when P < 0.05. Results Fifty-three of the 1529 (3.6%) CMRs were excluded because of either missing data or inadequate LGE image quality. Ten of the remaining 1476 (0.67%) displayed delayed myocardial enhancement, of which 90% were female (age 54 ± 8 years). Of these, three (0.2%) demonstrated sub-endocardial enhancement in a pattern consistent with UMIs. Of these three, Figure 1 Examples of myocardial enhancement patterns observed. (A) Sub-endocardial enhancement in the inferolateral wall of the basal left ventricle consistent with a myocardial infarct. (B) Mid-myocardial enhancement. (C) RV insertion point enhancement. one involved eight AHA segments in the left anterior descending (LAD) territory, with full thickness involvement of the apical segments and <50% thickness involvement in the mid-cavity segments, and an ejection fraction of 36.3%; the second involved six segments in the LAD territory, with full thickness involvement of the apical segments and <50% thickness involvement in the mid-cavity segments, associated with regional hypokinesia, but a preserved ejection fraction of 55%;

one involved eight AHA segments in the left anterior descending (LAD) territory, with full thickness involvement of the apical segments and <50% thickness involvement in the mid-cavity segments, and an ejection fraction of 36.3%; the second involved six segments in the LAD territory, with full thickness involvement of the apical segments and <50% thickness involvement in the mid-cavity segments, associated with regional hypokinesia, but a preserved ejection fraction of 55%; the final infarct involved two segments in the inferolateral wall basally, with <50% myocardial thickness scarring, associated with regional wall motion abnormality but a preserved ejection fraction (see Figure 1). Of the remaining seven (0.47%), one demonstrated epicardial enhancement involving one AHA segment; four had a mid-myocardial pattern of enhancement, all four of which involved a single AHA segment; two had right ventricular septal insertion point enhancement each involving one AHA segment. None of these were associated with regional wall motion abnormalities.

Of the remaining seven (0.47%), one demonstrated epicardial enhancement involving one AHA segment; four had a mid-myocardial pattern of enhancement, all four of which involved a single AHA segment; two had right ventricular septal insertion point enhancement each involving one AHA segment. None of these were associated with regional wall motion abnormalities. Of the three patients with UMI pattern LGE, only one had clinical symptoms describing shortness of breath on exertion, which had been labelled by their GP as chronic obstructive pulmonary disease because of a history of smoking and an obstructive pattern on spirometry, but on review was felt to more likely be owing to their systolic impairment. Single-photon emission computed tomography (SPECT) study confirmed a large irreversible perfusion defect in the LAD territory. Neither of the other two with UMI exhibited any chest pain or shortness of breath. Those with atypical LGE patterns were asymptomatic, with no prior recalled episodes of significant chest pain.

Single-photon emission computed tomography (SPECT) study confirmed a large irreversible perfusion defect in the LAD territory. Neither of the other two with UMI exhibited any chest pain or shortness of breath. Those with atypical LGE patterns were asymptomatic, with no prior recalled episodes of significant chest pain. The demographic data are described in Table 1. Those with UMIs had a significantly higher BNP [116 (31–133) vs. 22.6 (5–175) pg/mL, P = 0.015] compared with the group without the evidence of LGE but were otherwise similar in their baseline measures. Those with non-specific scarring had lower diastolic blood pressure [68 (54–70) vs. 72 (46–98)mmHg, P = 0.013], but were otherwise similar to those without scarring. Imaging findings of the two groups are described in Table 2. Compared with those without scarring, those with UMIs had significantly lower ejection fractions [54.6 (36–62) vs. 68.9 (38–89)%, P = 0.007] and higher end-systolic volumes [36.3 (27–61) vs. 21.7 (5–65) mL/m2, P = 0.014], whereas those with non-specific LGE demonstrated no significant differences in cardiac mass, volume, or function. Neither group demonstrated a significantly higher atheroma burden than those with no scarring (P = 0.27). Table 1 Comparison of cohort characteristics between those with and without late gadolinium enhancement

reas those with non-specific LGE demonstrated no significant differences in cardiac mass, volume, or function. Neither group demonstrated a significantly higher atheroma burden than those with no scarring (P = 0.27). Table 1 Comparison of cohort characteristics between those with and without late gadolinium enhancement No LGE UMI Non-specific LGE P n (%) 1463 (99.1) 3 (0.2) 7 (0.5) Male (%) 561 (38) 0 (0) 1 (14) 0.26 Age (years) 53.4 (40–80) 62 (53–66) 58.1 (50–67) 0.13 Pulse 62 (35–92) 67 (61–82) 62 (54–70) 0.44 Systolic BP 122 (80–145) 118 (104–134) 128 (102–140) 0.71 Diastolic BP 72 (46–98) 64 (58–76) 68 (58–70) 0.018 Total cholesterol 5.4 (2.7–9.5) 5.4 (5.2–6.4) 5.4 (4.6–7.2) 0.87 LDL-cholesterol (mmol/L) 3.4 (0.9–6.8) 3.3 (3.2–3.9) 3.2 (2.3–4.7) 0.96 HDL-cholesterol (mmol/L) 1.4 (0.4–2.6) 1.8 (1.3–1.8) 1.5 (0.9–2.3) 0.94 Triglycerides (mmol/L) 1.3 (0.5–7) 1.6 (0.5–1.7) 1.1 (0.5–2.5) 0.37 Random glucose 5.1 (3–11) 5.3 (5–6) 6.2 (5–8) 0.11 ASSIGN risk score 7.4 (0.9–48) 10.3 (10–13) 6 (3–17) 0.46 BMI (kg/m2) 26.1 (17–43) 27.6 (25–29) 22.9 (21–31) 0.09 Current/ex-smoker (%) 552 (38) 1 (33) 3 (43) 1 Smoking pack years 0 (0–105) 0 (0–30) 0 (0–57) 0.91 FH of CVD (%) 377 (26) 1 (33) 0 (0) 0.2 BNP (pg/mL) 22.7 (5–175) 116 (31–133) 22.6 (10–62) 0.05 ASSIGN, ASsessing cardiovascular risk using SIGN guidelines; BMI, body mass index; BNP, brain natriuretic peptide; BP, blood pressure; CVD, cardiovascular disease; FH, family history; HDL, high density lipoprotein; LDL, low density lipoprotein; LGE, late gadolinium enhancement; UMI, unrecognized myocardial infarction.

ASSIGN, ASsessing cardiovascular risk using SIGN guidelines; BMI, body mass index; BNP, brain natriuretic peptide; BP, blood pressure; CVD, cardiovascular disease; FH, family history; HDL, high density lipoprotein; LDL, low density lipoprotein; LGE, late gadolinium enhancement; UMI, unrecognized myocardial infarction. Bold indicates statistical significance. Table 2 Comparison of left ventricular measures and whole-body atheroma burden between the groups No LGE UMI Non-specific LGE P LVMa (g/m2) 53.6 (26–109) 64.1 (59–67) 47.7 (40–60) 0.08 LVEDVa (mL/m2) 70.5 (38–140) 79.9 (72–96) 69.6 (61–81) 0.23 LVESVa (mL/m2) 21.7 (5–66) 36.3 (27–61) 19.9 (13–28) 0.033 LVSVa (mL/m2) 48.3 (24–92) 43.7 (35–45) 49.7 (40–54) 0.20 LVEF (%) 68.9 (38–89) 54.7 (36–62) 71.4 (65–78) 0.019 LVMVR (g/mL) 0.76 (0.4–2.1) 0.74 (0.7–0.9) 0.70 (0.6–0.8) 0.30 SAS 0 (0–19.6) 4.2 (0–9) 0 (0–2) 0.27 LGE, late gadolinium enhancement; LVM, left ventricular mass; EDV, end diastolic volume; ESV, end-systolic volume; EF, ejection fraction; SV, stroke volume; LVMVR, left ventricular mass volume ratio; SAS, standardized atheroma score; UMI, unrecognized myocardial infarction. aNormalized to body surface area.

No LGE UMI Non-specific LGE P LVMa (g/m2) 53.6 (26–109) 64.1 (59–67) 47.7 (40–60) 0.08 LVEDVa (mL/m2) 70.5 (38–140) 79.9 (72–96) 69.6 (61–81) 0.23 LVESVa (mL/m2) 21.7 (5–66) 36.3 (27–61) 19.9 (13–28) 0.033 LVSVa (mL/m2) 48.3 (24–92) 43.7 (35–45) 49.7 (40–54) 0.20 LVEF (%) 68.9 (38–89) 54.7 (36–62) 71.4 (65–78) 0.019 LVMVR (g/mL) 0.76 (0.4–2.1) 0.74 (0.7–0.9) 0.70 (0.6–0.8) 0.30 SAS 0 (0–19.6) 4.2 (0–9) 0 (0–2) 0.27 LGE, late gadolinium enhancement; LVM, left ventricular mass; EDV, end diastolic volume; ESV, end-systolic volume; EF, ejection fraction; SV, stroke volume; LVMVR, left ventricular mass volume ratio; SAS, standardized atheroma score; UMI, unrecognized myocardial infarction. aNormalized to body surface area. Discussion In this study, we have demonstrated that those who are of middle-age with an intermediate or low cardiovascular risk have a very low risk of UMIs. The prevalence of UMIs of 0.2% in our study is similar to the 0.34% reported by Goehde et al.,12 although this prior study had focused on a younger age group than the current study with a mean age of 49.7 compared with our 54.2 years, 15.8% of their population being under 40 with under 40s excluded from recruitment within our study and finally only 2.6% of their population were over 65 compared with our 10.5%. This is important as two prior studies focusing on 70- and 75-year-olds, respectively, described a prevalence of 19.8% in the 70-year-old cohort and 30% in the 75-year-olds, consistent with a strong age-related association.4,13 The same group also demonstrated a lack of association between the presence of UMIs and traditional risk factors, carotid intima media thickness, high-sensitivity C-reactive protein, or whole-body atheroma burden.7 Although the average age of our study was significantly lower than that observed in these two prior studies, there were 331 participants over 60 years of age and 63 over the age of 70; thus, a significantly higher incidence of UMI would have been expected in our cohort on the basis of these prior studies. This suggests that the lack of observed differences between the population with and without UMI in these previous studies may have simply been due to being underpowered to detect differences, because when we have excluded those considered at high risk of CVD (calculated using standard risk factor measurements), the prevalence of UMIs is vanishingly low. A previous study by Schelbert et al. reported an incidence of UMIs of 17%, but this was again a much older population, including high-risk participants and those with known coronary artery disease.6 Contrary to Barbier's work and supported by our own study, they found a significantly greater prevalence of traditional CVD risk factors in those with UMIs compared with those without. They also found the incidence to be significantly higher in those with diabetes.

cipants and those with known coronary artery disease.6 Contrary to Barbier's work and supported by our own study, they found a significantly greater prevalence of traditional CVD risk factors in those with UMIs compared with those without. They also found the incidence to be significantly higher in those with diabetes. Indeed, in comparison to our current study, a healthy population with diabetes and similar demographic characteristics as our group demonstrated a prevalence of 6% UMIs.14

cipants and those with known coronary artery disease.6 Contrary to Barbier's work and supported by our own study, they found a significantly greater prevalence of traditional CVD risk factors in those with UMIs compared with those without. They also found the incidence to be significantly higher in those with diabetes. Indeed, in comparison to our current study, a healthy population with diabetes and similar demographic characteristics as our group demonstrated a prevalence of 6% UMIs.14 Knowledge of the prevalence and population likely to have UMIs is important for several reasons. First and most importantly is their significant prognostic implications, with those with UMIs at markedly increased risk of future cardiovascular events.1,6 Indeed in our study, UMIs had significant functional implications, with these associated with reduced ejection fraction, dilation of the LV cavity, and an elevated BNP. Those with UMIs have been shown to have a high prevalence of significant coronary stenosis, both globally within the coronary vessels and focally upstream from the lesion, particularly in UMIs occurring in the LAD territory.15,16 There are promising data showing the benefits of intervening in silent ischaemia,17 thus this population may benefit from more aggressive management. However, prior studies have focused on SPECT assessment, and thus further work on those with asymptomatic UMIs recognized by CMR is still required. Second, if a high prevalence of UMIs was to be observed in a young cohort, it would suggest significant missed opportunity for primary prevention. Our finding of a very low rate of UMIs in a predominantly middle-aged population without the high risk of CVD provides reassurance that this is not the case and that prevention efforts need not occur at excessively young ages, instead being better focused on middle and older ages.18

opportunity for primary prevention. Our finding of a very low rate of UMIs in a predominantly middle-aged population without the high risk of CVD provides reassurance that this is not the case and that prevention efforts need not occur at excessively young ages, instead being better focused on middle and older ages.18 More prevalent than UMIs in our current cohort, but still uncommon affecting only 0.47% of the group, was LGE in a pattern not typical of myocardial infarction. The most common of these was mid-myocardial enhancement, which is non-specific and can be seen in hypertensive cardiomyopathy, infiltrative cardiomyopathies, and myocarditis.19 Given that those with high blood pressure were excluded from the study and that the LV mass was not significantly elevated in these cases, the former two are less likely, leaving prior undiagnosed myocarditis as a possible source of this myocardial enhancement. In these participants with possible subclinical events, the lack of difference in ventricular function or volumes is reassuring that there do not appear to be any significant downstream sequelae of these scars. This is in agreement with recent work showing the lack of significant impact on prognosis between those with and without LGE in patients with clinically apparent myocarditis.20,21 The finding of a lower diastolic blood pressure in the cohort with non-specific LGE compared with the cohort without any LGE is uncertain but may be a manifestation of the high prevalence of females in the non-specific group, or possibly erroneous because of the small group size.

h clinically apparent myocarditis.20,21 The finding of a lower diastolic blood pressure in the cohort with non-specific LGE compared with the cohort without any LGE is uncertain but may be a manifestation of the high prevalence of females in the non-specific group, or possibly erroneous because of the small group size. There are several weaknesses in the current study. We acquired the LGE following a single-dose contrast injection, whereas clinical routine practice is for this to be performed following double dose. However, at 3 T, contrast exerts a greater effect on T1 relaxation, and single-dose LGE has been proved to be equivalent with double dose imaging for scar detection.22 Our LGE imaging was acquired only in the short-axis plane, thus apical lesions that are better appreciated on long-axis views may have been missed, although previous reports have shown this to be an uncommon location for UMIs.4 In addition, the lack of a second plane for correlation may have led to under-reporting of areas of indeterminate signal, potentially underestimating the true prevalence. The numbers of both the UMIs and non-specific LGE were small meaning only limited conclusions can be drawn from inter-group data comparison. Additionally, this was a single-centre study on a predominantly white population, thereby limiting the generalizability of our findings to other groups. Participants were recruited based on an old threshold for the diagnosis of hypertension (145/90mmHg) and using the ATP-III criteria, both of which were the reference standard at the point of study design, but have been superseded by more up-to-date risk criteria. Thus, our healthy cohort included some participants with what would now be considered hypertension. However, this would be expected to bias the results in favour of a higher rate of UMIs; thus, our findings of a low prevalence in this cohort remain valid. For the purposes of this paper for comparing risk between cohorts, we have instead adopted the ASSIGN score, as this has the greatest evidence basis in a Scottish cohort.23

this would be expected to bias the results in favour of a higher rate of UMIs; thus, our findings of a low prevalence in this cohort remain valid. For the purposes of this paper for comparing risk between cohorts, we have instead adopted the ASSIGN score, as this has the greatest evidence basis in a Scottish cohort.23 Conclusion Despite previous reports describing high prevalence of UMI in older populations, in a predominantly middle-aged cohort, those who are of intermediate or low cardiovascular risk have a very low risk of having an UMI. Funding The study was funded by the Souter Charitable Foundation and the Chest, Heart and Stroke Scotland Charity. J.R.W.M. is supported by the Wellcome Trust through the Scottish Translational Medicine and Therapeutics Initiative (grant no. WT 085664) in the form of a clinical research fellowship. Neither group had any role in the study design, collection, analysis, and interpretation of data; in the writing of the manuscript; or in the decision to submit the manuscript for publication. Conflict of interest: J.R.W.M. has received monies from Guerbet for attending symposia and for running educational meetings. J.G.H. is Director and Shareholder of Vascular Flow Technologies Ltd and has received research funds from Guerbet.

Introduction Aortic stenosis is not only characterized by progressive valve narrowing but also by the hypertrophic response of the left ventricle (LV) that ensues.1 Indeed the development of symptoms and adverse events appears as much related to events in the myocardium as the valve.2 New techniques for assessing adverse patterns of remodelling are therefore of major interest.3 Echocardiography is the most common imaging technique to assess patients with aortic stenosis and can provide assessments of wall thickness that can be used to calculate left ventricular mass index. Cardiovascular magnetic resonance is less widely employed but provides the gold standard assessment of left ventricular mass and wall thickness with the unique ability to identify myocardial fibrosis. Asymmetric wall thickening is most commonly associated with hypertrophic cardiomyopathy.4 However this form of remodelling has recently been described in patients with increased afterload such as hypertension and aortic stenosis.5–8 An initial magnetic resonance study suggested that asymmetric wall thickening could be observed in around a quarter of patients. However, this study was hampered by referral bias and did not involve detailed tissue characterization.6 Echocardiographic studies have suggested a lower prevalence of around 10%, with the prognostic implications of this observation remaining unclear.7,8

all thickening could be observed in around a quarter of patients. However, this study was hampered by referral bias and did not involve detailed tissue characterization.6 Echocardiographic studies have suggested a lower prevalence of around 10%, with the prognostic implications of this observation remaining unclear.7,8 In this study, we sought to assess asymmetric wall thickening in patients with aortic stenosis using both cardiovascular magnetic resonance and echocardiography. In particular, we aimed to investigate in depth the factors associated with an asymmetric pattern of wall thickening, its relationship with other markers of left ventricular remodelling and decompenzation and to assess its prognostic implications.

nosis using both cardiovascular magnetic resonance and echocardiography. In particular, we aimed to investigate in depth the factors associated with an asymmetric pattern of wall thickening, its relationship with other markers of left ventricular remodelling and decompenzation and to assess its prognostic implications. Methods Patient population We performed a prospective observational cohort study of stable subjects with mild, moderate, and severe aortic stenosis recruited from the Edinburgh Heart Centre. All patients who attended the institution between March 2012 and August 2014 were invited to participate. We excluded patients with other forms of valvular heart disease (moderate or severe in nature), end-stage heart failure, advanced malignancies or other comorbidities with a life expectancy < 2 years, cardiomyopathies (including previous myocarditis), and contraindications to gadolinium-enhanced magnetic resonance, such as ferromagnetic foreign bodies and estimated glomerular filtration rate < 30 mL/min/1.73 m2. Coronary artery disease was defined as previous myocardial infarction, documented evidence of myocardial ischaemia or a > 50% stenosis of the coronary artery lumen. The blood sampling and analysis protocols are described in the Supplementary data online.

ies and estimated glomerular filtration rate < 30 mL/min/1.73 m2. Coronary artery disease was defined as previous myocardial infarction, documented evidence of myocardial ischaemia or a > 50% stenosis of the coronary artery lumen. The blood sampling and analysis protocols are described in the Supplementary data online. Both the clinical and imaging assessments were carried out at the Clinical Research Facility and the Clinical Research Imaging Centre, Edinburgh. The study was conducted in accordance with the declaration of Helsinki and approved by the local research committee. Written informed consent was obtained from all enrolled patients.

clinical and imaging assessments were carried out at the Clinical Research Facility and the Clinical Research Imaging Centre, Edinburgh. The study was conducted in accordance with the declaration of Helsinki and approved by the local research committee. Written informed consent was obtained from all enrolled patients. Cardiovascular magnetic resonance Cardiovascular magnetic resonance was performed using a 3T scanner (MAGNETOM Verio, Siemens Healthcare GmbH, Erlangen, Germany). A balanced steady-state free precession sequence was used for short-axis cine imagine of the LV and assessment of left ventricular volumes, mass and ejection fraction (Argus software, Siemens AG, Healthcare Sector). LV longitudinal function was determined by measuring the difference in mitral annular displacement between end-systole and end-diastole. On short-axis cine images, epicardial and endocardial contours were carefully identified and planimetered in end-systole and end-diastole for left ventricular volume quantification. The left ventricular mass was calculated from the total myocardial volume (excluding trabeculations and papillary muscles) multiplied by the density of the myocardium (1.05 g/mL). All volumes and mass values were indexed to body surface area (calculated using the Du Bois formula). The left ventricular mass/volume ratio (M/V) was calculated by dividing the left ventricular mass by the left ventricular end-diastolic volume. This parameter indexes the left ventricular mass to cavity size, with M/V values > 1.16 g/mL identifying patients with a relative increase in wall thickness.6 Left ventricular hypertrophy was defined as an indexed left ventricular mass ≥ 95th centile of the normal range corrected for age and gender.9 Maximal wall thickness was evaluated in all 17 segments of the LV from cine images of the LV in end-diastole (again excluding ventricular trabeculations). Asymmetric left ventricular wall thickening was defined as a regional wall thickening ≥ 13 mm that was also ≥ 1.5-fold the thickness of the opposing myocardial segment.6 Such criteria had to be fulfilled on two adjacent short-axis magnetic resonance cine images.

nd-diastole (again excluding ventricular trabeculations). Asymmetric left ventricular wall thickening was defined as a regional wall thickening ≥ 13 mm that was also ≥ 1.5-fold the thickness of the opposing myocardial segment.6 Such criteria had to be fulfilled on two adjacent short-axis magnetic resonance cine images. Fibrosis assessment Late gadolinium enhancement (LGE) was used for detection of focal replacement fibrosis. Acquisition was performed between 8 and 15 min after gadobutrol (Gadovist, Bayer Pharma AG, Germany) administration using an inversion recovery fast gradient-echo sequence and a phase-sensitive inversion recovery sequence in two phase-encoding directions in order to distinguish true enhancement from artefact. The inversion time was optimized to achieve adequate nulling of the myocardium. The presence of LGE was evaluated visually by two experienced operators (Marc R. Dweck & Calvin W.L. Chin.). Where present, it was quantified using QMASS software (Medis, Leiden, The Netherlands) as an area of myocardium with a signal intensity exceeding the threshold of two standard deviations above the mean value of normal myocardium. All areas of inversion artefact or myocardial regions contaminated by blood pool or epicardial fat were excluded.

uantified using QMASS software (Medis, Leiden, The Netherlands) as an area of myocardium with a signal intensity exceeding the threshold of two standard deviations above the mean value of normal myocardium. All areas of inversion artefact or myocardial regions contaminated by blood pool or epicardial fat were excluded. T1 mapping was performed using the Modified Look-Locker Inversion recovery sequence for the assessment of diffuse myocardial fibrosis. Short-axis T1 maps of the mid-cavity were acquired in diastole before and 20 min after the administration of 0.1 mmol/kg of gadobutrol (Gadovist, Bayer Pharma AG, Germany) using a dedicated 32-channel anterior-posterior cardiac array. T1 maps were analysed using OsiriX 4.1.1 software (Geneva, Switzerland). For diffuse fibrosis, we calculated the extracellular volume fraction (ECV) derived from pre- and post-contrast myocardial T1 values corrected for blood-pool T1 and haematocrit. The ECV was calculated according to: ECV = partition coefficient ×  [1 − haematocrit], where partition coefficient = [ΔR1myocardium/ΔR1blood-pool] and ΔR1 = (1/post-contrast T1 − 1/pre-contrast T1). In order to evaluate the total amount of interstitial fibrosis in our study cohort, we calculated the indexed fibrosis volume in each patient using the following equation: ECV × left ventricular myocardial volume corrected for the body surface area.

ium/ΔR1blood-pool] and ΔR1 = (1/post-contrast T1 − 1/pre-contrast T1). In order to evaluate the total amount of interstitial fibrosis in our study cohort, we calculated the indexed fibrosis volume in each patient using the following equation: ECV × left ventricular myocardial volume corrected for the body surface area. Echocardiography All participants underwent a transthoracic echocardiographic examination for the assessment of aortic stenosis and cardiac function on iE33, Philips Medical Systems, The Netherlands. The severity of aortic stenosis was classified on the basis of the aortic jet peak velocity, the mean pressure gradient and the aortic valve area derived using the continuity equation. All assessments were conducted in accordance with European Association of Echocardiography/American Society of Echocardiography (ASE) guidelines.10 Transmitral early (E) and late diastolic velocities and deceleration time were measured at the tips of mitral valve leaflets using pulse wave Doppler. Pulse-wave tissue Doppler imaging was used to evaluate the early (e’) diastolic velocities of the medial and lateral mitral annulus. Diastolic function was determined using the E/e’ ratio. The left ventricular mass was calculated using wall thickness measurements and cavity dimensions (ASE formula) and indexed to body surface area.11 Cut-off values of 115 g/m2 for males and 95 g/m2 for females were used to distinguish subjects with left ventricular hypertrophy. Relative wall thickness (RWT) calculated according to the formula: RWT = 2PWTd/LVEDD (PWTd posterior wall thickness at end-diastole; LVEDD LV end-diastolic dimension) was used in a similar fashion to M/V to classify subjects into the different patterns of remodelling and hypertrophy (see Supplementary data online, Table S1). To assess the presence of asymmetric wall thickening, both long and short-axis images were screened by two experienced operators blinded to the magnetic resonance data (A.G.J. and J.K.). Similar to magnetic resonance assessments, asymmetric left ventricular wall thickening was defined as a regional wall thickening ≥ 13 mm that was also ≥ 1.5-fold the thickness of the opposing myocardial segment.

images were screened by two experienced operators blinded to the magnetic resonance data (A.G.J. and J.K.). Similar to magnetic resonance assessments, asymmetric left ventricular wall thickening was defined as a regional wall thickening ≥ 13 mm that was also ≥ 1.5-fold the thickness of the opposing myocardial segment. Patterns of left ventricular adaptation Using both echocardiography and Cardiovascular Magnetic Resonance (CMR), we categorized patients with aortic stenosis into six groups of anatomic adaptation based on the left ventricular mass index, the indexed left ventricular end-diastolic volume, M/V and the presence of asymmetric wall thickening.6Normal ventricular structure, concentric remodelling, asymmetric remodelling, concentric hypertrophy, asymmetric hypertrophy, and eccentric hypertrophy (see Supplementary data online, Table S1).

entricular mass index, the indexed left ventricular end-diastolic volume, M/V and the presence of asymmetric wall thickening.6Normal ventricular structure, concentric remodelling, asymmetric remodelling, concentric hypertrophy, asymmetric hypertrophy, and eccentric hypertrophy (see Supplementary data online, Table S1). Clinical endpoints The primary endpoint of the study was aortic valve replacement and all-cause mortality. In the subgroup of subjects that subsequently underwent aortic valve replacement, we determined the impact of asymmetric thickening on 30-day perioperative cardiovascular outcomes (myocardial infarctions, congestive heart failure, new episodes of atrial or ventricular arrhythmia, perivalvular leaks, permanent pacemaker insertion, cardiac tamponade). Patients that underwent AVR were censored for survival analysis and considered as withdrawn alive. All the mortality, surgery, and in-hospital complications data were obtained from the National Strategic Tracing Service, which is a national database for all National Health Service patients in UK.

n, cardiac tamponade). Patients that underwent AVR were censored for survival analysis and considered as withdrawn alive. All the mortality, surgery, and in-hospital complications data were obtained from the National Strategic Tracing Service, which is a national database for all National Health Service patients in UK. Statistical analysis We assessed the distribution of all continuous variables using the Shapiro–Wilk test, and presented them as mean ± standard deviation or median (interquartile range). Comparisons were made using one-way analysis-of-variance to compare continuous parametric data and the Kruskall–Wallis test for non-parametric data. Chi-square tests were used for categorical baseline characteristics. The association between biomarkers and asymmetric wall thickening was assessed using linear regression analyses with adjusting for potential confounders. Kaplan–Meier curves were used to elucidate the survival distributions with regard to all-cause mortality and AVR. Differences in the outcome of patients with and without asymmetric wall thickening were assessed using the log-rank test. A Cox proportional hazard regression with adjustment for potential confounders was performed to determine the predictors of worse outcome. A two-sided P < 0.05 was considered statistically significant. All statistical analysis has been performed using SPSS Version 20 (IBM Corp., Armonk, NY, USA) and GraphPad Prism Version 5.0 (GraphPad Software, San Diego, CA, USA).

ment for potential confounders was performed to determine the predictors of worse outcome. A two-sided P < 0.05 was considered statistically significant. All statistical analysis has been performed using SPSS Version 20 (IBM Corp., Armonk, NY, USA) and GraphPad Prism Version 5.0 (GraphPad Software, San Diego, CA, USA). Results Study population The study group comprised 166 patients with aortic stenosis (69 [63,75] years old, 68% males) and 37 age- and sex-matched healthy volunteers (68 [63,74] years; 65% males). A comprehensive overview of patients’ demographics, aortic stenosis severity, left ventricular characteristics, and co-morbidities as well as comparisons to healthy controls can be found in Table 1 and Supplementary data online. Table 1 Comparison of patient characteristics between those with concentric wall thickening and asymmetric wall thickening on magnetic resonance

ortic stenosis severity, left ventricular characteristics, and co-morbidities as well as comparisons to healthy controls can be found in Table 1 and Supplementary data online. Table 1 Comparison of patient characteristics between those with concentric wall thickening and asymmetric wall thickening on magnetic resonance Concentric wall thickening Asymmetric wall thickening (n = 67) (n = 43) P-value Baseline characteristics Age, years 70 [65,77] 72 [67,75] 0.56 Males, n (%) 52 (77) 31 (72) 0.52 CAD, n (%) 22 (33) 20 (47) 0.16 Diabetes, n (%) 9 (13) 7 (16) 0.80 Hyperlipidaemia, n (%) 38 (57) 22 (51) 0.84 Hypertension, n (%) 48 (72) 33 (77) 0.56 SBP, mmHg 150 ± 20 153 ± 22 0.53 Six minute walk, distance (m) 369 ± 96 358 ± 124 0.64 Symptomatic AS, n (%) 16 (24) 14 (32) 0.31 Echocardiography AVA, cm2 0.8 [0.7,1.1] 0.8 [0.7,1.0] 0.15 AVA indexed, cm2/m2 0.44 [0.38,0.58] 0.43 [0.36,0.50] 0.08 Dimensionless index 0.25 [0.21,0.30] 0.23 [0.19,0.28] 0.06 Vm, m/s 3.9 [3.4,4.5] 4.2 [3.8,4.8] 0.01 MPG, mmHg 35 [24,44] 41 [34,50] 0.01 Mild, n (%) 10 (15) 1 (2) 0.02 Moderate, n (%) 18 (27) 9 (22) 0.47 Severe, n (%) 39 (58) 33 (76) 0.05 Indexed SV < 35 mL/m2, n (%) 13 (19) 9 (21) 0.85 E/A 0.81 [0.68, 1.00] 0.82 [0.63, 1.16] 0.68 Deceleration time 206 [169, 254] 217 [196, 247] 0.41 E/e' 12.6 [9.8,16.7] 14.2 [11.5,18.5] 0.05 LVOT Vm, m/s 1.0 [0.9,1.1] 1.0 [0.9,1.2] 0.29 Bicuspid Aortic Valve n (%) 26 (39) 14 (33) 0.51 Cardiovascular magnetic resonance Indexed EDV, mL/m2 67 [60,74] 68 [62,78] 0.15 Indexed ESV, mL/m2 22 [16,26] 22 [18,26] 0.42 Indexed SV, mL/m2 46 ± 9 48 ± 9 0.26 Indexed SV < 35 mL/m2, n (%) 6 (9%) 4 (8%) 0.94 Max wall thickness, mm 12 [10,13] 16 [14,17] <0.001 Indexed left ventricular mass, g/m2 92 [81,103] 96 [80,106] 0.49 Left ventricular mass/EDV, g/mL 1.34 [1.24,1.56] 1.36 [1.21,1.50] 0.39 Mid-wall fibrosis, n (%) 25 (37) 21 (48) 0.24 Extracellular volume fraction, % 27.6 ± 2.9 28.1 ± 2.6 0.29 Indexed fibrosis volume, mL/m2 24.5 [20.7,29.4] 26.6 [21.1,30.4] 0.28 Ejection fraction, % 68 [64,72] 67 [64,73] 0.92 Longitudinal function, mm 11.8 ± 3.0 11.1 ± 2.6 0.18 Biomarkers HS-cTnI, ng/L 6.6 [4.3,10.38] 13.5 [8.1,32.8] <0.001 BNP, pg/mL 20.9 [8.1,51.8] 56.3 [25.5,112] <0.001 Outcomes Combined primary outcome, n (%) 34 (51) 35 (81) 0.001 AVR, n (%) 28 (42) 31 (72) 0.002 Aortic stenosis-related death, n (%) 2 (3) 2 (5) 0.78 All cause death, n (%) 7 (10) 6 (14) 0.58 CAD, coronary artery disease; SBP, systolic blood pressure; AS, Aortic Stenosis; MPG, mean pressure gradient; LVOT, l

12] <0.001 Outcomes Combined primary outcome, n (%) 34 (51) 35 (81) 0.001 AVR, n (%) 28 (42) 31 (72) 0.002 Aortic stenosis-related death, n (%) 2 (3) 2 (5) 0.78 All cause death, n (%) 7 (10) 6 (14) 0.58 CAD, coronary artery disease; SBP, systolic blood pressure; AS, Aortic Stenosis; MPG, mean pressure gradient; LVOT, l eft ventricular outflow track; EDV, end diastolic volume; ESV, end systolic volume; SV, stroke volume; HS-cTnI, high-sensitivity cardiac troponin I; AVR, aortic valve replacement. Cardiovasculur magnetic resonance Patterns of left ventricular adaptation Using magnetic resonance criteria, 39 patients with aortic stenosis (23%) had normal left ventricular structure. Thirty-four (20%) patients with aortic stenosis had left ventricular remodelling with 22 having a concentric pattern (13%) and 12 (7%) an asymmetric pattern. Among the 93 (58%) patients with left ventricular hypertrophy, the most frequently occurring adaptation pattern was concentric left ventricular hypertrophy detected in 45 individuals (27%), whilst asymmetric left ventricular hypertrophy was observed in 31 subjects (19%). Seventeen (10%) patients with aortic stenosis had an eccentric pattern of hypertrophy, with these patients often having associated aortic regurgitation (n = 13), mitral regurgitation (n = 2) or a history of myocardial infarction (n = 2). All 37 healthy volunteers that comprised the control group had normal left ventricular structure.

patients with aortic stenosis had an eccentric pattern of hypertrophy, with these patients often having associated aortic regurgitation (n = 13), mitral regurgitation (n = 2) or a history of myocardial infarction (n = 2). All 37 healthy volunteers that comprised the control group had normal left ventricular structure. Asymmetric wall thickening Overall 43 (26%) of our patients with aortic stenosis demonstrated evidence of asymmetric wall thickening on magnetic resonance (Figure 1). Twelve patients had asymmetric remodelling and 31 had asymmetric hypertrophy. Importantly none of the healthy volunteers exhibited such a pattern of left ventricular adaptation. The site of asymmetric wall thickening was almost universally in the septum: at the basal level in 33 patients (77% of those with asymmetric wall thickening) and at the mid-cavity in 27 (63%) (Figure 2A). In 2 patients (5%), the regional wall thickening was observed in the anterior wall, it was never observed in the lateral or inferior walls. In 13 patients (30%), regional thickening affected just 1 segment of the 17-segment model, 16 subjects (37%) had 2 segments of asymmetric thickening whilst 14 patients (33%) had 3 or 4 affected segments.

regional wall thickening was observed in the anterior wall, it was never observed in the lateral or inferior walls. In 13 patients (30%), regional thickening affected just 1 segment of the 17-segment model, 16 subjects (37%) had 2 segments of asymmetric thickening whilst 14 patients (33%) had 3 or 4 affected segments. Figure 1 Asymmetrical wall thickening on both magnetic resonance and echocardiography. Images demonstrating asymmetric wall thickening in patients with aortic stenosis. Cardiovascular magnetic resonance short-axis cine images showing an abnormally thickened septum: in a patient with asymmetric remodelling (A) and two subjects with asymmetric hypertrophy (B) and (C). Echocardiographic parasternal long-axis images demonstrating thickening of the septum in two further patients with asymmetric remodelling (D) and (E). Echocardiographic apical 4-chamber image in a subject with asymmetric hypertrophy (F). Figure 2 Prevalence, distribution, and resolution after aortic valve replacement of asymmetric wall thickening. (A) Seventeen segment model of the LV demonstrating the site of asymmetric wall thickening as detected by both magnetic resonance and echocardiography. Magnetic resonance was more sensitive in detecting asymmetric wall thickening (43 cases) than echocardiography (29 cases). On both modalities, asymmetric wall thickening was almost universally confined to the basal and mid-cavity segments of the septum. (B) Patient with asymmetric wall thickening at baseline, which resolved when magnetic resonance was repeated 1 year after aortic valve replacement.

cases) than echocardiography (29 cases). On both modalities, asymmetric wall thickening was almost universally confined to the basal and mid-cavity segments of the septum. (B) Patient with asymmetric wall thickening at baseline, which resolved when magnetic resonance was repeated 1 year after aortic valve replacement. We compared patients with asymmetric wall thickening on CMR to patients with concentric patterns of remodelling or hypertrophy (Table 1). As expected patients in the asymmetric group had increased maximal wall thickness compared to the concentric groups (16 [14,17] vs. 12 [10,13] mm, P < 0.001) but interestingly they also had similar aortic stenosis severity. Indeed, there was no difference between the groups in terms of the aortic valve area (0.8 [0.7,1.1] vs. 0.8 [0.7,1.0] cm2, respectively, P = 0.15), the indexed aortic valve area (0.43 [0.36,0.50] vs. 0.44 [0.38,0.58] cm2/m2, P = 0.08) nor the dimensionless index (0.25 [0.21,0.30] vs. 0.23 [0.19,0.28], P = 0.06). Whilst slightly higher peak aortic jet velocities (4.2 [3.8,4.8] vs. 3.9 [3.4,4.5] m/s, P = 0.01) and mean gradients (41([34,50] vs. 35 [24,44], P = 0.01) were observed in those with asymmetric wall thickening these differences were small. No differences were observed between the two groups in terms of comorbidities, the magnitude of the hypertrophic response nor the degree of myocardial fibrosis (all P > 0.15). Despite these similarities patients with an asymmetric pattern had double the plasma concentrations of troponin I (13.5 [8.1,32.8] vs. 6.6 [4.3,10.4] ng/L, P < 0.001) and brain natriuretic peptide (BNP) (56.3 [25.5,112.2] vs. 20.9 [8.1,51.8] pg/mL, P < 0.001) compared to subjects with concentric wall thickening. Indeed asymmetric wall thickening was associated with both troponin and BNP levels independent of age, sex, systolic blood pressure, aortic stenosis severity, and left ventricular mass index (P < 0.001) (Figure 3).

[25.5,112.2] vs. 20.9 [8.1,51.8] pg/mL, P < 0.001) compared to subjects with concentric wall thickening. Indeed asymmetric wall thickening was associated with both troponin and BNP levels independent of age, sex, systolic blood pressure, aortic stenosis severity, and left ventricular mass index (P < 0.001) (Figure 3). Figure 3 Characteristics of patients with asymmetric vs. concentric wall thickening. Boxplots presenting: aortic valve area (A), indexed left ventricular mass (B), high sensitivity cardiac troponin I (C) and brain natriuretic peptide (D) concentrations in aortic stenosis patients with asymmetric and concentric patterns of wall thickening. Despite no difference in AVA and left ventricular mass index (P = 0.15 and P = 0.49, respectively) patients with asymmetric wall thickening had higher cardiac troponin and BNP levels than those with concentric wall thickening (P < 0.001). Asymmetric wall thickening resolution Out of the 31 subjects with asymmetric wall thickening who underwent aortic valve replacement, 12 had a repeat magnetic resonance imaging 1-year following surgery. In these patients, the left ventricular mass index decreased on repeat imaging (from 97 [86,102] to 68 [65, 80] g/m2; P < 0.001) as did maximum wall thickness (from 15 [14,16] to 13 [11,14] mm; P = 0.006) with an observed tendency to reduced high-sensitivity troponin I levels (from 9.0 [4.9,20.8] to 4.0 [1.8,10.0] ng/L; P = 0.073) (Figure 2B). Overall 6 of the 12 patients had complete resolution of their regional wall thickening and no longer fulfilled criteria for asymmetric wall thickening.

o 13 [11,14] mm; P = 0.006) with an observed tendency to reduced high-sensitivity troponin I levels (from 9.0 [4.9,20.8] to 4.0 [1.8,10.0] ng/L; P = 0.073) (Figure 2B). Overall 6 of the 12 patients had complete resolution of their regional wall thickening and no longer fulfilled criteria for asymmetric wall thickening. Fibrosis Asymmetric wall thickening was associated with higher ECV and replacement fibrosis compared to healthy controls and AS subjects with a normal LV (28.1 ± 2.6% and 48% vs. 26.5 ± 1.3% and 0% and 27.2 ± 2.0 and 14% respectively, all P < 0.05). Interestingly there was no significant difference in the fibrosis burden between subjects with an asymmetric and concentric pattern of LV adaptation (28.1 ± 2.6% and 48% vs. 27.6 ± 2.9 and 37%, P = 0.29 and P = 0.24, respectively). Echocardiography Using echocardiographic criteria, 26 aortic stenosis patients (16%) had normal left ventricular structure. Thirty-two (19%) had left ventricular remodelling whilst 108 (65%) exhibited left ventricular hypertrophy, of whom 9 (6%) had eccentric hypertrophy. Overall echocardiography demonstrated good agreement with magnetic resonance in terms of determining the pattern of left ventricular adaptation (all P > 0.10 for a difference) (see Supplementary data online, Table S4).

ng whilst 108 (65%) exhibited left ventricular hypertrophy, of whom 9 (6%) had eccentric hypertrophy. Overall echocardiography demonstrated good agreement with magnetic resonance in terms of determining the pattern of left ventricular adaptation (all P > 0.10 for a difference) (see Supplementary data online, Table S4). We were interested whether echocardiography was similarly able to detect asymmetric wall thickening. Short-axis images of the LV were either unavailable or non-interpretable in 31 subjects, in whom analysis was based solely on the parasternal and apical long-axis images. Overall 29 of the 166 patients with aortic stenosis exhibited asymmetric wall thickening (17%), again it was not observed in the control patients. Compared to magnetic resonance (which served as the gold standard in this analysis) echocardiography was less sensitive (67%) but showed excellent specificity (100%) in detecting patients with asymmetric wall thickening. Half (n = 8, 57%) of the patients with asymmetric wall thickening on magnetic resonance that were missed by echocardiography had non-interpretable short-axis echo images. At the segment level echocardiography missed 47 (48%) of the segments with asymmetric thickening detected by magnetic resonance, although the distribution of this thickening was similar to magnetic resonance: confined to the septum with the exception of 1 patient with anterior wall involvement (Figure 2A). Importantly despite the relatively lower sensitivity, patients with asymmetric wall thickening on echocardiography again demonstrated troponin and BNP concentrations that were more than double the values in patients with concentric wall thickening (troponin I 13.5 [8.0,32.5] vs. 5.3 [3.6,11.2] ng/L, P = 0.001; and BNP 64.7 [28.1,130.5] vs. 24.3 [10.2,52.9] pg/mL, P < 0.001) (see Supplementary data online, Table S5).

chocardiography again demonstrated troponin and BNP concentrations that were more than double the values in patients with concentric wall thickening (troponin I 13.5 [8.0,32.5] vs. 5.3 [3.6,11.2] ng/L, P = 0.001; and BNP 64.7 [28.1,130.5] vs. 24.3 [10.2,52.9] pg/mL, P < 0.001) (see Supplementary data online, Table S5). Clinical outcomes Patients were followed up for a median of 28 [22,33] months and 86 events occurred (72 aortic valve replacements and 14 patients died). Using the magnetic resonance analysis, the primary end-point was higher in patients with asymmetric wall thickening (n = 35, 81%) compared to both patients with concentric patterns (n = 34, 51%) and to all patients who did not have asymmetric wall thickening (n = 51, 42%). Indeed asymmetric wall thickening was associated with worse outcomes independent of age, sex, left ventricular mass index, coronary artery disease, and importantly aortic stenosis severity assessed with the mean aortic valve gradient [hazard ratio (HR) = 2.15; 95% confidence interval (CI) 1.29–3.59; P = 0.003](Figure 4). Among patients that underwent aortic valve replacement, perioperative complications were also more frequently observed in individuals with asymmetric wall thickening than subjects without (55% vs. 13%; P = 0.004; Figure 4).

radient [hazard ratio (HR) = 2.15; 95% confidence interval (CI) 1.29–3.59; P = 0.003](Figure 4). Among patients that underwent aortic valve replacement, perioperative complications were also more frequently observed in individuals with asymmetric wall thickening than subjects without (55% vs. 13%; P = 0.004; Figure 4). Figure 4 Outcome data in aortic stenosis patients with and without asymmetric wall thickening. Kaplan–Meier event estimates by adaptation patterns for the occurrence of death and AVR in aortic stenosis patients. Asymmetric thickening was associated with worse cardiac outcomes both when detected using magnetic resonance (A) (HR = 2.15 (1.29–3.59); P = 0.003) and echocardiography (B) (HR = 1.79 (1.08–3.69); P = 0.021). Perioperative complications in aortic stenosis patients undergoing aortic valve replacement. Subjects with asymmetric wall thickening had more cardiac complications in the perioperative period than those without based upon both magnetic resonance (C) 55% vs. 13% (P = 0.004) and echocardiographic (D) 57% vs. 19% (P = 0.023) assessments.

plications in aortic stenosis patients undergoing aortic valve replacement. Subjects with asymmetric wall thickening had more cardiac complications in the perioperative period than those without based upon both magnetic resonance (C) 55% vs. 13% (P = 0.004) and echocardiographic (D) 57% vs. 19% (P = 0.023) assessments. Based upon the echocardiographic assessments, there were 23 outcome events (20 underwent aortic valve replacement and 3 died) in the 29 patients (79%) with asymmetrical wall thickening. This compared with 48 events in the 105 patients (46%) with concentric wall thickening. Using our combined primary outcome measure the asymmetric group on echo again had worse outcomes independent of age, sex, left ventricular mass index, coronary artery disease and the mean aortic valve gradient (HR = 1.79; 95% CI 1.08–3.69; P = 0.021). In patients that underwent aortic valve replacement, perioperative adverse outcomes were also more frequently reported in subjects with asymmetric wall thickening on echocardiographic (57% vs. 19%; P = 0.023).

ss index, coronary artery disease and the mean aortic valve gradient (HR = 1.79; 95% CI 1.08–3.69; P = 0.021). In patients that underwent aortic valve replacement, perioperative adverse outcomes were also more frequently reported in subjects with asymmetric wall thickening on echocardiographic (57% vs. 19%; P = 0.023). Discussion We here provide a comprehensive multimodality imaging assessment of asymmetric wall thickening in 166 patients with aortic stenosis. In a prospective consecutive cohort, we demonstrate that asymmetric wall thickening is common, affecting a quarter (n = 43) of patients with mild-to-severe aortic stenosis when assessed using magnetic resonance. Echocardiography is less sensitive missing a third of these cases so that asymmetric wall thickening was only identified in 17% (n = 29). Irrespective of the imaging modality used, patients with asymmetric wall thickening had evidence of more advanced left ventricular decompenzation with elevated myocardial injury and increased BNP concentrations compared to those with concentric wall thickening. This was despite the two groups having similar co-morbidities, valve narrowing, myocardial fibrosis, and left ventricular hypertrophy. Moreover patients with asymmetric wall thickening (on both magnetic resonance and echocardiography) were found to have an adverse prognosis, with this form of remodelling acting as an independent predictor of aortic valve replacement or death after correction for age, sex, left ventricular mass index, coronary artery disease, and aortic stenosis severity.

l thickening (on both magnetic resonance and echocardiography) were found to have an adverse prognosis, with this form of remodelling acting as an independent predictor of aortic valve replacement or death after correction for age, sex, left ventricular mass index, coronary artery disease, and aortic stenosis severity. Magnetic resonance has emerged as the gold standard non-invasive assessment of left ventricular mass and wall thickness. This is the first study to evaluate the true prevalence of asymmetric wall thickening in a cohort of patients with aortic stenosis, free from referral bias. This is in contrast to the only previous magnetic resonance study examining this question where patients were referred for magnetic resonance on clinical grounds. Unusual patterns of remodelling could therefore have potentially triggered the referral. Our data have demonstrated that asymmetric wall thickening is indeed common amongst patients with aortic stenosis affecting approximately a quarter of subjects (40% in those with severe stenosis), and characterized by advanced wall thickness measurements (16 [14,17] mm).

erefore have potentially triggered the referral. Our data have demonstrated that asymmetric wall thickening is indeed common amongst patients with aortic stenosis affecting approximately a quarter of subjects (40% in those with severe stenosis), and characterized by advanced wall thickness measurements (16 [14,17] mm). Echocardiography is less well suited to measuring wall thickness and limited by the availability of acoustic windows. In this study, it missed 14 cases of asymmetric wall thickening detected by magnetic resonance. In eight patients, this was because short-axis echocardiographic images were not interpretable. Overall our prevalence of asymmetric wall thickening on echocardiography is similar to that previously reported7,8 and to a recent analysis of the Intensive Lipid Lowering with Simvastatin and Ezetimibe in AS SEAS trial.5 Importantly patients with asymmetric wall thickening on echocardiography demonstrated the same characteristics as those detected by magnetic resonance including the increased troponin and BNP levels as well as the adverse prognosis (see Supplementary data online, Table S3). This is an important observation because echocardiography, unlike magnetic resonance, is performed routinely to assess all patients with aortic stenosis. It suggests that the presence of asymmetric wall thickening should therefore be actively looked for on echocardiography (not simply ignored and attributed to a ‘sigmoid septum’) and used to help identify patients likely to require aortic valve replacement rapidly and who might therefore benefit from more regular clinical follow-up.

hat the presence of asymmetric wall thickening should therefore be actively looked for on echocardiography (not simply ignored and attributed to a ‘sigmoid septum’) and used to help identify patients likely to require aortic valve replacement rapidly and who might therefore benefit from more regular clinical follow-up. Why do some patients develop asymmetric rather than concentric wall thickening in response to increased afterload? The explanation for this observation remains unclear. In this comprehensive evaluation, we did not observe any clear differences in patient demographics, co-morbidity, bicuspid aortic valve prevalence, or aortic stenosis severity between these two groups. Asymmetric wall thickening did appear reversible following surgery suggesting that it represents an adaptive response to an increased afterload. One potential theory relates to the bending radius of the septum and posterior wall of the LV. It has been suggested that the larger radius of the septum leads to greater myocardial tensions during contraction, promoting a more pronounced hypertrophic response.12,13 While this concept explains why asymmetric wall thickening is almost universally confined to the septum, it fails to account for inter-individual differences in the degree of asymmetric wall thickening. It is possible that patients may have a subtle genetic predisposition to regional wall thickening (similar to those driving the development of hypertrophic cardiomyopathy) that becomes clinically apparent with exposure to an increased left ventricular afterload.14 Unfortunately we do not have genetic data in this study to test this hypothesis. The mechanism for the adverse event rate in patients with asymmetric wall thickening also remains unclear, although it was predominantly driven by increased rates of aortic valve replacement. Asymmetric wall thickening was associated with evidence of increased myocardial injury (cTnI) and wall stress (BNP) suggesting that it is a marker of more advanced decompenzation and that patients will be more likely to develop symptoms and progress towards aortic valve replacement.15 The increased wall thickening may predispose patients to supply-demand ischaemia, increased myocyte injury and troponin levels, though this hypothesis requires confirmation. It is interesting that patients with asymmetric wall thickening did not have evidence of increased myocardial fibrosis, another useful marker of left ventricular decompenzation.

ay predispose patients to supply-demand ischaemia, increased myocyte injury and troponin levels, though this hypothesis requires confirmation. It is interesting that patients with asymmetric wall thickening did not have evidence of increased myocardial fibrosis, another useful marker of left ventricular decompenzation. Limitations During follow-up, a total of 14 patients died, limiting our mortality assessments, the outcome data are therefore predominantly driven by aortic valve replacement. A large multicentre study with longer follow-up is desirable. Further attention should also be paid to the mechanism underlying asymmetric wall thickening formation, including the underlying genetics and the explanation for the associated adverse prognosis. Conclusions Asymmetric wall thickening is a common adaptation pattern in patients with aortic stenosis, which can be detected using both cardiovascular magnetic resonance and echocardiography. Despite similar demographics, comorbidities, valve narrowing, myocardial hypertrophy, and fibrosis patients with asymmetric wall thickening have increased evidence of myocardial injury (with elevated cardiac troponin I) and BNP levels. Moreover, asymmetric wall thickening was associated with adverse outcomes acting as an independent predictor of aortic valve replacement or death in this population. Supplementary data Supplementary data are available at European Heart Journal—Cardiovascular Imaging online. Supplementary Material Supplementary Tables and Figures Click here for additional data file.

Conclusions Asymmetric wall thickening is a common adaptation pattern in patients with aortic stenosis, which can be detected using both cardiovascular magnetic resonance and echocardiography. Despite similar demographics, comorbidities, valve narrowing, myocardial hypertrophy, and fibrosis patients with asymmetric wall thickening have increased evidence of myocardial injury (with elevated cardiac troponin I) and BNP levels. Moreover, asymmetric wall thickening was associated with adverse outcomes acting as an independent predictor of aortic valve replacement or death in this population. Supplementary data Supplementary data are available at European Heart Journal—Cardiovascular Imaging online. Supplementary Material Supplementary Tables and Figures Click here for additional data file. Acknowledgements The authors thank Siemens Plc for the use of their T1 mapping work-in-progress software package (#448, Quantitative Cardiac Parameter Mapping). The Clinical Research Imaging Centre (Edinburgh) is supported by the National Health Service Research Scotland (NRS) through National Health Service Lothian Health Board. Conflict of interest: None declared.

Acknowledgements The authors thank Siemens Plc for the use of their T1 mapping work-in-progress software package (#448, Quantitative Cardiac Parameter Mapping). The Clinical Research Imaging Centre (Edinburgh) is supported by the National Health Service Research Scotland (NRS) through National Health Service Lothian Health Board. Conflict of interest: None declared. Funding This work was supported by the British Heart Foundation (CH/09/002 to D.E.N., CH/09/002 to M.R.D, and CH/09/002/26360 to R.J.E), the Sir Jules Thorn Biomedical Research Award 2015 (M.R.D.), and the National Research Foundation, Ministry of Health, Singapore (C.W.L.C). D.E.N. is the recipient of a Wellcome Trust Senior Investigator Award (WT103782AIA). N.M is supported by the BHF Butler Senior Clinical Research Fellowship (FS/16/04/32023). J.K. is supported by the Polish School of Medicine Memorial Fund.

Introduction Measurement of native T1 and extracellular volume (ECV) by cardiovascular magnetic resonance (CMR) allow quantification of diffuse myocardial fibrosis. The evidence is conflicting as to whether myocardial fibrosis increases with age with some studies suggesting it decreses1 and others pointing to different age related processes, e.g. myocardial lipofuschin or haemosiderin accumulation.2–4 Whether the ECV increases with age is unclear. The Multi-Ethnic Study of Atherosclerosis found a tiny correlation of ECV with age (R2 = 0.021, P = 0.012)5 but used region of interest (ROI) based on measurement rather than mapping, and studied a population with high rates of diabetes and hypertension. Others found no changes6,7 or found increases but with small populations8 or with significant comorbidities.9,10 Using best available mapping techniques and a cohort designed specifically to address healthy ageing curated by excluding established confounding comorbidities or cardiac disease, we sought to determine whether T1 and ECV increase with age—both in order to understand the aging biology and as a step towards developing normal reference ranges.

pping techniques and a cohort designed specifically to address healthy ageing curated by excluding established confounding comorbidities or cardiac disease, we sought to determine whether T1 and ECV increase with age—both in order to understand the aging biology and as a step towards developing normal reference ranges. Methods The study comply with the Declaration of Helsinki and obtained approval from the local research ethics committee. Healthy volunteers were recruited through advertising in the hospital. Pre-consent, we excluded patients with a history or symptoms of cardiovascular disease or diabetes. All participants provided written informed consent. A stratified approach was adopted for recruitment to ensure adequate representation of participants in each age decile. A blood sample was drawn from each subject the same day just before (approximately 30 min) the CMR examination to acquire the main biochemical parameters and in particular haematocrit (Hct) for ECV quantification. The scans were performed at 1.5-T (Magnetom Avanto; Siemens Medical Solutions, Erlangen, Germany) with 32-channel cardiac phased array receiver. The imaging protocol included cines, native T1 mapping, T2 mapping, late gadolinium enhancement, and post-contrast T1 mapping.

Methods The study comply with the Declaration of Helsinki and obtained approval from the local research ethics committee. Healthy volunteers were recruited through advertising in the hospital. Pre-consent, we excluded patients with a history or symptoms of cardiovascular disease or diabetes. All participants provided written informed consent. A stratified approach was adopted for recruitment to ensure adequate representation of participants in each age decile. A blood sample was drawn from each subject the same day just before (approximately 30 min) the CMR examination to acquire the main biochemical parameters and in particular haematocrit (Hct) for ECV quantification. The scans were performed at 1.5-T (Magnetom Avanto; Siemens Medical Solutions, Erlangen, Germany) with 32-channel cardiac phased array receiver. The imaging protocol included cines, native T1 mapping, T2 mapping, late gadolinium enhancement, and post-contrast T1 mapping. T1 mapping T1 maps were acquired using three different sequences: Shortened Modified Look-Locker Inversion recovery (ShMOLLI),11 a MOdified Look-Locker Inversion Recovery (MOLLI),12 and a saturation recovery single-shot acquisition (SASHA)13 pre and at approximately 15 min after the injection of 0.1 mmol/kg of Gadoterate meglumine (Gd-DOTA marketed as Dotarem, Guerbet S.A., Paris, France).

ned Modified Look-Locker Inversion recovery (ShMOLLI),11 a MOdified Look-Locker Inversion Recovery (MOLLI),12 and a saturation recovery single-shot acquisition (SASHA)13 pre and at approximately 15 min after the injection of 0.1 mmol/kg of Gadoterate meglumine (Gd-DOTA marketed as Dotarem, Guerbet S.A., Paris, France). For the ShMOLLI technique, pre- and post-contrast T1 maps were generated inline by merging images from three consecutive inversion recovery experiments as previously described.11 The typical acquisition parameters were: echo time = 1.05 ms; imaging duration = 210 ms; matrix = 192 × 140; phase partial Fourier 6/8; minimum TI= 110 ms; TI increment = 80 ms; flip angle 35°; slice thickness = 8 mm. For the MOLLI technique (work-in-progress 448b), sampling was in seconds and optimised for expected measured T1s [pre-contrast 5s(3s)3s, post-contrast 4s(1s)3s(1s)2s]14 with shortened inversion pulse for improved efficiency and reduced T2 dependence.15 The acquisition parameters were: pixel bandwidth 977 Hz/pixel; echo time = 1.14 ms; flip angle = 35°; matrix = 256 × 144; slice thickness = 6 mm. Inline motion correction and a non-linear least-square curve fitting were performed with the set of images acquired at different inversion times to generate a pixel-wise coloured T1 map.

ion parameters were: pixel bandwidth 977 Hz/pixel; echo time = 1.14 ms; flip angle = 35°; matrix = 256 × 144; slice thickness = 6 mm. Inline motion correction and a non-linear least-square curve fitting were performed with the set of images acquired at different inversion times to generate a pixel-wise coloured T1 map. For the SASHA technique,11 the acquisition parameters for pre and post contrast maps were: echo time = 1.36 ms; matrix = 256 × 149; flip angle = 70°; slice thickness = 8 mm, and with a variable flip angle readout. Saturation recovery images were acquired and reconstructed as previously described using a two parameter model12 with motion correction to generate pixel-wise coloured T1 map from the scanner.13 Further details of the various T1 acquisition parameters are available from the online appendix. A four chamber and a mid ventricular short axis slice were acquired for all the three T1 mapping sequences but only the latter was used for all analysis. MOLLI ECV maps The previously described and validated automated method for producing a pixel-wise ECV map was used for MOLLI T1 mapping.16 This method corrects for respiratory motion due to variation in breath-holding as well as patient movement between breath-holds and relies on co-registration of the native and post-contrast T1 pixel maps. An offline software (ECV Mapping Tool Version 1.1) subsequently generated pixel-wise ECV maps after adjusting for Hct using a variety of post processing steps as previously described.16

-holding as well as patient movement between breath-holds and relies on co-registration of the native and post-contrast T1 pixel maps. An offline software (ECV Mapping Tool Version 1.1) subsequently generated pixel-wise ECV maps after adjusting for Hct using a variety of post processing steps as previously described.16 T1 mapping analysis and ECV quantification All T1 maps were analysed using cvi42 software (Circle Cardiovascular Imaging Inc., Version 5.1.2[303], Calgary, Canada). For myocardial T1 analysis, manual epicardial and endocardial contours were drawn on the MOLLI mid ventricular short axis slice and segmented according to the American Heart Association (AHA) segmentation. The mid antero-septum (segment 8) was used for the analysis. In order to avoid confounders, with blood partial voluming being a particular concern, the effect of different degrees of endocardial and epicardial border erosion were initially assessed. The initial contours were drawn on what was visually considered to be the endocardial and epicardial boundaries on the maps and subsequently an off-set erosion of 10%, 20%, and 30% was applied on both the endocardial and epicardial borders using a function provided by the software as illustrated in Figure 1. With initial erosion (0–10%), measured myocardial T1 pre-contrast was lower and post-contrast increased suggesting partial voluming was being reduced. Above 10%, this reduction stopped (e.g. MOLLI: 10%, 20%, 30% erosion: native T1 1028 ms, 1019 ms, 1014 ms) so (given a predilection for fibrosis to be endocardial), a 10% erosion offset was used. These endocardial and epicardial borders were then copied on to the pre-contrast ShMOLLI and SASHA T1 maps and MOLLI ECV map and ShMOLLI and SASHA post-contrast T1 maps with manual adjustment if needed. An example of the myocardial and blood ROI is provided in Figure 1.

o be endocardial), a 10% erosion offset was used. These endocardial and epicardial borders were then copied on to the pre-contrast ShMOLLI and SASHA T1 maps and MOLLI ECV map and ShMOLLI and SASHA post-contrast T1 maps with manual adjustment if needed. An example of the myocardial and blood ROI is provided in Figure 1. Figure 1 Example of myocardial and blood ROIs in ShMOLLI native (left) and post-contrast (right) T1 maps. Manual epicardial (green) and endocardial (red) contours with 10% erosion offset (white borders) drawn in the native ShMOLLI T1 map (left) and exported on to the post-contrast map (right) are shown. The myocardial ROI is segmented according to the AHA model. A blood ROI was also drawn in the native T1 map and copied on to the post-contrast map. For blood analysis, a ROI was drawn in the left ventricular (LV) blood pool of the mid ventricular short axis slice of the pre-contrast MOLLI T1 map and care was taken to avoid papillary muscles. This ROI was then copied on to the native ShMOLLI and SASHA T1 maps. Very limited adjustments were needed due to changes in cardiac position, movement/different breath-holds and slightly smaller cavity in some ShMOLLI images. Extracellular volume was calculated using the mean segmental pixel value from the MOLLI ECV maps and using the formula ECV =  (Δ[1/T1myo]/Δ[1/T1blood]) * [1-Hct]) for ShMOLLI and SASHA.17

For blood analysis, a ROI was drawn in the left ventricular (LV) blood pool of the mid ventricular short axis slice of the pre-contrast MOLLI T1 map and care was taken to avoid papillary muscles. This ROI was then copied on to the native ShMOLLI and SASHA T1 maps. Very limited adjustments were needed due to changes in cardiac position, movement/different breath-holds and slightly smaller cavity in some ShMOLLI images. Extracellular volume was calculated using the mean segmental pixel value from the MOLLI ECV maps and using the formula ECV =  (Δ[1/T1myo]/Δ[1/T1blood]) * [1-Hct]) for ShMOLLI and SASHA.17 Statistical analysis Statistical analysis was performed using R (version 3.0.1, 2013) and SPSS (Version 22, IBM Corporation, IL, USA). A priori, we set out to define normal values for native T1 and ECV in health by MOLLI, ShMOLLI, and SASHA as the range of values containing the central 95% (2 SD) of native T1 and ECV readings in the ‘healthy’ population, with reference limits of 2.5% and 97.5%. This definition results in 5% of the ‘healthy’ population being classified as ‘abnormal’ or as high native T1/ECV candidates. The study was therefore designed to enrol a representative group of healthy volunteers which would allow at least two subjects (1 male, 1 female) for each 2.5% interval of native T1/ECV for testing that is n = 80. A priori, we inflated recruitment to 105 to account for potential participant exclusion (in fact, we excluded 11). Our final reference group of n = 94 thus allowed us close to three subjects for testing across the majority of 2.5% native T1/ECV intervals for understanding normal variation in health.

ing that is n = 80. A priori, we inflated recruitment to 105 to account for potential participant exclusion (in fact, we excluded 11). Our final reference group of n = 94 thus allowed us close to three subjects for testing across the majority of 2.5% native T1/ECV intervals for understanding normal variation in health. Shapiro–Wilk test was used to assess for normality. Normally distributed continuous variables were presented as mean ± standard deviation. Categorical data were reported as frequencies and percentages. A two-sample independent t-test was used to compared normally distributed continuous variables. To adjust for clustering among and by groups, a varying intercept model was fitted using R package ‘lme4’ to provide a mixed-effect modelling framework for studying the impact of age and gender on group-level variation in native T1 and ECV in study subjects. For each of T1 and ECV we compared the full model (with the fixed effects of age and gender) against a reduced model without the fixed effects to determine whether the fixed effects were significant based on the difference between the likelihood of these two models by conducting the likelihood ratio test using the ANOVA function to determine χ2 values. All statistical tests were two-tailed, and P-values of less than 0.05 were considered statistically significant.

To adjust for clustering among and by groups, a varying intercept model was fitted using R package ‘lme4’ to provide a mixed-effect modelling framework for studying the impact of age and gender on group-level variation in native T1 and ECV in study subjects. For each of T1 and ECV we compared the full model (with the fixed effects of age and gender) against a reduced model without the fixed effects to determine whether the fixed effects were significant based on the difference between the likelihood of these two models by conducting the likelihood ratio test using the ANOVA function to determine χ2 values. All statistical tests were two-tailed, and P-values of less than 0.05 were considered statistically significant. Results One hundred and five healthy volunteers were recruited and underwent blood tests, 12-lead electrocardiogram (ECG) and CMR. Any subject found with abnormalities on ECG or CMR (with the exception of inconsequential extra-cardiac findings such as liver cysts which were allowed) was excluded. There were 11 such exclusions: 3 not completing the CMR (1 contrast reaction resulting in premature termination of the scan, 2 for claustrophobia); 4 for incidental findings (1 CMR finding of pulmonary valve stenosis, 1 CMR finding of biventricular impairment, 1 blood test high glucose—subsequently confirmed diabetes, 1 high blood pressure and ECG criteria for LV hypertrophy); 2 for post-enrolment disclosures (1 on beta-blockers therapy for atrial ectopics, 1 due to previous breast cancer treatment including left sided radiotherapy), and 2 due to scanner crashes with post-contrast images not available. The result was a final study population of 94 volunteers (Figure 2). As expected, these had some cardiovascular risk factors: smoking [1 (1%) active, 17 (18% ex-smokers); hypertension [none diagnosed; 7 (7%) had blood pressure >140/90 mmHg on attendance); family history: 1 (1%) of coronary artery disease in a first degree relative (mother suffering a myocardial infarction aged 64), 1 (1%) of premature sudden cardiac death <40 year old in first degree relative. Five (5%) had hypercholesterolemia and were on statins for primary prevention and 15 (16%) had a total cholesterol (here unfasted) above 6.2 mmol/L (240 mg/dL). The average body mass index (BMI) was 24.4 ± 3.7 with nine volunteers having BMI > 30 with one having the highest value of 36. No one had peripheral vascular disease. Ethnicity was 68 (72%), Caucasian, 14 (15%) Asian, 7 (7%) Black/Afro-caribbean, 5 (5%) other. Six (6%) had medications for primary prevention (five on statins, one on aspirin). All volunteers had a normal 12-lead ECG and all the CMR scans were reported as normal by experienced Level 3 CMR physicians. Mean age was 50 ± 14 years, range 20–76 years, 52% male. The age decile distribution was: 20–29 (n = 10), 30–39 (n = 14), 40–49 (n = 24), 50–59 (n = 21), 60–69 (n = 16), and 70–79 (n = 9).

irin). All volunteers had a normal 12-lead ECG and all the CMR scans were reported as normal by experienced Level 3 CMR physicians. Mean age was 50 ± 14 years, range 20–76 years, 52% male. The age decile distribution was: 20–29 (n = 10), 30–39 (n = 14), 40–49 (n = 24), 50–59 (n = 21), 60–69 (n = 16), and 70–79 (n = 9). Clinical characteristics of the overall population and according to gender are described in Table 1. There were no gender differences in age (males 51 ± 14 years and females 49 ± 15 years, P = 0.55). Heart rate decreased slightly with age (R2 = 0.075, P = 0.008) (Figure 2, top right). There was no relationship between age and other blood variables that have influences on blood T1, in particular Hct (R2 = 0.008, P = 0.402), iron bound to transferrin (R2 = 0.003, P = 0.617) and high density lipoprotein-cholesterol (R2 = 0.020, P = 0.182) (Supplementary data online, Figure S1, top left and bottom right and left). Table 1 Clinical characteristics of the 94 healthy volunteers Overall population (n = 94) Males (n = 49, 52%) Females (n = 45, 48%) Age (years) 50 ± 14 51 ± 14 49 ± 15 SBP (mmHg) 122 ± 13 125 ± 12 120 ± 13 DBP (mmHg) 76 ± 9 77 ± 7 75 ± 10 EDV (mL) 132 ± 32 149 ± 34 115 ± 19 ESV (mL) 44 ± 13 51 ± 14 37 ± 9 LV mass (g) 123 ± 34 144 ± 30 100 ± 20 LVSV (mL) 88 ± 21 97 ± 24 77 ± 12 LVEF (%) 67 ± 4 66 ± 4 68 ± 4 LAAi (cm2/m2) 11 ± 2 11 ± 2 11 ± 2 Hct (L/L) 0.42 ± 0.04 0.44 ± 0.03 0.39 ± 0.03 Data reported as mean ± SD.

20 ± 13 DBP (mmHg) 76 ± 9 77 ± 7 75 ± 10 EDV (mL) 132 ± 32 149 ± 34 115 ± 19 ESV (mL) 44 ± 13 51 ± 14 37 ± 9 LV mass (g) 123 ± 34 144 ± 30 100 ± 20 LVSV (mL) 88 ± 21 97 ± 24 77 ± 12 LVEF (%) 67 ± 4 66 ± 4 68 ± 4 LAAi (cm2/m2) 11 ± 2 11 ± 2 11 ± 2 Hct (L/L) 0.42 ± 0.04 0.44 ± 0.03 0.39 ± 0.03 Data reported as mean ± SD. DBP, diastolic blood pressure; EDV, end-diastolic volume; ESV, end-systolic volume; Hct, haematocrit; LAAi, left atrial area indexed; LVEF, left ventricular ejection fraction; SBP, systolic blood pressure; SD: standard deviation; LVSV, left ventricular stroke volume. Figure 2 Population selection process. In the flow chart is illustrated the final population selection process. CV, cardiovascular. Acquisition of the mapping sequences was particularly careful in the context of the research project and the need for repeat acquisition was limited to <5% of healthy volunteers. This was typically one additional breath-hold per slice, although there were two outlier subjects where multiple breath-holds were needed. There was one power-cut (requiring restart) and one period where the SASHA sequence did not work for post-contrast acquisitions. Artefacts were typically breathing artefacts for ShMOLLI (a non-MOCO sequence, albeit shorter breath-hold) with misgating more common in MOLLI and SASHA. In three subjects, the source images were good, but the ECV map failed—these required post-acquisition individualised post-processing (author P.K.) for reconstruction.

rtefacts were typically breathing artefacts for ShMOLLI (a non-MOCO sequence, albeit shorter breath-hold) with misgating more common in MOLLI and SASHA. In three subjects, the source images were good, but the ECV map failed—these required post-acquisition individualised post-processing (author P.K.) for reconstruction. Considering all study subjects of all ages and across both genders, overall mean values of native T1 and ECV (reported in the order of ShMOLLI, MOLLI, and SASHA, respectively throughout the manuscript) were T1: 957 ± 30 ms, 1025 ± 38 ms, 1144 ± 45 ms and ECV: 28.4 ± 3.0% [95% confidence interval (CI) 27.8–29.0], 27.3 ± 2.7 (95% CI 26.8–27.9), 24.1 ± 2.9% (95% CI 23.5–24.7), Table 2. Table 2 T1 mapping data in the overall population of 94 healthy volunteers. Overall (n = 94) Males (n = 49, 52%) Females (n = 45, 48%) P-value Myocardial native T1 ShMOLLI (ms) 957 ± 30 948 ± 26 966 ± 31 0.003 MOLLI (ms) 1024 ± 39 1008 ± 33 1043 ± 37 <0.0001 SASHA (ms) 1144 ± 45 1120 ± 35 1171 ± 41 <0.0001 ECV ShMOLLI (%) 28.4 ± 3.0 27.1 ± 2.7 29.8 ± 2.7 <0.0001 MOLLI (%) 27.3 ± 2.7 26.1 ± 2.3 28.7 ± 2.6 <0.0001 SASHA (%) 24.1 ± 2.9 22.6 ± 2.3 26.0 ± 2.4 <0.0001 ECV, extra-cellular volume; MOLLI, MOdified Look-Locker Inversion recovery; ShMOLLI, Shortened MOdified Look-Locker Inversion recovery; SASHA, saturation recovery single-shot acquisition. Gender: For both native T1 and ECV, females had higher values by all techniques (T1 +18 ms, +35 ms, +51 ms; ECV +2.7%, +2.6%, +3.4%, P-value being statistically significant for all parameters), Table 2.

Overall (n = 94) Males (n = 49, 52%) Females (n = 45, 48%) P-value Myocardial native T1 ShMOLLI (ms) 957 ± 30 948 ± 26 966 ± 31 0.003 MOLLI (ms) 1024 ± 39 1008 ± 33 1043 ± 37 <0.0001 SASHA (ms) 1144 ± 45 1120 ± 35 1171 ± 41 <0.0001 ECV ShMOLLI (%) 28.4 ± 3.0 27.1 ± 2.7 29.8 ± 2.7 <0.0001 MOLLI (%) 27.3 ± 2.7 26.1 ± 2.3 28.7 ± 2.6 <0.0001 SASHA (%) 24.1 ± 2.9 22.6 ± 2.3 26.0 ± 2.4 <0.0001 ECV, extra-cellular volume; MOLLI, MOdified Look-Locker Inversion recovery; ShMOLLI, Shortened MOdified Look-Locker Inversion recovery; SASHA, saturation recovery single-shot acquisition. Gender: For both native T1 and ECV, females had higher values by all techniques (T1 +18 ms, +35 ms, +51 ms; ECV +2.7%, +2.6%, +3.4%, P-value being statistically significant for all parameters), Table 2. Age: native myocardial T1 was slightly lower with increasing age (R2 = 0.042, P = 0.048; R2 = 0.131, P < 0.0001), on average by 8 and 11 ms/decade by ShMOLLI and MOLLI but not by SASHA (R2 = 0.033, P = 0.083) (Figure 3). This was in both males (R2 = 0.130, P = 0.011) and females (R2 = 0.150, P = 0.009). ECV did not change significantly with age by any technique: (R2 = 0.003, P = 0.582; R2 = 0.001, P = 0.733 and R2 = 0.003, P = 0.615, Figure 4). Figure 3 Relationship between age and native myocardial T1 according to MOLLI, ShMOLLI and SASHA. Native myocardial T1 decreased slightly with age by (A) MOLLI (R2 = 0.131, P < 0.0001) and (B) ShMOLLI (R2 = 0.042, P = 0.048), while this was not the case for (C) SASHA (R2 = 0.033, P = 0.083).

Age: native myocardial T1 was slightly lower with increasing age (R2 = 0.042, P = 0.048; R2 = 0.131, P < 0.0001), on average by 8 and 11 ms/decade by ShMOLLI and MOLLI but not by SASHA (R2 = 0.033, P = 0.083) (Figure 3). This was in both males (R2 = 0.130, P = 0.011) and females (R2 = 0.150, P = 0.009). ECV did not change significantly with age by any technique: (R2 = 0.003, P = 0.582; R2 = 0.001, P = 0.733 and R2 = 0.003, P = 0.615, Figure 4). Figure 3 Relationship between age and native myocardial T1 according to MOLLI, ShMOLLI and SASHA. Native myocardial T1 decreased slightly with age by (A) MOLLI (R2 = 0.131, P < 0.0001) and (B) ShMOLLI (R2 = 0.042, P = 0.048), while this was not the case for (C) SASHA (R2 = 0.033, P = 0.083). Figure 4 Relationship between ECV and age according to MOLLI, ShMOLLI and SASHA. R2 = 0.001, P = 0.733 by MOLLI (A), R2 = 0.003, P = 0.582 by ShMOLLI (B) and R2 = 0.003, P = 0.615 by SASHA (C). Adjusting for clustered data, analysis for T1 by the three sequences using linear models with age and gender as fixed effects, confirmed how female gender affected T1 (χ2 (1)= 62.87, P =<0.0001), increasing it by about 33 ms and how increasing age also affected T1 (χ2 (1)= 25.62, P = 0.0001), decreasing it by 28, 25, 26, 33, and 41 ms across 2nd, 3rd, 4th, 5th, and 6th decades, respectively.

linear models with age and gender as fixed effects, confirmed how female gender affected T1 (χ2 (1)= 62.87, P =<0.0001), increasing it by about 33 ms and how increasing age also affected T1 (χ2 (1)= 25.62, P = 0.0001), decreasing it by 28, 25, 26, 33, and 41 ms across 2nd, 3rd, 4th, 5th, and 6th decades, respectively. Similar analysis for ECV by the three sequences confirmed how female gender affected ECV (χ2 (1)= 82.07, P = <0.0001), increasing it by about 2.9 ECV points and how increasing age did not significantly affect ECV (χ2 (1)= 10.98, P = 0.05), describing small changes of the order −0.83, −0.13, 0.28, 0.62 and −0.79 ECV points across 2nd, 3rd, 4th, 5th, and 6th decades, respectively.

female gender affected ECV (χ2 (1)= 82.07, P = <0.0001), increasing it by about 2.9 ECV points and how increasing age did not significantly affect ECV (χ2 (1)= 10.98, P = 0.05), describing small changes of the order −0.83, −0.13, 0.28, 0.62 and −0.79 ECV points across 2nd, 3rd, 4th, 5th, and 6th decades, respectively. Discussion Native myocardial T1 and ECV allow quantification of myocardial fibrosis and are increasingly used in the clinical practice for the diagnosis of several cardiac disorders, such as cardiomyopathies, myocarditis, iron overload and ischaemic heart disease,18,19 so the need for accurately establishing normal reference ranges and understanding variations with physiological parameters is clear. We, therefore, prospectively aimed to understand changes in these parameters with healthy aging and gender, using best available technology including three T1 mapping sequences and two ECV analysis techniques. We found that native T1 and ECV were higher in females, as previously found20,21 but, that native T1 was slight lower with age (ShMOLLI and MOLLI), and ECV does not change. The known measurement differences by ShMOLLI, MOLLI, and SASHA were present for T1 and although reduced for ECV were still present. The meticulous selection of the cohort to be healthy, its size, the prospective nature of the study, the way blood dependent variables were checked as confounders, the use of three sequences and their quality control by the physicists that wrote them and the use of two analysis techniques (both drawn regions of interest and ECV mapping), are strengths of the study and suggest the results are robust.

ture of the study, the way blood dependent variables were checked as confounders, the use of three sequences and their quality control by the physicists that wrote them and the use of two analysis techniques (both drawn regions of interest and ECV mapping), are strengths of the study and suggest the results are robust. There is a belief that myocardial fibrosis increases with age. There is little supporting evidence and some studies say the opposite with an increase in myocyte size and volume fraction with a decrease in the relative amount of interstitium.1 This study explored healthy aging—our subjects had no comorbidities (no diabetes, no hypertension), both of which increased in prevalence with age. A recent group similarly showed in a smaller cohort of 44 individuals free from cardiovascular disease, diabetes, hypertension, and not on any cardiovascular medication, that native myocardial T1 measured by SASHA was higher in women compared to men and did not vary significantly with age (P = 0.59) while ECV did not vary significantly with age (P = 0.20) or gender (P = 0.14).22

ividuals free from cardiovascular disease, diabetes, hypertension, and not on any cardiovascular medication, that native myocardial T1 measured by SASHA was higher in women compared to men and did not vary significantly with age (P = 0.59) while ECV did not vary significantly with age (P = 0.20) or gender (P = 0.14).22 Other authors point to other age related changes e.g. myocardial lipofuschin or haemosiderin accumulation.2–4 These could well account for the native T1 being lower with age with constant ECV. The native T1 technique measures signal from both myocytes and interstitum, and the ECV measurement includes plasma volume. There are plausible reasons that, for example capilliary density or resting recruitment could be different with age, but this is not convincing as vasodilatation would increase both native T1 and ECV. Little data is available in this area on healthy ageing.23,24 The gender differences found are similar to all other literature. Partial voluming of blood pool has always been a plausible reason for this with the thinner (on average) female heart more predisposed to this. This, and the longer blood T1 in capilliaries within myocardium25 remain plausible but look increasingly unlikely to be the whole explanation.

similar to all other literature. Partial voluming of blood pool has always been a plausible reason for this with the thinner (on average) female heart more predisposed to this. This, and the longer blood T1 in capilliaries within myocardium25 remain plausible but look increasingly unlikely to be the whole explanation. As with other studies (e.g. reference ranges for volume analysis), the data here is described as being related to ageing. This is not strictly true—they are all single timepoint study of different birth-cohorts. There is the potential that 60 and 70 year olds, whose fetal development and early growth occurred in the immediate post war period could be different to that of 20–30 year olds. Long term follow-up studies with paired data decades apart are extraordinarily difficult to do as the technology changes so fast. These will however be needed in the future. We used three sequences: one Saturation recovery based (SASHA) and two inversion recovery based sequences (ShMOLLI and MOLLI). We did this for generalizability of results but also because the accuracy and precision of the sequences are not known, particularly when applied in vivo. There are also differences in what is being measured—in particular SASHA is less magnetization transfer sensitive than infrared based sequences. We were looking for the biology of healthy ageing. Trends (or lack of trends) appearing in all three sequences would therefore be robust.

n, particularly when applied in vivo. There are also differences in what is being measured—in particular SASHA is less magnetization transfer sensitive than infrared based sequences. We were looking for the biology of healthy ageing. Trends (or lack of trends) appearing in all three sequences would therefore be robust. Limitations The slice thickness for MOLLI was 6 mm (Avanto 6 mm by Siemens) compared to 8 mm for ShMOLLI and SASHA. Corresponding automated ECV map for ShMOLLI and SASHA were not available at the time of this study. We adopted a stratified approach to recruiting patients but this was done to ensure satisfactory distribution of participants in each age decile. Conclusion In health, gender influences native myocardial T1 and ECV with women having a higher native T1 and ECV compared with males. Native myocardial T1 is slightly lower with increasing age while ECV is not influenced by aging. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online. Conflict of interest: None declared.

Conclusion In health, gender influences native myocardial T1 and ECV with women having a higher native T1 and ECV compared with males. Native myocardial T1 is slightly lower with increasing age while ECV is not influenced by aging. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online. Conflict of interest: None declared. Funding S.R. is supported by Borse di studio SIC e MSD Italia-Merck Sharp & Dohme. M.F is supported by Clinical Research Training Fellowship from the British Heart Foundation (FS/12/56/29723). T.A.T is supported by Doctoral Research Fellowship from the NIHR, UK (NIHR-DRF-2013-06-102). A.A.G. is supported by the Rosetrees Trust. J.C.M. has received grant funding from GlaxoSmithKline. This work was undertaken at University College London Hospital, which received a proportion of funding from the UK Department of Health National Institute for Health Research Biomedical Research Centers funding scheme. SKP is supported by the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre based at The Oxford University Hospitals Trust at the University of Oxford. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. Supplementary Material Supplementary Figure 1S Click here for additional data file. Supplementary Data Click here for additional data file.

Introduction Aortic stenosis (AS) is a common valvular heart disease1 in which echocardiographic assessment is a key part of judgment of timing procedural intervention.2,3 Although other imaging modalities can provide useful complementary information,4 echocardiography is the standard technique for serial monitoring.2 Nevertheless, it can be challenging to avoid random variation between assessments on separate visits. Echocardiography provides a range of parameters5–8 including peak aortic velocity, aortic velocity time integral (VTI) and, by using measurements of the left ventricular outflow tract (LVOT) and the continuity equation, aortic valve (AV) area. Dimensionless index (DI) is calculated as the Doppler measurement made in the LVOT divided by the Doppler measurement made at the AV. It avoids measurements of the size of the LVOT.5–8 One of the many sources of variation between visits is differences that arise between different observers reading the same trace. Previous studies have evaluated the reproducibility of peak and VTI measurements in patients with AS.9–12 However, only two operators were studied. Whether the results from a single pair of operators can be generalized to a larger group of operators is unknown. Moreover, computer technology now allows research to probe deeper into the causation of variability in human measurement processes, in order to provide mechanistic information to those developing clinical protocols to perform efficiently and consistently.

One of the many sources of variation between visits is differences that arise between different observers reading the same trace. Previous studies have evaluated the reproducibility of peak and VTI measurements in patients with AS.9–12 However, only two operators were studied. Whether the results from a single pair of operators can be generalized to a larger group of operators is unknown. Moreover, computer technology now allows research to probe deeper into the causation of variability in human measurement processes, in order to provide mechanistic information to those developing clinical protocols to perform efficiently and consistently. In this study, we asked 25 operators to make measurements from 40 cases, which, unbeknown to operators, were 20 cases shown twice, enabling assessment of intra-operator reproducibility. The aims of our study were to quantify, across a broader range of operators and cases of AS, the intra-operator and inter-operator reproducibility of measurements made from Doppler traces. Methods We reviewed our clinical imaging database to identify 20 consecutive patients undergoing transthoracic echocardiography in the Echocardiography Department at St. Mary’s Hospital in which AS of any severity [defined as peak continuous wave (CW) Doppler velocity of greater than 2 m/s across AV] had been identified. All patients had undergone standard Doppler examination of flow in the LVOT and AV as recommended by guidelines.7,8 Images were acquired by accredited echocardiographers who were free to optimize sweep speed, scale, gain, and filters as they wished.

Methods We reviewed our clinical imaging database to identify 20 consecutive patients undergoing transthoracic echocardiography in the Echocardiography Department at St. Mary’s Hospital in which AS of any severity [defined as peak continuous wave (CW) Doppler velocity of greater than 2 m/s across AV] had been identified. All patients had undergone standard Doppler examination of flow in the LVOT and AV as recommended by guidelines.7,8 Images were acquired by accredited echocardiographers who were free to optimize sweep speed, scale, gain, and filters as they wished. The LVOT and AV Doppler trace images for each patient were exported and anonymized. We used custom-designed software (Matlab and Statistics Toolbox Release Matlab R2015b, The MathWorks, Inc., Natick, MA, USA) to mask all but one beat. Whilst in normal clinical practice operators would be free to make measurements from any beat, we wanted to specifically study the process of measuring from the trace to specifically compare peak and VTI. To enable assessment of intra-operator as well as inter-operator variability, the 40 images (20 CW and 20 PW images) were then duplicated (creating 80 images in total) and presented in random order. Twenty-five operators from three different hospitals, unaware of the duplication and blinded to the results of their or others’ measurements and study hypothesis, were asked to view the images. For each image, they were asked to measure both the peak velocity and the VTI using custom-designed software. The VTI traces were stored for comparison.

three different hospitals, unaware of the duplication and blinded to the results of their or others’ measurements and study hypothesis, were asked to view the images. For each image, they were asked to measure both the peak velocity and the VTI using custom-designed software. The VTI traces were stored for comparison. Reversal of traces for VTI To further investigate variability in tracing the VTI, we attempted to isolate the variability arising from (i) the challenge facing an operator when deciding where to begin a trace and (ii) the challenge of tracing the steep gradients that arise at both the beginning and end of the trace. To do this, images were reversed along the horizontal axis, so that the beginning of the trace was now at the end and vice versa. Ten operators unware of the image reversal, or the reasons for it, were asked to review the 80 reversed images and again make measurements of VTI. Statistical analysis Statistical analysis was performed using ‘The R project for statistical computing’ with the package lme4.13 Figures were prepared using the package ggplot2.14 Continuous data are expressed as mean ± standard deviation. Categorical variables are summarized as percentages. A P-value of <0.05 was considered significant. To quantify intra- and inter-observer variability, we use a mixed-model analysis.13 To calculate the variability as a percentage of the measurement, we calculated the variability of the log transformed measurement and then back-transformed the variance.

zed as percentages. A P-value of <0.05 was considered significant. To quantify intra- and inter-observer variability, we use a mixed-model analysis.13 To calculate the variability as a percentage of the measurement, we calculated the variability of the log transformed measurement and then back-transformed the variance. For each measurement, the percentage difference, scaled to the mean of all operators’ measurements, was calculated to quantify the tendency for individual operators to over-read or under-read. To quantify the sources of variability arising from tracing a VTI, mean consensus curves were identified and divided into five equal vertical strips. For each strip, the mean standard deviation from the consensus curve was calculated. Analysis of variance (ANOVA) and the post-hoc Tukey Honest Significant Difference test were used to perform a comparison between the mean variability at different segments of the tracings. Results Cases The average age of patients was 79 ±10 years. Eight (40%) were male and 12 (60%) were female. The severity (as described by the reporting physician) was mild in 1 (5%), moderate in 12 (60%), and severe in 7 (35%) patients. The indications for echocardiography were: AS follow-up in 12 (60%) patients, to investigate the cause of shortness of breath in 5 (25%) patients, to investigate the presence of a systolic murmur in 2 (10%) patients, and preoperative evaluation in 1 (5%) patient.

), moderate in 12 (60%), and severe in 7 (35%) patients. The indications for echocardiography were: AS follow-up in 12 (60%) patients, to investigate the cause of shortness of breath in 5 (25%) patients, to investigate the presence of a systolic murmur in 2 (10%) patients, and preoperative evaluation in 1 (5%) patient. Operator characteristics Twenty-five operators from our institution reviewed 80 Doppler images in random order. These consisted of 40 pairs of images: 20 continuous-wave traces acquired through the AV and 20 paired pulsed-wave traces acquired from the LVOT. In fact, the 40 pairs were 20 pairs shown twice, but operators viewing the sequence of 80 randomly ordered traces were unaware of this duplication. Mean experience of echocardiography was 5.8 ± 6.4 years. Twelve (48%) held formal accreditation. Variability in VTI and peak measurements The distribution of VTI measurements is shown in the left panel of Figure 1. Across all measurements made by all operators in all cases, the overall mean VTI was 70.1 ±18.6 cm for CW through the AV and 18.7 ± 4.7 cm for PW in the left ventricular tract. Across all cases, the coefficient of variation was 18.0% for pulsed-wave Doppler traces in the LVOT, made up of an intra-operator coefficient of variation of 11.9% and an inter-operator coefficient of variation of 12.9%. Across all cases, the coefficient of variation was 10.2% for CW Doppler traces through the AV, made up of an intra-operator coefficient of variation of 7.3% and an inter-operator coefficient of variation of 6.9%.

n intra-operator coefficient of variation of 11.9% and an inter-operator coefficient of variation of 12.9%. Across all cases, the coefficient of variation was 10.2% for CW Doppler traces through the AV, made up of an intra-operator coefficient of variation of 7.3% and an inter-operator coefficient of variation of 6.9%. Figure 1 Variation in VTI (left panel) and peak (right panel) measurements. Each column represents a different case, ordered from the smallest average measurement on the left to largest on the right. Each point represents an operator’s measurement for that case. The upper group are measurements from a CW acquisition through the AV. The lower group are measurements from a pulsed-wave acquisition in the LVOT. The distribution of peak measurements is shown in the right panel of Figure 1. Across all measurements made by all operators in all cases, the overall mean peak velociy was 346.5 ± 62.0 cm/s for CW through the AV and 95.8 ± 25.0 cm/s for PW in the left ventricular tract. Across all cases, the coefficient of variation was 10.1% for pulsed-wave Doppler traces in the LVOT, made up of an intra-operator coefficient of variation of 5.6% and an inter-operator coefficient of variation of 8.2%. Across all cases, the coefficient of variation was 4.0% for CW Doppler traces through the AV, made up of an intra-operator coefficient of variation of 2.5% and an inter-operator coefficient of variation of 3.1%.

f an intra-operator coefficient of variation of 5.6% and an inter-operator coefficient of variation of 8.2%. Across all cases, the coefficient of variation was 4.0% for CW Doppler traces through the AV, made up of an intra-operator coefficient of variation of 2.5% and an inter-operator coefficient of variation of 3.1%. As can be seen from Figure 1, peak values were more tightly clustered than VTI values. The coefficient of variation was significantly smaller (P < 0.001 by ANOVA, and the post-hoc Tukey Honest Significant Difference test). Tendency for an operator to over-read or under-read on repeated viewing of identical images Across all measurements, operators showed a tendency to consistently make measurements which over-read or under-read the average of all operators when they were unknowingly represented with the same image again, as shown in Figure 2. Proportional over-measurement or under-measurement was strongly correlated for first and second VTI measurements of AV CW traces (Pearson’s r = 0.48; P < 0.001). It was also strongly correlated for first and second VTI measurements of LVOT PW traces (r = 0.78; P < 0.001). Even stronger correlations were observed for first and second peak measurements of AV CW traces (Pearson’s r = 0.78: P < 0.001) and first and second peak measurements of LVOT PW traces (Pearson’s r = 0.89: P < 0.001).

lso strongly correlated for first and second VTI measurements of LVOT PW traces (r = 0.78; P < 0.001). Even stronger correlations were observed for first and second peak measurements of AV CW traces (Pearson’s r = 0.78: P < 0.001) and first and second peak measurements of LVOT PW traces (Pearson’s r = 0.89: P < 0.001). Figure 2 Consistency in operators’ behaviour when reassessing the same images. Each point represents a measurement made by one operator viewing one case. The position on the horizontal axis represents whether the operator over-read or under-read on the first viewing and scaled to the average of all operators for that case. The position on the vertical axis represents whether the operator over-read or under-read on the second viewing and again scaled to the average of all operators for that case. Operators consistently over-reading on both viewings lie in the top-right quadrant, whereas operators consistently under-reading on both viewings lie in the bottom-left quadrant. The upper panel shows measurements of VTI. The lower panel shows measurements of peak velocity. The left panel shows pulsed-wave LVOT measurements. The right panel showed CW AV measurements.

ngs lie in the top-right quadrant, whereas operators consistently under-reading on both viewings lie in the bottom-left quadrant. The upper panel shows measurements of VTI. The lower panel shows measurements of peak velocity. The left panel shows pulsed-wave LVOT measurements. The right panel showed CW AV measurements. Tendency for an operator to over-read or under-read in general for any case When an operator’s measurements of all the cases were considered, some operators had a tendency to under-read and over-read across cases in general. For VTI, the operator with the largest tendency to under-read did so by −19.6 ± 10.2%, whilst the operator with the largest tendency to over-read did so by +12.8 ± 10.4%. For peak, the operator with the largest tendency to under-read did so by −12.6 ± 15.9%, whilst the operator with the largest tendency to over-read did so by +10.8 ± 6.3%. The distribution of the tendency to under-read or over-read by individual operators is shown in Figure 3. Figure 3 Tendency of operators to under-read or over-read relative to the average for that case. Each column represents a different operator, ordered from the operator under-reading by the largest proportion on the left to the operator over-reading by the largest proportion on the right. The values have been scaled to the average for that case. The left panel shows VTI measurement. The right panel shows peak measurement.

lumn represents a different operator, ordered from the operator under-reading by the largest proportion on the left to the operator over-reading by the largest proportion on the right. The values have been scaled to the average for that case. The left panel shows VTI measurement. The right panel shows peak measurement. Tendency to over-read and under-read CW and PW images from the same case When considering an individual patient, an operator making an AV CW VTI measurement higher than the average from all operators was also likely to make an LVOT VTI measurement higher than the average from all operators (Pearson’s r = 0.39; P < 0.001) (Figure 4, left panel). Similarly, operators making an AV CW peak velocity measurement higher than the average from all operators were also likely to make a LVOT PW peak velocity measurement higher than the average from all operators (Pearson’s r = 0.41; P < 0.001) (Figure 4, right panel). Variability in dimensionless index measurements Across all measurements made by all operators in all cases, the overall mean DI was 0.280 ± 0.077 and 0.282 ± 0.080 using VTIs and peak velocities, respectively. Across all cases, the coefficient of variation was 9.3% for DI using peak, made up of an intra-operator coefficient of variation of 6.2% and an inter-operator coefficient of variation of 6.7%. Across all cases, the coefficient of variation was 17.1% for DI using VTI, made up of an intra-operator coefficient of variation of 13.9% and an inter-operator coefficient of variation of 9.3%.

ng peak, made up of an intra-operator coefficient of variation of 6.2% and an inter-operator coefficient of variation of 6.7%. Across all cases, the coefficient of variation was 17.1% for DI using VTI, made up of an intra-operator coefficient of variation of 13.9% and an inter-operator coefficient of variation of 9.3%. Steep slopes are the main source of VTI tracing variability The variability in tracing followed the same pattern for both CW and pulsed wave (PW) Doppler traces, as shown in Figure 5. The standard deviation of the trace from the consensus is low for segments of the trace which have either no or shallow slopes. The standard deviation becomes much higher when the slopes are steeper. Figure 4 Relationship between under-reading and over-reading for pulsed-wave and continuous wave traces from the same patient for VTI (left panel) and peak measurements (right panel). Each point represents a case reviewed by a single operator. The tendency to over-read or under-read the pulsed wave LVOT measurement is represented on the horizontal axis. The tendency to over-read or under-read the continuous wave AV measurement is represented on the vertical axis. When variability is considered across five equal vertical sections of the trace (as shown in Figure 6), the highest variability was seen in the first part of the trace [standard deviation (SD) 41.1 ± 12.6 cm for CW, SD 12.4 ± 2.3 cm for PW]. The lowest variability was seen in the middle parts of the trace. The last part of the trace showed intermediate variability (SD 32.4 ± 9.5 cm for CW, SD 10.3 ± 4.7 cm for PW).

e 6), the highest variability was seen in the first part of the trace [standard deviation (SD) 41.1 ± 12.6 cm for CW, SD 12.4 ± 2.3 cm for PW]. The lowest variability was seen in the middle parts of the trace. The last part of the trace showed intermediate variability (SD 32.4 ± 9.5 cm for CW, SD 10.3 ± 4.7 cm for PW). Figure 5 Disagreement with the consensus of all operators at steep angles. Each point represents a small portion of an individual operator's trace for an individual case. Deviation from the consensus is low at shallow angles but far greater when the slope is steep, whether it be downwards from the baseline (negative angle, left of the diagram) or back towards the baseline (positive angle, right of the diagram). Steep slopes rather than initiating the trace is the source of variability To test the hypothesis that variability arises from the difficulty in reliably tracing the steep part of the curve rather than the act of deciding where to begin the trace, we represented the images to 10 observers a third and a fourth time, but with the images flipped horizontally (i.e. the time-axis reversed). When tracings were reversed (right panels on Figure 6), the trend in variability was also reversed. The highest variability was seen in the last part of the trace (SD 45.3 ± 12.2 cm for CW, SD 11.4 ± 2.3 cm for PW). The lowest variability was again seen in the middle part of the trace. The first part of the trace showed intermediate variability (SD 29.9 ± 7.4 cm for CW, SD 9.5 ± 4.1 cm for PW).

ity was also reversed. The highest variability was seen in the last part of the trace (SD 45.3 ± 12.2 cm for CW, SD 11.4 ± 2.3 cm for PW). The lowest variability was again seen in the middle part of the trace. The first part of the trace showed intermediate variability (SD 29.9 ± 7.4 cm for CW, SD 9.5 ± 4.1 cm for PW). Discussion This study shows that measuring the peak of a Doppler trace is a far more reproducible strategy than measuring the VTI, with, on average, a 2.5-fold reduction in coefficient of variation. In research, the resulting six-fold reduction in the number of patients required to power a study using peak velocity rather than VTI has huge financial and logistical implications. In clinical practice, a patient with the average peak velocity from our study of 346.5 cm/s, a change of 38.5 cm/s (11.1%) could be detected with 95% confidence. In clinical practice, a patient with the average VTI from our study of 70.1 cm, a change of 19.8 cm (28.3%) could be detected with 95% confidence. The management of AS depends on accurate quantification of severity.2,3 The AV and LVOT VTIs are routinely used to calculate aortic valve area (AVA) by the continuity equation, but peak velocities are often substituted2 based on the evidence that both AVA and DI, derived interchangeably from either VTIs or peak velocities, correlated well with the gold standard catheterization-derived AV area.15–22 However, in order to be clinically useful, a parameter must be accurate and reproducible.5,6 In this study, we show that peak velocity is considerably more reproducible than VTI.

DI, derived interchangeably from either VTIs or peak velocities, correlated well with the gold standard catheterization-derived AV area.15–22 However, in order to be clinically useful, a parameter must be accurate and reproducible.5,6 In this study, we show that peak velocity is considerably more reproducible than VTI. Comparison with previous studies The reproducibility of peak velocity and VTI has been previously studied with two operators.9–12 Just as a study that attempted to assess the average height of a population would measure more than two people, a study measuring the average performance of operators should ideally measure more than two operators. Our study is unique in testing reproducibility across a much larger group. The other benefit of measuring more than two operators is that it is worthwhile setting up a data collection system that allowed operators to make blinded reassessments, allowing us to study both intra-operator and inter-operator reproducibility and the mechanism of disagreement when tracing a VTI.

a much larger group. The other benefit of measuring more than two operators is that it is worthwhile setting up a data collection system that allowed operators to make blinded reassessments, allowing us to study both intra-operator and inter-operator reproducibility and the mechanism of disagreement when tracing a VTI. We found that the CW aortic peak velocity was more reproducible, with an intra-operator and inter-operator variability of 2.5% and 3.1% leading to an overall coefficient of variation of 4.0%. This is consistent with the values previously reported.9–12 Our results for PW LVOT peak velocity were less reproducible than the previous literature10 with an intra-operator and inter-operator variability of 5.6% and 8.2%, respectively leading to an overall coefficient of variation of 10.1%. CW VTI had intra-operator and inter-operator variability of 7.3% and 6.9% leading to an overall coefficient of variation of 10.2%. For PW VTI, this study shows worse reproducibility than previous studies; PW VTI had intra-operator and inter-operator variability of 11.9% and 12.9% leading to an overall coefficient of variation of 18.0% which is higher than the intra-operator and inter-operator variability previously demonstrated.10

iation of 10.2%. For PW VTI, this study shows worse reproducibility than previous studies; PW VTI had intra-operator and inter-operator variability of 11.9% and 12.9% leading to an overall coefficient of variation of 18.0% which is higher than the intra-operator and inter-operator variability previously demonstrated.10 A reason why dimensionless index works: systematic under-reading or over-reading of both AV and LVOT traces by individual operators This study confirms that DI shows better reproducibility across operators than would be expected from two peak or VTI measurements made in isolation. Some operators demonstrated a tendency to consistently make measurements which were smaller than or larger than the average for that case, as shown in Figure 3. When they interpret PW and CW images from the same patient, they show a consistent tendency to make measurements smaller than or larger than the average for that image, as shown in Figure 4. This is important, because it underlies some of the benefit of DI, which arises because an under-read in one image matched by an under-read in the other image will tend to cancel out and lead to an comparatively consistent DI.

ke measurements smaller than or larger than the average for that image, as shown in Figure 4. This is important, because it underlies some of the benefit of DI, which arises because an under-read in one image matched by an under-read in the other image will tend to cancel out and lead to an comparatively consistent DI. Figure 6 Variation arising from tracing of continuous wave (upper panels) and pulsed-wave (lower panels) velocity time integral. Each beat is divided into five columns of equal width. The variability is highest in the columns at the beginning and end of the traces. The left panels show the standard deviation for traces presented normally. The right panels show the standard deviation for traces when the horizontal (time) axis is reversed. Based on the mathematical principle of propagation of errors, the coefficient variation for DI can be estimated as the square root of the summed squares of the coefficients of variation of the two measurements forming the ratio.23 For peak measurements, the coefficient of variation of individual AV and LVOT measurements was 4.0% and 10.1%, respectively. The coefficient of variation of the resulting DI would therefore be expected to be ∼√(4.02 + 10.12), which is 10.9%, but we discovered it to be only 9.3%. Similarly, one might expect the VTI-derived DI based to have a coefficient of variation of 20.7%, but our data showed it to be only 17.1%. As shown in Figure 7, DI produces smaller coefficients of variation than would be expected from combining the two measurements.

which is 10.9%, but we discovered it to be only 9.3%. Similarly, one might expect the VTI-derived DI based to have a coefficient of variation of 20.7%, but our data showed it to be only 17.1%. As shown in Figure 7, DI produces smaller coefficients of variation than would be expected from combining the two measurements. Figure 7 Expected and observed proportional variance in dimensionless index. The area of the square represents the proportional variance. The orange area is the CW AV variance and the blue area the pulsed-wave LVOT variance. The top panel shows that the expected variance can be calculated from CW and PW variances (combined orange and blue). The actually observed variance (white) indicates the benefit that arises from dimensionless index. The middle panel shows the benefit of dimensionless index in VTI (coefficient of variation 20.7–17.1%). The bottom panel shows the benefit of dimensionless index in peak measurements (coefficient of variation 10.9–9.3%). Source of variability in tracing VTI Our analysis shows that most of the noise arising when measuring the VTI occurs at the beginning and the end of the Doppler trace. When compared to the middle of the wave, the variation from the consensus curve is larger at the beginning and end of the wave. Our experiment of reversing the images showed a corresponding reversal in the pattern of variability: it is not the act of deciding where to start tracing, but the steep slope of the Doppler trace which hinders reproducibility.

wave, the variation from the consensus curve is larger at the beginning and end of the wave. Our experiment of reversing the images showed a corresponding reversal in the pattern of variability: it is not the act of deciding where to start tracing, but the steep slope of the Doppler trace which hinders reproducibility. Limitations In this analysis, 25 operators viewed the same images. There is not the same as 25 operators acquiring their own images and then making measurements from them. The variability we demonstrate in this study is a lower limit estimate, since the acquisition of different images would add further variability but could not reduce it. We also selected only one beat for each patient, excluding beat-to-beat variability, which is another reason our result is a lower limit estimate. However, this study indicates that further work to characterize the variability arising from different operators making measurements or different operators choosing different beats should take place using peak rather than VTI measurements. Assessment of AS severity includes more than Doppler measurements. In the real world, clinicians integrate other imaging findings (such as the morphological appearances of the AV) and clinical information in their assessment. The relationship between the number of different pieces of information provided to operators and variability in their overall assessment of severity remains unknown.

he real world, clinicians integrate other imaging findings (such as the morphological appearances of the AV) and clinical information in their assessment. The relationship between the number of different pieces of information provided to operators and variability in their overall assessment of severity remains unknown. Conclusions Measuring the peak of a Doppler trace is a more reproducible strategy than measuring the VTI. The inferiority of VTI reproducibility arises mainly because of disagreement at the beginning and end of the tracing where the slope of the Doppler trace is steep. Individual operators show a tendency to over-read or under-read, which is responsible for some of the benefit of dimensionless index. The extent of superiority of peak over VTI for an individual patient is non-trivial: an average operator would be 95% sure of detecting a difference of 11.1% difference in peak velocity between two different images. For VTI, the same confidence would only arise with a much larger 28.3% change. Similarly, a clinical trial using a VTI as the endpoint would have to be more than six times larger than one using peak velocity. Acknowledgements We are grateful for colleagues at Imperial College Healthcare Trust who volunteered time to take part in this study.

The extent of superiority of peak over VTI for an individual patient is non-trivial: an average operator would be 95% sure of detecting a difference of 11.1% difference in peak velocity between two different images. For VTI, the same confidence would only arise with a much larger 28.3% change. Similarly, a clinical trial using a VTI as the endpoint would have to be more than six times larger than one using peak velocity. Acknowledgements We are grateful for colleagues at Imperial College Healthcare Trust who volunteered time to take part in this study. Funding S.S. was supported by a grant from SIC-MSD Italia-Merck Sharp and Dohme Corporation. N.D. and M.Z. are funded by the European Research Council (281524). M.J.S. is funded by the British Heart Foundation (FS/14/27/30752). D.P.F. is funded by the British Heart Foundation (FS/010/038). G.D.C. is funded by the British Heart Foundation (FS/12/12/29294). Conflict of interest: None declared.

Introduction Left ventricular (LV) thrombus (LVT) remains a life-threatening complication of myocardial infarction (MI), being associated with a five-fold increased risk of systemic embolism.1 The risk for LVT is greater with anterior MI, low ejection fraction (EF), LV aneurysms, and apical akinesis or dyskinesis,1,2 but LVT formation can also be found in patients with smaller infarcts, inferior infarcts, and only mild to moderate LV systolic dysfunction.3 The development of LVT is a complex process involving substrates of the Virchow's triad: disturbance of flow (stasis or turbulence), hypercoagulability, and endothelial injury/dysfunction. Early echocardiographic studies have demonstrated that abnormal flow patterns are associated with LVT.4–6 However, comprehensive insight into flow changes in post-MI patients with LVT is lacking, partly because tests capable of examining the complex 3D intra-cavity flow have not been available in the past. The development of four-dimensional (4D) flow cardiovascular magnetic resonance imaging (CMR) now allows mapping and quantification of intra-cavity LV flow kinetic energy (KE).7–12 LV blood flow KE appears to be reduced in patients with heart failure,13 and has the potential to provide new mechanistic insights into the pathophysiology of LVT formation in patients with ischaemic cardiomyopathy by detecting specific signatures of flow disturbance associated with LV flow stasis in LVT.

rgy (KE).7–12 LV blood flow KE appears to be reduced in patients with heart failure,13 and has the potential to provide new mechanistic insights into the pathophysiology of LVT formation in patients with ischaemic cardiomyopathy by detecting specific signatures of flow disturbance associated with LV flow stasis in LVT. The aim of this study was to use 4D flow CMR to map LV flow KE and characterize flow changes in patients with MI associated ischaemic cardiomyopathy with and without LVT. We hypothesized that patients with LVT show a re-distribution of LV flow KE resulting in reduced wash-in and wash-out of the LV. Furthermore, we aimed to investigate if LV flow KE mapping parameters are better associated with presence of LVT than the traditional risk factors for the development of LVT.

and without LVT. We hypothesized that patients with LVT show a re-distribution of LV flow KE resulting in reduced wash-in and wash-out of the LV. Furthermore, we aimed to investigate if LV flow KE mapping parameters are better associated with presence of LVT than the traditional risk factors for the development of LVT. Methods Study population This was a prospective cohort study of patients with MI and matched controls (Figure 1). Controls were recruited from two centres (Leeds and Leiden). They had no history or symptoms of cardiovascular disease, were on no cardiovascular or other relevant medication and had no contraindications to CMR. MI patients were recruited in Leeds and included both acute ST-elevation MI (STEMI) and chronic MI patients (MI > 3 months). Because of the relatively low incidence of LVT, we identified patients with LVT from routine clinical echocardiography lab and clinical CMR lists between January 2015 and April 2017 and in parallel recruited age and gender matched patients with MI but without LVT. Patients identified for this study were offered research CMR bolt-on scans immediately after clinical CMR or full protocol CMR if identified from echocardiography lab. All patients presenting to our CMR lab gave prospective consent prior to their clinical scans for the bolt-on research protocol. Figure 1 Study design. A, acute reperfused ST-elevation myocardial infarction; C, chronic myocardial infarction; CMR, cardiovascular magnetic resonance imaging; LGE, late gadolinium enhancement imaging; MI, myocardial infarction.

Methods Study population This was a prospective cohort study of patients with MI and matched controls (Figure 1). Controls were recruited from two centres (Leeds and Leiden). They had no history or symptoms of cardiovascular disease, were on no cardiovascular or other relevant medication and had no contraindications to CMR. MI patients were recruited in Leeds and included both acute ST-elevation MI (STEMI) and chronic MI patients (MI > 3 months). Because of the relatively low incidence of LVT, we identified patients with LVT from routine clinical echocardiography lab and clinical CMR lists between January 2015 and April 2017 and in parallel recruited age and gender matched patients with MI but without LVT. Patients identified for this study were offered research CMR bolt-on scans immediately after clinical CMR or full protocol CMR if identified from echocardiography lab. All patients presenting to our CMR lab gave prospective consent prior to their clinical scans for the bolt-on research protocol. Figure 1 Study design. A, acute reperfused ST-elevation myocardial infarction; C, chronic myocardial infarction; CMR, cardiovascular magnetic resonance imaging; LGE, late gadolinium enhancement imaging; MI, myocardial infarction. The inclusion criteria for acute STEMI patients were: first-time acute STEMI revascularized by primary percutaneous coronary intervention within 12 h of onset of chest pain. Acute STEMI patients were scheduled for CMR imaging within 72 h of indexed presentation. The inclusion criteria for chronic MI patients were: previous history of MI and presence of scar on late gadolinium enhancement (LGE) imaging. Exclusion criteria for all included the following: cardiomyopathy, atrial fibrillation, haemodynamic instability, and any contraindications to CMR.

72 h of indexed presentation. The inclusion criteria for chronic MI patients were: previous history of MI and presence of scar on late gadolinium enhancement (LGE) imaging. Exclusion criteria for all included the following: cardiomyopathy, atrial fibrillation, haemodynamic instability, and any contraindications to CMR. Ethical approval The study protocol was approved by the National Research Ethics Service (12/YH/0169) in the United Kingdom and the institutional Medical Ethical Committee (P11.136) in Leiden, The Netherlands. The study complied with the Declaration of Helsinki and all patients gave written informed consent. CMR examination All controls and patients underwent CMR imaging on identical 1.5 T systems at the two study sites (Ingenia, Philips, Best, The Netherlands) with 28-channel coils and digitization of the magnetic resonance signal in the receiver coil. CMR protocol and image acquisition The CMR protocol included survey, cines, early gadolinium enhancement imaging, LGE imaging, and at the end 4D flow CMR.14 4D flow acquisition For 4D flow, the field-of-view was planned in the trans-axial plane ensuring full cardiac coverage by adjusting the number of slices. A free-breathing, non-respiratory navigated, Echo-Planer Imaging (EPI)- accelerated 4D flow sequence was used.14 4D flow data reconstruction and error corrections are detailed in the Supplementary data online, Document S1.

lanned in the trans-axial plane ensuring full cardiac coverage by adjusting the number of slices. A free-breathing, non-respiratory navigated, Echo-Planer Imaging (EPI)- accelerated 4D flow sequence was used.14 4D flow data reconstruction and error corrections are detailed in the Supplementary data online, Document S1. Image analysis All images were evaluated offline using in-house developed research software (MASS; Version 2017-EXP, Leiden University Medical Center, Leiden, The Netherlands). For functional and flow analysis, anonymised cine/4D flow CMR data were shared with the core lab at Leiden (Figure 1) and tissue characterisation was done at the main clinical acquisition site (Leeds) and remained blinded to core lab functional/flow analysis. Methods of analysis are descripted in the Supplementary data online, Document S1. KE mapping LV was contoured manually in the images of the short-axis cine acquisition (Figure 2A). Correction for translational and rotational misalignment between the short-axis cine and the 4D flow CMR acquisition was performed using automated image registration as previously described.15 Figure 2 (A) Left ventricular short-axis endocardial segmentation in patient with LV thrombus. Intra-cavity thrombus was manually contoured (orange contour) to avoid under-estimation of LV KE parameters. Intra-cavity KE of blood is demonstrated at peak late ventricular filling (peak A-wave). (B) Illustration of KE curve demonstrating majority of the KE parameters studied.

egmentation in patient with LV thrombus. Intra-cavity thrombus was manually contoured (orange contour) to avoid under-estimation of LV KE parameters. Intra-cavity KE of blood is demonstrated at peak late ventricular filling (peak A-wave). (B) Illustration of KE curve demonstrating majority of the KE parameters studied. For calculation of LV blood flow KE parameters, the LV volumetric mesh was resliced into short-axis sections of 2 mm thickness and pixel spacing equal to the original reconstructed pixel size of the short-axis cine acquisition (1.0–1.2 mm). For each volumetric element (voxel) the KE was computed as KE=12 ρblood . Vvoxel . v2, with ρblood being the density of blood (1.06 g/cm3), Vvoxel the voxel volume, and v the velocity magnitude. For each phase, the total KE within the LV was obtained by summation of the KE of every voxel. In addition, the KE was computed for basal, mid and apical LV level by dividing the LV into equal thirds. Similarly, the in-plane component of KE (i.e. the short-axis plane) was computed, by taking for v the magnitude of the in-plane component of velocity. KE parameters were normalized to the LV end-diastolic volume and reported in μJ/mL (KEiEDV).

ted for basal, mid and apical LV level by dividing the LV into equal thirds. Similarly, the in-plane component of KE (i.e. the short-axis plane) was computed, by taking for v the magnitude of the in-plane component of velocity. KE parameters were normalized to the LV end-diastolic volume and reported in μJ/mL (KEiEDV). In patients with LVT, LVT was excluded from KE mapping to reduce under-estimation of LV flowing blood KE (Figure 2A). The global, wash-in, and wash-out components of the LV KE that were mapped are described in Figure 2B. During diastolic filling, the flow velocity and its associated KE of early (E-wave) and late (A-wave) filling decrease from base to apex. In this study, this drop has been measured as a relative drop in percentage from each level (base to mid-ventricle; mid-ventricle to apex) for both E-wave flow KE and A-wave flow KE. Higher relative-drop of LV in-flow KE signifies reduced wash-in of the LV. In-plane KE The in-plane KE is the sum of all KE in the x-y direction, in the short-axis LV from base to apex. In this study, the in-plane KE is represented as a percentage of the total LV KE. This parameter was computed mainly to better understand the in-plane flow dynamics within the LV cavity.

In patients with LVT, LVT was excluded from KE mapping to reduce under-estimation of LV flowing blood KE (Figure 2A). The global, wash-in, and wash-out components of the LV KE that were mapped are described in Figure 2B. During diastolic filling, the flow velocity and its associated KE of early (E-wave) and late (A-wave) filling decrease from base to apex. In this study, this drop has been measured as a relative drop in percentage from each level (base to mid-ventricle; mid-ventricle to apex) for both E-wave flow KE and A-wave flow KE. Higher relative-drop of LV in-flow KE signifies reduced wash-in of the LV. In-plane KE The in-plane KE is the sum of all KE in the x-y direction, in the short-axis LV from base to apex. In this study, the in-plane KE is represented as a percentage of the total LV KE. This parameter was computed mainly to better understand the in-plane flow dynamics within the LV cavity. Time difference We also computed the time difference (TD) to peak early mitral in-flow velocity (E-wave) from the base of the LV to mid-ventricle. This transit time or TD should be higher if the mitral valve propagation velocity (Vp), as measured by M-mode echocardiography is lower. Hence, the transit time of the peak KE from base to mid-ventricle, described as the TD in this study, may represent a novel marker of delayed filling. A detailed description of the CMR protocol, pulse sequences, and the intra-/inter-observer reproducibility test are given in the Supplementary data online, Document S1.

Time difference We also computed the time difference (TD) to peak early mitral in-flow velocity (E-wave) from the base of the LV to mid-ventricle. This transit time or TD should be higher if the mitral valve propagation velocity (Vp), as measured by M-mode echocardiography is lower. Hence, the transit time of the peak KE from base to mid-ventricle, described as the TD in this study, may represent a novel marker of delayed filling. A detailed description of the CMR protocol, pulse sequences, and the intra-/inter-observer reproducibility test are given in the Supplementary data online, Document S1. Statistical analysis Statistical analysis was performed using IBM SPSS® Statistics 23.0. Quantitative parameters are presented as mean ± standard deviation or median and interquartile ranges, where appropriate. Demographic comparisons were performed with post hoc analysis of variance (ANOVA) with Bonferroni corrections. Imaging data was handled as non-parametric. Step-wise multivariate logistic regression was used for clinical, functional, and KE parameters with statistical significance from one-way analysis (P < 0.1). Diagnostic performance tests were done using the receiver-operator characteristic. To avoid collinearity issues within volumetric parameters, only LV EF was included in the multivariate analysis. A P-value <0.05 was considered statistically significant. Sample size calculations are described in the Supplementary data online, Document S1.

Statistical analysis Statistical analysis was performed using IBM SPSS® Statistics 23.0. Quantitative parameters are presented as mean ± standard deviation or median and interquartile ranges, where appropriate. Demographic comparisons were performed with post hoc analysis of variance (ANOVA) with Bonferroni corrections. Imaging data was handled as non-parametric. Step-wise multivariate logistic regression was used for clinical, functional, and KE parameters with statistical significance from one-way analysis (P < 0.1). Diagnostic performance tests were done using the receiver-operator characteristic. To avoid collinearity issues within volumetric parameters, only LV EF was included in the multivariate analysis. A P-value <0.05 was considered statistically significant. Sample size calculations are described in the Supplementary data online, Document S1. Results Demographic characteristics We identified 135 subjects for this study, 23 did not meet the eligibility criteria and 4 were claustrophobic. Hence, 108 subjects completed the study (Figure 1). These included 40 controls (Leiden = 13, Leeds = 27), 36 LVT− patients and 32 LVT+ patients. From the 32 LVT+ patients, 5 LVT+ patients (16%) were identified at echocardiography lab and the rest were identified at the CMR lab. Patients on anti-coagulation had already been diagnosed with LVT by echocardiography.

ure 1). These included 40 controls (Leiden = 13, Leeds = 27), 36 LVT− patients and 32 LVT+ patients. From the 32 LVT+ patients, 5 LVT+ patients (16%) were identified at echocardiography lab and the rest were identified at the CMR lab. Patients on anti-coagulation had already been diagnosed with LVT by echocardiography. All subjects had comparable heart rates (P > 0.05) (Table 1). Heart failure status was comparable between the two patient groups. Patients with LVT were more likely to have diabetes (P = 0.01) and to be on anti-coagulation than patients without LVT (P = 0.04). Table 1 Study demographics (study population = 108)

ure 1). These included 40 controls (Leiden = 13, Leeds = 27), 36 LVT− patients and 32 LVT+ patients. From the 32 LVT+ patients, 5 LVT+ patients (16%) were identified at echocardiography lab and the rest were identified at the CMR lab. Patients on anti-coagulation had already been diagnosed with LVT by echocardiography. All subjects had comparable heart rates (P > 0.05) (Table 1). Heart failure status was comparable between the two patient groups. Patients with LVT were more likely to have diabetes (P = 0.01) and to be on anti-coagulation than patients without LVT (P = 0.04). Table 1 Study demographics (study population = 108) Younger controls (n = 24) Age-matched controls (n = 16) LVT− (n = 36) LVT+ (n = 32) P-valuea P-valueb P-valuec Baseline characteristics Age (years) 30 ± 10 57 ± 7 60 ± 9 61 ± 13 <0.01 0.7 1 Sex (female) 7 8 8 3 0.2 0.3 0.4 Body surface area (m²) 1.9 ± 0.2 1.8 ± 0.2 1.9 ± 0.2 2 ± 0.2 0.2 0.03 0.67 Smoker 0 0 23 13 <0.01 0.07 Hypertension 0 0 8 9 0.02 1 Hypercholesterolaemia 0 0 16 8 <0.01 0.14 Diabetes 0 0 2 8 0.48 0.01 Baseline clinical parameters Systolic BP (mmHg) 131 ± 34 137 ± 16 0.5 Heart rate (b.p.m.) 64 ± 7 65 ± 15 65.5 ± 9 66 ± 12 1 0.10 0.3 KC 1 33 25 0.12 KC 2 3 7 0.12 Medical therapy at the time of recruitment ACE-inhibitor 35 23 0.11 HMG-CoA reductase inhibitors 35 20 0.07 β-blockers 35 23 0.11 Aspirin 35 20 0.07 Anti-coagulation 1 4 0.04 Blood results Haemoglobin (g/dL) 144 ± 14 141 ± 12 0.99 eGFR 83.7 ± 9 77 ± 17 0.08 C-reactive protein (mg/L) 36 ± 38 24 ± 22 0.55 HBA1c (mmol/mol) 41.5 ± 11 47 ± 18 0.28 Data are presented as mean ± standard deviation or count (n).

inhibitors 35 20 0.07 β-blockers 35 23 0.11 Aspirin 35 20 0.07 Anti-coagulation 1 4 0.04 Blood results Haemoglobin (g/dL) 144 ± 14 141 ± 12 0.99 eGFR 83.7 ± 9 77 ± 17 0.08 C-reactive protein (mg/L) 36 ± 38 24 ± 22 0.55 HBA1c (mmol/mol) 41.5 ± 11 47 ± 18 0.28 Data are presented as mean ± standard deviation or count (n). BP, blood pressure; CAD, coronary artery disease; KC, heart failure Killip class. a Younger controls vs. older age-matched to patient controls. b Age-matched controls vs. LVT−. c LVT− vs. LVT+. Baseline CMR Patients with LVT were more likely to have anterior MI than those without LVT (87% vs. 61%, P = 0.01). Infarct size was comparable between patients with/without LVT (P = 0.6) (Figure 3 and Table 2). Also, for all four apical segments, scar transmurality was not different between patients with/without LVT (P > 0.05) (Supplementary data online, Table S4). Apical RWM-abnormality score showed a lower trend in LVT+ vs. LVT− but did not achieve statistical significance (P = 0.055). EF was significantly lower in patients with LVT and end-diastolic/end-systolic volumes and mass were significantly increased in patients with LVT compared with patients without LVT. Table 2 Left ventricular baseline volumetric and haemodynamic study

VT+ vs. LVT− but did not achieve statistical significance (P = 0.055). EF was significantly lower in patients with LVT and end-diastolic/end-systolic volumes and mass were significantly increased in patients with LVT compared with patients without LVT. Table 2 Left ventricular baseline volumetric and haemodynamic study Younger controls (n = 24) Age-matched controls (n = 16) LVT− (n = 36) LVT+ (n = 32) P-valuea P-valueb P-valuec Volumetric assessment Anterior infarction (n) 22 28 0.01 Infarct size (% of LV) 21 ± 14 24 ± 12 0.6 Apical RWMA score 3 ± 1 4 ± 0.5 0.055 Stroke volume (mL) 105 ± 30 93 ± 28 77 ± 26 75 ± 31 <0.01 0.06 0.42 Ejection fraction (%) 62 ± 7 64 ± 5 44 ± 11 32 ± 17 0.13 <0.01 <0.01 LVEDVi (mL/m2) 91 ± 19 75 ± 21 91 ± 16 100 ± 43 <0.01 <0.01 0.01 LVESVi (mL/m2) 35 ± 11 27 ± 6 52 ± 15 67 ± 41 <0.01 <0.01 <0.01 LVMi (g/m2) 52 ± 15 49 ± 9 56 ± 13 68 ± 21 0.44 0.03 <0.01 Haemodynamic assessment E-wave velocity (cm/s)d 59 ± 44 48 ± 22 26 ± 9 27 ± 10 0.08 <0.01 0.66 A-wave velocity (cm/s)d 32 ± 19 34 ± 13 24 ± 8 24 ± 8 0.65 <0.01 0.21 E/A ratio 2 ± 0.6 1.3 ± 0.4 1.3 ± 0.7 1.2 ± 0.8 0.09 0.71 0.71 MDT (ms) 147 ± 35 140 ± 26 146 ± 30 137 ± 42 0.7 0.4 0.81 Mitral regurgitation (mL) 0 ± 0.01 0 ± 0.02 3 ± 4 4 ± 3 0.33 0.01 0.09 Data are presented as median ± interquartile range or count (n). LV measurements are indexed to body surface area.

65 <0.01 0.21 E/A ratio 2 ± 0.6 1.3 ± 0.4 1.3 ± 0.7 1.2 ± 0.8 0.09 0.71 0.71 MDT (ms) 147 ± 35 140 ± 26 146 ± 30 137 ± 42 0.7 0.4 0.81 Mitral regurgitation (mL) 0 ± 0.01 0 ± 0.02 3 ± 4 4 ± 3 0.33 0.01 0.09 Data are presented as median ± interquartile range or count (n). LV measurements are indexed to body surface area. LVEDVi, left ventricular end-diastolic volume (indexed); LVESVi, left ventricular end-systolic volume (indexed); LVMi, left ventricular mass (indexed); MDT, mitral deceleration time; MV, mitral valve; RWMA, regional-wall motion abnormality score (1 = normal, 2 = hypokinaesia, 3 = akinetic, 4 = diskinetic). a Younger controls vs. older age-matched to patient controls. b Age-matched controls vs. LVT−. c LVT− vs. LVT+. d These peak inflow velocities are average peak velocities for the full mitral annular flow. Figure 3 Late gadolinium enhancement imaging of two case examples from the study population with anterior MI. Infarct characteristics alone did not differentiate the two cases for LVT. Left ventricular KE flow analysis demonstrated rise in in-plane, rotational KE, and reduced A-wave wash-in of the LV. EF, ejection fraction; KE, kinetic energy; LVT, left ventricular thrombus. Haemodynamic analysis Patients with infarct were more likely to have mitral regurgitation (MR) (P = 0.01). However, MR was similar in both patient groups (P = 0.09). No other mitral in-flow diastolic function parameter demonstrated any significant difference between the two MI groups.

Figure 3 Late gadolinium enhancement imaging of two case examples from the study population with anterior MI. Infarct characteristics alone did not differentiate the two cases for LVT. Left ventricular KE flow analysis demonstrated rise in in-plane, rotational KE, and reduced A-wave wash-in of the LV. EF, ejection fraction; KE, kinetic energy; LVT, left ventricular thrombus. Haemodynamic analysis Patients with infarct were more likely to have mitral regurgitation (MR) (P = 0.01). However, MR was similar in both patient groups (P = 0.09). No other mitral in-flow diastolic function parameter demonstrated any significant difference between the two MI groups. Thrombus characteristics In the 36 patients recruited with LVT, 26 (72%) patients had mural thrombus, 6 (17%) had mobile thrombus and 10 (27%) had protruding thrombus. Average indexed volume of thrombus was 4.9 ± 10 mL/m2. Thrombus characteristics and its association to flow mapping are detailed in the Supplementary data online, Table S5.

tics In the 36 patients recruited with LVT, 26 (72%) patients had mural thrombus, 6 (17%) had mobile thrombus and 10 (27%) had protruding thrombus. Average indexed volume of thrombus was 4.9 ± 10 mL/m2. Thrombus characteristics and its association to flow mapping are detailed in the Supplementary data online, Table S5. KE mapping LV KEiEDV averaged over the complete cardiac cycle was significantly lower in both patient groups vs. healthy controls (P < 0.05) (Figure 4). Similarly, average systolic KEiEDV and peak E-wave KEiEDV were significantly lower in patients than in age-matched controls (Table 3). Peak late filling (A-wave) KEiEDV was not different between the three groups (P > 0.05). The proportion of in-plane KE of the LV was not different between controls and LVT− patients (P = 0.82). However, patients with LVT demonstrated a significantly higher proportion of in-plane KE vs. LVT− (40% vs. 36%, P = 0.02) (Figure 5). Table 3 Detailed mapping of left ventricular kinetic energy

ee groups (P > 0.05). The proportion of in-plane KE of the LV was not different between controls and LVT− patients (P = 0.82). However, patients with LVT demonstrated a significantly higher proportion of in-plane KE vs. LVT− (40% vs. 36%, P = 0.02) (Figure 5). Table 3 Detailed mapping of left ventricular kinetic energy Younger controls (n = 24) Age-matched controls (n = 16) LVT− LVT+ P-valuea P-valueb P-valuec Global LV kinetic energy LVd 10 ± 4 8.3 ± 1.5 6.3 ± 2 5.6 ± 2 0.63 0.01 0.19 Minimald 1 ± 0.4 0.9 ± 0.5 0.7 ± 0.56 0.8 ± 0.61 0.64 0.07 0.19 Systolicd 10 ± 3 9 ± 4 6.6 ± 2 6.7 ± 3 0.98 <0.01 0.65 Diastolicd 10 ± 4 8 ± 2 5.9 ± 3 5.4 ± 2 0.47 0.07 0.31 Peak E-waved 25 ± 13 22 ± 12 12.4 ± 7 10.8 ± 8 0.01 <0.01 0.54 Peak A-waved 9 ± 5 13 ± 10 11 ± 5 9.5 ± 5 0.01 0.15 0.08 In-plane KE (%) 33 ± 10 37 ± 6 36 ± 7 40 ± 5 0.27 0.82 0.02 Relative KE drop from base to apex (%) E-wave (B→M) 60 ± 18 52 ± 12 56 ± 16 65 ± 27 0.18 0.42 0.22 A-wave (B→M) 69 ± 12 68 ± 2 64 ± 17 60 ± 33 0.36 0.69 0.08 E-wave (M→A) 88 ± 8 89 ± 5 89 ± 9 89 ± 10 0.44 0.7 0.38 A-wave (M→A) 78 ± 10 78 ± 11 78 ± 14 87 ± 9 0.34 0.27 <0.01 Data are presented as median ± interquartile range or count (n). Unit of normalized kinetic energy: μJ/mL. B→M, base to mid-ventricle; KE, kinetic energy; M→A, mid-ventricle to apex. a Younger controls vs. older age-matched to patient controls. b Age-matched controls vs. LVT−. c LVT− vs. LVT+. d KEiEDV variables (μJ/mL).

Younger controls (n = 24) Age-matched controls (n = 16) LVT− LVT+ P-valuea P-valueb P-valuec Global LV kinetic energy LVd 10 ± 4 8.3 ± 1.5 6.3 ± 2 5.6 ± 2 0.63 0.01 0.19 Minimald 1 ± 0.4 0.9 ± 0.5 0.7 ± 0.56 0.8 ± 0.61 0.64 0.07 0.19 Systolicd 10 ± 3 9 ± 4 6.6 ± 2 6.7 ± 3 0.98 <0.01 0.65 Diastolicd 10 ± 4 8 ± 2 5.9 ± 3 5.4 ± 2 0.47 0.07 0.31 Peak E-waved 25 ± 13 22 ± 12 12.4 ± 7 10.8 ± 8 0.01 <0.01 0.54 Peak A-waved 9 ± 5 13 ± 10 11 ± 5 9.5 ± 5 0.01 0.15 0.08 In-plane KE (%) 33 ± 10 37 ± 6 36 ± 7 40 ± 5 0.27 0.82 0.02 Relative KE drop from base to apex (%) E-wave (B→M) 60 ± 18 52 ± 12 56 ± 16 65 ± 27 0.18 0.42 0.22 A-wave (B→M) 69 ± 12 68 ± 2 64 ± 17 60 ± 33 0.36 0.69 0.08 E-wave (M→A) 88 ± 8 89 ± 5 89 ± 9 89 ± 10 0.44 0.7 0.38 A-wave (M→A) 78 ± 10 78 ± 11 78 ± 14 87 ± 9 0.34 0.27 <0.01 Data are presented as median ± interquartile range or count (n). Unit of normalized kinetic energy: μJ/mL. B→M, base to mid-ventricle; KE, kinetic energy; M→A, mid-ventricle to apex. a Younger controls vs. older age-matched to patient controls. b Age-matched controls vs. LVT−. c LVT− vs. LVT+. d KEiEDV variables (μJ/mL). Figure 4 LV flow component KE curves in different study subjects. (A) KE curves for total and in-plane LV KE. The minimal KE is marked by orange arrows. (B) Regional KE curves for base, mid, and apex of the LV. Time differences to peak E-wave KE propagation also demonstrated progressive increase (gapes between dotted lines marked by red arrows). Blue arrows point to drop in A-wave KE from mid-ventricle to apex. KE, kinetic energy; LV, left ventricle; LVT−, patients without LV thrombus; LVT+, patients with LV thrombus; MI, myocardial infarction.

peak E-wave KE propagation also demonstrated progressive increase (gapes between dotted lines marked by red arrows). Blue arrows point to drop in A-wave KE from mid-ventricle to apex. KE, kinetic energy; LV, left ventricle; LVT−, patients without LV thrombus; LVT+, patients with LV thrombus; MI, myocardial infarction. Figure 5 Bar graph of the main LV blood flow kinetic energy parameters which demonstrated significant changes in patients with/without thrombus (whiskers: interquartile range). KE, kinetic energy; LVT−, patients without LV thrombus; LVT+, patients with LV thrombus. The relative drop of proximal and distal, intra-ventricular E-wave KE, did not differ across the three groups (P > 0.05) (Table 3). Peak A-wave KE drop from mid-ventricle to apex was not different in controls and LVT− patients (P = 0.69). However, LVT+ patients demonstrated a significantly higher drop in A-wave KE from mid to apex when compared with LVT− patients (87% vs. 78%, P < 0.01) (Figure 5). The TD of peak E-wave KE propagation from base to mid-ventricle demonstrated rise between all the four groups of subjects—younger controls = 14 ± 15 ms vs. age-matched older controls = 18 ± 28 ms (P = 0.52); age-matched older controls = 18 ± 28 ms vs. LVT− = 38 ± 38 ms (P = 0.07); LVT− = 38 ± 38 ms vs. LVT+ = 62 ± 56 ms (P = 0.04) with an overall ANOVA between groups of <0.001) (Figure 5).

ated rise between all the four groups of subjects—younger controls = 14 ± 15 ms vs. age-matched older controls = 18 ± 28 ms (P = 0.52); age-matched older controls = 18 ± 28 ms vs. LVT− = 38 ± 38 ms (P = 0.07); LVT− = 38 ± 38 ms vs. LVT+ = 62 ± 56 ms (P = 0.04) with an overall ANOVA between groups of <0.001) (Figure 5). Logistic regression In univariate analysis, the following parameters were associated with the presence of LVT: history of diabetes, anterior MI, EF, drop of A-wave KE from mid-ventricle to apex, proportion of LV in-plane KE, and TD of peak early-filling (E-wave) KE from base to mid-ventricle (Table 4). In multivariate analysis, only distal drop of A-wave KE (beta = 11.5, P = 0.002) and EF (beta = −0.08, P = 0.01) demonstrated independent association with LVT. A combined CMR model of EF and relative drop in A-wave KE demonstrated significantly larger area under the curve than LV EF [difference in AUC = 0.11, 95% confidence interval (CI) 0.1–0.23; P = 0.02] and infarct size (difference in AUC = 0.26, 95% CI 0.1–0.4; P = 0.02) (Figure 6). Table 4 Logistic regression analysis of variables which influence presence of LVT Univariate Multivariate Beta SD P-value Beta SD P-value DM −1.8 0.8 0.03 0.14 Anterior MI −1.5 0.6 0.02 0.13 Ejection fraction −0.81 0.03 <0.01 −0.08 0.03 0.01 Distal A-wave KE drop 8.4 2.9 <0.01 11.5 3.7 0.002 In-plane KE 0.13 0.06 0.02 0.19 Time difference 0.011 0.006 0.04 0.74 Bold results are highlight independent predictors in the regression model. DM, diabetes mellitus; KE, kinetic energy; MI, myocardial infarction.

Univariate Multivariate Beta SD P-value Beta SD P-value DM −1.8 0.8 0.03 0.14 Anterior MI −1.5 0.6 0.02 0.13 Ejection fraction −0.81 0.03 <0.01 −0.08 0.03 0.01 Distal A-wave KE drop 8.4 2.9 <0.01 11.5 3.7 0.002 In-plane KE 0.13 0.06 0.02 0.19 Time difference 0.011 0.006 0.04 0.74 Bold results are highlight independent predictors in the regression model. DM, diabetes mellitus; KE, kinetic energy; MI, myocardial infarction. Figure 6 Receiver operating curve analysis for the presence of LVT. Intra-/inter-observer reliability Global LV KE parameters demonstrated very low bias (intra: average 2%; inter: average 4%) and good precision (intra: −16% to −20%; inter: −21% to −13%). Inter-rater reliability of main KE parameters thresholds were good (in-plane KE >37%; weighted-kappa = 1, distal A-wave KE drop >85%; weighted-kappa = 0.63, and TD from base to mid >31 ms; weighted-kappa = 0.67). Comprehensive results for individual parameters can be found in the Supplementary data online, Document S1. Discussion This study provides mechanistic insights into intra-cavity LV flow disturbances in MI patients with and without LVT. Firstly, we demonstrate that global LV KEiEDV is reduced in MI patients compared with healthy, age-matched controls. Secondly, MI patients with LVT demonstrated reduced wash-in of blood to the distal LV during late diastole, detected by the prominent drop of A-wave KE from the mid-ventricle to the apex. This parameter of LV blood flow disturbance was most strongly associated with the presence of LVT.

th healthy, age-matched controls. Secondly, MI patients with LVT demonstrated reduced wash-in of blood to the distal LV during late diastole, detected by the prominent drop of A-wave KE from the mid-ventricle to the apex. This parameter of LV blood flow disturbance was most strongly associated with the presence of LVT. Precursor to LV flow stasis Blood stasis in the LV is the hallmark of LVT formation. In the detailed mapping of LV flow KE, we noted that the global LV KE parameters are significantly altered in health vs. MI patients. LV KEiEDV averaged over the complete cardiac cycle, systolic KEiEDV and peak E-wave KEiEDV were all significantly reduced in patients with MI. Even though these global LV flow KE parameters were not significantly different in patients with/without thrombus, they are likely to be the key initial substrates of flow stasis post-MI. Reduced and delayed diastolic wash-in of the LV During early and late diastole, blood flow into the LV cavity takes very little time due to intra-ventricular pressure gradients. In this study, the TD showed an increase from healthy controls to LVT− and further to LVT+ patients, demonstrating that patients with LVT have significantly delayed wash-in of the LV.

the LV During early and late diastole, blood flow into the LV cavity takes very little time due to intra-ventricular pressure gradients. In this study, the TD showed an increase from healthy controls to LVT− and further to LVT+ patients, demonstrating that patients with LVT have significantly delayed wash-in of the LV. Additionally, the relative drop of distal A-wave KE was significantly higher in MI patients with LVT vs. patients without LVT. This finding suggests that there is a reduction in distal intra-ventricular pressure gradients due to a relative increase in distal (apical) pressures within the LV cavity in patients with LVT and that the late filling phase of diastole plays an important role in reduced wash-in of the LV. The in-plane KE of blood flow was significantly higher in MI patients with LVT than those without LVT. Our data support the notion that an increase in the in-plane flow will reduce the proportion of through-plane flow in the LV cavity, and thus less blood will pass through the ventricle per-unit-time resulting in reduced global wash-in and wash-out of the LV. In addition to lower wash-in and wash-out, such non-physiological in-plane flow may exert strain on the LV wall, resulting in more dilatation and increase in endothelial dysfunction in the endocardium, similar to the vascular system.16 An increased in-plane rotational component of the intra-cavity LV flow may also increase the shear stress on the platelets and activate them, which would promote thrombosis.17

train on the LV wall, resulting in more dilatation and increase in endothelial dysfunction in the endocardium, similar to the vascular system.16 An increased in-plane rotational component of the intra-cavity LV flow may also increase the shear stress on the platelets and activate them, which would promote thrombosis.17 Traditional risk factors Akin to published studies in patients with LVT, our study also demonstrated the association of infarct location and depressed EF to LVT.2,18,19 However, this study failed to associate infarct size with presence of LVT (P = 0.82). This is possibly because we included patients with chronic infarction in both the MI groups. In chronic infarcts, the infarct size substantially decreases from the acute stage, which may lessen the overall impact of infarct size. In addition, even though there was a trend of apical regional wall motion score to be higher in LVT− patients, this study did not demonstrate any significant changes in LVT+ patients. This may be explained by the fact that in a previous study by Keren et al.,20 none of the inferior MI patients had thrombus, whereas this study recruited 12.5% inferior/posterior MI who had LVT. In this study, MI patients with diabetes were more likely to have LVT (P < 0.01). This finding is likely due to under representation of patients with diabetes in the LVT− MI cohort as observational studies have demonstrated prevalence of diabetes is around 24–36% in MI.21

ruited 12.5% inferior/posterior MI who had LVT. In this study, MI patients with diabetes were more likely to have LVT (P < 0.01). This finding is likely due to under representation of patients with diabetes in the LVT− MI cohort as observational studies have demonstrated prevalence of diabetes is around 24–36% in MI.21 LVT characteristics and associated flow changes LVT volume was the only parameter which had some association to flow characterisation (minimal KE, in-plane KE, and TD flow parameters). Mobility, produrance, and murality of the thrombus did not demonstrate any significant flow association. We speculate this may be because thrombus characteristics change rapidly after MI and probably depend on the timing of the imaging.

ciation to flow characterisation (minimal KE, in-plane KE, and TD flow parameters). Mobility, produrance, and murality of the thrombus did not demonstrate any significant flow association. We speculate this may be because thrombus characteristics change rapidly after MI and probably depend on the timing of the imaging. Limitations This study was a prospective cohort study, hence our results cannot be used to determine the prevalence of thrombus in MI. Additionally, we studied differences in flow patterns in the presence of LVT and not prior to its genesis, and a prospective evaluation of the parameters tested in this study is required. Arrhythmias can introduce errors in 4D flow analysis. To reduce these errors, we performed robust quality checks on all the data. Additionally, we used retrospectively gated acquisition sequence for 4D flow to reduce time blurring.14 The LV geometry was defined by LV cine stack which was done using breath-hold technique while the 4D flow was done using free breathing. Hence, although spatial miss-registration was corrected for, other issues still remain including difference in heart rate and physiological conditions. This may have impact on the time-varying flow characteristics which could not be corrected for. The temporal resolution of the 4D flow was 40 ms, which may affect the overall quality of TD assessment and make them less reliable.

other issues still remain including difference in heart rate and physiological conditions. This may have impact on the time-varying flow characteristics which could not be corrected for. The temporal resolution of the 4D flow was 40 ms, which may affect the overall quality of TD assessment and make them less reliable. Conclusions This study provides mechanistic insights into disturbed flow patterns in MI patients with and without thrombus. Patients with MI have reduced global LV KE and MI patients with LVT have evidence of reduced wash-in of the LV. Among all imaging biomarkers, the relative drop of distal intra-ventricular A-wave KE, which represents the distal late-diastolic wash-in of the LV, was most strongly associated with the presence of LVT. Future studies need to evaluate the prognostic significance of blood flow KE changes in the LV in patients with LVT. Supplementary Material Supplementary Data Click here for additional data file. Acknowledgements We thank Gavin Bainbridge, Caroline Richmond, Margaret Saysell, and Petra Bijsterveld for their assistance in recruitment. Funding This work was supported by the British Heart Foundation [FS/10/62/28409 to S.P.] and Dutch ZonMw [104003001 to J.W]. Conflict of interest: None declared.

[Eur Heart J Cardiovasc Imaging 2019;20:108–17] There were three errors in the originally published version: 1. Affiliation 3 was incorrect in the original version and should have read ‘Radboudumc, Department of Cardiology, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.’ 2. Under the heading ‘CMR examination’, the abbreviation ‘MR’ has been spelt out as ‘mitral regurgitation’ but should have read ‘magnetic resonance’. 3. Under the heading ‘Haemodynamic analysis’, the abbreviation ‘MR’ stands for ‘mitral regurgitation’.

This editorial refers to ‘Intramyocardial hemorrhage and prognosis after ST-elevation myocardial infarction’ by S.J. Reinstadler et al., pp. 138–146. Restoration of coronary blood flow with primary percutaneous coronary intervention (PCI) is an effective treatment for ST-segment elevation myocardial infarction (STEMI), and primary PCI is the evidence-based standard of care for STEMI patients presenting within 12 h of symptom onset.1 On the other hand, restoration of epicardial blood flow results in reperfusion injury with failed myocardial perfusion in approximately 50% of patients,2 typically in the context of a successful primary PCI procedure. Procedure success defined as normal antegrade coronary blood flow is achieved in >95% of patients during daily practice.3,4 Failed myocardial reperfusion is a complex, heterogeneous microvascular problem. Several mechanisms have been implicated, including intra-vascular problems, such as distal embolization of thrombus/atheroma and extravascular problems, such as extrinsic microvascular compression due to intracellular (e.g. cardiomyocyte) and extracellular oedema.5 Taken together, these pathologies manifest clinically as microvascular obstruction (MVO).

ated, including intra-vascular problems, such as distal embolization of thrombus/atheroma and extravascular problems, such as extrinsic microvascular compression due to intracellular (e.g. cardiomyocyte) and extracellular oedema.5 Taken together, these pathologies manifest clinically as microvascular obstruction (MVO). Endothelial cells may be more resistant to ischaemia than the cardiac myocte,6 but eventually sustained ischaemia leads to endothelial dysfunction. Endothelial damage leads to impaired capillary integrity, tissue oedema and extravasation of red bloods cells into the extracellular space. Multiple studies have shown that MVO and intramyocardial haemorrhage (IMH) are closely related. In general, IMH does not occur in the absence of MVO but, on the other hand, MVO commonly occurs in the absence of IMH.2 The dynamic nature of MVO supports the concept that it may be reversible and thus a therapeutic target. On the other hand, IMH is a downstream pathological consequence of irreversible microvascular damage.7 The occurrence of IMH therefore represents failed myocardial reperfusion, and a failure of the therapeutic strategy.

dynamic nature of MVO supports the concept that it may be reversible and thus a therapeutic target. On the other hand, IMH is a downstream pathological consequence of irreversible microvascular damage.7 The occurrence of IMH therefore represents failed myocardial reperfusion, and a failure of the therapeutic strategy. MVO is a predictor of poor outcome independent of infarct size.8 Patients with MVO are more likely to develop heart failure post-MI with increased mortality. The prognostic significance of IMH has been the subject of much debate. In a study of 286 patients presenting with acute STEMI, we found that myocardial haemorrhage (identified by T2* imaging) was more closely associated with all-cause death and heart failure during 2.3 years follow-up when compared with MVO alone.2

y. The prognostic significance of IMH has been the subject of much debate. In a study of 286 patients presenting with acute STEMI, we found that myocardial haemorrhage (identified by T2* imaging) was more closely associated with all-cause death and heart failure during 2.3 years follow-up when compared with MVO alone.2 The pathophysiological mechanisms linking IMH with worse outcomes independent of infarct size and MVO are incompletely understood. Key to this may be persistent local tissue inflammation within the infarct core in response to persistence of haemoglobin breakdown products and accumulation of deoxidized iron residues and tissue fibrosis. These pathologies prevent the natural healing process that otherwise would normally occur in reperfused myocardium in the absence of MVO and IMH. Cigarette smoking and a history of hypertension are risk factors for IMH. Carberry et al. demonstrated that persistent iron affected one in five patients who survived through to 6 months post-STEMI and was associated with adverse LV remodelling, worsening ejection fractions at 6 months. Systemic inflammation at baseline, reflected by the neutrophil count, was a univariable associate of persistent iron at 6 months, and presenting heart rate and a history of hypertension were multivariable associates.9 Additionally, iron deposition within the infarcted myocardium may have deleterious effects on the electrical stability of the heart and so may increase the likelihood of compromising ventricular arrhythmias and sudden cardiac death post-MI.10

ting heart rate and a history of hypertension were multivariable associates.9 Additionally, iron deposition within the infarcted myocardium may have deleterious effects on the electrical stability of the heart and so may increase the likelihood of compromising ventricular arrhythmias and sudden cardiac death post-MI.10 Cardiovascular magnetic resonance (CMR) is the only method available to clinicians to detect this problem in vivo. T2* imaging is generally accepted as the reference method for the assessment of IMH in STEMI patients,11 and T2* imaging is increasingly available as an option in standard CMR protocols. Blood degradation products such as deoxyhaemoglobin exert a paramagnetic effect, reducing the T2* signal, represented by hypointense areas within the infarct core. Still, local signal loss due to artefact can complicate the imaging read-out, especially if supporting features such as reduced wall motion and infarction are absent.

adation products such as deoxyhaemoglobin exert a paramagnetic effect, reducing the T2* signal, represented by hypointense areas within the infarct core. Still, local signal loss due to artefact can complicate the imaging read-out, especially if supporting features such as reduced wall motion and infarction are absent. Reinstadler et al.12 provide additional evidence for the clinical importance of IMH characterized by T2* imaging post-STEMI. They conducted a prospective multicentre study of 264 STEMI patients presenting within 12 h of symptom onset undergoing primary PCI. The primary endpoint was a composite of death, reinfarction, and new congestive heart failure at 12 months. Sixty patients had IMH, of these, 9 (15%) had major adverse cardiac events (MACE), whereas only 10 (4.9%) patients without IMH experienced a MACE. IMH was independently associated with MACE, and IMH increased the prognostic value of a model which included MVO. This study adds to the previous work by Carrick et al.,2 reaffirming IMH as a determinative pathological complication post-STEMI.

ents (MACE), whereas only 10 (4.9%) patients without IMH experienced a MACE. IMH was independently associated with MACE, and IMH increased the prognostic value of a model which included MVO. This study adds to the previous work by Carrick et al.,2 reaffirming IMH as a determinative pathological complication post-STEMI. The study by Reinstadler et al.12 did have some limitations. The number of MACE events (n = 19) was modest and in isolation the results have qualified significance. On the other hand, these data are consistent with other studies.2,13 Reinstadler et al. highlighted five patients with IMH but no MVO which is not consistent with previous studies. In a serial imaging time course sub-study of 30 patients, MVO had resolved by day 10 in 44% of affected patients, with persistence of IMH in 25% of these.2 One potential explanation is the differing time-course of these pathologies with resolution of MVO in the presence of persistent IMH not disclosed by imaging at a single time-point up to 7 days. Imaging artefact may also be relevant. A small amount of MVO may not be visible within a zone of late gadolinium enhancement imaging and T2* artefact occurs at a tissue–air interface such as the infero-lateral wall of the left ventricle which may be mistaken for haemorrhage.

imaging at a single time-point up to 7 days. Imaging artefact may also be relevant. A small amount of MVO may not be visible within a zone of late gadolinium enhancement imaging and T2* artefact occurs at a tissue–air interface such as the infero-lateral wall of the left ventricle which may be mistaken for haemorrhage. Therefore, IMH represents a target for preventive therapy, however, aside from timely reperfusion, there are no specific treatments for this problem. Favourable results in preclinical studies have not translated when assessed in patients.14 Randomized controlled clinical trials of novel therapeutic approaches designed to reduce the extent and severity of infarction, including novel cardioprotective interventions such as intra-venous beta-blocker therapy before reperfusion (EARLY-BAMI),15 intravenous inhibitors of mitochondria-mediated reperfusion injury, i.e. cyclosporine (CIRCUS),16 TRO40303 (MITOCARE),17 post-ischaemic conditioning (DANAMI-3 iPOST),18 and deferred stenting (DANAMI-3-DEFER),19 have not proven beneficial and intra-coronary vasodilator therapy with adenosine was actually harmful (REFLO-STEMI).20 A post hoc analysis of the Phase 2 METOCARD-CNIC trial21 indicated intravenous beta-blocker therapy might reduce the risk of MVO through inhibition of neutrophil recruitment and platelet activation. MVO is now identified as a therapeutic target in practice guidelines,22 but the gaps in evidence on the causes and treatment of IMH highlight the need for more research.

trial21 indicated intravenous beta-blocker therapy might reduce the risk of MVO through inhibition of neutrophil recruitment and platelet activation. MVO is now identified as a therapeutic target in practice guidelines,22 but the gaps in evidence on the causes and treatment of IMH highlight the need for more research. We have recently conducted a Phase 2 clinical trial of low dose adjunctive intracoronary fibrinolysis with alteplase in reperfused STEMI (ClinicalTrials.gov: NCT02257294). The clinical strategy involved identifying patients in whom initial coronary angiography identified occluded infarct-related artery and/or with a high thrombus burden. These characteristics place the participants at an increased risk of MVO. By targeting thrombus within the infarct-related artery and microcirculation with fibrinolytic therapy the aim was to restore microvascular blood flow at the earliest point after coronary reperfusion. On the other hand, the intervention has the potential to promote bleeding within the infarct zone and systemically. The risk of IMH was purposefully mitigated by selecting patients presenting with a comparatively short ischaemic time (<6 h) in whom radial artery access was used. The overall objective was to conduct a safety, efficacy and mechanisms evaluation within the context of a trial that was sufficiently large to give definitive results.

IMH was purposefully mitigated by selecting patients presenting with a comparatively short ischaemic time (<6 h) in whom radial artery access was used. The overall objective was to conduct a safety, efficacy and mechanisms evaluation within the context of a trial that was sufficiently large to give definitive results. In conclusion, given that failed myocardial reperfusion occurs in one in every two patients undergoing primary PCI for acute STEMI, and IMH is an independent driver of prognosis, can primary PCI really be considered successful when these eventualities routinely occur? We think not. We propose that successful primary PCI is redefined as restoration of normal coronary blood flow in the absence of MVO and IMH. However, until specific evidence-based treatments for MVO and IMH become available, routine imaging with CMR to assess for these pathologies cannot be justified on economic grounds, and clinicians should follow optimal guideline-directed management for their post-MI patients. Funding This work was supported by the British Heart Foundation (BHF) Centre of Research Excellence Award (RE/13/5/30177) and a research grant from the National Institute of Health Research Efficacy and Mechanism Evaluation Programme (12/170/45). Conflict of interest: The University of Glasgow holds a research agreement with Siemens Healthcare. Footnotes The opinions expressed in this article are not necessarily those of the Editors of EHJCI, the European Heart Rhythm Association or the European Society of Cardiology.

Introduction The overall prevalence of mitral regurgitation (MR) in the general population is ∼2% and its aetiology may be primary (or organic) or secondary (or functional). Secondary MR is a consequence of annular dilatation and geometrical distortion of the sub-valvular apparatus secondary to left ventricular (LV) remodelling associated with cardiomyopathy or coronary artery disease. Severe secondary MR is associated with a poor prognosis in patients with chronic heart failure and reduced left ventricular ejection fraction (LVEF). Percutaneous mitral valve repair using the MitraClip device has been proposed to correct secondary MR. Recently, the results of two randomized controlled trials, that is MITRA-FR (Percutaneous Repair with the MitraClip Device for Severe Functional/Secondary Mitral Regurgitation) and COAPT (Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional Mitral Regurgitation), assessing the efficacy and safety of MitraClip in patients with systolic heart failure and severe secondary MR were published in the New England Journal of Medicine.1,2 A priori, these two trials targeted the same patient populations with the same disease using the same device but the results of these trials were diametrically opposed, MITRA-FR being neutral and COAPT being highly positive with respect to efficacy of the MitraClip procedure.

in the New England Journal of Medicine.1,2 A priori, these two trials targeted the same patient populations with the same disease using the same device but the results of these trials were diametrically opposed, MITRA-FR being neutral and COAPT being highly positive with respect to efficacy of the MitraClip procedure. The objectives of this viewpoint are: (i) to highlight not only the similarities but also the differences between MITRA-FR and COAPT, which may explain the strikingly different results and conclusions between these two trials and (ii) to derive from these results, implications with regards to the application of the MitraClip procedure in clinical practice.

(i) to highlight not only the similarities but also the differences between MITRA-FR and COAPT, which may explain the strikingly different results and conclusions between these two trials and (ii) to derive from these results, implications with regards to the application of the MitraClip procedure in clinical practice. Summary of the design and results of MITRA-FR and COAPT The MITRA-FR study randomized 304 patients with symptomatic systolic heart failure and severe secondary MR defined as an effective regurgitant orifice area (EROA) >20 mm2 and/or a regurgitant volume >30 mL, and LVEF between 15% and 40%, in a 1:1 ratio, to percutaneous mitral valve repair with MitraClip in addition to optimized medical therapy (intervention group) or to optimized medical therapy alone (control group) (Tables 1and2).1 The primary efficacy endpoint was a composite of death from any cause or hospitalization for heart failure at 1 year. There was no difference between the intervention vs. control groups for the rate of the primary composite endpoint (54.6% vs. 51.3%, respectively; P = 0.53), the rate of mortality (24.3% vs. 22.4%) or the rate of unplanned heart failure hospitalization (48.7% vs. 47.4%) (Table 3). The authors concluded that MitraClip is safe and effective in reducing secondary MR but does not improve prognosis (as compared with optimized medical therapy) in patients with secondary MR and systolic heart failure. Table 1 Similarities and differences among MITRA-FR, COAPT, and RESHAPE-HF2 with respect to study design and endpoints

Summary of the design and results of MITRA-FR and COAPT The MITRA-FR study randomized 304 patients with symptomatic systolic heart failure and severe secondary MR defined as an effective regurgitant orifice area (EROA) >20 mm2 and/or a regurgitant volume >30 mL, and LVEF between 15% and 40%, in a 1:1 ratio, to percutaneous mitral valve repair with MitraClip in addition to optimized medical therapy (intervention group) or to optimized medical therapy alone (control group) (Tables 1and2).1 The primary efficacy endpoint was a composite of death from any cause or hospitalization for heart failure at 1 year. There was no difference between the intervention vs. control groups for the rate of the primary composite endpoint (54.6% vs. 51.3%, respectively; P = 0.53), the rate of mortality (24.3% vs. 22.4%) or the rate of unplanned heart failure hospitalization (48.7% vs. 47.4%) (Table 3). The authors concluded that MitraClip is safe and effective in reducing secondary MR but does not improve prognosis (as compared with optimized medical therapy) in patients with secondary MR and systolic heart failure. Table 1 Similarities and differences among MITRA-FR, COAPT, and RESHAPE-HF2 with respect to study design and endpoints MITRA-FR COAPT RESHAPE HF2 Study design Prospective, randomized Prospective, randomized Prospective, randomized Randomization 1:1 in: Intervention arm MitraClip + GDMT MitraClip + GDMT MitraClip + GDMT Control arm GDMT GDMT GDMT Patientsrecruitment Total no. of patients 304 614 420 No. of patients in intervention/control groups 152/152 302/312 Enrolment period, year 3.2 4.8 No. of sites 22 85 No. of patients/site 8.2 7.8 No. of patients/site/year 2.6 1.6 Inclusion/exclusion criteria By European Guidelines6 By American Guidelines8,12 By EACVI recommendations12 ≥ Moderate-to-severe (3+) MR EROA >20 mm2 and/or RV >30 mL EROA ≥30 mm2 and/or RV >45 mL EROA >30 mm2 and/or RV >45 mL LV end-systolic diameter, mm ≤70 mm LV ejection fraction, % ≥15 and ≤40 ≥20 and ≤50 ≥15 and ≤35 if NYHA II≥15 and ≤45 if NYHA III or IV GDMT at baseline GDMT variable adjustment in each group per ‘real-world’ practice Stable maximal doses of GDMT and cardiac resynchronization therapy (if appropriate) Stable optimal GDMT and revascularization or cardiac resynchronization therapy (if appropriate) Symptomatic status NYHA class: II, III, IV NYHA class: II, III, IVa (ambulatory) NYHA class: II, III, IV Surgical risk Not candidate for mitral-valve surgery Not candidate for mitral-valve surgery Mitral-valve surgery is not the preferred option Primary endpoint Death or HF hospitalization at 1 year HF hospitalization at 1 year Composite rate of recurrent HF hospitalizations and cardiovascular death at 2 years EACVI, European Association of Cardiovascular Imaging; EROA, effective regurgitant orifice area; GDMT, guideline-directed medical therapy; HF, heart failure; MR, mitral regurgitation; NYHA, New York Heart Association; RV, regurgitant volume.

site rate of recurrent HF hospitalizations and cardiovascular death at 2 years EACVI, European Association of Cardiovascular Imaging; EROA, effective regurgitant orifice area; GDMT, guideline-directed medical therapy; HF, heart failure; MR, mitral regurgitation; NYHA, New York Heart Association; RV, regurgitant volume. Table 2 Similarities and differences between MITRA-FR and COAPT with respect to baseline characteristics of the study populations MITRA-FR COAPT Baseline clinical characteristics Age, year 70 ± 10 72 ± 11 NYHA class, % I 0 0.2 II 32.9 39.0 III 58.5 52.5 IV 8.6 8.3 Surgical risk STS score ≥8% 42.7% EuroSCORE II, median and IQR 6.2 (3.5–11.0) Baseline echocardiographic characteristics MR severity, % Moderate (EROA 20-29 mm2) 52 14 Moderate-to-severe (EROA 30-39 mm2) 32 46 Severe (EROA ≥ 40 mm2) 16 41 EROA, mm2 31 ± 10 41 ± 15 LV end-diastolic volume index, mL/m2 135 ± 35 101 ± 34 LV ejection fraction, % 33 ± 7 31 ± 9 IQR, inter-quartile range; STS, Society of Thoracic Surgery. Other abbreviations as in Table 1. Table 3 Similarities and differences between MITRA-FR and COAPT with respect to study outcomes

MITRA-FR COAPT Baseline clinical characteristics Age, year 70 ± 10 72 ± 11 NYHA class, % I 0 0.2 II 32.9 39.0 III 58.5 52.5 IV 8.6 8.3 Surgical risk STS score ≥8% 42.7% EuroSCORE II, median and IQR 6.2 (3.5–11.0) Baseline echocardiographic characteristics MR severity, % Moderate (EROA 20-29 mm2) 52 14 Moderate-to-severe (EROA 30-39 mm2) 32 46 Severe (EROA ≥ 40 mm2) 16 41 EROA, mm2 31 ± 10 41 ± 15 LV end-diastolic volume index, mL/m2 135 ± 35 101 ± 34 LV ejection fraction, % 33 ± 7 31 ± 9 IQR, inter-quartile range; STS, Society of Thoracic Surgery. Other abbreviations as in Table 1. Table 3 Similarities and differences between MITRA-FR and COAPT with respect to study outcomes MITRA-FR COAPT Procedural characteristics and outcomesa Procedural success, %a 96 98 Procedural complications, %a 14.6 8.5 Number of clips, %b 1 Clip 46 36 2 Clips 45 55 3 Clips 9 8 4 Clips 0 0.3 Post-procedural MR ≥ moderate-to-severe (3+), %a End of procedure 9 5 1 year post-procedure 17 5 2 years post-procedure 0.9 1-Year outcomes 1-Year mortality, % Intervention 24.2 19.1 (P < 0.001) Control 22.4 23.2 1-Year heart failure hospitalization, % Primary outcome Intervention 48.7 35.8 (P < 0.001)c Control 47.4 67.9c 1-Year mortality or heart failure hospitalization Primary outcome Intervention 54.6 (P = 0.53) 33.9 (P < 0.001) Control 51.3 46.5 Abbreviations as in Table 1. a Data are from the intervention group only. b Data are from the intervention group with procedural success. c Annualized rate (in % per year) within 2-year follow-up.

MITRA-FR COAPT Procedural characteristics and outcomesa Procedural success, %a 96 98 Procedural complications, %a 14.6 8.5 Number of clips, %b 1 Clip 46 36 2 Clips 45 55 3 Clips 9 8 4 Clips 0 0.3 Post-procedural MR ≥ moderate-to-severe (3+), %a End of procedure 9 5 1 year post-procedure 17 5 2 years post-procedure 0.9 1-Year outcomes 1-Year mortality, % Intervention 24.2 19.1 (P < 0.001) Control 22.4 23.2 1-Year heart failure hospitalization, % Primary outcome Intervention 48.7 35.8 (P < 0.001)c Control 47.4 67.9c 1-Year mortality or heart failure hospitalization Primary outcome Intervention 54.6 (P = 0.53) 33.9 (P < 0.001) Control 51.3 46.5 Abbreviations as in Table 1. a Data are from the intervention group only. b Data are from the intervention group with procedural success. c Annualized rate (in % per year) within 2-year follow-up. The COAPT trial randomly assigned 614 patients with symptomatic heart failure and moderate-to-severe or severe secondary MR defined as an EROA >30 mm2 and/or regurgitant volume >45 mL, and LVEF ≥20%, in a 1:1 ratio, to percutaneous mitral valve repair with MitraClip plus optimized medical therapy (intervention group) or to optimized medical therapy alone (control group) (Tables 1 and 2).3 The primary efficacy endpoint was all hospitalizations within 2-year follow-up. The annualized rate of all hospitalizations for heart failure within 2 years was 35.8% per patient-year in the intervention group as compared with 67.9% per patient-year in the control group (P < 0.001) (Table 3). Death from any cause within 2 years occurred in 29.1% of the patients in the intervention group vs. 46.1% in the control group (P < 0.001). The authors concluded that among patients with heart failure and ≥ moderate-to-severe secondary MR who remained symptomatic despite the use of optimal guideline-directed medical therapy (GDMT), the MitraClip procedure reduces the rates of hospitalization for heart failure and all-cause mortality within 2 years of follow-up than medical therapy alone. The number needed to treat to prevent one hospitalization for heart failure within 24 months was 3.1.

e use of optimal guideline-directed medical therapy (GDMT), the MitraClip procedure reduces the rates of hospitalization for heart failure and all-cause mortality within 2 years of follow-up than medical therapy alone. The number needed to treat to prevent one hospitalization for heart failure within 24 months was 3.1. Similarities and differences between MITRA-FR and COAPT Tables 1–3 present the main similarities and differences in study design and results between the MITRA-FR and COAPT trials. The sample size was about 2-fold larger in the COAPT study vs. the MITRA-FR trial and the primary efficacy endpoint was the rate of the composite of death from any cause and unplanned hospitalization for heart failure at 1 year in the MITRA-FR trial vs. the rate of all hospitalizations for heart failure within 2-year follow-up in the COAPT study.

fold larger in the COAPT study vs. the MITRA-FR trial and the primary efficacy endpoint was the rate of the composite of death from any cause and unplanned hospitalization for heart failure at 1 year in the MITRA-FR trial vs. the rate of all hospitalizations for heart failure within 2-year follow-up in the COAPT study. Extent of LV damage and MR severity Compared with patients in the COAPT trial, those enrolled in the MITRA-FR trial had substantially more LV damage. The patients had larger LV end-diastolic volumes (MITRA-FR: 135 ± 35 mL/m2 vs. COAPT: 101 ± 34 mL/m2) suggesting a more advanced stage of the LV disease (Table 2). This difference is likely related to the fact that COAPT excluded patients with very severe LV dilation/dysfunction (LV end-systolic diameter <70 mm), whereas MITRA-FR had no LV dimension limit. Also in COAPT, the inclusion criteria for LVEF were 20–50% vs. 15–40% in MITRA-FR. Several studies have reported that in heart failure patients with ischaemic MR, severe LV dilation (LV end-diastolic diameter >65 mm) and LV dysfunction (LVEF < 20%, LV end-systolic diameter >55 mm) are associated with high rates of persistent/recurrent MR, less reverse LV remodelling, and worse outcomes after surgical correction of ischaemic MR.4,5

hat in heart failure patients with ischaemic MR, severe LV dilation (LV end-diastolic diameter >65 mm) and LV dysfunction (LVEF < 20%, LV end-systolic diameter >55 mm) are associated with high rates of persistent/recurrent MR, less reverse LV remodelling, and worse outcomes after surgical correction of ischaemic MR.4,5 MITRA-FR patients also had less severe MR (EROA: 31 ± 10 mm2) as compared with COAPT (41 ± 15 mm2) (Table 2). Although the inclusion criteria were at least moderate-to-severe (3+) secondary MR in both trials, MITRA-FR actually used the 2012 European guidelines criteria,6 that is EROA ≥20 mm2 and/or regurgitant volume ≥30 mL, whereas COAPT used the 2006/2008 American guidelines criteria,7,8 that is EROA ≥30 mm2 and/or regurgitant volume ≥45 mL. The criteria used in MITRA-FR, correspond to ≥ moderate (or 2+) MR according to American guidelines criteria.7,8 The European guidelines6 as well as the 2014 American guidelines9 recommended to use 2-fold lower cut-off values of EROA (20 vs. 40 mm2) and regurgitant volume (30 vs. 60 mL) to define severe MR in secondary vs. primary MR. This was based on the rationale that the risk of mortality rises significantly at a lower level of MR severity (EROA ≥20 vs. 40 mm2) in secondary vs. primary MR.10,11 However, ischaemic MR is a complex and multifaceted disease and it is unclear whether a volumetrically moderate MR is truly an actor or simply a marker of the LV adverse remodelling/dysfunction and of the heart failure symptoms; in other words whether it is primarily a valvular disease or a myocardial (LV) disease. If one applies the same criteria of EROA to define the severity of MR, there appears to be major difference in the distribution of baseline MR severity between MITRA-FR and COAPT: only 16% of MITRA-FR patients had severe MR as defined by EROA ≥40 mm2 vs. 41% of COAPT patients (Table 2).

se or a myocardial (LV) disease. If one applies the same criteria of EROA to define the severity of MR, there appears to be major difference in the distribution of baseline MR severity between MITRA-FR and COAPT: only 16% of MITRA-FR patients had severe MR as defined by EROA ≥40 mm2 vs. 41% of COAPT patients (Table 2). It could be that these differences in the inclusion/exclusion criteria for MR severity, LV dimensions and dysfunction are the main reasons for the discrepancies in the outcomes observed between MITRA-FR and COAPT (Table 1). In MITRA-FR, the underlying cardiomyopathy (myocardial or LV disease) was likely the predominant cause of the heart failure and thus the main determinant of the poor clinical outcome. And in this context, the MR was probably more a bystander than an actor of the heart failure. On the other hand, in COAPT, heart failure was, in large part, related to the valvular disease (the MR was more severe), while LV disease (smaller size and higher LVEF) was less advanced. Hence in COAPT, MR was an important contributor to the heart failure and the clinical outcomes, whereas in MITRA-FR the LV disease (dysfunction) was the main determinant of clinical outcomes. Possibly, because of these differences in baseline characteristics, the COAPT patients were more likely to benefit from the MitraClip procedure compared with the MITRA-FR patients.

he heart failure and the clinical outcomes, whereas in MITRA-FR the LV disease (dysfunction) was the main determinant of clinical outcomes. Possibly, because of these differences in baseline characteristics, the COAPT patients were more likely to benefit from the MitraClip procedure compared with the MITRA-FR patients. Optimization of medical therapy at baseline and during follow-up To confirm patient eligibility, both trials required that patients remained symptomatic (NYHA class 2 or more) despite the use of GDMT for chronic systolic heart failure (Table 1). However, COAPT imposed a more demanding criteria for inclusion of patients: that is use of maximal tolerated doses of GDMT, and treatment with cardiac resynchronization therapy, defibrillators, and revascularization, if appropriate. Hence in COAPT, medical treatment was optimized prior to randomization and only a few major adjustments in treatment occurred during follow-up. On the other hand in MITRA-FR, medical therapy was not optimized in all patients at baseline and multiple adjustments in medical treatment were allowed in each arm during follow-up per ‘real-world’ practice. This issue may also have decreased the ability to reveal a beneficial effect of the intervention in MITRA-FR.

p. On the other hand in MITRA-FR, medical therapy was not optimized in all patients at baseline and multiple adjustments in medical treatment were allowed in each arm during follow-up per ‘real-world’ practice. This issue may also have decreased the ability to reveal a beneficial effect of the intervention in MITRA-FR. Efficacy in the correction of MR A more aggressive strategy for correction of MR was applied in COAPT, as suggested by the larger number of clips implanted per patient in COAPT vs. in MITRA-FR (Table 3). Furthermore, the rate of sustained reduction of MR was higher in COAPT than in MITRA-FR. At 1 year, 17% of the MITRA-FR patients randomized to MitraClip had ≥ moderate-to-severe (3+) residual MR compared with only 5% in COAPT. The lower sustained efficacy of the MitraClip procedure may also have contributed to the lack of benefit of the intervention in MITRA-FR. Conclusions and implications for the management of patients with secondary MR MITRA-FR and COAPT targeted the same disease entity with the same device, the MitraClip. However, COAPT enrolled a subset of patients who had more severe MR and less advanced LV disease (dilation/dysfunction) compared with MITRA-FR patients. These differences may explain the different outcomes observed in COAPT vs. MITRA-FR. Indeed, patients with too severe LV dilation/dysfunction (i.e. too extensive LV myocardial damage) may not benefit from the MitraClip procedure (Figure 1).

ess advanced LV disease (dilation/dysfunction) compared with MITRA-FR patients. These differences may explain the different outcomes observed in COAPT vs. MITRA-FR. Indeed, patients with too severe LV dilation/dysfunction (i.e. too extensive LV myocardial damage) may not benefit from the MitraClip procedure (Figure 1). Figure 1 Utility vs. futility of MitraClip procedure according to severity of MR and LV systolic dysfunction. EROA, effective regurgitant orifice area; LVEF, left ventricular ejection fraction; LVESD, left ventricular end-systolic diameter; MR, mitral regurgitation; RV, regurgitant volume. In light of the results of the MITRA-FR and COAPT trials, it thus appears reasonable to conclude that the MitraClip procedure reduces heart failure hospitalization and mortality in patients meeting the following criteria (Figure 1): ≥ moderate-to-severe (3+) secondary MR defined as EROA ≥30 mm2 and/or regurgitant volume >45 mL; LVEF between 20% and 50% and LV end-systolic diameter <70 mm; Persistent heart failure symptoms (NYHA ≥ II) despite optimal (maximally tolerated) GDMT with cardiac resynchronization and coronary revascularization if appropriate. Furthermore, the goal of the procedure should be to obtain an acute reduction of the MR severity to ≤ mild (1+) and the implantation of additional clips should be considered to achieve this goal.

Persistent heart failure symptoms (NYHA ≥ II) despite optimal (maximally tolerated) GDMT with cardiac resynchronization and coronary revascularization if appropriate. Furthermore, the goal of the procedure should be to obtain an acute reduction of the MR severity to ≤ mild (1+) and the implantation of additional clips should be considered to achieve this goal. Further insight will come from the results of the ReshapeHF2 trial [A Clinical Evaluation of the Safety and Effectiveness of the MitraClip System in the Treatment of Clinically Significant Functional Mitral Regurgitation (Reshape-HF2) (https://clinicaltrials.gov/ct2/show/NCT 02444338], which has the same inclusion criteria as those of the COAPT trial in terms of MR severity, with intermediary criteria COAPT and MITRA-FR in terms of LV dysfunction severity (Table 1). Funding P.P. holds the Canada Research Chair in Valvular Heart Disease, Canadian Institutes of Health Research (CIHR), Ottawa Canada and received Foundation Scheme research grant #FDN-143225) from CIHR. The department of Cardiology of the Leiden University Medical Center receives unrestricted research grants from Biotronik, Boston Scientific, Medtronic, Edwards Lifesciences and GE Healthcare. Conflict of interest: P.P. received funding from Edwards Lifesciences and Cardiac Phoenix for echocardiography corelab analyses with no personal compensation. J.J.B. received speaker fees from Abbot Vascular and Boehringer Ingelheim. V.D. received speaker fees from Abbott Vascular.

Introduction Hypertrophic cardiomyopathy (HCM) is the most common inherited heart muscle disorder and an important cause of heart failure (HF) with preserved left ventricular (LV) ejection fraction.1,2 Diastolic dysfunction is a major contributor to the pathophysiology of HF in HCM, which encompasses a complex sequence of interrelated mechanisms including LV hypertrophy, intraventricular obstruction, microvascular ischaemia, and myocardial fibrosis.3 Evidence of diastolic dysfunction is frequently observed in HCM4 and is known to have a major impact on symptom severity, functional capacity, medical treatment, and prognosis.5–7 Echocardiography is the non-invasive modality of choice for the semi-quantitative evaluation of LV diastolic function. Using echocardiography, a comprehensive multi-parameter algorithm is recommended for the estimation of elevated mean left atrial pressure (LAP).8,9

Introduction Hypertrophic cardiomyopathy (HCM) is the most common inherited heart muscle disorder and an important cause of heart failure (HF) with preserved left ventricular (LV) ejection fraction.1,2 Diastolic dysfunction is a major contributor to the pathophysiology of HF in HCM, which encompasses a complex sequence of interrelated mechanisms including LV hypertrophy, intraventricular obstruction, microvascular ischaemia, and myocardial fibrosis.3 Evidence of diastolic dysfunction is frequently observed in HCM4 and is known to have a major impact on symptom severity, functional capacity, medical treatment, and prognosis.5–7 Echocardiography is the non-invasive modality of choice for the semi-quantitative evaluation of LV diastolic function. Using echocardiography, a comprehensive multi-parameter algorithm is recommended for the estimation of elevated mean left atrial pressure (LAP).8,9 Despite this, the identification of accurate and reproducible indices for quantitative assessment of haemodynamic congestion in HCM is still an unmet clinical need, in particular, to allow for early detection of disease progression and treatment guidance. An alternative approach to assess the presence and severity of diastolic dysfunction would be by measuring to what extent the elevated LV filling pressure is transmitted retrogradely into the pulmonary circulation leading to an increase in central transit time,10 pulmonary blood volume,11 and eventually to increased pulmonary capillary hydrostatic pressure (haemodynamic congestion). Pulmonary blood volume index (PBVI) by first-pass perfusion cardiovascular magnetic resonance (CMR) imaging has been shown to differentiate between stages of diastolic dysfunction in patients with HF and reduced LV ejection fraction and has been proposed as a quantitative marker of HF useful for quantification and monitoring of haemodynamic congestion.11,12 In this study, we aimed to assess the diagnostic accuracy of PBVI by CMR for the detection and grading of haemodynamic congestion in patients with HCM and preserved LV ejection fraction.

action and has been proposed as a quantitative marker of HF useful for quantification and monitoring of haemodynamic congestion.11,12 In this study, we aimed to assess the diagnostic accuracy of PBVI by CMR for the detection and grading of haemodynamic congestion in patients with HCM and preserved LV ejection fraction. Methods Study design and population This is a retrospective study of consecutive patients recruited into the Barts BioResource medical research programme complying with the 1975 Declaration of Helsinki and approved by our Institutional Ethics Committee. The Barts BioResource is a joint partnership between Barts Health NHS Trust, the William Harvey Research Institute/Queen Mary University of London, and University College London with support from the National Institute for Health Research Barts Biomedical Research Centre aimed to establish a research repository of consented patients clinically managed by Cardiac Services at Barts Health NHS Trust. All patients gave written informed consent and received standard optimal medical therapy according to the current guidelines. We identified all consecutive HCM outpatients undergoing CMR stress perfusion imaging for clinical purposes between January 2015 and September 2017.

y Cardiac Services at Barts Health NHS Trust. All patients gave written informed consent and received standard optimal medical therapy according to the current guidelines. We identified all consecutive HCM outpatients undergoing CMR stress perfusion imaging for clinical purposes between January 2015 and September 2017. Exclusion criteria were as follows: (i) age <18 years; (ii) presence of any contra-indication to CMR and/or to administration of gadolinium-based contrast agents; (iii) atrial fibrillation or frequent ventricular ectopy during image acquisition; (iv) any hospitalization, worsening of the functional status, changes in medical treatment occurred in the time interval between transthoracic echocardiography (TTE) and CMR; (v) at least moderate tricuspid valve regurgitation by TTE; (vi) known significant coronary artery disease; and (vii) LV ejection fraction <50%. All patients underwent a standard clinical assessment including medical history, physical examination, routine blood tests, transthoracic echocardiogram, and contrast-enhanced CMR.

Exclusion criteria were as follows: (i) age <18 years; (ii) presence of any contra-indication to CMR and/or to administration of gadolinium-based contrast agents; (iii) atrial fibrillation or frequent ventricular ectopy during image acquisition; (iv) any hospitalization, worsening of the functional status, changes in medical treatment occurred in the time interval between transthoracic echocardiography (TTE) and CMR; (v) at least moderate tricuspid valve regurgitation by TTE; (vi) known significant coronary artery disease; and (vii) LV ejection fraction <50%. All patients underwent a standard clinical assessment including medical history, physical examination, routine blood tests, transthoracic echocardiogram, and contrast-enhanced CMR. Transthoracic echocardiography All patients underwent both standard TTE (Vivid E9TM, General Electric) and CMR examinations [median time interval 5 days, interquartile range (IQR) 3–10]. A comprehensive echocardiographic assessment of LV diastolic function was performed in agreement with the current international guidelines,8 including structural assessment of LV size and mass, left atrial (LA) volume, mitral inflow and pulmonary venous flow patterns, pulsed-wave tissue-Doppler imaging of mitral annular velocities, peak velocity of tricuspid regurgitation, and systolic pulmonary artery pressure (sPAP). LV diastolic dysfunction with increased LA pressures was defined as the presence of a Grade II or III diastolic dysfunction by Doppler echocardiography, as described previously.8

tissue-Doppler imaging of mitral annular velocities, peak velocity of tricuspid regurgitation, and systolic pulmonary artery pressure (sPAP). LV diastolic dysfunction with increased LA pressures was defined as the presence of a Grade II or III diastolic dysfunction by Doppler echocardiography, as described previously.8 Cardiovascular magnetic resonance All patients were examined on two Magnetom Aera 1.5 T scanners (Siemens Healthineers, Erlangen, Germany) using an 18-channel cardiac phased array anterior coil in combination with the spine coil (up to 12-channel coil elements). Biventricular function was assessed by breath-hold, steady-state free precession (SSFP) cine short-axis imaging covering both ventricles. Typical scan parameters were: field of view 360 × 290 mm2, 8 mm slice thickness with a 2 mm inter-slice gap, repetition time 2.7 ms, echo time 1.14 ms, flip angle 80°, acquisition matrix 208 × 169, reconstruction pixel size 1.73 × 1.73 mm2, acquired temporal resolution 35.1 ms (13 views per k-space segment), and 30 calculated cardiac phases. CMR perfusion imaging was performed with the first-pass technique as described previously.11 Briefly, a bolus of gadolinium-based contrast agent (Dotarem, Guerbet, France, 0.05 mmol/kg at an injection rate of 4 mL/s) was injected into a peripheral vein, followed by a saline flush (injection rate 4 mL/s). Adenosine was given intravenously at 140 μg/kg/min for 4 min. First-pass images were obtained using cardiac-triggered gradient echo-train SSFP pulse sequences. Images from three short-axis planes (basal, middle, and apical) were acquired with the following parameters: typical field of view 360 × 270 mm2, slice thickness 8 mm, slice gap 12 mm, repetition time 2.5 ms, echo time 1.04 ms, flip angle 50°, bandwidth 1085 Hz/pixel, 60 dynamics, acquisition matrix 111 × 192, and a parallel acceleration factor R = 3 (TGRAPPA).13 Image acquisition was begun at the same time as the contrast media injection.

of view 360 × 270 mm2, slice thickness 8 mm, slice gap 12 mm, repetition time 2.5 ms, echo time 1.04 ms, flip angle 50°, bandwidth 1085 Hz/pixel, 60 dynamics, acquisition matrix 111 × 192, and a parallel acceleration factor R = 3 (TGRAPPA).13 Image acquisition was begun at the same time as the contrast media injection. Image analysis All CMR studies were analysed off-line using cvi42 post-processing software (Version 5.1.1, Circle Cardiovascular Imaging Inc., Calgary, Canada) by two EACVI CMR Level 3 certified operators (F.R. and N.A.), blinded to clinical and echocardiographic data. Ventricular volumes and function, as well as LV mass and atrial dimensions, were determined using balanced SSFP cine images as described previously.14

5.1.1, Circle Cardiovascular Imaging Inc., Calgary, Canada) by two EACVI CMR Level 3 certified operators (F.R. and N.A.), blinded to clinical and echocardiographic data. Ventricular volumes and function, as well as LV mass and atrial dimensions, were determined using balanced SSFP cine images as described previously.14 Pulmonary transit time Pulmonary transit time (PTT) has been measured as the time interval required for a bolus of contrast agent to pass from the right ventricle to LV. First-pass perfusion imaging allows us to obtain an image for each heartbeat for 60 consecutive beats and can, therefore, be used to measure PTT. Briefly, in the basal slice of the first-pass perfusion imaging short-axis view stack, a region of interest was placed in the right ventricular (RV) cavity and copied in the same position for the entire stack of images. The average signal intensity (SI) of the region of interest was measured in every image and a SI/time curve was generated (Figure 1). A second SI/time curve was generated from another region of interest placed in the basal LV cavity. The PTT was measured as the peak-to-peak time interval between the two curves, both under resting and stress conditions. The number of cardiac cycles between the peaks of the two SI/time curves defines the number of pulmonary transit beats (PTB). Bazett’s formula was used to correct the PTT interval for the heart rate: PTTc (s) = PTT (s)/√R-R interval(s). In a preliminary study, we compared the measurement of PTT and PTB as onset-to-onset, peak-to-peak and half-width time interval of the wash-in phase of SI/heartbeat curves.11 We observed no significant differences among such methods, and we finally chose the peak-to-peak time interval as the most reproducible, yielding intra-class correlation coefficient values of 0.99 and 0.97 for intra-observer and inter-observer reliability, respectively.

al of the wash-in phase of SI/heartbeat curves.11 We observed no significant differences among such methods, and we finally chose the peak-to-peak time interval as the most reproducible, yielding intra-class correlation coefficient values of 0.99 and 0.97 for intra-observer and inter-observer reliability, respectively. Figure 1 Pulmonary transit time and PBVI. SI/time curves of two regions of interest drawn in the RV cavity (red) and in the LV cavity (green), respectively. PTT is defined by the time interval between the peaks of the two SI/time curves. PBVI is obtained by the product between the RVSVI and the pulmonary transit time normalized by the R-R interval (see Methods section). Gd, gadolinium; LV, left ventricle; PBVI, pulmonary blood volume index; PTT, pulmonary transit time; RV, right ventricle; RVSVI, right ventricular stroke volume index; SI, signal intensity. Pulmonary blood volume index Resting pulmonary blood volume indexed to body surface area (PBVI, mL/m2) was measured by the product of PTB and the RV stroke volume index (RVSVI) (Figure 1). This is better understood if one considers PBVI as being analogous to a cargo train, with the number of wagons equivalent to the number of cardiac cycles for an intravenous bolus of gadolinium contrast to pass through the pulmonary circulation, and the content of each wagon equal to the RVSVI.

index (RVSVI) (Figure 1). This is better understood if one considers PBVI as being analogous to a cargo train, with the number of wagons equivalent to the number of cardiac cycles for an intravenous bolus of gadolinium contrast to pass through the pulmonary circulation, and the content of each wagon equal to the RVSVI. Statistical analysis We used the Kolmogorov–Smirnov test for normal distribution. Continuous variables were expressed as mean ± standard deviation or median and IQR, if not normally distributed. Categorical variables were expressed as frequency and percentage. The Student’s independent t-test, Mann–Whitney U test, analysis of variance (ANOVA), or Kruskal–Wallis test were used to compare continuous variables, while categorical variables were compared using the χ2 or Fisher’s exact test. Diagnostic accuracy of PBVI for the diagnosis of diastolic dysfunction was assessed by receiver operating characteristic (ROC) curve analysis. Uni- and multivariable logistic regression analyses were performed to determine predictors of diastolic dysfunction. The adjusted model was built by adding age, sex, and covariates found to be clinically relevant and/or statistically significant at univariate analysis. The significance level of all analyses was set at P < 0.05. Statistical analyses were performed using MedCalc for Windows, version 12.5 (MedCalc Software, Ostend, Belgium) and Wizard for Mac, version 1.7.8 (Boston, MA, USA).

riates found to be clinically relevant and/or statistically significant at univariate analysis. The significance level of all analyses was set at P < 0.05. Statistical analyses were performed using MedCalc for Windows, version 12.5 (MedCalc Software, Ostend, Belgium) and Wizard for Mac, version 1.7.8 (Boston, MA, USA). Results Baseline characteristics of the study population are summarized in Table 1. All patients (mean age 58 ± 11 years; 83% of men) were diagnosed with HCM based on clinical, electrocardiographic, echocardiographic and CMR findings. Asymmetric septal (48%) and apical (30%) hypertrophy were the most frequent morphologic HCM phenotypes. Overall, 32 (46%) patients were in New York Heart Association (NYHA) Class I, 29 (42%) in Class II, and 8 (12%) in Class III. Of 23 patients with diastolic dysfunction and increased LAP, 6 (26%) showed a restrictive filling pattern (Grade III). Table 1 Baseline characteristics of HCM outpatients by severity of diastolic dysfunction as assessed by transthoracic echocardiography

Results Baseline characteristics of the study population are summarized in Table 1. All patients (mean age 58 ± 11 years; 83% of men) were diagnosed with HCM based on clinical, electrocardiographic, echocardiographic and CMR findings. Asymmetric septal (48%) and apical (30%) hypertrophy were the most frequent morphologic HCM phenotypes. Overall, 32 (46%) patients were in New York Heart Association (NYHA) Class I, 29 (42%) in Class II, and 8 (12%) in Class III. Of 23 patients with diastolic dysfunction and increased LAP, 6 (26%) showed a restrictive filling pattern (Grade III). Table 1 Baseline characteristics of HCM outpatients by severity of diastolic dysfunction as assessed by transthoracic echocardiography Characteristics HCM (n = 69) Increased LAP Grade II–III (n = 23) Normal LAP Grade I (n = 46) P-value Age (years) 58 ± 11 57 ± 11 58 ± 11 NSa Gender (male), n (%) 57 (83) 18 (78) 39 (85) NSb BMI (kg/m2), median (IQR) 29 (6) 29 (8) 28.5 (6) NSc BSA (m2), median (IQR) 1.99 (0.30) 2.00 (0.39) 1.98 (0.31) NSc Diabetes, n (%) 12 (17) 3 (13) 9 (19) NSb Active smoker, n (%) 29 (42) 10 (43) 19 (41) NSb Hypertension, n (%) 28 (40) 10 (43) 18 (39) NSb NYHA class, n (%) I 32 (46) 6 (26) 26 (57) 0.017b II 29 (42) 10 (44) 19 (41) NSb III 8 (12) 7 (30) 1 (2) <0.001b IV 0 (0) 0 (0) 0 (0) Resting cardiac morphology and function by CMR imaging LVSVI (mL/m2), median (IQR) 62 (17) 61 (14) 50 (14) <0.001c LVEF (%) 73 ± 9 75 ± 8 72 ± 9 NSa LVEDVI (mL/m2), median (IQR) 74 (22) 95 (23) 78 (14) 0.006c LVESVI (mL/m2) 21 ± 10 21 ± 9 21 ± 10 NSa RVSVI (mL/m2) 51 (15) 55 (12) 47 (12) 0.004c RVEF (%) 68 ± 8 70 ± 9 67 ± 8 NSa RVEDVI (mL/m2), median IQR 73 (15) 78 (14) 68 (19) 0.005c RVESVI (mL/m2) 23 ± 10 24 ± 8 23 ± 11 NSa LA area index (cm2/m2) 14 ± 4 15 ± 4 14 ± 4 NSa LA volume index (mL/m2) 56 ± 12 69 ± 13 50 ± 12 <0.001a AP-LA diameter (mm) 40 ± 8 40 ± 12 40 ± 7 NSa Maximal wall thickness (mm) 18 ± 5 18 ± 4 18 ± 5 NSa LV mass index (g/m2) 91 ± 25 93 ± 32 90 ± 22 NSa SAM, n (%) 20 (30) 10 (43) 10 (22) NSb LVOTO, n (%) 11 (16) 4 (17) 7 (15) NSb MR,dn (%) 8 (12) 6 (26) 2 (4) 0.015b LGE, n (%) 48 (70) 19 (83) 29 (63) NSb Adenosine-induced perfusion defects, n (%) 52 (75) 19 (82) 33 (72) NSb Heart rate (bpm), rest 68 ± 11 67 ± 10 68 ± 11 NSa Heart rate (bpm), stress 84 ± 13 81 ± 4.5 87 ± 5.8 NSa Diastolic function assessment PTT (s) 6.7 ± 1.7 7.2 ± 1.5 6.4 ± 1.7 0.042a PTTc (s) 6.8 ± 1.6 7.7 ± 1.4 6.5 ± 1.6 0.014a PTTc stress (s) 7.5 ± 2 7.7 ± 2.1 7.4 ± 2.1 NSa PTB (bpm) 7.4 ± 1.8 8.0 ± 1.5 7.1 ± 1.8 0.047a PBVI (mL/m2) 361 ± 124 463 ± 127 310 ± 86 <0.001a PFRE (mL/s) 641 ± 201 666 ± 189 628 ± 207 NSa PFRE/BSA (mL/m2) 333 ± 101 352 ± 107 323 ± 98 NSc PFRE/EDV (/s) 4.5 ± 1.5 4.4 ± 1.2 4.7 ± 1.6 NSc PFRA (mL/s) 505 ± 196 563 ± 243 477 ± 165 NSa PFRA/BSA (mL/m2) 261 ± 91 293 ± 107 245 ± 79 NSc PFRA/EDV (/s) 3.5 ± 1.1 3.7 ± 1.2 3.5 ± 1.1 NSc PFRE/PFRA 1.4 ± 0.4 1.3 ± 0.4 1.4 ± 0.5 NSc sPAP (mmHg)

666 ± 189 628 ± 207 NSa PFRE/BSA (mL/m2) 333 ± 101 352 ± 107 323 ± 98 NSc PFRE/EDV (/s) 4.5 ± 1.5 4.4 ± 1.2 4.7 ± 1.6 NSc PFRA (mL/s) 505 ± 196 563 ± 243 477 ± 165 NSa PFRA/BSA (mL/m2) 261 ± 91 293 ± 107 245 ± 79 NSc PFRA/EDV (/s) 3.5 ± 1.1 3.7 ± 1.2 3.5 ± 1.1 NSc PFRE/PFRA 1.4 ± 0.4 1.3 ± 0.4 1.4 ± 0.5 NSc sPAP (mmHg) ,e median (IQR) 29 (13) 37 (16) 27 (9) 0.003c E wave (cm/s),e median (IQR) 70 (26) 78 (17) 62 (26) 0.009c E/A,e median (IQR) 1.2 (0.5) 1.2 (0.4) 1.1 (0.5) NSc LAVI (mL/m2),e median (IQR) 49 (26) 68 (45) 47 (14) 0.011c E/E',e median (IQR) 12 (5) 15 (3.5) 11 (4) <0.001c Data are presented as mean ± SD unless otherwise indicated.

n (IQR) 29 (13) 37 (16) 27 (9) 0.003c E wave (cm/s),e median (IQR) 70 (26) 78 (17) 62 (26) 0.009c E/A,e median (IQR) 1.2 (0.5) 1.2 (0.4) 1.1 (0.5) NSc LAVI (mL/m2),e median (IQR) 49 (26) 68 (45) 47 (14) 0.011c E/E',e median (IQR) 12 (5) 15 (3.5) 11 (4) <0.001c Data are presented as mean ± SD unless otherwise indicated. AP, anteroposterior; BMI, body mass index; BSA, body surface area; EDV, end-diastolic volume; HCM, hypertrophic cardiomyopathy; IQR, interquartile range; LA, left atrium; LAP, left atrial pressure; LAVI, left atrial volume index; LGE, late gadolinium enhancement; LV, left ventricle; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; LVESVI, left ventricular end-systolic volume index; MR, mitral regurgitation; NS, not significant; PBVI, pulmonary blood volume index; PFRA, atrial (active) peak filling rate; PFRE, early (passive) peak filling rate; PTB, pulmonary transit beats; PTT, pulmonary transit time; PTTc, pulmonary transit time corrected by Bazett's formula; RA, right atrium; RVEDVI, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction; RVESVI, right ventricular end-systolic volume index; SAM, systolic anterior movement; SD, standard deviation; sPAP, systolic pulmonary artery pressure. a The Student's t-test for unpaired data. b The χ2 test. c The Mann–Whitney U test. d At least moderate. e Assessed by transthoracic echocardiography.

AP, anteroposterior; BMI, body mass index; BSA, body surface area; EDV, end-diastolic volume; HCM, hypertrophic cardiomyopathy; IQR, interquartile range; LA, left atrium; LAP, left atrial pressure; LAVI, left atrial volume index; LGE, late gadolinium enhancement; LV, left ventricle; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; LVESVI, left ventricular end-systolic volume index; MR, mitral regurgitation; NS, not significant; PBVI, pulmonary blood volume index; PFRA, atrial (active) peak filling rate; PFRE, early (passive) peak filling rate; PTB, pulmonary transit beats; PTT, pulmonary transit time; PTTc, pulmonary transit time corrected by Bazett's formula; RA, right atrium; RVEDVI, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction; RVESVI, right ventricular end-systolic volume index; SAM, systolic anterior movement; SD, standard deviation; sPAP, systolic pulmonary artery pressure. a The Student's t-test for unpaired data. b The χ2 test. c The Mann–Whitney U test. d At least moderate. e Assessed by transthoracic echocardiography. Compared with patients without elevated LAP, patients with increased LAP showed significantly larger biventricular end-diastolic volumes, stroke volumes and left atrium volume index, longer PTTc, larger PBVI, and more severe mitral regurgitation (Table 1). Presence and extent of myocardial fibrosis, adenosine inducible perfusion defects, and prevalence of systolic anterior movement and LV outflow obstruction were similar between the two groups (P = not significant).

Compared with patients without elevated LAP, patients with increased LAP showed significantly larger biventricular end-diastolic volumes, stroke volumes and left atrium volume index, longer PTTc, larger PBVI, and more severe mitral regurgitation (Table 1). Presence and extent of myocardial fibrosis, adenosine inducible perfusion defects, and prevalence of systolic anterior movement and LV outflow obstruction were similar between the two groups (P = not significant). Compared with resting measurement, PTTc significantly increased during adenosine stress in patients with normal LAP (P = 0.045), remaining unchanged in the subgroup of patients with elevated LAP. PBVI increased progressively with worsening NYHA functional class (Figure 2) and echocardiographic stages of diastolic dysfunction (ANOVA; P < 0.001 for both) (Figure 3). In the subgroup of asymptomatic or mildly symptomatic patients (NYHA functional Class I or II), patients with increased LAP showed higher PBVI compared with patients with normal LAP (PBVI 419 ± 125 vs. 308 ± 88 mL/m2; P < 0.0001). Figure 2 PBVI by NYHA functional class status. NYHA, New York Heart Association; PBVI, pulmonary blood volume index. Figure 3 PBVI and echocardiographic stages of diastolic dysfunction. PBVI, pulmonary blood volume index.

PBVI increased progressively with worsening NYHA functional class (Figure 2) and echocardiographic stages of diastolic dysfunction (ANOVA; P < 0.001 for both) (Figure 3). In the subgroup of asymptomatic or mildly symptomatic patients (NYHA functional Class I or II), patients with increased LAP showed higher PBVI compared with patients with normal LAP (PBVI 419 ± 125 vs. 308 ± 88 mL/m2; P < 0.0001). Figure 2 PBVI by NYHA functional class status. NYHA, New York Heart Association; PBVI, pulmonary blood volume index. Figure 3 PBVI and echocardiographic stages of diastolic dysfunction. PBVI, pulmonary blood volume index. PBVI yielded overall good diagnostic accuracy for diastolic dysfunction [area under the ROC curve 0.83, 95% confidence interval (CI) 0.73–0.94], with the best cut-off point corresponding to 413 mL/m2 and leading to a sensitivity of 60% (95% CI 38.5–80.3) and a specificity of 91% (95% CI 78.8–97.5) (Figure 4). E/e’, LA volume index, and sPAP were significantly higher in patients with PBVI >413 mL/m2 as compared to those below this threshold (Figure 5). PBVI was significantly associated with several echocardiographic indices of diastolic dysfunction and showed significant positive correlation with tissue-Doppler E/e’ ratio (r = 0.275, P = 0.026), sPAP (r = 0.284, P = 0.02), and LA volume index (r = 0.506, P < 0.001). On multivariate logistic regression analysis, PBVI was an independent predictor of LV diastolic dysfunction with elevated LAP (Table 2).

dysfunction and showed significant positive correlation with tissue-Doppler E/e’ ratio (r = 0.275, P = 0.026), sPAP (r = 0.284, P = 0.02), and LA volume index (r = 0.506, P < 0.001). On multivariate logistic regression analysis, PBVI was an independent predictor of LV diastolic dysfunction with elevated LAP (Table 2). Figure 4 Diagnostic accuracy of PBVI for the diagnosis of diastolic dysfunction with increased mean left atrial pressure. AUROC, area under the receiver operating characteristic curve; CI, confidence interval; PBVI, pulmonary blood volume index. Figure 5 Echocardiographic parameters of diastolic function in patients with higher and lower PBVI. Patients with PBVI >413 mL/m2 (corresponding to the best cut-off point derived from receiver operating characteristic curve analysis) had significantly higher values of E/e’, LAVI, and sPAP than the other subgroup of patients with PBVI ≤413 mL/m2. AUROC, area under the ROC curve; E/A, mitral valve E velocity divided by A-wave velocity; E/e’, mitral valve E velocity divided by mitral annular e’ velocity; PBVI, pulmonary blood volume index; LAVI, left atrial volume index; sPAP, systolic pulmonary artery pressure. Table 2 Univariate and multivariate logistic regression analysis of CMR predictors of haemodynamic congestion

Figure 5 Echocardiographic parameters of diastolic function in patients with higher and lower PBVI. Patients with PBVI >413 mL/m2 (corresponding to the best cut-off point derived from receiver operating characteristic curve analysis) had significantly higher values of E/e’, LAVI, and sPAP than the other subgroup of patients with PBVI ≤413 mL/m2. AUROC, area under the ROC curve; E/A, mitral valve E velocity divided by A-wave velocity; E/e’, mitral valve E velocity divided by mitral annular e’ velocity; PBVI, pulmonary blood volume index; LAVI, left atrial volume index; sPAP, systolic pulmonary artery pressure. Table 2 Univariate and multivariate logistic regression analysis of CMR predictors of haemodynamic congestion Characteristics Univariate Multivariatea Odds ratio (95% CI) P-value Odds ratio (95% CI) P-value Age (years) 1.01 (0.95–1.04) 0.748 0.98 (0.88–1.09) 0.679 Male sex 0.65 (0.18–2.32) 0.502 0.04 (0.0–5.95) 0.204 NYHA class 3.96 (1.67–9.39) 0.002 0.95 (0.16–5.52) 0.305 LVEDVI (mL/m2) 1.04 (1.01–1.07) 0.022 0.7 (0.32–1.53) 0.366 LVSVI (mL/m2) 1.09 (1.03–1.14) 0.002 1.63 (0.54–4.9) 0.386 LVEF (%) 1.05 (0.98–1.12) 0.144 0.66 (0.26–1.66) 0.375 RVEDVI (mL/m2) 1.06 (1.01–1.11) 0.043 0.65 (0.37–1.16) 0.146 RVSVI (mL/m2) 1.13 (1.03–1.23) 0.008 2.19 (0.85–5.59) 0.103 RVEF (%) 1.04 (0.96–1.12) 0.327 0.58 (0.29–1.14) 0.115 MRb (binary) 2.92 (1.02–8.34) 0.046 0.27 (0.25–6.34) 0.770 LGE (binary) 2.88 (0.84–9.92) 0.093 4.0 (0.24–65.44) 0.332 Adenosine inducible perfusion defects (binary) 1.93 (0.55–6.78) 0.305 0.35 (0.02–7.62) 0.507 LV mass indexc (g/m2) 1.01 (0.84–1.2) 0.932 0.86 (0.55–1.33) 0.511 PBVIc (mL/m2) 1.77 (1.34–2.33) <0.001 1.97 (1.06–3.68) 0.032 CI, confidence interval; LGE, late gadolinium enhancement; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; PBVI, pulmonary blood volume index.

4–1.2) 0.932 0.86 (0.55–1.33) 0.511 PBVIc (mL/m2) 1.77 (1.34–2.33) <0.001 1.97 (1.06–3.68) 0.032 CI, confidence interval; LGE, late gadolinium enhancement; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; PBVI, pulmonary blood volume index. a Adjusted for all covariates (CI, LGE, LVEDVI, LVEF, NYHA, and PBVI.) listed in the tables. b At least moderate. c Per 10% increase. Discussion In the present study, we used first-pass perfusion CMR imaging to estimate pulmonary blood volume and transit time in patients with HCM. The major findings of this study were that: (i) increases in PBVI yield good diagnostic accuracy for the detection of diastolic dysfunction with elevated LAP (as defined by echocardiography), providing excellent specificity at the best cut-off level of 413 mL/m2; (ii) HCM patients with Grade II or III diastolic dysfunction also showed significantly larger PBVI and longer PTTc as compared to patients with normal LAP; and (iii) the magnitude of PBVI correlates with the severity of diastolic dysfunction and worsening functional capacity.

ity at the best cut-off level of 413 mL/m2; (ii) HCM patients with Grade II or III diastolic dysfunction also showed significantly larger PBVI and longer PTTc as compared to patients with normal LAP; and (iii) the magnitude of PBVI correlates with the severity of diastolic dysfunction and worsening functional capacity. The symptoms in patients with HCM largely result from elevated mean LAP (‘backward’ HF).1 However, in many cases backward and forward failure coexist in a complex interplay of pathophysiological mechanisms, including coronary microvascular dysfunction with reduced arteriolar density and structural abnormalities of intramural coronary arterioles, replacement fibrosis, myocyte disarray, mitral regurgitation associated with systolic anterior movement, and LV outflow tract obstruction as well as elevated LV end-diastolic pressure, eventually leading to the elevation of pulmonary capillary hydrostatic pressure (or haemodynamic congestion) and, at an advanced stage, to overt pulmonary congestion.15–18

ay, mitral regurgitation associated with systolic anterior movement, and LV outflow tract obstruction as well as elevated LV end-diastolic pressure, eventually leading to the elevation of pulmonary capillary hydrostatic pressure (or haemodynamic congestion) and, at an advanced stage, to overt pulmonary congestion.15–18 Since haemodynamic derangements usually precede clinical symptoms, the ability to detect and grade haemodynamic congestion from non-invasive measurements is a highly desirable though unmet clinical need.19 Echocardiography is the recommended non-invasive imaging modality for the evaluation of LV diastolic function, but may be sometimes misleading, with key indices yielding discrepant information, especially in HCM,20 and the use of many parameters is perceived too complex and not practical to perform and repeat on every patient.8 Furthermore, diagnostic accuracy of the 2009 and 2016 echocardiographic grading algorithms for the diagnosis of LV diastolic dysfunction has been recently questioned by the results of the multicentre EACVI Euro-Filling study.21 CMR is the gold-standard imaging technique for the non-invasive quantification of blood flow and volumes22 and has been increasingly recognized as a valuable complement to TTE for the clinical evaluation of HCM, providing a number of unique strengths which make it particularly well suited to detailed characterization of the HCM phenotype, and therefore, an important aid for diagnosis and potentially prognosis.23–26

olumes22 and has been increasingly recognized as a valuable complement to TTE for the clinical evaluation of HCM, providing a number of unique strengths which make it particularly well suited to detailed characterization of the HCM phenotype, and therefore, an important aid for diagnosis and potentially prognosis.23–26 Furthermore, comprehensive CMR imaging protocols have been advocated to improve diagnosis and targeted management of HF with preserved LV ejection fraction.27 While a non-invasive gold standard for the quantification of diastolic function in HCM is still an unmet clinical need,28 CMR has been proposed as an alternative non-invasive modality for the evaluation of LV diastolic function, although the relatively low temporal resolution—i.e. 20–30 ms—limited its clinical application so far.29,30 More recently, PBVI by CMR has proved to be helpful to quantitatively determine haemodynamic congestion—meaning a greater volume of blood in the pulmonary circulation leading to higher hydrostatic pressure—in HF outpatients with reduced ejection fraction11 and adults with repaired congenital heart disease.31 In this study, we confirmed previous findings11,32 and reported a positive relationship between PBVI and echocardiographic stages of diastolic dysfunction in a cohort of both asymptomatic and symptomatic HCM patients. Furthermore, increasing values of PBVI were significantly associated with worsening functional status, i.e. higher NYHA functional class and lower exercise time duration.

sitive relationship between PBVI and echocardiographic stages of diastolic dysfunction in a cohort of both asymptomatic and symptomatic HCM patients. Furthermore, increasing values of PBVI were significantly associated with worsening functional status, i.e. higher NYHA functional class and lower exercise time duration. Overall, PBVI might be considered as a quantitative and reproducible biomarker of haemodynamic congestion—a well-recognized unmet clinical need19—and might represent a specific tool to unmask poor haemodynamic reserve, eventually useful for the clarification of the pathophysiology underlying impaired functional capacity and HF symptoms in HCM patients. Assessment of resting PBVI and transit time kinetics may compliment functional analysis and tissue characterization and can be ideally performed on a routine basis without affecting total scan time duration.

he clarification of the pathophysiology underlying impaired functional capacity and HF symptoms in HCM patients. Assessment of resting PBVI and transit time kinetics may compliment functional analysis and tissue characterization and can be ideally performed on a routine basis without affecting total scan time duration. After adenosine infusion, PTTc significantly increased in patients with better diastolic function, while it remained unchanged in patients with diastolic dysfunction and elevated LAP. We know that similar to the systemic and coronary circulation, the pulmonary arteries dilate in response to purine nucleoside adenosine which has a direct endothelium-independent effect on the A2b receptor in vascular smooth muscle.33 We may argue that reduced pulmonary resistance in response to the endothelium-independent vasodilatory effect of adenosine would lead to an increased transit time of the contrast agent (PTT = PBV/cardiac output) due to the distention of the pulmonary vascular bed and recruitment of the pulmonary microcirculation, overall expanding the pulmonary capillary surface area and volume.34 Conversely, pulmonary vasodilatory reserve may have been exhausted in our HCM patients with elevated LAP. However, other authors would consider constant PTTc at rest and during stress in keeping with an intact vasoreactive response to accommodate increased cardiac output during adenosine infusion.33 A number of factors may act on endothelium-independent function of the pulmonary vasculature and accuracy of PTT measurement, namely active cigarette smoking, which has been associated with impaired vasoreactivity during CMR adenosine stress testing, adenosine dosing and delay between stress and rest perfusion imaging acquisition. Nevertheless, the comparison of pulmonary blood volume variation over the cardiac cycle, as previously described,12,35 at rest and during adenosine stress would best explain the differences in central transit time kinetics observed between patients with increased and normal LAP.

and rest perfusion imaging acquisition. Nevertheless, the comparison of pulmonary blood volume variation over the cardiac cycle, as previously described,12,35 at rest and during adenosine stress would best explain the differences in central transit time kinetics observed between patients with increased and normal LAP. Further studies are warranted in order to: (i) assess the clinical utility and prognostic value of serial PBVI analysis in HCM population as well as in other clinical settings; (ii) investigate the association between PBVI and serum levels of natriuretic peptides; (iii) investigate the relationship between symptoms and PBVI before and after structural interventions, i.e. alcohol septal ablation; (iv) test the effect of adenosine—and/or exercise stress CMR—on pulmonary vascular function and right heart-pulmonary circulation unit; (v) test the potential usefulness PBVI in the assessment of HF with preserved LV ejection fraction, specifically in the subgroup in which diastolic function assessment according to the current guidelines is normal or indeterminate; and (vi) test the diagnostic accuracy of PBVI with established gold-standard invasive techniques for the measurement of pulmonary artery wedge pressure and LV end-diastolic pressure.

tion, specifically in the subgroup in which diastolic function assessment according to the current guidelines is normal or indeterminate; and (vi) test the diagnostic accuracy of PBVI with established gold-standard invasive techniques for the measurement of pulmonary artery wedge pressure and LV end-diastolic pressure. Finally, from a wider perspective, adenosine stress CMR imaging might represent a unique tool for simultaneous assessment of myocardial ischaemia—also able to distinguish between obstructive epicardial CAD and microvascular coronary dysfunction36—and diastolic function reserve, which are both involved in a vicious cycle harbinger of future risk of HF with preserved ejection fraction hospitalization.37

ique tool for simultaneous assessment of myocardial ischaemia—also able to distinguish between obstructive epicardial CAD and microvascular coronary dysfunction36—and diastolic function reserve, which are both involved in a vicious cycle harbinger of future risk of HF with preserved ejection fraction hospitalization.37 Limitations A few limitations should be addressed. Firstly, due to the retrospective design, the referral to stress perfusion CMR imaging may have introduced a selection bias of patients with higher burden of symptoms, however, most patients presented with NYHA functional III or better. Secondly, right ventricular stroke volume was calculated as the difference between end-diastolic and end-systolic volume. For the purpose of this study, measurement of forward flow by phase-contrast velocity mapping would have been more precise; however, exclusion of patients with moderate or severe tricuspid regurgitation attenuated possible inaccuracy. Thirdly, we could not exclude coexisting coronary artery disease as a possible mechanism underlying diastolic dysfunction in all patients, however, this was out of the main scope of our analysis. Fourthly, because of the limited sample size, the multivariable analysis is likely to be underpowered to derive a robust predictive model and our findings should only be intended as hypothesis-generating. Fifthly, the accuracy of measuring PBVI may be limited due to the uncertain relationship between the bolus concentration of contrast material and the SI on CMR images used for PTT measurements. Sixthly, we did not obtain cardiac volumes during stress and could not assess changes in PBVI and cardiac output after adenosine administration. Furthermore, invasive haemodynamic measurements derived from right and/or left heart catheterization were not available in the current study, however it has been recently showed that PTT prolongation was significantly associated with haemodynamic abnormalities at invasive haemodynamic testing by right and left heart catheterization such as elevated pulmonary capillary wedge pressure, LV end-diastolic pressure, reduced cardiac index, and oxygen saturation, of which increased pulmonary capillary wedge pressure demonstrated the strongest association.10 Finally, serum levels of natriuretic peptides, cardiopulmonary exercise testing data, and T1 mapping indices were not routinely available in our series.

V end-diastolic pressure, reduced cardiac index, and oxygen saturation, of which increased pulmonary capillary wedge pressure demonstrated the strongest association.10 Finally, serum levels of natriuretic peptides, cardiopulmonary exercise testing data, and T1 mapping indices were not routinely available in our series. Conclusions PBVI analysis is a promising application for assessment of haemodynamic congestion which may effectively compliment the functional assessment and tissue characterization by CMR in patients with HCM. Future prospective studies should investigate the prognostic role and the clinical utility of PBVI in targeted management of patients with HCM, fostering the wider objective of imaging-guided precision medicine in HF. Acknowledgements EACVI (F.R./S.E.P.): F.R. and S.E.P. acknowledge the support through a grant by the European Association of Cardiovascular Imaging. S.E.P. acknowledges support from the NIHR Barts Biomedical Research Centre and the ‘SmartHeart’ EPSRC programme grant (EP/P001009/1). K.F. acknowledges support by The Medical College of Saint Bartholomew’s Hospital Trust, an independent registered charity that promotes and advances medical and dental education and research at Barts and The London School of Medicine and Dentistry. This work uses data provided by patients and collected by the NHS as part of their care and support.

The Medical College of Saint Bartholomew’s Hospital Trust, an independent registered charity that promotes and advances medical and dental education and research at Barts and The London School of Medicine and Dentistry. This work uses data provided by patients and collected by the NHS as part of their care and support. Funding This work was supported by a Wellcome Trust Research Training Fellowship [203553/Z/Z to N.A.]. This work was directly funded by the National Institute for Health Research Biomedical Research Centre at Barts Heart Centre. Conflict of interest: S.E.P. provides Consultancy to Circle Cardiovascular Imaging Inc., Calgary, Alberta, Canada and has options for shares. And all other authors have no conflict of interest to declare.

Introduction Ischaemic heart disease is the leading cause of morbidity and mortality in the Western world. The number of patients who developed an acute myocardial infarction has been increasing exponentially. It was recently estimated that 935 000 new cases of acute coronary occlusion are identified each year in the USA.1 Several experimental studies have demonstrated that the region supplied by the occluded coronary artery, also known as myocardium at risk (MaR), is a major independent variable that determines the final infarct size, which in turn is closely related to the clinical outcome.2–7 The difference between MaR and infarct size is used to calculate the myocardial salvage index, which is a measurement of the effectiveness of acute interventions aimed at reducing the extent of myocardial infarction.8,9 Thus, a feasible and accurate method for determination of MaR is of significant value in clinical studies evaluating the efficiency of cardioprotective therapies.

te the myocardial salvage index, which is a measurement of the effectiveness of acute interventions aimed at reducing the extent of myocardial infarction.8,9 Thus, a feasible and accurate method for determination of MaR is of significant value in clinical studies evaluating the efficiency of cardioprotective therapies. Cardiovascular magnetic resonance (CMR) imaging is currently considered the reference standard for in vivo assessment of myocardial infarction using late gadolinium-enhanced (LGE) imaging.10 Recently, CMR has also been introduced as a promising method for assessing MaR using T2-weighted imaging11–14 up to 1 week after the acute event.15 Thus, CMR can be used to calculate the myocardial salvage index by a single CMR imaging session several days after the acute event in a stable clinical setting. Even though T2-weighted imaging is promising for the assessment of MaR, this technique still has certain limitations such as difficulties in distinguishing blood pool from the myocardium, especially in the apical parts of the left ventricle (LV) where hypokinesia and trabeculation cause stagnant blood flow.16

setting. Even though T2-weighted imaging is promising for the assessment of MaR, this technique still has certain limitations such as difficulties in distinguishing blood pool from the myocardium, especially in the apical parts of the left ventricle (LV) where hypokinesia and trabeculation cause stagnant blood flow.16 More recently, a new CMR method for assessment of MaR, referred to as contrast-enhanced steady-state free precession (CE-SSFP), has been introduced.17 This technique is based on acquisition of time-resolved steady-state free precession images after injection of the gadolinium-based contrast agent and was recently validated against myocardial perfusion single-photon emission computed tomography (SPECT).17 Still, CE-SSFP has not been compared head-to-head with T2-weighted imaging. Therefore, the aim of the present study was to explore the relationship between CE-SSFP  and T2-weighted imaging with regard to MaR in patients with first-time reperfused acute myocardial infarction.

More recently, a new CMR method for assessment of MaR, referred to as contrast-enhanced steady-state free precession (CE-SSFP), has been introduced.17 This technique is based on acquisition of time-resolved steady-state free precession images after injection of the gadolinium-based contrast agent and was recently validated against myocardial perfusion single-photon emission computed tomography (SPECT).17 Still, CE-SSFP has not been compared head-to-head with T2-weighted imaging. Therefore, the aim of the present study was to explore the relationship between CE-SSFP  and T2-weighted imaging with regard to MaR in patients with first-time reperfused acute myocardial infarction. Methods Study population The protocol and procedures were approved by the regional research ethics committee and all patients gave their written consent. Between April 2009 and October 2010, 21 patients (age 59 ± 10 years, 17 males) presenting with first-time acute ST-elevation myocardial infarction (STEMI), due to an occluded coronary artery confirmed by angiography, were prospectively included in the study. All patients were treated with primary percutaneous coronary intervention with coronary stenting, resulting in TIMI grade III flow in the culprit artery. Five patients have been included in an earlier study.18

STEMI), due to an occluded coronary artery confirmed by angiography, were prospectively included in the study. All patients were treated with primary percutaneous coronary intervention with coronary stenting, resulting in TIMI grade III flow in the culprit artery. Five patients have been included in an earlier study.18 Cardiovascular magnetic resonance One week after admission, CMR was performed using a 1.5 T Philips Intera CV (Philips, Best, The Netherlands). Images were obtained using a five-element chest array surface coil with two anterior and three posterior elements. All subjects were placed in supine position and images were acquired at end-expiratory breath hold with vectrocardiographic gating. Initial scout images were acquired to locate the heart, and a dark-blood T2-weighted triple inversion turbo spin-echo sequence (T2-STIR) was employed to depict the MaR. T2-weighted images were acquired in the short-axis view, covering the LV from the base to the apex. Imaging parameters for the T2-weighted sequence were: echo time, 100 ms; repetition time, 2 heart beats; echo train length, 33; number of averages, 2; inversion time, 180 ms; image resolution, 1.5 × 1.5 × 8 mm; slice gap, 0 mm. No parallel imaging was performed, but surface coil intensity correction was performed to minimize the signal inhomogeneities due to differences in receiver coil sensitivity.

n time, 2 heart beats; echo train length, 33; number of averages, 2; inversion time, 180 ms; image resolution, 1.5 × 1.5 × 8 mm; slice gap, 0 mm. No parallel imaging was performed, but surface coil intensity correction was performed to minimize the signal inhomogeneities due to differences in receiver coil sensitivity. Acquisition of short-axis, retrospectively gated SSFP cine images, covering the LV from the base to the apex, were initiated within 8 min after the administration of 0.2 mmol/kg extracellular gadolinium-based contrast agent (gadoteric acid, Gd-DOTA; Guerbet, Gothia Medical AB, Billdal, Sweden). These images were referred to as CE-SSFP images (see Supplementary data online, Movie S1). Typical image parameters were: echo time, 1.4 ms; repetition time, 2.9 ms; flip angle, 60°; image resolution, 1.5 × 1.5 × 8 mm; slice gap, 0 mm. No parallel imaging was performed to maximize the signal-to-noise ratio (SNR). Long- and short-axis LGE images covering the LV were then acquired ∼15 min after injection of gadolinium. The LGE images were acquired with an inversion-recovery sequence with following image parameters: slice thickness, 8 mm; field of view, 340 mm; flip angle, 15°; repetition time, 4.2 ms; echo time, 1.3 ms; image resolution, 1.4 × 1.4 × 8 mm; slice gap, 0 mm; parallel imaging with a SENSE factor of 2. The inversion time was adjusted to null the signal from viable myocardium.19 Image analysis All CMR images were analysed using the freely available software Segment v1.8 (http://segment.heiberg.se).20

Long- and short-axis LGE images covering the LV were then acquired ∼15 min after injection of gadolinium. The LGE images were acquired with an inversion-recovery sequence with following image parameters: slice thickness, 8 mm; field of view, 340 mm; flip angle, 15°; repetition time, 4.2 ms; echo time, 1.3 ms; image resolution, 1.4 × 1.4 × 8 mm; slice gap, 0 mm; parallel imaging with a SENSE factor of 2. The inversion time was adjusted to null the signal from viable myocardium.19 Image analysis All CMR images were analysed using the freely available software Segment v1.8 (http://segment.heiberg.se).20 The MaR derived from T2-weighted imaging was assessed according to a previously described methodology.15 In short, endocardial and epicardial borders of the LV were traced in all short-axis slices, followed by manual delineation of the hyperintense regions by two blinded observers. The papillary muscles were excluded from the myocardium. The MaR was then defined as the total amount of hyperintense myocardium in all short-axis slices and expressed as percentage of LV. If present, hypo-intense myocardium within the area of increased signal intensity (microvascular obstruction) was included in the MaR.

papillary muscles were excluded from the myocardium. The MaR was then defined as the total amount of hyperintense myocardium in all short-axis slices and expressed as percentage of LV. If present, hypo-intense myocardium within the area of increased signal intensity (microvascular obstruction) was included in the MaR. The MaR derived from CE-SSFP was also assessed according to previously described methodology.17 In short, endocardial and epicardial borders of the LV were traced in all short-axis slices in end-diastole and end-systole, followed by manual delineation of the hyperintense regions in both end-diastole and end-systole, by two observers blinded to both the T2-weighted and LGE images. The values of MaR in end-diastole and end-systole were averaged and expressed as a percentage of the LV. The contrast ratio (CR) for the CE-SSFP images was determined for each patient as the mean signal intensity in the MaR divided by the mean signal intensity in remote myocardium. A slice-by-slice comparison of MaR between T2-weighted imaging and CE-SSFP was also performed for corresponding slice positions.

The MaR derived from CE-SSFP was also assessed according to previously described methodology.17 In short, endocardial and epicardial borders of the LV were traced in all short-axis slices in end-diastole and end-systole, followed by manual delineation of the hyperintense regions in both end-diastole and end-systole, by two observers blinded to both the T2-weighted and LGE images. The values of MaR in end-diastole and end-systole were averaged and expressed as a percentage of the LV. The contrast ratio (CR) for the CE-SSFP images was determined for each patient as the mean signal intensity in the MaR divided by the mean signal intensity in remote myocardium. A slice-by-slice comparison of MaR between T2-weighted imaging and CE-SSFP was also performed for corresponding slice positions. The infarcted myocardium was automatically quantified from the short-axis LGE images according to a previously described method.21 In short, the endocardial and epicardial borders were traced manually with exclusion of the papillary muscles. The LGE myocardium was then defined using a computer algorithm that takes into consideration partial volume effects within the infarcted region. Manual adjustments were made when image artefacts caused misinterpretation by the computer algorithm. If present, a hypointense signal within the area of LGE (microvascular obstruction) was included in the analysis as 100% infarction. Finally, the myocardial infarct size was expressed as a percentage of the LV.

region. Manual adjustments were made when image artefacts caused misinterpretation by the computer algorithm. If present, a hypointense signal within the area of LGE (microvascular obstruction) was included in the analysis as 100% infarction. Finally, the myocardial infarct size was expressed as a percentage of the LV. The myocardial salvage index was defined as: 100*[(MaR – infarct size)/MaR], where MaR was assessed by both T2-weighted imaging and CE-SSFP. The SNR ratio, contrast-to-noise (CNR) ratio and contrast ratio (CR) were determined for T2-weighted, CE-SSFP, and LGE images, respectively. The SNR was calculated as the mean signal intensity within the affected region (MaR or infarcted myocardium) divided by the standard deviation of signal intensities within a background region of interest. The CNR was calculated as the SNR in the affected region (MaR or infarcted myocardium) – SNR within remote myocardium and CR was calculated as the mean signal intensity in the affected region (MaR or infarcted myocardium) divided by the mean signal intensity in remote myocardium.

within a background region of interest. The CNR was calculated as the SNR in the affected region (MaR or infarcted myocardium) – SNR within remote myocardium and CR was calculated as the mean signal intensity in the affected region (MaR or infarcted myocardium) divided by the mean signal intensity in remote myocardium. Statistics Continuous variables are presented as mean ± SD if nothing else specified. Pearson's correlation was used to determine the relationship between T2-weighted imaging and CE-SSFP with regard to MaR and myocardial salvage index. The agreement between T2-weigthed imaging and CE-SSFP was expressed as mean difference ± SD, and the limits of agreement were shown in a Bland–Altman graph as mean ± 2 SD. The inter-observer variability was expressed as mean difference ± SD. A paired t-test was used to compare the means of the contrast ratio between MaR and remote myocardium for both T2-weighted imaging and CE-SSFP as well as difference in MaR by T2-weighted imaging and CE-SSFP. SPSS version 17.0 software package (Chicago, IL, USA) was used for analysis. Results with a P-value of <0.05 were considered statistically significant.

means of the contrast ratio between MaR and remote myocardium for both T2-weighted imaging and CE-SSFP as well as difference in MaR by T2-weighted imaging and CE-SSFP. SPSS version 17.0 software package (Chicago, IL, USA) was used for analysis. Results with a P-value of <0.05 were considered statistically significant. Results In 59% (12 of 21) of the patients, the right coronary artery (RCA) was the culprit vessel and in 29% (6 of 21) of the patients, the left anterior descending coronary artery (LAD) was the culprit vessel. Furthermore, 2 patients presented with an occlusion of the left circumflex coronary artery and 1 patient had a left main occlusion. In all patients, T2-weighted imaging and CE-SSFP identified the MaR in the same perfusion territory and in concordance with angiography. Figure 1A shows an example of multi-slice short-axis images from the base to the apex demonstrating MaR by T2-weighted imaging and CE-SSFP as well as myocardial infarction by LGE, in one patient with an occlusion in the left anterior descending coronary artery. For one patient, myocardial infarct size could not be determined due to poor LGE image quality resulting from frequent arrhythmias. Another patient had a clinical history of a prior infarction in a coronary territory different from that supplied by the current culprit vessel. This patient had no signs of infarction by LGE imaging in the part of the myocardium previously reported to be infarcted. Therefore, this patient was included in the present study. Figure 1 The myocardium at risk by T2-weighted imaging and CE-SSFP, and the infarct size by late gadolinium enhancement. (A) T2-weighted, CE-SSFP, and late gadolinium-enhanced short-axis images at the corresponding LV levels in a patient with a reperfused left anterior descending coronary artery occlusion. (B) Single corresponding mid-ventricular short-axis images from a patient with an occlusion in the left anterior descending coronary artery (LAD), left circumflex coronary artery (LCx), and the right coronary artery (RCA), respectively. The epicardium is traced in green and the endocardium is traced in red. The hyperenhanced regions constituting the myocardium at risk (dashed arrows) and the infarcted myocardium (solid arrows) are traced in white. Note the similarity in location and the extent of the affected region between T2-weighted imaging and CE-SSFP.

picardium is traced in green and the endocardium is traced in red. The hyperenhanced regions constituting the myocardium at risk (dashed arrows) and the infarcted myocardium (solid arrows) are traced in white. Note the similarity in location and the extent of the affected region between T2-weighted imaging and CE-SSFP. Also note the significantly smaller infarction compared with the myocardium at risk indicating a significant myocardial salvage accomplished by the acute reperfusion therapy. CE-SSFP, contrast-enhanced steady-state free precession. Myocardium at risk A region with increased signal intensity by T2-weighted imaging and CE-SSFP was observed in all patients, yielding a mean MaR of 29 ± 11% (range 12–65) and 32 ± 12% (range 8–70) of the LV, respectively (Table 1). Table 1 Myocardium at risk, infarct size and myocardial salvage index for each patient

Also note the significantly smaller infarction compared with the myocardium at risk indicating a significant myocardial salvage accomplished by the acute reperfusion therapy. CE-SSFP, contrast-enhanced steady-state free precession. Myocardium at risk A region with increased signal intensity by T2-weighted imaging and CE-SSFP was observed in all patients, yielding a mean MaR of 29 ± 11% (range 12–65) and 32 ± 12% (range 8–70) of the LV, respectively (Table 1). Table 1 Myocardium at risk, infarct size and myocardial salvage index for each patient Case no. Culprit vessel Myocardium at risk by T2W (%) Myocardium at risk by CE-SSFP (%) Infarct size by LGE (%) Myocardial salvage index (%) by Obs 1 Obs 2 Mean Obs 1 Obs 2 Mean Obs 1 Obs 2 Mean T2W CE-SSFP 1 LM 65 60 63 70 66 68 49 43 46 26 32 2 LAD 33 27 30 43 34 39 9 10 10 70 76 3 LAD 19 20 20 35 28 32 5 6 6 72 83 4 RCA 39 23 31 24 37 31 23 24 24 24 23 5 RCA 22 20 21 19 23 21 6 6 6 71 71 6 RCA 38 38 38 33 44 39 33 32 33 15 16 7 LAD 43 42 43 42 46 44 —a —a —a —a —a 8 RCA 27 28 28 32 32 32 16 14 15 44 52 9 RCA 22 12 17 21 29 25 1 1 1 93 95 10 LAD 47 44 46 46 41 44 25 22 24 49 46 11 RCA 33 31 32 30 30 30 12 10 11 65 63 12 RCA 14 12 13 8 10 9 5 9 7 47 23 13 LAD 38 31 35 44 40 42 21 21 21 38 49 14 LAD 36 28 32 38 35 37 3 3 3 92 93 15 LCX 17 14 16 17 18 18 8 7 8 52 57 16 RCA 30 27 29 30 28 29 11 12 12 60 60 17 LCX 32 12 22 31 19 25 4 4 4 82 84 18 RCA 26 25 26 29 25 27 11 13 12 52 55 19 RCA 30 20 25 27 29 28 7 7 7 73 76 20 RCA 37 31 34 36 38 37 19 16 18 49 53 21 RCA 23 20 22 26 27 27 12 14 13 38 50 CE-SSFP, contrast-enhanced steady state free precession; LAD, left anterior descending; LGE, late gadolinium enhancement; LM, left main; LV, left ventricle; Obs, observer; RCA, right coronary artery; T2W, T2-weighted imaging.

28 7 7 7 73 76 20 RCA 37 31 34 36 38 37 19 16 18 49 53 21 RCA 23 20 22 26 27 27 12 14 13 38 50 CE-SSFP, contrast-enhanced steady state free precession; LAD, left anterior descending; LGE, late gadolinium enhancement; LM, left main; LV, left ventricle; Obs, observer; RCA, right coronary artery; T2W, T2-weighted imaging. aInfarct size could not be assessed due to poor image quality related to frequent arrhythmias. Figure 1B shows three examples of the agreement of the MaR by T2-weighted imaging and CE-SSFP. Another example including the cine CE-SSFP can be seen in Supplementary data online, Movie S1. Figure 2A shows a scatter plot indicating the relationship between T2-weighted imaging and CE-SSFP (mean of two observers). There was a strong correlation between the two methods (r2 = 0.89, P < 0.01). Figure 2B shows the limits of agreement between T2-weighted imaging and CE-SSFP, demonstrating a bias of −3.0 ± 4.0% of the LV (P < 0.01). For the slice-by-slice comparison, r2 was 0.69 (P < 0.01; y = 0.98x + 0.61) with a bias of −1.8 ± 16% per slice. Figure 2 The myocardium at risk by T2-weighted imaging and CE-SSFP. (A) MaR by T2-weighted imaging (T2W) vs. CE-SSFP. The solid line denotes the line of identity. (B) The Bland–Altman graph showing the difference between the MaR quantified by T2W and CE-SSFP vs. the mean of the two methods. The difference between T2W and CE-SSFP was −3.0 ± 3.9%. CE-SSFP, contrast-enhanced steady-state free precession; LV, left ventricle; the solid line denotes mean difference; dashed lines denote ±2SD.

Bland–Altman graph showing the difference between the MaR quantified by T2W and CE-SSFP vs. the mean of the two methods. The difference between T2W and CE-SSFP was −3.0 ± 3.9%. CE-SSFP, contrast-enhanced steady-state free precession; LV, left ventricle; the solid line denotes mean difference; dashed lines denote ±2SD. Figure 3A shows the relationship between time after contrast agent administration at which the CE-SSFP image acquisition was commenced (2–8 min) and the difference between T2-weighted imaging and CE-SSFP for the assessment of MaR. The CE-SSFP imaging lasted for 2–4 min and the latest image was acquired 12 min after contrast agent administration. There was no change in the relationship between the two techniques as a function of time after contrast agent administration. This is illustrated in a patient with an LAD occlusion in Figure 3B. Figure 3 The MaR by CE-SSFP as a function of time after Gd administration. (A) The difference between MaR by T2-weighted imaging and CE-SSFP as a function of time after contrast agent administration at which the CE-SSFP imaging was commenced. The relationship between MaR by T2-weighted imaging and CE-SSFP was not affected by increased time after contrast injection at which the CE-SSFP was commenced. (B) An example of a patient with an infarction within the LAD territory. It is shown that no hyperenhancement could be seen prior to Gd administration (A), whereas a hyperenhanced region was seen 2, 8, and 30 min after Gd administration (B–D). Note that the extent of hyperenhancement did not change during the first 30 min after Gd injection (arrows). CE-SSFP, contrast-enhanced steady-state free precession; Gd-inj, gadolinium contrast agent injection; T2W, T2-weighted imaging; solid line, mean difference; dashed lines =± 2SD.

in after Gd administration (B–D). Note that the extent of hyperenhancement did not change during the first 30 min after Gd injection (arrows). CE-SSFP, contrast-enhanced steady-state free precession; Gd-inj, gadolinium contrast agent injection; T2W, T2-weighted imaging; solid line, mean difference; dashed lines =± 2SD. In six patients, cine SSFP short-axis imaging was also performed prior to contrast agent administration. In five of six patients, no hyperenhancement was seen, which is illustrated in Figure 3B. For one patient, a slight hyperenhancement with indistinct borders was found within the MaR. SNR within the MaR was 156 ± 7 and 132 ± 10 for the T2-weighted imaging and CE-SSFP, respectively (mean ± SEM). The CNR was 58 ± 3 and 27 ± 6 for the T2-weighted imaging and CE-SSFP, respectively (mean ± SEM). The contrast ratio between MaR and remote myocardium for T2-weighted imaging was 1.7 ± 0.3 compared with 1.5 ± 0.4 for CE-SSFP, which was not statistically significant different (P > 0.05). The interobserver variability was 5.0 ± 5.4% of the LV for T2-weighted imaging and 0.1 ± 6.2% of the LV for CE-SSFP. Myocardial salvage index The mean infarct size by LGE was 14 ± 11% (range 1–49) of the LV. The interobserver variability was 0.3 ± 2.2% of the LV. The infarct size was smaller in all patients when compared with MaR assessed by T2-weighted imaging (P < 0.01) and CE-SSFP (P < 0.01).

The interobserver variability was 5.0 ± 5.4% of the LV for T2-weighted imaging and 0.1 ± 6.2% of the LV for CE-SSFP. Myocardial salvage index The mean infarct size by LGE was 14 ± 11% (range 1–49) of the LV. The interobserver variability was 0.3 ± 2.2% of the LV. The infarct size was smaller in all patients when compared with MaR assessed by T2-weighted imaging (P < 0.01) and CE-SSFP (P < 0.01). Comparison of the infarct size by LGE in relation to MaR determined by T2-weighted imaging and CE-SSFP yielded a myocardial salvage index of 56 ± 22% (range: 15–93) and 58 ± 23% (range: 16–95), respectively (Figure 4, Table 1). There was a significant correlation (r2 = 0.89, P < 0.01) between the myocardial salvage index measured by the two methods (Figure 5), with an insignificant bias of −2.3 ± 7.4% of the LV (P = 0.20). Figure 4 Graphical display of myocardial salvage. Mid-ventricular short-axis slices in a patient with a right coronary artery occlusion. The epicardium is traced in green, the endocardium is traced in red, and the affected region is traced in white. The infarcted myocardium on the late gadolinium-enhanced image is superimposed (orange) on the myocardium at risk within the T2-weighted image and the CE-SSFP image. The non-coloured part of the myocardium at risk defines the salvaged myocardium. For this patient, T2-weighted imaging and CE-SSFP showed 52 and 55% myocardial salvage, respectively. CE-SSFP, contrast-enhanced steady-state free precession.

e) on the myocardium at risk within the T2-weighted image and the CE-SSFP image. The non-coloured part of the myocardium at risk defines the salvaged myocardium. For this patient, T2-weighted imaging and CE-SSFP showed 52 and 55% myocardial salvage, respectively. CE-SSFP, contrast-enhanced steady-state free precession. Figure 5 Myocardial salvage by T2-weighted imaging and CE-SSFP. (A) The myocardial salvage index measured by T2-weighted imaging vs. myocardial salvage index measured by CE-SSFP; the solid line denotes the line of identity. (B) The Bland–Altman graph showing the difference between myocardial salvage index quantified by T2W and CE-SSFP vs. the mean of the two methods. The difference between T2W and CE-SSFP was −2.3 ± 7.4%; the solid line denotes the mean difference; dashed lines represent ±2SD. CE-SSFP, contrast-enhanced steady-state free precession; LV, left ventricle; MaR, myocardium at risk; T2W, T2-weighted imaging. For the LGE images, the SNR within the infarction was 115 ± 12 and the CNR was 73 ± 8. Note that the LGE images were acquired using parallel imaging increasing the noise levels in these images. The contrast ratio between remote and infarcted myocardium was 2.9 ± 0.8, which was significantly higher than for T2-weighted imaging and CE-SSFP (P < 0.001). Discussion This study explored the agreement between MaR assessed by T2-weighted imaging and MaR assessed by CE-SSFP. The agreement was shown to be good, both for the assessment of MaR and myocardial salvage index.

For the LGE images, the SNR within the infarction was 115 ± 12 and the CNR was 73 ± 8. Note that the LGE images were acquired using parallel imaging increasing the noise levels in these images. The contrast ratio between remote and infarcted myocardium was 2.9 ± 0.8, which was significantly higher than for T2-weighted imaging and CE-SSFP (P < 0.001). Discussion This study explored the agreement between MaR assessed by T2-weighted imaging and MaR assessed by CE-SSFP. The agreement was shown to be good, both for the assessment of MaR and myocardial salvage index. Both T2-weighted imaging and CE-SSFP have previously been validated against MaR assessed by myocardial perfusion SPECT, with a mean difference of −2.3 ± 5.7 and 0.5 ± 5.1% of the LV, respectively.17,22 In accordance with these results, the present study found a small bias of −3.0 ± 4.0% when T2-weighted imaging was compared with CE-SSFP. This is, however, the first time these two techniques have been compared head-to-head which is important in order to ensure that both techniques can be used interchangeably in clinical cardioprotection trials using myocardial salvage as endpoint.

l bias of −3.0 ± 4.0% when T2-weighted imaging was compared with CE-SSFP. This is, however, the first time these two techniques have been compared head-to-head which is important in order to ensure that both techniques can be used interchangeably in clinical cardioprotection trials using myocardial salvage as endpoint. The correlation between MaR by T2-weighted imaging and CE-SSFP was not as strong for the slice-by-slice comparison as for the global measure of MaR normalized to the entire LV. As the heart rate varies, the timing within the cardiac cycle at which the T2-weighted image is acquired varies resulting in difficulties with local registration of the myocardium when comparing the two techniques. Even though the slice position was approximately the same for the two acquisitions, the part of the myocardium depicted may differ due to AV-plane movement during the cardiac cycle. This is not a problem when comparing global measures of MaR normalized to the entire LV.

the myocardium when comparing the two techniques. Even though the slice position was approximately the same for the two acquisitions, the part of the myocardium depicted may differ due to AV-plane movement during the cardiac cycle. This is not a problem when comparing global measures of MaR normalized to the entire LV. More recently, bright-blood T2-weighted sequences have been developed to increase the diagnostic performance of T2 CMR for oedema depiction.11,13 In a recent study by Payne et al.,14 dark-blood T2-STIR was shown to underestimate the MaR with ∼9% units compared with the recently introduced bright-blood T2-weighted sequence (ACUT2E).11 It was concluded that ACUT2E was more accurate for the determination of MaR and myocardial salvage than was dark-blood T2-STIR. Thus, CE-SSFP might be more accurate than dark-blood T2 in the present study since CE-SSFP showed larger MaR and, previously, showed no bias compared with myocardial perfusion SPECT.17

T2E).11 It was concluded that ACUT2E was more accurate for the determination of MaR and myocardial salvage than was dark-blood T2-STIR. Thus, CE-SSFP might be more accurate than dark-blood T2 in the present study since CE-SSFP showed larger MaR and, previously, showed no bias compared with myocardial perfusion SPECT.17 The pathophysiological basis for the enhanced myocardium observed using T2-weighted imaging is still not completely understood. Following acute coronary occlusion, the ischaemic myocardium shifts from aerobic metabolism to anaerobic glycolysis and ceases to contract. This failure of the energy-regulated membrane channels results in swelling of the myocytes due to influx of water and sodium.23 Furthermore, reperfusion leads to inflammatory-like response, increasing the amount of extracellular fluid.24 This increased water content in the affected myocardium is likely to explain the increased signal intensity compared with the non-affected myocardium as seen by T2-weighted imaging. Whether the increased water content is predominantly intracellular or extracellular remains to be determined. The ischaemic episode causes post-ischaemic stunning,25 associated with a decreased contractility in the previously ischaemic myocardium. This decreased contractility is likely associated with a decreased lymphatic drainage from this part of the myocardium, which may also contribute to residual increased water content 1 week after the acute event.

post-ischaemic stunning,25 associated with a decreased contractility in the previously ischaemic myocardium. This decreased contractility is likely associated with a decreased lymphatic drainage from this part of the myocardium, which may also contribute to residual increased water content 1 week after the acute event. The mechanisms behind the enhanced myocardium observed in CE-SSFP are not completely known either. The contrast in SSFP images is dependent on the T2/T1 ratio.22 In the presence of paramagnetic gadolinium, the T1 for the surrounding tissue is shortened. This is utilized for infarct visualization in T1-weighted inversion-recovery LGE imaging, where the concentration of an extracellular gadolinium-based contrast agent is increased due to an increased distribution volume in ischaemically injured myocardium.26–29 It has been shown that even reversibly injured myocardium within the MaR has an increased distribution volume in the acute phase after an ischaemic episode.28,29 Hence, the T2/T1 ratio in the entire MaR, including both reversible and irreversible injured myocardium, is affected by the presence of gadolinium. This might explain the increased signal intensity in the MaR seen by CE-SSFP. Furthermore, it has been shown that the relationship between the change in T1-relaxation rates before and after contrast agent administration (ΔR1) in different parts of the myocardium (remote, salvaged, and infarcted) in relation to ΔR1 in blood (ΔR1 ratios) remains constant from 4 to 29 min after contrast agent administration in acute myocardial infarction.29 These earlier findings indicate that the rate of exchange of contrast agent between the myocardium (normal and injured) and the blood pool is constant and much faster than the clearance rate in the kidneys during the first 30 min after contrast agent administration. Ugander et al.30 recently showed similar findings supporting this concept using T1 mapping before and after contrast administration in experimentally induced infarction. Thus, this can explain why the relationship between MaR assessed by T2-weighted imaging and CE-SSFP did not change with time after contrast agent administration in the present study (Figure 3) and why the timing of CE-SSFP after contrast agent administration is not critical. Therefore, CE-SSFP could potentially be added to clinical protocols so that cine imaging is acquired after contrast agent administration for the assessment of LV function and MaR.

rast agent administration in the present study (Figure 3) and why the timing of CE-SSFP after contrast agent administration is not critical. Therefore, CE-SSFP could potentially be added to clinical protocols so that cine imaging is acquired after contrast agent administration for the assessment of LV function and MaR. One advantage with CE-SSFP imaging is that this technique is based on a multi-phase acquisition throughout the cardiac cycle. This enables tracking of the MaR and myocardial borders in multiple time frames (see Supplementary data online, Movie S1), making delineation of both MaR and myocardial borders more robust. In the present protocol, MaR in the CE-SSFP images were traced in both end-systole and end-diastole and averaged. Another situation where CE-SSFP could be advantageous is when limited time for scanning is available, due to heavy clinical load or an unstable patient. In such a situation, gadolinium can be injected prior to the examination and the imaging protocol can be shortened since LV dimensions/function and MaR can be assessed from the same set of images. On the other hand, T2-weighted imaging for the determination of MaR can be performed in those patients where administration of a gadolinium-based contrast agent is contraindicated. Thus, there are several advantages of having access to more than one method for determination of MaR by CMR and we recommend both T2-weighted imaging and CE-SSFP to be implemented in the imaging protocol. Owing to the relatively small differences in signal intensity between MaR and remote myocardium, we recommend not to use parallel imaging when acquiring CE-SSFP images as this decreases the SNR. No significant difference in the contrast ratio between T2-weighted imaging and CE-SSFP imaging was found in the present study where no parallel imaging was performed.

in signal intensity between MaR and remote myocardium, we recommend not to use parallel imaging when acquiring CE-SSFP images as this decreases the SNR. No significant difference in the contrast ratio between T2-weighted imaging and CE-SSFP imaging was found in the present study where no parallel imaging was performed. Study limitations The present study was performed on a limited number of STEMI patients, all undergoing successful reperfusion. Thus, how this would translate to a non-STEMI population or to patients treated with thrombolytic therapy is not known. Since only a few female patients where included, gender perspective cannot be evaluated. No semi-quantitative method was used to determine the MaR by both T2-weighted imaging and CE-SSFP, since the signal intensities of the MaR and the remote myocardium varied between slices and between patients, making it difficult to choose a fixed standard deviation of signal intensities to differentiate MaR from remote myocardium. Conclusions There is a good agreement between MaR assessed by T2-weighted imaging and MaR assessed by CE-SSFP in patients with reperfused acute myocardial infarction 1 week after the acute event. Thus, both methods can be used to determine MaR and subsequently myocardial salvage at this point in time. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Conclusions There is a good agreement between MaR assessed by T2-weighted imaging and MaR assessed by CE-SSFP in patients with reperfused acute myocardial infarction 1 week after the acute event. Thus, both methods can be used to determine MaR and subsequently myocardial salvage at this point in time. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online. Funding This study was funded by the Swedish Research Council (521-2008-2461), the Swedish Heart and Lung Foundation, the Medical Faculty at Lund University, Sweden, European Commission FP7 Consortium CardioCell, Region of Scania, Sweden, the Stockholm County Council and Karolinska Institutet/Stockholm County Council Strategic Cardiovascular Program. Supplementary Material Supplementary Data Acknowledgements The authors would like to acknowledge Ann-Helen Arvidsson and Christel Carlander, both with the Lund Cardiac MR Group, for skilful assistance with image acquisition. Conflict of interest: none declared.

An 18-year-old male was referred to us for chest discomfort. A physical examination showed normal findings. The chest X-ray was also normal. ECG showed sinus rhythm and ischaemic changes in apico-infero-posterior myocardium. Two-dimensional-TTE (2DTTE) (Supplementary data online, Movie S1) showed a single, large oval cyst (39 × 56 × 38 mm). The edge was well defined (Panel A), it was intramyocardial and fluid-filled. It was located in the infero-postero-apical segment, protruding into the pericardium but mainly bulging into the left ventricular cavity (Supplementary data online, Movie S2). L3DTTE (Supplementary data online, Movie S3) allowed cropping of this cystic mass in transverse, sagittal (Panel B), and oblique planes, showing an apico-infero-posterior intramyocardial ellipsoid cystic mass with liquid content and well-defined edges. These data showed a life-threatening mass, mainly due to the possibility of intrapericardial and/or left ventricular rupture as well as systemic embolization. Chest computed tomography (CT) (Panel C) confirmed the presence of a left ventricular intramyocardial cyst and also a multivesicular hepatic cyst. The patient was referred to cardiac surgery. Surgical findings (Panel D) confirmed the accuracy of the preoperative echocardiography. The postoperative course was uneventful and the patient was discharged in stable condition. Three months later, his hepatic cyst was removed with no complications.

esicular hepatic cyst. The patient was referred to cardiac surgery. Surgical findings (Panel D) confirmed the accuracy of the preoperative echocardiography. The postoperative course was uneventful and the patient was discharged in stable condition. Three months later, his hepatic cyst was removed with no complications. Two-dimensional TTE remain the widely used imaging modality in cardiac echinococcosis, L3DTTE was shown to complement 2DTTE by displaying the 3D volume of the cyst and spatial relation to the neighbouring cardiac structures. This information is vital, especially since surgical removal of the cyst is the treatment of choice for this uncommon, but life-threatening disease. (Panel A) TTE: zoomed apex with egg-shaped hydatid mass. (Panel B) Live three-dimensional transthoracic echocardiogram, zoomed and cropped view. (Panel C) CT scan revealed a well defined, non-enhancing, oval hypodense mass. (Panel D) Intraoperative situs. The cystic mass is seen in the muscle of the left ventricular apex. The cyst was opened and its content aspirated. Cystic material removed. A large defect before its reconstruction. Supplementary data Supplementary data are available at European Journal of Echocardiography online. Supplementary Data

A 39-year-old severely obese male (body mass index 39 kg/m²) patient was referred to our department with suspected endocarditis. He presented with dyspnoea and recurrent fever of unknown origin. Infection parameters were found to be positive (C-reactive protein 24 mg/dL, reference <0.5 mg/dL; white blood cells 12 900/nL) and platelet count was low (16 000/µL). Other blood tests including troponin and blood cultures were normal. Transthoracic echocardiography showed a complex and irregular mass adherent to the endocardial surface of the right outflow tract and smaller floating structures of the right atrium (RA) and right ventricle (RV). Transoesophageal echocardiogram certified three masses: one was connected with the RA septum near the vena cava superior (Panel A), hyperechogenic and extending to the RV in diastole; another one was attached to the free RV wall; and the third one was adherent to the wall of the RV outflow tract and intermittently prolapsed through the pulmonary valve. A cardiac computed tomography confirmed the pseudotumours with partial calcification and the absence of contrast enhancement (Panel B). Hereditary thrombophilia and heparin-induced thrombocytopenia as well as rheumatological and serological testings were negative. A coronary angiography was performed due to intermittent chest pain and in preparation for a planned surgical treatment. The diagnostic procedure ruled out coronary heart disease and verified normal systolic left ventricular function. Cardiothoracic surgery was then performed due to the obstruction of RV outflow tract and to prevent further embolic complications. Surgical resection confirmed the thrombotic material with focal regions of calcification (Panel C). The diagnosis of antiphospholipid syndrome was confirmed by an anticardiolipin antibody test. Despite of continuous effective anticoagulation therapy, the patient died 2 weeks after surgical resection. An autopsy demonstrated disseminated intracardial thrombosis in all four chambers that led to multiple organ infarctions and vascular occlusive events (Panel D).

was confirmed by an anticardiolipin antibody test. Despite of continuous effective anticoagulation therapy, the patient died 2 weeks after surgical resection. An autopsy demonstrated disseminated intracardial thrombosis in all four chambers that led to multiple organ infarctions and vascular occlusive events (Panel D). Supplementary data are available at European Heart Journal – Cardiovascular Imaging online. Panel A. Transoesophageal echocardiography showing the right (RA) and left atrium (LA). Intracardiac mass with partial calcification was detected in the RA. Panel B. Cardiac multislice computed tomography demonstrating intra-atrial mass (arrowheads) with partial calcification and lack of contrast enhancement. Panel C. Resected right atrial thrombus. Panel D. Disseminated intracardiac thrombosis (asterisk) with extensive intracardiac thrombosis in the right atrium (RA) and right ventricular (RV) apex and destruction of the tricuspid valve with disseminated thrombosis. Supplementary Material Supplementary Data

Introduction Despite developments in coronary intervention and effective medications, coronary artery disease is the leading cause of death in the elderly worldwide. However, the majority of these individuals do not develop coronary symptoms before the onset of acute myocardial infarction or sudden cardiac death. Therefore, screening of patients with vulnerable plaque is important for the prevention of the onset of acute cardiovascular events.

use of death in the elderly worldwide. However, the majority of these individuals do not develop coronary symptoms before the onset of acute myocardial infarction or sudden cardiac death. Therefore, screening of patients with vulnerable plaque is important for the prevention of the onset of acute cardiovascular events. Plaque rupture or erosion of the endothelial surface with subsequent thrombus formation is currently recognized as the most important mechanism for acute coronary syndromes.1,2 Recently, magnetic resonance imaging technology has reached a level of spatial resolution that is sufficient for plaque visualization of large and static arteries, such as carotid arteries.3,4 Moreover, following the introduction of plaque imaging with non-contrast T1-weighted imaging (T1WI), many investigators have speculated that a hyperintense plaque (HIP) on non-contrast T1WI indicates the presence of mural or intraplaque haemorrhage containing methemoglobin.5,6 However, because of their small dimensions and their continuous motion during data acquisition, the visualization of coronary plaques is challenging,7−9 and the clinical significance of coronary HIP on T1WI remains unknown. Optical coherence tomography (OCT) is a new intravascular imaging modality with a high resolution of approximately 10–20 µm, which is 10-fold higher than that of intravascular ultrasound (IVUS). This new modality allows us to identify not only plaque rupture, but also fibrous cap thickness and intracoronary thrombus in vivo.10–14 The aim of this study was to compare HIP on T1WI in cardiac magnetic resonance (CMR) with coronary plaque morphology assessed by OCT in patients with angina pectoris.

trasound (IVUS). This new modality allows us to identify not only plaque rupture, but also fibrous cap thickness and intracoronary thrombus in vivo.10–14 The aim of this study was to compare HIP on T1WI in cardiac magnetic resonance (CMR) with coronary plaque morphology assessed by OCT in patients with angina pectoris. Methods Patients Thirty-nine patients with either stable or unstable angina pectoris, who had not undergone previous percutaneous transluminal coronary intervention (PCI) or coronary artery bypass grafting, were prospectively enrolled in this study between September 2010 and October 2011. Patients eligible for an early invasive strategy according to the ACCF/AHA Guideline (elevated levels of cardiac biomarkers, signs or symptoms of heart failure, and reduced left ventricular function) and those with contraindications to CMR were excluded from the study. All patients underwent CMR within 24 h before the day on which invasive coronary angiography (CAG) and OCT were performed. Of these patients, five patients who did not have significant stenoses after CAG and eight patients who did not undergo OCT examination before PCI were excluded from this analysis (five HIP and eight non-HIP).

underwent CMR within 24 h before the day on which invasive coronary angiography (CAG) and OCT were performed. Of these patients, five patients who did not have significant stenoses after CAG and eight patients who did not undergo OCT examination before PCI were excluded from this analysis (five HIP and eight non-HIP). Thus, 26 lesions from 26 patients, who had angiographically documented narrowing of at least 50% of the luminal diameter of a major coronary artery on CAG, were examined in this study. Pre-interventional OCT images were obtained for all patients with significant stenoses. In all patients, the procedure was performed on a native ‘de novo’ atherosclerotic lesion considered to be the culprit lesion. Unstable angina pectoris was diagnosed in 20 patients. Unstable angina pectoris was defined either as (i) new onset angina within 2 months after a previous bout; (ii) angina with a progressive crescendo pattern, with the anginal episodes increasing in frequency and/or duration; (iii) angina that occurred at rest; or (iv) angina occurring in the immediate post-infarct period. Stable angina pectoris was diagnosed in another six patients and defined as chest pain typical of cardiac ischaemia on exertion. Oral aspirin (100 mg) and clopidogrel (75 mg) were administered on admission. Moreover, patients with high risk were also treated with intravenous heparin, but no one had received thrombolytic agents.

gina pectoris was diagnosed in another six patients and defined as chest pain typical of cardiac ischaemia on exertion. Oral aspirin (100 mg) and clopidogrel (75 mg) were administered on admission. Moreover, patients with high risk were also treated with intravenous heparin, but no one had received thrombolytic agents. The following data were collected: age, sex, and presence of risk factors [smoking and hypertension, as defined by the Joint National Committee VII; diabetes mellitus, as defined by the World Health Organization (WHO) Study Group; or hypercholesterolaemia, as defined by the Japan Atherosclerosis Society Guidelines]. The study was approved by the hospital ethics committee, and informed consent was obtained from all patients before the study.

National Committee VII; diabetes mellitus, as defined by the World Health Organization (WHO) Study Group; or hypercholesterolaemia, as defined by the Japan Atherosclerosis Society Guidelines]. The study was approved by the hospital ethics committee, and informed consent was obtained from all patients before the study. CMR coronary plaque image acquisition Coronary plaque imaging was performed using a 1.5-T MR imager (Achieva, Philips Medical Systems, Best, the Netherlands) with a 5-element cardiac coil. Nitroglycerin (0.3 mg) was administered sublingually immediately before taking images to obtain high-quality CMR images. First, to obtain detailed anatomic information, free-breathing steady-state free precession whole-heart coronary MR angiographic images were acquired, according to the method described in previous reports.15,16 Briefly, initial survey images were focused around the heart, and the reference images were then obtained for sensitivity of parallel imaging. Transaxial cine MR images were then acquired using a steady-state free precession sequence with breath-holding (repetition time, 2.2 ms; echo time, 1.1 ms; flip angle, 60°; field of view, 350 × 350 × 10 mm; acquisition matrix, 160 × 160; cardiac phases, 100; SENSE factor, 3.0 in the anteroposterior direction; imaging time, 4.9 s) to determine the trigger delay time when the motion of the right coronary artery is minimum. Patient-specific acquisition windows were set during either the diastolic or systolic phase, depending on the phase of minimal motion of the right coronary artery. Coronary MR angiography was performed while patients were free-breathing by using a three-dimensional segmented steady-state free precession sequence with T2 preparation and radial k-space sampling in the Y–Z plane [repetition time, 3.7 ms; echo time, 1.8 ms; flip angle, 80°; SENSE factor, 2.0; number of excitations, 1; navigator gating window of ±2.0 mm with diaphragm drift correction; field of view, 300 × 255 × 120 mm (rectangular field of view, 85%); acquisition matrix, 240 × 240; reconstruction matrix, 512 × 512 × 160, resulting in an acquired spatial resolution of 1.25 × 1.25 × 1.5 mm reconstructed to 0.6 × 0.6 × 0.75 mm]. Diaphragm drift due to irregular respiration was corrected automatically by the diaphragm drift correction system provided by the MR system. The same value was set for the acquisition window as in the coronary plaque imaging.

an acquired spatial resolution of 1.25 × 1.25 × 1.5 mm reconstructed to 0.6 × 0.6 × 0.75 mm]. Diaphragm drift due to irregular respiration was corrected automatically by the diaphragm drift correction system provided by the MR system. The same value was set for the acquisition window as in the coronary plaque imaging. Next, coronary plaque images were obtained when patients were breathing freely by using a three-dimensional T1WI inversion-recovery gradient-echo sequence with black-blood condition (inversion time 450 ms), fat-suppressed and radial k-space sampling in the Y–Z plane [repetition time, 4.4 ms; echo time, 2.0 ms; flip angle, 20°; SENSE factor, 2.5; number of excitations, 2; navigator gating window of ±1.5 mm with diaphragm drift correction; field of view, 300 × 240 × 120 mm (rectangular field of view, 80%); acquisition matrix, 224 × 224; reconstruction matrix, 512 × 512 × 140, resulting in an acquired spatial resolution of 1.34 × 1.34 × 1.7 mm reconstructed to 0.6 × 0.6 × 0.85 mm].5,8,9

; navigator gating window of ±1.5 mm with diaphragm drift correction; field of view, 300 × 240 × 120 mm (rectangular field of view, 80%); acquisition matrix, 224 × 224; reconstruction matrix, 512 × 512 × 140, resulting in an acquired spatial resolution of 1.34 × 1.34 × 1.7 mm reconstructed to 0.6 × 0.6 × 0.85 mm].5,8,9 CMR coronary plaque image analysis Coronary CMR image analysis was performed by a single experienced cardiologist who was blinded to the plaque information obtained by OCT. If the target lesion was confirmed in the coronary MR angiography, the areas corresponding to the above site in the coronary T1WI CMR image obtained were carefully matched according to the surrounding cardiac and chest wall structures. Finally, the signal intensity of coronary plaque to cardiac muscle ratio (PMR; PMR was defined as the signal intensity of the coronary plaque divided by the signal intensity of the left ventricular muscle near the coronary plaque), as measured by placing a circular region of interest on a standard console of the clinical MR system, was calculated. Areas with PMR >1.0 were defined as HIP, whereas areas with PMR ≤1.0 were defined as non-HIP, according to the method described in a previous report.8 The interobserver variability for measurement of the PMR performed in a random sample of patients previously was 5.8 ± 3.9% (R2 = 0.968, P < 0.0001) and the κ-value for interobserver agreement in categorization of coronary plaque as high or non-high signal intensity was 0.88 (substantial).

cribed in a previous report.8 The interobserver variability for measurement of the PMR performed in a random sample of patients previously was 5.8 ± 3.9% (R2 = 0.968, P < 0.0001) and the κ-value for interobserver agreement in categorization of coronary plaque as high or non-high signal intensity was 0.88 (substantial). OCT image acquisition OCT imaging was performed before intervention and only after administration of 0.2 mg of intracoronary nitroglycerin. Thrombolysis or thrombectomy was not performed for any patient. The culprit vessel was identified on the basis of clinical, scintigram stress test, and angiographic data. In all SAP patients, the culprit vessel was considered to be the ischaemia-related vessel, which was shown ischaemia by exercise scintigram stress test. The culprit lesion site selected for the analysis was the image slice with the smallest lumen cross-sectional area. A 0.016-inch OCT catheter (ImageWire; LightLab Imaging, Westford, MA, USA) was advanced to the culprit lesion through a 3-F occlusion balloon catheter. In order to remove the blood from the field of view, the occlusion balloon was inflated to 0.6atm at the proximal site of the culprit lesion, and Lactated Ringer's solution was infused into the coronary artery from the distal tip of the occlusion balloon catheter at 0.5 mL/s. The entire length of the culprit lesion was imaged using an automatic pullback device moving at 1 mm/s, and the OCT image clearly visualized the culprit lesion.

e of the culprit lesion, and Lactated Ringer's solution was infused into the coronary artery from the distal tip of the occlusion balloon catheter at 0.5 mL/s. The entire length of the culprit lesion was imaged using an automatic pullback device moving at 1 mm/s, and the OCT image clearly visualized the culprit lesion. OCT image analysis OCT image analysis was performed by two experienced observers blinded to the clinical information by using previously established criteria for OCT plaque characterization.10–14 The presence of lipid, thin-cap fibroatheroma (TCFA), plaque rupture, calcification, and thrombus were evaluated within the 10 mm long culprit lesion segment (5 mm proximal and 5 mm distal to the culprit lesion site), according to the previous reports.17,18 If there was discordance of diagnosis between the two observers, a consensus diagnosis was obtained using repeated off-line readings. When lipid was present in ≥2 quadrants in any of the images within a plaque, it was considered a lipid-rich plaque. Thin-cap fibroatheromas were defined as a lipid-rich plaque with a fibrous cap thickness measuring ≤65 µm. Plaque rupture was defined as an intimal interruption and cavity formation in the plaque. Calcification was defined as well-delineated, signal-poor regions with sharp borders. Thrombus was identified as an irregular high- or low-backscattering mass protruding into the lumen. Furthermore, thrombus type was classified as red or white thrombus by OCT, according to the previous reports.12 Red thrombus was defined as high-backscattering protrusions with signal-free shadowing in OCT images. White thrombus was defined as signal-rich, low-backscattering projections in OCT images.

g into the lumen. Furthermore, thrombus type was classified as red or white thrombus by OCT, according to the previous reports.12 Red thrombus was defined as high-backscattering protrusions with signal-free shadowing in OCT images. White thrombus was defined as signal-rich, low-backscattering projections in OCT images. Statistical analyses Continuous data are presented as mean ± SD. In case the data were normally distributed, the two groups were compared with a two-tailed unpaired Student's t-test. Otherwise, a Mann–Whitney U-test was used. Categorical variables were compared by two-sided Fisher's exact test. Concordance between investigators was assessed by κ-statistics. By using intracoronary thrombus detected by OCT as a gold standard, we calculated the sensitivities, specificities, and positive and negative predictive values of the HIP lesions on T1WI, according to the standard methods. All calculations were performed using SPSS software (version 11.5, SPSS Inc., Chicago, IL, USA) and P-values of <0.05 were considered significant. Results Table 1 shows the baseline clinical characteristics and angiographic findings in patients with HIP and non-HIP. Of 26 lesions from 26 patients, 16 (62%) had HIP lesions and 10 (38%) had non-HIP lesions. The PMR in HIP lesions was significantly higher than that in non-HIP lesions. There were no significant differences in patient clinical characteristics and angiographic findings between HIP and non-HIP lesions. Table 1 Clinical characteristics and angiographic findings

had HIP lesions and 10 (38%) had non-HIP lesions. The PMR in HIP lesions was significantly higher than that in non-HIP lesions. There were no significant differences in patient clinical characteristics and angiographic findings between HIP and non-HIP lesions. Table 1 Clinical characteristics and angiographic findings HIP (n= 16) Non-HIP (n = 10) P-value PMR 1.85 ± 0.89 0.76 ± 0.15 0.0001 Age (years) 70 ± 11 65 ± 10 0.261 Male 12 (75%) 4 (40%) 0.109 Diagnosis 0.644 SAP 3 (19%) 3 (30%) UAP 13 (81%) 7 (70%) Hypertension 11 (69%) 7 (70%) 0.999 Hypercholesterolaemia 10 (63%) 7 (70%) 0.999 Diabetes mellitus 6 (38%) 3 (30%) 0.999 Smoking 6 (38%) 2 (20%) 0.420 Culprit vessel LAD 9 (56%) 8 (80%) LCx 0 (0%) 0 (0%) RCA 7 (44%) 2 (20%) Percent diameter stenosis 75 ± 7 74 ± 10 0.698 Values are mean ± SD or n (percentage). HIP, hyperintense plaque; PMR, signal intensity of coronary plaque to cardiac muscle ratio; SAP, stable angina pectoris; UAP, unstable angina pectoris; LAD, left anterior descending coronary artery; LCx, left circumflex coronary artery; RCA, right coronary artery.

HIP (n= 16) Non-HIP (n = 10) P-value PMR 1.85 ± 0.89 0.76 ± 0.15 0.0001 Age (years) 70 ± 11 65 ± 10 0.261 Male 12 (75%) 4 (40%) 0.109 Diagnosis 0.644 SAP 3 (19%) 3 (30%) UAP 13 (81%) 7 (70%) Hypertension 11 (69%) 7 (70%) 0.999 Hypercholesterolaemia 10 (63%) 7 (70%) 0.999 Diabetes mellitus 6 (38%) 3 (30%) 0.999 Smoking 6 (38%) 2 (20%) 0.420 Culprit vessel LAD 9 (56%) 8 (80%) LCx 0 (0%) 0 (0%) RCA 7 (44%) 2 (20%) Percent diameter stenosis 75 ± 7 74 ± 10 0.698 Values are mean ± SD or n (percentage). HIP, hyperintense plaque; PMR, signal intensity of coronary plaque to cardiac muscle ratio; SAP, stable angina pectoris; UAP, unstable angina pectoris; LAD, left anterior descending coronary artery; LCx, left circumflex coronary artery; RCA, right coronary artery. The relationship between HIP/non-HIP on T1WI and plaque morphology assessed by OCT is shown in Table 2. There were no significant differences in the frequency of lipid-rich plaque, TCFA, plaque rupture, and calcification between HIP and non-HIP lesions. In contrast, the frequency of thrombus was significantly higher in HIP lesions than in non-HIP lesions (P = 0.004). Thrombus on OCT was observed in 12 (75%) of the 16 lesions with HIP as opposed to 1 (10%) of the 10 lesions with non-HIP. Among 12 HIP lesions with thrombus, 7 had a red thrombus and 5 had a white thrombus. In contrast, one non-HIP lesion with thrombus had a white thrombus. Hyperintense plaque on T1WI as an indicator of intracoronary thrombus on OCT had a sensitivity of 92%, specificity of 69%, positive predictive value of 75%, and negative predictive value of 90%. Regarding OCT plaque characterization, interobserver agreement measured as a κ-statistic for TCFA (κ = 0.77), plaque rupture (κ = 0.89), calcification (κ = 0.88), and thrombus (κ = 0.89) was substantial. In contrast, interobserver agreement showed slightly lower concordance for lipid-rich plaques (κ = 0.62). A representative case of HIP lesion on T1WI compared with plaque morphology on OCT is shown in Figure 1. Whole-heart coronary MR angiography and CAG show severe coronary stenosis in the proximal left anterior descending coronary artery. The area corresponding to the stenotic lesion shows hyperintensity on T1WI. Intracoronary thrombus was detected by OCT in the culprit lesion. Furthermore, Figure 2 shows one case of both HIP and non-HIP lesions within a single patient, although one lesion in the right coronary artery was not included in the overall statistical analysis. In this case, the HIP lesion in the left anterior descending coronary artery contained thrombus as well as TCFA and plaque rupture, and the other non-HIP lesion in the right coronary artery did not show those plaque morphologies. Table 2 OCT findings in hyperintense and non-hyperintense plaque

all statistical analysis. In this case, the HIP lesion in the left anterior descending coronary artery contained thrombus as well as TCFA and plaque rupture, and the other non-HIP lesion in the right coronary artery did not show those plaque morphologies. Table 2 OCT findings in hyperintense and non-hyperintense plaque HIP (n = 16) Non-HIP (n = 10) P-value Lipid-rich plaque 12 (75%) 5 (50%) 0.234 TCFA 6 (38%) 2 (20%) 0.420 Plaque rupture 7 (44%) 3 (30%) 0.683 Calcification 9 (56%) 7 (70%) 0.683 Thrombus 12 (75%) 1 (10%) 0.004 Red thrombus 7 (58%) 0 (0%) White thrombus 5 (42%) 1 (100%) Values are n (percentage). OCT, optical coherence tomography; HIP, hyperintense plaque; TCFA, thin-cap fibroatheroma (lipid, ≥2 quadrants and fibrous cap thickness, ≤65 µm). Figure 1 A representative case of HIP lesion on T1WI compared with plaque morphology on OCT. (A) Whole-heart coronary MR angiography shows severe coronary stenosis (indicated by arrow) in the proximal left anterior descending coronary artery (LAD). (B and C) Coronary T1WI CMR images (B: horizontal, C: sagittal). The area corresponding to the stenotic lesion shows hyperintensity (indicated by arrow). (D) CAG shows severe coronary stenosis (indicated by arrow) in the proximal LAD. (E and F) Images (E) and (F) show intracoronary thrombus (indicated by arrowhead) that was detected by OCT in the proximal (E) and mid (F) sites of the culprit lesion.

nding to the stenotic lesion shows hyperintensity (indicated by arrow). (D) CAG shows severe coronary stenosis (indicated by arrow) in the proximal LAD. (E and F) Images (E) and (F) show intracoronary thrombus (indicated by arrowhead) that was detected by OCT in the proximal (E) and mid (F) sites of the culprit lesion. Figure 2 A case of both HIP and non-HIP lesions in a single patient. (A and H) Whole-heart coronary MR angiographic images show severe coronary stenoses (indicated by arrow) in the proximal LAD (A), and mid right coronary artery (RCA) (H). (B, C, I) Coronary T1WI CMR images. The area corresponding to the LAD lesion shows hyperintensity [indicated by arrow in B (horizontal) and C (coronal), and the other area corresponding to the RCA lesion shows non-hyperintensity (indicated by arrow in I)]. (D and J) CAG showing severe coronary stenoses (indicated by arrow) in the proximal LAD (D) and, the mid and distal RCA (J). (E–G) Images (E) to (G) show the OCT images in the LAD lesion. Thin-cap fibroatheroma (E), plaque rupture (F), and intracoronary thrombus (G) (indicated by arrowhead) were observed in the corresponding lesion with HIP. Images (K) (mid RCA) and (L) (distal RCA) show the OCT images in the RCA lesions. Thrombus was not found in the corresponding lesions with non-HIP. Discussion To the best of our knowledge, this is the first study to show that HIP on T1WI is directly associated with intracoronary thrombus detected by OCT imaging in patients with angina pectoris.

Figure 2 A case of both HIP and non-HIP lesions in a single patient. (A and H) Whole-heart coronary MR angiographic images show severe coronary stenoses (indicated by arrow) in the proximal LAD (A), and mid right coronary artery (RCA) (H). (B, C, I) Coronary T1WI CMR images. The area corresponding to the LAD lesion shows hyperintensity [indicated by arrow in B (horizontal) and C (coronal), and the other area corresponding to the RCA lesion shows non-hyperintensity (indicated by arrow in I)]. (D and J) CAG showing severe coronary stenoses (indicated by arrow) in the proximal LAD (D) and, the mid and distal RCA (J). (E–G) Images (E) to (G) show the OCT images in the LAD lesion. Thin-cap fibroatheroma (E), plaque rupture (F), and intracoronary thrombus (G) (indicated by arrowhead) were observed in the corresponding lesion with HIP. Images (K) (mid RCA) and (L) (distal RCA) show the OCT images in the RCA lesions. Thrombus was not found in the corresponding lesions with non-HIP. Discussion To the best of our knowledge, this is the first study to show that HIP on T1WI is directly associated with intracoronary thrombus detected by OCT imaging in patients with angina pectoris. While coronary wall imaging by CMR is challenging due to the small size of coronary arteries and cardiac/respiratory motion, it has been applied in patients using breath-hold or free-breathing techniques. The introduction of plaque imaging with black-blood non-contrast T1WI has especially encouraged researchers in this field. There have been a few CMR studies of coronary plaque vulnerability compared with IVUS, multislice computed tomography (MSCT), or invasive CAG. Kawasaki et al.8 reported that typical coronary HIP on T1WI was associated with a high frequency of ultrasound attenuation and positive remodelling, remarkably low CT density, and a high incidence of transient slow-flow phenomena by using both MSCT and IVUS. These features seem to represent vulnerable plaques. However, in their study, the presence of plaque rupture and thrombus was not assessed, because current IVUS technology does not allow a definitive distinction among some plaque morphologies, such as lipid core, thrombus, and plaque rupture. More recently, Jansen et al.9 showed that HIP on T1WI identified intracoronary thrombus as detected by invasive CAG in patients with acute myocardial infarction. The diagnosis of thrombus by CAG is generally made on the basis of presumptive evidence; therefore, they would also have underestimated intracoronary thrombus. Intravascular OCT, on the other hand, was recently developed as a high-resolution imaging device for plaque characterization that provides additional morphological information beyond that of IVUS images.10–14 Several studies have already shown that OCT allows the identification of not only plaque rupture, but also TCFA and intracoronary thrombus in vivo more frequently compared with IVUS and angioscopy.10–14 By using this technology, we have shown that the factor associated with HIP lesions on T1WI was intracoronary thrombus. The predictive power of HIP on T1WI in the detection of intracoronary thrombus on OCT was considerably substantial. However, four false-positive cases were observed in lesions with HIP. There was a time-lag of ≤24 h between CMR and OCT procedures in this study.

iated with HIP lesions on T1WI was intracoronary thrombus. The predictive power of HIP on T1WI in the detection of intracoronary thrombus on OCT was considerably substantial. However, four false-positive cases were observed in lesions with HIP. There was a time-lag of ≤24 h between CMR and OCT procedures in this study. The possibility cannot be excluded that intracoronary thrombus had disappeared during the time-lag. Nevertheless, our present findings add more detailed information on HIP lesions by CMR to the previous data by using IVUS or CAG. In the present study, the relation between HIP on T1WI and lipid-rich plaque, TCFA, or plaque rupture was not found. Further studies with more cases will be needed to confirm the present results.

theless, our present findings add more detailed information on HIP lesions by CMR to the previous data by using IVUS or CAG. In the present study, the relation between HIP on T1WI and lipid-rich plaque, TCFA, or plaque rupture was not found. Further studies with more cases will be needed to confirm the present results. Previous studies have shown that HIP formation on T1WI scans is due to methemoglobin production in the early stages of thrombus formation. When the thrombus is formed, red blood cells are trapped within a mesh of platelets and fibrin. Kelly et al.19 reported that red blood cells containing methemoglobin produced T1 shortening, the extent of which was in proportion to the level of methemoglobin. This signal persists for several weeks, although the overall period is less than 6 months. The information obtained from the signal might provide data about thrombus characteristics, such as age and volume. At this stage, it is understood that OCT does not allow accurate quantification of thrombus volume, because some intracoronary thrombi produce extensive signal-free shadowing, which makes it impossible (or at least unreliable) to assess thrombus volume. In this patient population, among 12 HIP lesions with thrombus, 7 had a red thrombus and 5 had a white thrombus. In contrast, one non-HIP lesion with thrombus had a white thrombus. Further studies are needed to clarify the relationship between thrombus characteristics and HIP on T1WI.

liable) to assess thrombus volume. In this patient population, among 12 HIP lesions with thrombus, 7 had a red thrombus and 5 had a white thrombus. In contrast, one non-HIP lesion with thrombus had a white thrombus. Further studies are needed to clarify the relationship between thrombus characteristics and HIP on T1WI. What is the clinical implication of identifying intracoronary thrombus by non-contrast T1WI in CMR in human atherosclerotic lesions? This method is completely non-invasive and easily repeatable because of lack of ionizing radiation, contrast injections, or a vascular access. From this point of view, T1WI in CMR is best suited for the detection of patients with subclinical vulnerable plaques. Moreover, there is increasing evidence that atherosclerotic disease is a chronic inflammatory process with the involvement of many arterial segments, including coronary and carotid arteries, despite the fact that a single localized culprit lesion may cause an acute cardiovascular event. T1-weighted imaging in CMR is a promising tool for the detection of multiple vulnerable plaques associated with thrombus. One of the goals of our future studies will be to investigate whether the presence of HIP on T1WI in non-ischaemic-related coronary or carotid arteries is associated with subsequent development of acute cardiovascular events. Finally, if HIP on T1WI can be shown to be limited to a finite time span, its presence could be used to accurately identify recent plaque thrombosis. This information may have several novel clinical implications in the field of coronary intervention, e.g. it can be used to predict slow-flow/no-flow phenomena or the age of the chronic total occlusion. Furthermore, although patients with elevated levels of cardiac biomarkers, signs, or symptoms of heart failure, or ECG changes should be considered for early invasive intervention (early CAG), differential diagnosis and treatment of the remaining patients remain challenging in emergency triage. This non-invasive thrombus-detection technique may be useful for further risk stratification and for obtaining prognostic information in the high-risk group.

be considered for early invasive intervention (early CAG), differential diagnosis and treatment of the remaining patients remain challenging in emergency triage. This non-invasive thrombus-detection technique may be useful for further risk stratification and for obtaining prognostic information in the high-risk group. Limitations This study has several limitations. First, coronary thrombectomy was not performed for the identification and examination of intracoronary thrombus. Therefore, there was no evidence based on pathohistological findings. For coronary MR, however, OCT is the generally accepted method and is acknowledged as one of the most reliable tools for assessment of coronary plaque characterization, although this modality is not routinely used in the clinical setting. We consider, therefore, that the quality of both coronary MR and OCT data obtained by this approach is sufficiently high to validate our conclusion. Secondly, patients eligible for an early invasive strategy and without significant stenoses were excluded from this study. Moreover, there were no patients with the culprit lesion in the left circumflex coronary artery. Therefore, our results were limited by selection bias and may not apply to such patients. Thirdly, in the present study, a fixed inversion time (450 ms) was used for black-blood condition. Strictly speaking, however, the patient-specific (heart rate-specific) inversion time should be adjusted to null blood signal. Finally, in the present CMR and OCT analyses, the number of patients examined was very small. This may have limited the statistical power, making all comparisons descriptive.

k-blood condition. Strictly speaking, however, the patient-specific (heart rate-specific) inversion time should be adjusted to null blood signal. Finally, in the present CMR and OCT analyses, the number of patients examined was very small. This may have limited the statistical power, making all comparisons descriptive. Conclusions This study shows that the HIP lesions on T1WI in patients with angina pectoris relate to the presence of intracoronary thrombus as detected by OCT imaging. This non-invasive technology appears to be a promising tool for identifying vulnerable plaques associated with thrombus. Funding This work was supported by JSPS KAKENHI (23591057). Conflicts of interest: none declared.

A 25-year-old male suffered from shortness of breath during normal physical activity. Auscultation showed a 4/6 grade systolic and diastolic murmur on the left lower sternal border. A previous two-dimensional transthoracic echocardiogram (2D TTE) revealed enhanced mitral valve thickening with severe stenosis and moderate regurgitation (mitral valve area: 0.9 cm2, transmitral gradient: 18 mmHg, E peak: 180 cm/s, A peak: 240 cm/s, regurgitation jet area: 5.0 cm2, and vena contracta width: 4.0 mm), left atrial enlargement (50 × 56 × 67 mm), and moderate pulmonary hypertension (pulmonary artery systolic pressure 64 mmHg). As 2D TTE images were suboptimal, real-time three-dimensional transoesophageal echocardiography (3D TEE) was performed using a Philips X7-2T probe. The TEE demonstrated mitral valve thickening and restricted motion (Panels A and B), with absent chordae tendinae. The 3D TEE showed that the mitral leaflets were directly adherent to the papillary muscles (Panel C), presenting a funnel-shaped stenosis during diastole (Panel D). At surgical inspection, the mitral valve was obviously thickened and porcelain white (Panel E) with no anterior or posterior mitral leaflet chordae tendinae. There was severe subvalvular stenosis with direct leaflet adherence to the two papillary muscles (Panel F).

A 25-year-old male suffered from shortness of breath during normal physical activity. Auscultation showed a 4/6 grade systolic and diastolic murmur on the left lower sternal border. A previous two-dimensional transthoracic echocardiogram (2D TTE) revealed enhanced mitral valve thickening with severe stenosis and moderate regurgitation (mitral valve area: 0.9 cm2, transmitral gradient: 18 mmHg, E peak: 180 cm/s, A peak: 240 cm/s, regurgitation jet area: 5.0 cm2, and vena contracta width: 4.0 mm), left atrial enlargement (50 × 56 × 67 mm), and moderate pulmonary hypertension (pulmonary artery systolic pressure 64 mmHg). As 2D TTE images were suboptimal, real-time three-dimensional transoesophageal echocardiography (3D TEE) was performed using a Philips X7-2T probe. The TEE demonstrated mitral valve thickening and restricted motion (Panels A and B), with absent chordae tendinae. The 3D TEE showed that the mitral leaflets were directly adherent to the papillary muscles (Panel C), presenting a funnel-shaped stenosis during diastole (Panel D). At surgical inspection, the mitral valve was obviously thickened and porcelain white (Panel E) with no anterior or posterior mitral leaflet chordae tendinae. There was severe subvalvular stenosis with direct leaflet adherence to the two papillary muscles (Panel F). Isolated mitral chordae tendinae absence is very rare and often missed. Three-dimensional TEE can look from the left atrium and ventricle, to clearly demonstrate all mitral valve apparatus components and provide optimal visualization and detailed information on abnormal mitral valve anatomic structures.

At surgical inspection, the mitral valve was obviously thickened and porcelain white (Panel E) with no anterior or posterior mitral leaflet chordae tendinae. There was severe subvalvular stenosis with direct leaflet adherence to the two papillary muscles (Panel F). Isolated mitral chordae tendinae absence is very rare and often missed. Three-dimensional TEE can look from the left atrium and ventricle, to clearly demonstrate all mitral valve apparatus components and provide optimal visualization and detailed information on abnormal mitral valve anatomic structures. Panel A. Two-dimensional TEE showed restricted motion of the mitral valve (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet; PM, papillary muscles; LV, left ventricle; RV, right ventricle). Panel B. Real-time 3D-colour flow imaging showed subvalvular colourful mitral flow in the diastole (arrow = colourful mitral flow; LV, left ventricle; LA, left atrium). Panel C. Three-dimensional TEE showed that mitral leaflets were directly adherent to the papillary muscles (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet; PM, papillary muscles; LV, left ventricle). Panel D. Three-dimensional TEE looking from the left atrium showed the mitral valve presenting a funnel-shaped stenosis during diastole (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet). Panel E. At surgical inspection, the mitral valve was thickened and had a porcelain white appearance (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet).

Panel D. Three-dimensional TEE looking from the left atrium showed the mitral valve presenting a funnel-shaped stenosis during diastole (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet). Panel E. At surgical inspection, the mitral valve was thickened and had a porcelain white appearance (AMVL, anterior mitral valve leaflet; PMVL, posterior mitral valve leaflet). Panel F. At surgical inspection, the leaflet directly adhered to the posteromedial papillary muscles (MV, mitral valve; PM, papillary muscles; LVIT, left ventricular inflow tract). Supplementary data are available at European Heart Journal – Cardiovascular Imaging online. Supplementary Material Supplementary Data

Introduction An early and correct diagnosis is a crucial step in the treatment of patients. A delayed or wrong diagnosis may delay the treatment, complicate inpatient workflow, and may in worst case scenario have a lethal outcome.1 During the recent two decades, the development of new digital technology and miniaturization of ultrasound scanners have moved these scanners from the echo-labs into the white coat pocket.2,3 This makes them an excellent clinical tool, available for any physician in different clinical settings as a point-of-care ultrasonography.4 These newly developed scanners have been studied in several clinical settings. In the hands of experienced users, pocket-size hand-held echocardiographic (PHHE) devices offer high-quality semi-quantitative assessment of cardiac structures, abdominal great vessels, and the pleural space at the physicians' point-of-care with a demonstrable clinical benefit.5–11 Medical history-taking and physical examination of most patients are performed by the residents in the emergency departments or bed wards. Few of these are skilled in ultrasonography and given the cost and the availability of the PHHE devices, non-expert users will frequently have such technology available for diagnostic use. Thus, we aimed to study the feasibility and reliability of PHHE in the hands of medical residents after a targeted training period in cardiovascular ultrasound.

ltrasonography and given the cost and the availability of the PHHE devices, non-expert users will frequently have such technology available for diagnostic use. Thus, we aimed to study the feasibility and reliability of PHHE in the hands of medical residents after a targeted training period in cardiovascular ultrasound. Methods Study population This prospective observational study included 199 patients admitted to the medical department at Levanger Hospital, Norway. The patients were included in the period 4 April to 23 June 2011. The examination was performed by six medical residents taking part in the study. At study start, 12 medical residents were employed at the department, and half of them were randomized to participate in the study. During the study, another two residents joined the department, but they did not participate in the study. The residents have in-house call 24×7. Thus, the six participating residents covered ∼42% of the total period of inclusion. All emergency admissions during the time these six residents were on call were included in the study. There were no other criteria of inclusion. Only patients who did not consent to participate or did not stay long enough in the department to enable the necessary diagnostic procedures for the study were excluded. Due to logistic reasons, inclusion of patients was restricted to 199 of 446 available patients as standard diagnostic procedures and treatment had first priority.

s who did not consent to participate or did not stay long enough in the department to enable the necessary diagnostic procedures for the study were excluded. Due to logistic reasons, inclusion of patients was restricted to 199 of 446 available patients as standard diagnostic procedures and treatment had first priority. The patients were admitted to the emergency room in a standard way. After having been triaged according to their symptoms, they were examined by the resident. Based on the medical history, physical examination, and supplemental tests, a preliminary diagnosis was made. Thus, usual care diagnostics were done prior to the examination with PHHE. All patients had standard follow-up according to their symptoms and findings. Patients, in whom pathology was suggested either by PHHE or by the standard clinical care, were referred for relevant gold-standard diagnostic follow-up. To improve the reliability of the sensitivity and specificity of the data, approximately 10 negatively described PHHE examinations per resident were randomly selected by the study committee and referred for reference imaging procedures as well. The study was approved by the Regional Committee for Medical Research Ethics, and conducted according to the second Helsinki Declaration. All the patients gave their informed consent to participate in the study.

esident were randomly selected by the study committee and referred for reference imaging procedures as well. The study was approved by the Regional Committee for Medical Research Ethics, and conducted according to the second Helsinki Declaration. All the patients gave their informed consent to participate in the study. Education of residents The residents underwent a brief training program covering both the examination with PHHE and interpretation of the recordings. The program consisted of 4 h of lectures dealing with the theoretical basics and pitfalls of cardiovascular ultrasonography. Normal and pathological findings were demonstrated, and they were all provided with access to a virtual ultrasound-imaging library. All participating residents had a personal supervisor. Subsequently, the residents underwent 3 months of practical training, initially together with the supervisors in the echo-lab and in the radiology department, then using PHHE in the medical department with close connection to experienced ultrasonographers, having the opportunity to discuss their findings. They were encouraged to perform at least 100 examinations during the tutorial period. The actual numbers performed were median (interquartile range) 95 (80–225) examinations.

en using PHHE in the medical department with close connection to experienced ultrasonographers, having the opportunity to discuss their findings. They were encouraged to perform at least 100 examinations during the tutorial period. The actual numbers performed were median (interquartile range) 95 (80–225) examinations. Pocket-size echocardiographic examination The residents performed the PHHE examinations using a Vscan (version 1.2; GE Vingmed Ultrasound, Horten, Norway). This device offers B-mode and colour flow (CF) imaging. The total weight is 390 g including the phased array probe with bandwidth of 1.7–3.8 MHz. It provides two dimensional (2D) imaging and real time colour-Doppler within a sector that has fixed size, but is movable throughout the 2D sector. An algorithm enables automatic storage and loop recording of a cardiac cycle without ECG signal.12 Patient identification was performed by voice recording and the automatically assigned examination number. All images and recordings were saved on the device's micro-SD card and later transferred to a computer by commercial software (Gateway; GE Vingmed Ultrasound).

op recording of a cardiac cycle without ECG signal.12 Patient identification was performed by voice recording and the automatically assigned examination number. All images and recordings were saved on the device's micro-SD card and later transferred to a computer by commercial software (Gateway; GE Vingmed Ultrasound). The pocket-size echocardiographic examinations were performed bedside, and when possible with the patients in the left-lateral decubitus position. The examinations included parasternal long- and short-axis views and apical four-chambers, two-chambers, and long-axis views. All views contained 2D and CF recordings. The patients were turned to supine position when examining the abdominal great vessels. The pleural space was recorded from supine or upright position. A standard examination protocol was used. Assessment of left- and right-ventricular function were done semi-quantitatively from the parasternal and apical positions, classified as normal/near normal, moderate, or severe dysfunction. The quantification was based on the systolic excursion of the atrioventricular plane for both ventricles. In addition, eye-balling of the left-ventricular ejection fraction as ≥45, 30–45, or <30% corresponded to normal/near normal, moderate, or severe dysfunction, respectively. With respect to the assessment of right-ventricular function, dilatation of the ventricle and/or diastolic shift to the left of the intraventricular septum was also included in the judgement. Severe regional dysfunction was classified as present or not. Valvular pathology and dysfunction was classified semi-quantitatively as mild, moderate, or severe. Quantification of stenosis was based on the amount of calcification and the movement of the cusps/leaflets. Quantification of the regurgitations was based on the CF jet and size and function of the adjacent chambers. The size of the left atrium (LA) was measured online from the parasternal position and quantified as normal (<40 mm), moderately dilated (40–50 mm), or severely dilated (>50 mm). Pericardial effusion was if present classified as significant or not based on visual judgement of the influence of the adjacent chambers. The inferior vena cava diameter was assessed from the subcostal position at the end expiration within 2 cm from the right atrial orifice. The size of the abdominal aorta was determined by the largest measured diameter. It was classified as aneurysmatic if the diameter exceeded 30 mm. Both pleural cavities were examined.

he inferior vena cava diameter was assessed from the subcostal position at the end expiration within 2 cm from the right atrial orifice. The size of the abdominal aorta was determined by the largest measured diameter. It was classified as aneurysmatic if the diameter exceeded 30 mm. Both pleural cavities were examined. If pleural effusion was present, this was graded as small or large amount. A large amount of pleural effusion was registered if the diameter between the thoracic wall and the lung exceeded 5 and 4.5 cm in the left or the right pleural cavity, respectively. The examinations of the different structures were judged by the residents as feasible if they were able to quantify the specific cardiac structures or function indices based on their recordings.

diameter between the thoracic wall and the lung exceeded 5 and 4.5 cm in the left or the right pleural cavity, respectively. The examinations of the different structures were judged by the residents as feasible if they were able to quantify the specific cardiac structures or function indices based on their recordings. Validation of point-of-care pocket-size echocardiography Standard echocardiography was performed in the hospital's echo-lab, under optimal conditions. The system used was a Vivid 7 scanner (GE Vingmed Ultrasound, Horten, Norway) using a 2.0 MHz phased-array transducer (M3S) with bandwidth 1.5–3.6 MHz. Second harmonic imaging was used. The recording of a cardiac cycle was ECG triggered. The standard examinations were performed independently by one of four experienced cardiologists blinded to the results of PHHE with a median (range) time delay of 21.1 (0.4–166) h. A complete echocardiographic examination was performed. Dimensions were measured from a parasternal view. Ejection fraction was measured by Simpson's rule from apical four- and two-chamber views.13 Valvular pathology was graded according to the recommendations from the European Association of Cardiovascular Imaging (EACVI) [former European Association of Echocardiography (EAE)].14–16 For the analyses of the patients who underwent both echocardiographic and radiographic examinations, the radiologists' classifications of pleural effusion [computer tomography (CT) or ultrasound] and the size of the abdominal aorta were preferred.

ng (EACVI) [former European Association of Echocardiography (EAE)].14–16 For the analyses of the patients who underwent both echocardiographic and radiographic examinations, the radiologists' classifications of pleural effusion [computer tomography (CT) or ultrasound] and the size of the abdominal aorta were preferred. Statistical analysis As the different echocardiographic and anthropometric measures partly were skewed compared with normal distribution, the basic characteristics are presented as mean ± standard deviation (SD) and (interquartile) range. Spearman's rho (r) was used for comparison of the ranking of pathology between the PHHE and the high-end echocardiographic examinations. Data are presented as r [95% confidence interval (CI)] with the 95% CI computed using bootstrapping. For comparison of continuous variables, Pearson's rho (r) and Bland–Altman statistics were used. Statistical analyses were performed using SPSS for Windows version 20.0 (SPSS, Inc., Chicago, IL, USA).

phic examinations. Data are presented as r [95% confidence interval (CI)] with the 95% CI computed using bootstrapping. For comparison of continuous variables, Pearson's rho (r) and Bland–Altman statistics were used. Statistical analyses were performed using SPSS for Windows version 20.0 (SPSS, Inc., Chicago, IL, USA). Results Study population Table 1 shows the baseline data of the 199 patients included in the study (107 men and 92 women). Mean ± SD (range) age was 65.6 ± 18.2 (17.1–98.5) years. The distribution of age was positively skewed compared with a normal distribution. The mean height was 170.9 ± 9.7 cm and the body mass index was 26.4 ± 5.6 kg/m2. At admission, atrial fibrillation was present in 33 (17%) patients, hypertension was present in 67 (34%) patients, 36 (18%) had known diabetes mellitus, and 20 (10%) had established heart failure. In total, cardiovascular disease defined as either angina pectoris, prior myocardial infarction, prior stroke, or established peripheral arterial disease was present in 71 (36%) of the patients. There were no significant differences in the basic characteristics of the 199 participants included in the study and the 247 participants not included in the study, but who were admitted to the hospital the days when the six residents performing PHHE examinations were on duty. Table 1 Basic characteristics of the 199 study participants

nificant differences in the basic characteristics of the 199 participants included in the study and the 247 participants not included in the study, but who were admitted to the hospital the days when the six residents performing PHHE examinations were on duty. Table 1 Basic characteristics of the 199 study participants Mean ± SD (range)a Age, years 65.6 ± 18.2 (17.1–98.5) Male, n (%) 107 (53.8) Height, cm 170.9 ± 9.7 (150–196) Body mass index, kg/m2 26.4 ± 5.6 (12–45) Systolic blood pressure, mmHg 143.9 ± 28.6 (74–245) Diastolic blood pressure, mmHg 75.0 ± 15.6 (24–120) Heart rate, bpm 82.8 ± 22.6 (40–160) Atrial fibrillation, n (%) 33 (16.6) Known hypertension, n (%) 67 (33.7) Known diabetes, n (%) 36 (18.1) Known myocardial infarction, n (%) 32 (16.1) Known angina, n (%) 17 (8.5) Known heart failure, n (%) 20 (10.1) Known peripheral vessel disease, n (%) 7 (3.5) Known stroke, n (%) 35 (17.6) Known cardiovascular disease, n (%) 71 (35.7) Known cancer, n (%) 16 (8.0) aData are presented as mean ± SD (range) unless otherwise specified.

rdial infarction, n (%) 32 (16.1) Known angina, n (%) 17 (8.5) Known heart failure, n (%) 20 (10.1) Known peripheral vessel disease, n (%) 7 (3.5) Known stroke, n (%) 35 (17.6) Known cardiovascular disease, n (%) 71 (35.7) Known cancer, n (%) 16 (8.0) aData are presented as mean ± SD (range) unless otherwise specified. Pocket-size hand-held echocardiography The time consumption of the examination, including large vessels, was median (range) 5.7 (1.6–19.9) min. Each resident performed a median (interquartile range) of 27 (19–46) examinations. Table 2 shows the feasibility of PHHE. The left-ventricular (LV) function was assessed to satisfaction in nearly all of the patients (97%) and the pericardial space in all patients. The aortic and atrioventricular valves were assessed in at least 76% and the pulmonary valve in <50% of the patients. The vena cava inferior was assessed to satisfaction in 77% and the abdominal aorta in 50% of the population. This is also illustrated in Figure 1. Table 2 Feasibility of point-of-care pocket-size echocardiography Anatomic structure Assessed to satisfaction (%) Left ventricle 194 (97) Right ventricle 172 (86) Pericardium 199 (100) Left atrium 173 (87) Mitral valve 177 (89) Aortic valve 171 (86) Pulmonary valve 97 (49) Tricuspid valve 152 (76) Abdominal aorta 99 (50) Vena cava inferior 154 (77) Pleura 190 (95)

Pocket-size hand-held echocardiography The time consumption of the examination, including large vessels, was median (range) 5.7 (1.6–19.9) min. Each resident performed a median (interquartile range) of 27 (19–46) examinations. Table 2 shows the feasibility of PHHE. The left-ventricular (LV) function was assessed to satisfaction in nearly all of the patients (97%) and the pericardial space in all patients. The aortic and atrioventricular valves were assessed in at least 76% and the pulmonary valve in <50% of the patients. The vena cava inferior was assessed to satisfaction in 77% and the abdominal aorta in 50% of the population. This is also illustrated in Figure 1. Table 2 Feasibility of point-of-care pocket-size echocardiography Anatomic structure Assessed to satisfaction (%) Left ventricle 194 (97) Right ventricle 172 (86) Pericardium 199 (100) Left atrium 173 (87) Mitral valve 177 (89) Aortic valve 171 (86) Pulmonary valve 97 (49) Tricuspid valve 152 (76) Abdominal aorta 99 (50) Vena cava inferior 154 (77) Pleura 190 (95) Figure 1 Feasibility of point-of-care pocket-size echocardiography. Feasibility (%) of the different cardiovascular structures when pocket-size echocardiography was performed by residents. The examinations of the different structures were judged by the residents as feasible if they were able to quantify the specific cardiac structures or function indices based on the recordings.

raphy. Feasibility (%) of the different cardiovascular structures when pocket-size echocardiography was performed by residents. The examinations of the different structures were judged by the residents as feasible if they were able to quantify the specific cardiac structures or function indices based on the recordings. A total of 133 and 74 patients underwent high-end echocardiography and radiographic (CT or ultrasound) reference imaging, respectively. In total, 186 (93%) patients underwent reference imaging (Figure 2). For the different indices of cardiac structure or function, the available numbers of validated examinations are shown in Tables 3 and 4. Table 3 shows the correlations of semi-quantitative assessment of cardiovascular structures and function indices between PHHE and standard echocardiography. The classification of global left-ventricular function, pleural, and pericardial effusion showed very strong correlation with standard diagnostic procedures (Spearman's r ≥ 0.83, with variations between residents 0.70–0.93, 0.54–1.0, and 0.81–1.0, respectively). Regional left-ventricular function showed moderate correlation, r = 0.60 (variation between residents 0.53–0.61). The classification of aortic valve calcification/stenosis and regurgitation showed strong correlation with r = 0.67 (variation between residents 0.29–0.93) and r = 0.68 (variation between residents 0.33–1.0), respectively. Regurgitation of the atrioventricular valves showed moderate-to-strong correlations, r = 0.53 (variation between residents 0.34–0.80) for mitral and r = 0.61 (variation between residents 0.21–0.78) for tricuspid regurgitation, so did the degree of dilatation of the LA (r = 0.61) (variation between residents 0.23–0.76). No serious findings were missed. PHHE correlated strongly with standard diagnostics with respect to detect abdominal aortic aneurysms, r = 0.70. No aneurysms were missed, but there was one false positive diagnosis where the measurement of the aorta was 32 mm by PHHE and 28 mm by the abdominal CT. Figure 3 illustrates the reproducibility data of the abdominal aortic diameter. The maximal diameter of the inferior vena cava correlated only moderately with high-end echocardiography, Pearson's r = 0.45. Figure 4 illustrates the total number of misclassifications of global and regional ventricular and valvular pathology by PHHE compared with the reference.

data of the abdominal aortic diameter. The maximal diameter of the inferior vena cava correlated only moderately with high-end echocardiography, Pearson's r = 0.45. Figure 4 illustrates the total number of misclassifications of global and regional ventricular and valvular pathology by PHHE compared with the reference. For the quantification of LV global function, LA size, and aortic stenosis, respectively, 7, 2, and 5% of the misclassifications were two degrees; all other misclassifications were only one degree. Figure 5 shows clinical examples of PHHE compared with reference method, and a clinical example is given in Supplementary material online, Videos S1 and S2. Table 3 Correlations of semi-quantitative classification of echocardiographic indices of pocket-size echocardiography and reference method n total n pathology R 95% CI Global systolic function, left ventricle 129 26 0.83 0.71–0.93 Apparent regional dysfunction, left ventricle 129 22 0.60 0.39–0.78 Global systolic function, right ventricle 115 10 0.44 0.10–0.72 Size of left atrium 117 68 0.61 0.48–0.72 Aortic calcification and stenosis 119 37 0.67 0.52–0.80 Aortic regurgitation 117 27 0.68 0.52–0.82 Mitral regurgitation 123 54 0.53 0.37–0.68 Tricuspid regurgitation 107 49 0.61 0.45–0.74 Pericardial effusion 131 4 0.86 0.57–1.00 Pleural effusion 151 20 0.83 0.67–0.94 Abdominal aorta 52 2 0.70 0.49–1.00 Inferior vena cavaa 94 0.45 0.24–0.62 Data presented as correlation coefficient (r) with 95% confidence interval achieved by bootstrapping.

54 0.53 0.37–0.68 Tricuspid regurgitation 107 49 0.61 0.45–0.74 Pericardial effusion 131 4 0.86 0.57–1.00 Pleural effusion 151 20 0.83 0.67–0.94 Abdominal aorta 52 2 0.70 0.49–1.00 Inferior vena cavaa 94 0.45 0.24–0.62 Data presented as correlation coefficient (r) with 95% confidence interval achieved by bootstrapping. n total, the total number who underwent both PHHE and reference imaging; n pathology, total number with the described pathology. aContinuous variable, analysed by Pearson's correlation, all others analysed by Spearman's rank correlation. Table 4 Sensitivity, specificity, positive, and negative predictive value of point-of-care pocket-size echocardiography to detect at least moderate pathology compared with reference method n total n pathology Sensitivity Specificity PPV NPV LV dysfunction 129 30 92 94 80 98 RV dysfunction 115 10 40 97 57 94 LA enlargement 117 68 62 94 93 64 Aortic regurgitation 117 27 82 89 69 94 Aortic stenosis/calcification 119 37 76 88 74 89 Mitral regurgitation 123 48 71 81 71 81 Tricuspid regurgitation 107 49 65 90 84 75 n total, the total number who underwent both PHHE and reference imaging; n pathology, total number with the described pathology; LV, left ventricle; RV, right ventricle; LA, left atrium; PPV, positive predictive value; NPV, negative predictive value. Figure 2 Validation of PHHE. Illustration of the number of patients that were validated with reference imaging (left) and by what kind of reference imaging (right). Echo, echocardiography.

n total n pathology Sensitivity Specificity PPV NPV LV dysfunction 129 30 92 94 80 98 RV dysfunction 115 10 40 97 57 94 LA enlargement 117 68 62 94 93 64 Aortic regurgitation 117 27 82 89 69 94 Aortic stenosis/calcification 119 37 76 88 74 89 Mitral regurgitation 123 48 71 81 71 81 Tricuspid regurgitation 107 49 65 90 84 75 n total, the total number who underwent both PHHE and reference imaging; n pathology, total number with the described pathology; LV, left ventricle; RV, right ventricle; LA, left atrium; PPV, positive predictive value; NPV, negative predictive value. Figure 2 Validation of PHHE. Illustration of the number of patients that were validated with reference imaging (left) and by what kind of reference imaging (right). Echo, echocardiography. Figure 3 Bland–Altman plot for the assessment of the abdominal aortic diameter using PHHE and reference imaging. Reproducibility for the assessment of the diameter of the abdominal aorta. Bland-Altman plot of difference between PHHE and reference imaging by the mean of the measurements. Figure 4 Classification of ventricular and valvular pathology by PHHE compared with reference echocardiography. The agreement of PHHE and reference echocardiography in the quantification of ventricular and valvular pathology is illustrated. Over- and underestimation is the total numbers of misclassifications. In total, only 2% were misclassified by two degrees, the rest by one degree. LV, left ventricle; N, numbers; regurg, regurgitation.

ent of PHHE and reference echocardiography in the quantification of ventricular and valvular pathology is illustrated. Over- and underestimation is the total numbers of misclassifications. In total, only 2% were misclassified by two degrees, the rest by one degree. LV, left ventricle; N, numbers; regurg, regurgitation. Figure 5 Cases illustrating the comparison of PHHE with reference method. (A) shows images from the pocket-size device, while (B) shows images from the high-end Vivid 7 scanner (GE Vingmed Ultrasound). 1 (A and B): 54-year-old man with principal diagnosis of liver cirrhosis changed to dilated cardiomyopathy after PHHE. 2 (A and B): 70-year-old man with known heart failure concluded to be decompensated after finding the shown significant amount of pleural effusion, dilated vena cava inferior, and reduced LV function. 3 (A and B): 75-year-old man referred with stroke where PHHE revealed an unknown moderate aortic regurgitation (without importance for the acute treatment). 4 (A and B): 88-year-old woman admitted with heart failure. PHHE revealed dilated ventricles, the shown large tricuspid regurgitation, pleural effusion, and ascites due to hypervolaemia.

d man referred with stroke where PHHE revealed an unknown moderate aortic regurgitation (without importance for the acute treatment). 4 (A and B): 88-year-old woman admitted with heart failure. PHHE revealed dilated ventricles, the shown large tricuspid regurgitation, pleural effusion, and ascites due to hypervolaemia. Table 4 shows the sensitivity, specificity, positive, and negative predictive values of PHHE to detect at least moderate pathology. There was high specificity and negative predictive values of detecting left- and right-ventricular dysfunction and aortic-valve pathology. On the contrary, the lower sensitivity and positive predictive values for the assessment of right-ventricular function and left-atrial size are mainly caused by some underestimation of pathology. Discussion Our study demonstrates that medical residents in <6 min can perform a bedside ultrasound examination of the heart, pleural space, and the abdominal great vessels after a 3 months training period and get reliable and clinically important diagnostic information beyond the standard physical examination. The patients were included solely during the time when the participating residents were on call and represent otherwise an unselected population in our department. The population characteristics are also in line with patient characteristics from previous studies in similar settings.10,17,18

Discussion Our study demonstrates that medical residents in <6 min can perform a bedside ultrasound examination of the heart, pleural space, and the abdominal great vessels after a 3 months training period and get reliable and clinically important diagnostic information beyond the standard physical examination. The patients were included solely during the time when the participating residents were on call and represent otherwise an unselected population in our department. The population characteristics are also in line with patient characteristics from previous studies in similar settings.10,17,18 PHHE has in several studies showed a high feasibility and accuracy when performed by experts.6–9 Galderisi et al.8 showed slightly lower sensitivity and specificity when trainees performed PHHE compared with experts. Panoulas et al.19 showed improved diagnostic accuracy when medical students and junior doctors added a PHHE examination to history, physical examination, and ECG findings. Our results are in line with their findings when PHHE is performed by non-experts. The feasibility is overall very good, 75–100% for all structures except the pulmonic valve and the abdominal aorta which were assessed to satisfaction in approximately one-half of the patients. Inexperienced users may be less able to provide optimal image quality and need better image quality to be able to interpret the recordings compared with expert users, but we have no data to support this hypothesis. The abdominal aorta was assessed in a relatively small number of patients compared with expert studies.7,20 This may partly be explained by the fact that the residents did not register the aorta as assessed unless the entire length of the aorta was satisfactorily assessed. Secondly, patients were non-fasting, thereby reducing abdominal image quality, and BMI was ∼2 kg/m2 higher in whom the abdominal aorta was not assessed (P < 0.001). Nonetheless, there may have been too little focus on examining the great vessels during the training period.

ngth of the aorta was satisfactorily assessed. Secondly, patients were non-fasting, thereby reducing abdominal image quality, and BMI was ∼2 kg/m2 higher in whom the abdominal aorta was not assessed (P < 0.001). Nonetheless, there may have been too little focus on examining the great vessels during the training period. The assessment of the global left-ventricular function and the pericardial and pleural space compared excellently with standard diagnostics. These are crucial issues in the cardiovascular ultrasound examination.21 The classification/assessment of valvular function showed moderate-to-strong correlation and we found high specificity and high negative predictive values for detecting at least moderate valvular pathology. Importantly, no serious findings were missed, neither according to aortic valve pathology or regurgitation of the atrioventricular valves. However, there was some under- and overestimation of both ventricular dysfunction and valvular pathology. This may be explained by less experienced users, a very sensitive colour mode, and the lack of spectral Doppler in the PHHE devices. We find the presented degree of misclassification of aortic stenosis, in line with the presented, but less pronounced overestimation of aortic stenosis related to the lack of spectral Doppler in recent studies.6,7 No moderate or severe aortic stenosis was missed. Atrioventricular valves regurgitations were missed more often compared with aortic regurgitations and this may be related to the higher number of atrioventricular regurgitations in the presented population. Due to moderate feasibility, the correlation of the aortic diameter was tested in only 52 patients and in these patients there was a strong agreement, and in the one misclassified, the difference was 4 mm. No aneurysms were missed by PHHE. The moderate agreement between PHHE and standard diagnostics in the assessment of the inferior vena cava may be explained by the period of time between PHHE and the standard echocardiography of median 21 h. Physiological variations and treatment effects may have influenced the results.22 In addition, measurements of the size of the LA and vena cava inferior may be influenced by the fact that the pocket-size device lacks ECG-cables and there are limited opportunities to ensure the correct timing in the cardiac or respiratory cycles.

31P MRS,24 we have previously determined that the reproducibility coefficient of the PCr/ATP ratio was 0.28 or 15% in four control subjects, on two separate occasions. All analysis was performed using SPSS version 17 (SPSS Inc, Chicago, IL). All tests were two-sided and statistical significance was assumed at P < 0.05. Results Patient population Baseline characteristics of 22 patients (19 probands and 3 individual family members, unrelated to each other) are summarized in Table 1. Disease burden was mild or moderate in all patients (11 patients had MIDD, 10 had myopathic phenotype and 1 patient had MELAS). Eleven patients had diabetes mellitus and 5 had treated hypertension. There were no significant differences in the current systolic or diastolic blood pressures. The frequencies of specific clinical features are summarized in Table 2, with individual clinical features and medications in the Supplementary date online, Table S1. The eGFR was >60 mL/min/1.73 m2 in all patients. Table 1 Baseline characteristics

cal variations and treatment effects may have influenced the results.22 In addition, measurements of the size of the LA and vena cava inferior may be influenced by the fact that the pocket-size device lacks ECG-cables and there are limited opportunities to ensure the correct timing in the cardiac or respiratory cycles. Taking a thorough medical history and performing a physical examination will remain the cornerstones in the diagnostic procedure, but there is a need for improvement in diagnostic accuracy to decrease medical errors.1,23 PHHE is an excellent tool to provide further diagnostic information. As stated by the EACVI (former EAE) the users level of competence is very important in these devices.24 Experienced ultrasonographers can start using PHHE without limitations. In less-experienced users, targeted education and a training period are necessary and PHHE should be used only for targeted examinations depending on the skills of the user. Even in the hands of relatively inexperienced residents, PHHE provides feasible and reliable information at the point-of-care and improves the diagnostic precision without significant time delay. However, it is important to state that PHHE cannot replace the standard echocardiographic examination performed by experts in the echo lab. It should remain a bedside imaging tool which allows for quick and important information without losing valuable time.

oves the diagnostic precision without significant time delay. However, it is important to state that PHHE cannot replace the standard echocardiographic examination performed by experts in the echo lab. It should remain a bedside imaging tool which allows for quick and important information without losing valuable time. Limitations In the study period, 1076 emergency admissions to the medical department were recorded and 84 of these patients declined consent. Out of the 446 patients randomized to receive PHHE examination, only 199 actually received it. This is mainly explained by busy working hours, hospital logistics, and the residents being informed to have a priority on standard diagnostics and treatment of patients. The study was a single-centre study with a limited number of participating residents and patients. Consecutive patients were included and critical diagnosis such as aortic dissection and cardiac tamponade were not registered during the inclusion period. It is important to emphasize that in such cases, PHHE may offer a fast track to the correct diagnosis,10 but negative findings must not rule out further diagnostic tests if the clinician still suspects specific conditions. Conclusion By adding a point-of-care PHHE examination lasting <6 min, medical residents were able to obtain reliable information of important cardiac structures and great vessels in patients admitted to a medical department. Thus, a focused examination with PHHE performed by residents, after a targeted training period have the potential to improve in-hospital diagnostics and care.

6 min, medical residents were able to obtain reliable information of important cardiac structures and great vessels in patients admitted to a medical department. Thus, a focused examination with PHHE performed by residents, after a targeted training period have the potential to improve in-hospital diagnostics and care. Supplementary material Supplementary material is available at European Heart Journal – Cardiovascular Imaging online. Funding This study is funded by the Nord-Trøndelag Health Trust, Norway and the Norwegian University of Science and Technology, Norway. Conflict of interest: none declared. Supplementary Material Supplementary Data Acknowledgements We thank all the participating doctors, nurses, and secretarial staff at Levanger hospital for their invaluable assistance with the inclusion and data collection for this study. OCM, GNA, HD, and BOH hold positions at MI Lab, a Centre of Research-based Innovation that is funded by the Research Council of Norway and industry. One of the industry partners is GE Vingmed Ultrasound. The Centre has a total budget of 124 million NOK for the 8 years period from 2007 to 2014, and the contribution from GE Vingmed Ultrasound to this budget is 7 million NOK (6%).

Introduction Mitochondrial diseases are an important group of genetic disorders, primarily affecting tissues with high-energy requirements. The m.3243A>G mitochondrial DNA (mtDNA) mutation in the mt-tRNALeu(UUR) gene is the most common pathogenic mutation present in ∼1 in 300 of the general population and causing disease in ∼1 in 6000 individuals.1 Originally described in an individual with mitochondrial encephalopathy, lactic acidosis and stroke-like episodes (MELAS), the m.3243A>G mutation also causes maternally inherited diabetes and deafness (MIDD), myopathy, ophthalmoplegia and cardiomyopathy, as isolated clinical features or part of multisystem disease.

s.1 Originally described in an individual with mitochondrial encephalopathy, lactic acidosis and stroke-like episodes (MELAS), the m.3243A>G mutation also causes maternally inherited diabetes and deafness (MIDD), myopathy, ophthalmoplegia and cardiomyopathy, as isolated clinical features or part of multisystem disease. Cardiomyopathy, commonly with a hypertrophic phenotype, occurs in 20–40% of patients carrying the m.3243A>G mutation,2–6 and is an independent predictor of morbidity and early mortality;5,7 many patients die from cardiac arrhythmias or congestive heart failure.4 Echocardiography and 12-lead ECG are recommended screening strategies as early identification of asymptomatic cardiac structural defects (stage B cardiomyopathy) allows the initiation of cardioprotective therapies.8 In other neuromuscular diseases associated with cardiomyopathy such interventions slow cardiac remodelling and reduce symptoms.9 Two-dimensional echocardiography has limited sensitivity to detect small changes in left ventricular (LV) mass, particularly in asymptomatic cases,10 and previous studies used normal reference ranges, rather than comparison with age- and gender-matched healthy controls, further decreasing sensitivity.3,5,6 Magnetic resonance imaging (MRI) may reveal cardiac involvement in multisystem disease when standard evaluation is unremarkable,11 and may provide novel therapeutic targets, where the efficacy of early intervention remains to be determined.12

ge- and gender-matched healthy controls, further decreasing sensitivity.3,5,6 Magnetic resonance imaging (MRI) may reveal cardiac involvement in multisystem disease when standard evaluation is unremarkable,11 and may provide novel therapeutic targets, where the efficacy of early intervention remains to be determined.12 MRI is the gold standard investigation of cardiac morphology and function. Cardiac tagging enables the detection of early defects in myocardial deformation by analysis of circumferential strain and torsion.13 Torsion describes the twisting motion of the heart due to opposite rotation of base and apex, and maintains homogeneity of strain across the myocardial wall. The torsion to endocardial circumferential strain ratio (TSR), a sensitive marker of altered epicardial–endocardial interactions, is constant among healthy individuals of the same age but increases with normal ageing.13 Elevated torsion and/or TSR have been demonstrated in patients with left ventricular hypertrophy (LVH) caused by increased haemodynamic loading,14 in hypertrophic cardiomyopathy (HCM) patients,15 and recently, in HCM mutation carriers without hypertrophy, perhaps providing an early phenotypic marker.16

normal ageing.13 Elevated torsion and/or TSR have been demonstrated in patients with left ventricular hypertrophy (LVH) caused by increased haemodynamic loading,14 in hypertrophic cardiomyopathy (HCM) patients,15 and recently, in HCM mutation carriers without hypertrophy, perhaps providing an early phenotypic marker.16 Phosphorus-31 (31P) magnetic resonance spectroscopy (MRS) permits the evaluation of myocardial bioenergetics by calculation of the phosphocreatine (PCr)/adenosine triphosphate (ATP) ratio.17 PCr/ATP ratio is reduced in systolic dysfunction and in HCM with normal systolic function.18 Impaired cardiac bioenergetics occur in mutation carriers of both HCM17 and mitochondrial disease,19 without echocardiographic evidence of hypertrophy, suggesting a potential role in early detection of disease. Using these modalities, we sought to characterize the cardiac phenotype in a clinically and genetically well-characterized cohort with reference to age- and gender-matched healthy controls. Based on studies in patients with mitochondrial disease,19 and in HCM mutation carriers without hypertrophy,16,17 our hypotheses were that abnormalities of LV mechanics and bioenergetics would be detectable in patients carrying the m.3243A>G mutation without known cardiac involvement, and that such abnormalities would be related to markers of disease burden. We provide a comprehensive MRI/31P MRS evaluation of cardiac changes in this population, with important implications for future screening and management strategies.

e in patients carrying the m.3243A>G mutation without known cardiac involvement, and that such abnormalities would be related to markers of disease burden. We provide a comprehensive MRI/31P MRS evaluation of cardiac changes in this population, with important implications for future screening and management strategies. Methods Subjects Twenty-two adult patients with mitochondrial disease due to the m.3243A>G mutation, but no known cardiac involvement, were recruited from a mitochondrial disease specialist clinic. The absence of cardiac involvement was determined using screening strategies commonly employed in patients with mitochondrial disease, namely clinical history and examination with a normal ECG, echocardiogram (documenting no significant valvular disease, a maximum LV wall thickness ≤12 mm, an ejection fraction ≥55% and the LV end-diastolic dimension ≤32 mm/m2), and exercise stress ECG (documenting no symptoms or ECG changes suggestive of underlying coronary artery disease); patients were excluded using these criteria (n = 3) and the presence of contra-indications to MRI (n = 1; claustrophobia). All 22 patients were matched with respect to age and gender with healthy controls, with a normal ECG and no history of cardiovascular or metabolic disease, recruited through advertisement. Institutional ethical approval and informed consent were obtained and the study complied with the Declaration of Helsinki.

phobia). All 22 patients were matched with respect to age and gender with healthy controls, with a normal ECG and no history of cardiovascular or metabolic disease, recruited through advertisement. Institutional ethical approval and informed consent were obtained and the study complied with the Declaration of Helsinki. Clinical assessment Subjects underwent physical examination by an experienced clinician. Disease burden was assessed using the Newcastle Mitochondrial Disease Adult Scale (NMDAS), a validated scoring system.20 The m.3243A>G mutation load was determined in urinary epithelial cells.21 The estimated glomerular filtration rate (eGFR) was calculated in all patients using the Modified Diet in Renal Disease equation, prior to administration of gadolinium. Cardiac magnetic resonance imaging Using a 3-Tesla Intera Achieva scanner (Philips, Best, NL), subjects underwent (i) 31P MRS, (ii) cine imaging, (ii) cardiac tagging, and (iv) late gadolinium enhancement (LGE) imaging.

icant differences in the current systolic or diastolic blood pressures. The frequencies of specific clinical features are summarized in Table 2, with individual clinical features and medications in the Supplementary date online, Table S1. The eGFR was >60 mL/min/1.73 m2 in all patients. Table 1 Baseline characteristics Characteristic Carriers (n = 22) Controls (n = 22) P value Age (years) 42.5 ± 12.2 42.8 ± 13.4 ns Male sex, n (%) 11 (50) 11 (50) ns Height (cm) 168 ± 11 171 ± 11 ns Weight (kg) 65.7 ± 16.2 77.1 ± 13.4 ns BMI (kg/m2) 23.1 ± 5.1 26.6 ± 4.8 ns Body surface area (m2) 1.74 ± 0.23 1.84 ± 0.19 ns Diabetes, n (%) 11 (50) 0 (0) N/A Hypertension, n (%) 5 (23) 0 (0) N/A Cardiac parameters Sinus rhythm, n (%) 24 (100) 24 (100) ns Heart rate (min−1) 77 ± 13 59 ± 9 <0.0001 SBP (mmHg) 116 ± 13 123 ± 11 ns DBP (mmHg) 77 ± 8 75 ± 9 ns Selected medications ACE inhibitor/ARB 9 (41) 0 (0) N/A Beta blocker 2 (9) 0 (0) N/A Calcium channel blocker 2 (9) 0 (0) N/A Statin 8 (36) 0 (0) N/A Insulin 6 (27) 0 (0) N/A ns, not significant (P > 0.05); N/A, not applicable; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index. Table 2 Frequency of clinical features

Clinical assessment Subjects underwent physical examination by an experienced clinician. Disease burden was assessed using the Newcastle Mitochondrial Disease Adult Scale (NMDAS), a validated scoring system.20 The m.3243A>G mutation load was determined in urinary epithelial cells.21 The estimated glomerular filtration rate (eGFR) was calculated in all patients using the Modified Diet in Renal Disease equation, prior to administration of gadolinium. Cardiac magnetic resonance imaging Using a 3-Tesla Intera Achieva scanner (Philips, Best, NL), subjects underwent (i) 31P MRS, (ii) cine imaging, (ii) cardiac tagging, and (iv) late gadolinium enhancement (LGE) imaging. Cardiac spectroscopy Subjects were scanned prone, using a 10-cm diameter 31P surface coil (Pulseteq, UK). A cardiac gated one-dimensional (1-D) chemical shift imaging (CSI) sequence was used with spatial pre-saturation of the skeletal muscle. A 7-cm slice was selected using a ‘spredrex’-type pulse of 2.3 ms duration to eliminate liver contamination.22 Negligible liver contamination was assured by 1-D foot-head oriented CSI experiments in phantoms: <1% of the total phosphorus signal originated from outside the prescribed volume. Sixteen coronal phase-encoding steps yielded spectra from 10-mm slices (TR = heart rate, 192 averages, acquisition time approximately 20 min). The first spectrum arising entirely beyond the chest wall was analysed using the AMARES time domain fit23 to quantify PCr, the γ resonance of adenosine triphosphate (ATP), and 2,3-diphosphoglycerate (DPG). The ATP peak area was corrected for blood contamination by 1/6th combined 2,3-DPG peak, and PCr/ATP ratios were corrected for T1 saturation and local flip angle.24

est wall was analysed using the AMARES time domain fit23 to quantify PCr, the γ resonance of adenosine triphosphate (ATP), and 2,3-diphosphoglycerate (DPG). The ATP peak area was corrected for blood contamination by 1/6th combined 2,3-DPG peak, and PCr/ATP ratios were corrected for T1 saturation and local flip angle.24 Cine imaging Subjects were scanned supine using a 6-channel cardiac coil and ECG gating (Philips). The short-axis balanced steady-state free precession images were obtained covering the entire left ventricle [field of view (FOV) 350 × 350 mm2, repetition time/echo time (TR/TE) = 3.7/1.9 ms, turbo factor 17, flip angle (FA) 40°, slice thickness 8 mm, 25 phases, resolution 1.37 mm]; long-axis images were acquired. Endocardial and epicardial borders were traced manually on short-axis slices throughout the cardiac cycle using the ViewForum workstation (Philips).The LV mass, and systolic and diastolic parameters, including the ratio of early to late ventricular filling velocity (E/A ratio) and early filling percentage, were calculated as previously described.24 The ratio of the LV mass to end-diastolic volume (M/V ratio) was calculated as an index of concentric hypertrophy.25 Longitudinal shortening was determined in the four-chamber view as the percentage difference in distance from the mitral valve plane to the apex at end-systole and end-diastole. The myocardial wall thickness was determined at the same level as tagging, and the percentage increase from diastole to systole (radial thickening) was calculated.

ng was determined in the four-chamber view as the percentage difference in distance from the mitral valve plane to the apex at end-systole and end-diastole. The myocardial wall thickness was determined at the same level as tagging, and the percentage increase from diastole to systole (radial thickening) was calculated. Cardiac tagging MR signal from the myocardium in diastole was cancelled in a rectangular grid pattern and tags were tracked through the cardiac cycle (Figure 1A).13 A multi-shot turbo-field echo sequence was used (TR/TE/FA/number of averages = 4.9/3.1/10°/1, turbo factor 9, SENSE factor 2, FOV 350 × 350 mm2, voxel size 1.37 mm, tag spacing 7 mm, 12 phases). Two adjacent short-axis slices of 10 mm thickness were acquired at the mid-ventricle with a 2-mm gap. The Cardiac Image Modeling package (Auckland UniServices Ltd, Auckland, New Zealand) was used to align a mesh on the tags between the endocardial and epicardial contours. Peak circumferential strain for both the whole myocardial wall and the endocardial third were calculated. Peak torsion between the two slices was calculated as the circumferential-longitudinal shear angle, defined on the epicardial surface (Figure 1B).13 Figure 1 Cardiac tagging analysis. (A) Cine imaging (top panels) and tagging (bottom) at end-diastole (left panels) and end-systole (right). A rectangular grid of nulled myocardium applied in the diastole enables tracking of myocardial deformation. (B) Tagging of two parallel short-axis slices allows the calculation of torsion, the longitudinal-circumferential sheer angle (θ) as shown.

tagging (bottom) at end-diastole (left panels) and end-systole (right). A rectangular grid of nulled myocardium applied in the diastole enables tracking of myocardial deformation. (B) Tagging of two parallel short-axis slices allows the calculation of torsion, the longitudinal-circumferential sheer angle (θ) as shown. LGE imaging Gadolinium-DTPA of 0.2 mmol/kg (Dotarem, Guerbet, France) was administered intravenously to patients only. LGE images were obtained at 10 min using a breath-held, cardiac-triggered three-dimensional phase-sensitive inversion recovery sequence (multi-shot gradient echo TR/TE = 5/2.4, FA = 15°/5°, acceleration factor 31, parallel imaging factor 2, 1.8-mm resolution zero-filled to 1.3 mm) with the inversion time to null normal myocardium determined from a prior multi-slice 2D Look-Locker experiment (multi-shot EPI with an EPI factor 5, acceleration factor 2, TR/TE = 7.3/2.8, 3-mm resolution). Qualitative analysis for the presence and distribution of LGE in a 17-segment model was performed by two experienced, independent observers, blinded to patient status and MRI/MRS findings as per Society for Cardiovascular Magnetic Resonance guidelines26: consensus agreement was planned in the case of initial disagreement between observers.

for the presence and distribution of LGE in a 17-segment model was performed by two experienced, independent observers, blinded to patient status and MRI/MRS findings as per Society for Cardiovascular Magnetic Resonance guidelines26: consensus agreement was planned in the case of initial disagreement between observers. Statistical analysis Data are presented as means ± SD for continuous data and as percentages and numbers for categorical data. Continuous data were tested for normality using the Shapiro–Wilk test, and group comparisons were made using two-tailed Student's t-tests or Mann–Whitney U tests. Categorical variables were compared using Fisher's exact test and the correlations were executed using Pearson's method. The Bonferroni adjustment for multiple comparisons was used for both group comparisons and correlations. The reliability of MRI measures of myocardial strains and diastolic function were assessed with the Bland–Altman analysis by comparing the values derived from the contours redrawn after 1 month (M.G.D.B.), and by two independent observers (M.G.D.B. and K.G.H.), in four randomly selected patients and four controls. Inter-observer and intra-observer limits of agreement were, respectively, −0.19 ± 0.31° and 0.06 ± 0.51° for torsion, 0.74 ± 1.31 and 1.53 ± 1.08% for endocardial circumferential strain, 0.001 ± 0.11 and 0.08 ± 0.16 for E/A ratio and –0.54 ± 1.58 and 0.68 ± 2.84% for early filling percentage. Using our implementation of cardiac 31P MRS,24 we have previously determined that the reproducibility coefficient of the PCr/ATP ratio was 0.28 or 15% in four control subjects, on two separate occasions. All analysis was performed using SPSS version 17 (SPSS Inc, Chicago, IL). All tests were two-sided and statistical significance was assumed at P < 0.05.

Characteristic Carriers (n = 22) Controls (n = 22) P value Age (years) 42.5 ± 12.2 42.8 ± 13.4 ns Male sex, n (%) 11 (50) 11 (50) ns Height (cm) 168 ± 11 171 ± 11 ns Weight (kg) 65.7 ± 16.2 77.1 ± 13.4 ns BMI (kg/m2) 23.1 ± 5.1 26.6 ± 4.8 ns Body surface area (m2) 1.74 ± 0.23 1.84 ± 0.19 ns Diabetes, n (%) 11 (50) 0 (0) N/A Hypertension, n (%) 5 (23) 0 (0) N/A Cardiac parameters Sinus rhythm, n (%) 24 (100) 24 (100) ns Heart rate (min−1) 77 ± 13 59 ± 9 <0.0001 SBP (mmHg) 116 ± 13 123 ± 11 ns DBP (mmHg) 77 ± 8 75 ± 9 ns Selected medications ACE inhibitor/ARB 9 (41) 0 (0) N/A Beta blocker 2 (9) 0 (0) N/A Calcium channel blocker 2 (9) 0 (0) N/A Statin 8 (36) 0 (0) N/A Insulin 6 (27) 0 (0) N/A ns, not significant (P > 0.05); N/A, not applicable; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index. Table 2 Frequency of clinical features Clinical feature Number of patients Percentage of all patients Hearing loss 20 91 Exercise intolerance 17 77 Ataxia 14 64 Constipation 13 59 Myopathy 12 55 Diabetes mellitus 11 50 Migraine 10 45 Depression 10 45 Retinopathy 8 36 Fatigue 6 27 Short stature 5 23 Dysarthria 4 18 Underweight (BMI <18.5) 4 18 Myalgia 4 18 Epilepsy 2 9 Neuropathy 2 9 Hypothyroidism 2 9 Ophthalmoplegia 2 9 Ptosis 2 9 Cataracts 1 5 Encephalopathy 1 5 Dysphagia 1 5 Stroke-like episodes 1 5 BMI, body mass index (kg/m2).

0 Migraine 10 45 Depression 10 45 Retinopathy 8 36 Fatigue 6 27 Short stature 5 23 Dysarthria 4 18 Underweight (BMI <18.5) 4 18 Myalgia 4 18 Epilepsy 2 9 Neuropathy 2 9 Hypothyroidism 2 9 Ophthalmoplegia 2 9 Ptosis 2 9 Cataracts 1 5 Encephalopathy 1 5 Dysphagia 1 5 Stroke-like episodes 1 5 BMI, body mass index (kg/m2). Cardiac morphology and global function Table 3 summarizes the morphological and functional parameters for patient and control groups with all subjects completing the scan protocol (total duration of 77 ± 13 min). The means and ranges of control group parameters are in agreement with a large cohort study using quantitative cardiac MRI.27 Table 3 Cardiac morphology and function Parameter Patients Controls P value End-diastolic volume (mL) 93 ± 17 138 ± 26 <0.0001 End-diastolic index (mL/m2) 54 ± 9 73 ± 15 <0.0001 End-systolic volume (mL) 36 ± 10 57 ± 15 <0.0001 End-systolic index (mL/m2) 21 ± 6 31 ± 8 <0.001 Stroke volume (mL) 57 ± 10 82 ± 15 <0.0001 Stroke index (mL/m2) 33 ± 5 45 ± 8 <0.0001 Cardiac output (l/min) 4.4 ± 1.0 4.7 ± 0.6 ns Cardiac index (l/min/m2) 2.5 ± 0.5 2.6 ± 0.4 ns Ejection fraction (%) 62 ± 7 59 ± 6 ns LV mass (g) 119 ± 28 109 ± 20 ns Wall thickness in systole (mm) 13.5 ± 3.1 10.9 ± 1.9 <0.01 Wall thickness in diastole (mm) 9.8 ± 2.9 6.9 ± 1.1 <0.001 LVMI (LV mass/BSA) (g/m2) 70 ± 12 59 ± 7 <0.01 M/V ratio (LV mass/end-diastolic volume) (g/mL) 1.28 ± 0.33 0.81 ± 0.14 <0.0001 ns, not significant (P > 0.05); LV, left ventricular; LVMI, left ventricular mass index; BSA, body surface area; M/V, LV mass/end-diastolic volume.

s in diastole (mm) 9.8 ± 2.9 6.9 ± 1.1 <0.001 LVMI (LV mass/BSA) (g/m2) 70 ± 12 59 ± 7 <0.01 M/V ratio (LV mass/end-diastolic volume) (g/mL) 1.28 ± 0.33 0.81 ± 0.14 <0.0001 ns, not significant (P > 0.05); LV, left ventricular; LVMI, left ventricular mass index; BSA, body surface area; M/V, LV mass/end-diastolic volume. The end-systolic and end-diastolic volumes, both as raw values and when indexed to body surface area (BSA), were significantly decreased in patients compared with controls (Table 3). A proportional decrease in these parameters (33 and 37%, respectively) ensured no significant difference in ejection fraction. Stroke volume and stroke index were significantly decreased in patients: this occurred in association with a significant increase in the heart rate (r = −0.65, P < 0.01) with no significant difference in cardiac output or cardiac index between the groups.

ctively) ensured no significant difference in ejection fraction. Stroke volume and stroke index were significantly decreased in patients: this occurred in association with a significant increase in the heart rate (r = −0.65, P < 0.01) with no significant difference in cardiac output or cardiac index between the groups. The LV mass was not significantly different between the patient and control groups (Table 3). However, when indexed to BSA, the LV mass index (LVMI) was significantly increased in patients. A significant increase in the M/V ratio (58%) suggested that concentric remodelling had occurred. Consistent with this, the radial wall thicknesses in both diastole and systole were significantly increased in patients. Within the control group, there were significant correlations between the systolic blood pressure and both LV mass (r = 0.57, P < 0.02) and the radial wall thickness in diastole (r = 0.54, P < 0.03). Within the patient group, no significant correlations were shown between systolic or diastolic blood pressure and any markers of LV mass (Supplementary data online, Figure S1). Similarly fasting blood glucose and HbA1C in patients did not correlate significantly with the LV mass, LVMI, radial wall thicknesses or M/V ratio. The exclusion of diabetic or treated hypertensive patients from statistical analyses did not affect significantly increases in LVMI or radial wall thicknesses. Both the independent observers were in agreement that no patients displayed evidence of focal intramyocardial fibrosis on LGE imaging.

The LV mass was not significantly different between the patient and control groups (Table 3). However, when indexed to BSA, the LV mass index (LVMI) was significantly increased in patients. A significant increase in the M/V ratio (58%) suggested that concentric remodelling had occurred. Consistent with this, the radial wall thicknesses in both diastole and systole were significantly increased in patients. Within the control group, there were significant correlations between the systolic blood pressure and both LV mass (r = 0.57, P < 0.02) and the radial wall thickness in diastole (r = 0.54, P < 0.03). Within the patient group, no significant correlations were shown between systolic or diastolic blood pressure and any markers of LV mass (Supplementary data online, Figure S1). Similarly fasting blood glucose and HbA1C in patients did not correlate significantly with the LV mass, LVMI, radial wall thicknesses or M/V ratio. The exclusion of diabetic or treated hypertensive patients from statistical analyses did not affect significantly increases in LVMI or radial wall thicknesses. Both the independent observers were in agreement that no patients displayed evidence of focal intramyocardial fibrosis on LGE imaging. Cardiac tagging and myocardial strains Longitudinal shortening was significantly decreased (17%) in patients (Table 4) and correlated significantly with an increased LVMI (r = −0.52, P < 0.03). Peak torsion (36%) and TSR (35%) were significantly increased in patients. No significant differences were detected in the rates of systolic and diastolic torsion, after correction for peak torsion, or in the diastolic function represented by the E/A ratio and early filling percentage. Table 4 Cardiac tagging and diastolic function

(36%) and TSR (35%) were significantly increased in patients. No significant differences were detected in the rates of systolic and diastolic torsion, after correction for peak torsion, or in the diastolic function represented by the E/A ratio and early filling percentage. Table 4 Cardiac tagging and diastolic function Parameter Patients Controls P value Longitudinal shortening (%) 15.1 ± 1.5 18.2 ± 2.3 <0.0001 Radial wall thickening (%) 65.6 ± 17.0 60.1 ± 16.0 ns Peak torsion (°) 8.0 ± 2.7 5.9 ± 1.4 <0.03 Systolic torsion rate (°/s) 37 ± 13 24 ± 10 <0.02 Diastolic torsion rate (°/s) 23 ± 10 17 ± 10 ns Systolic torsion rate/peak torsion (s−1) 4.7 ± 1.3 4.1 ± 1.8 ns Diastolic torsion rate/peak torsion (s−1) 3.0 ± 1.3 2.8 ± 1.6 ns Peak whole wall circumferential strain (%) 16.7 ± 2.2 17.8 ± 2.5 ns Peak endocardial circumferential strain (%) 22.6 ± 2.6 24.4 ± 2.5 ns TSR (rad) 0.62 ± 0.21 0.46 ± 0.10 <0.03 E/A ratio 1.53 ± 0.55 1.75 ± 0.62 ns Early filling (%) 72.4 ± 8.9 73.2 ± 8.3 ns ns, not significant (P > 0.05); E/A ratio, ratio of early to late ventricular filling velocity; TSR, torsion to endocardial strain ratio. No significant differences were found in radial thickening or in the circumferential whole wall or endocardial strain (Table 4).

Parameter Patients Controls P value Longitudinal shortening (%) 15.1 ± 1.5 18.2 ± 2.3 <0.0001 Radial wall thickening (%) 65.6 ± 17.0 60.1 ± 16.0 ns Peak torsion (°) 8.0 ± 2.7 5.9 ± 1.4 <0.03 Systolic torsion rate (°/s) 37 ± 13 24 ± 10 <0.02 Diastolic torsion rate (°/s) 23 ± 10 17 ± 10 ns Systolic torsion rate/peak torsion (s−1) 4.7 ± 1.3 4.1 ± 1.8 ns Diastolic torsion rate/peak torsion (s−1) 3.0 ± 1.3 2.8 ± 1.6 ns Peak whole wall circumferential strain (%) 16.7 ± 2.2 17.8 ± 2.5 ns Peak endocardial circumferential strain (%) 22.6 ± 2.6 24.4 ± 2.5 ns TSR (rad) 0.62 ± 0.21 0.46 ± 0.10 <0.03 E/A ratio 1.53 ± 0.55 1.75 ± 0.62 ns Early filling (%) 72.4 ± 8.9 73.2 ± 8.3 ns ns, not significant (P > 0.05); E/A ratio, ratio of early to late ventricular filling velocity; TSR, torsion to endocardial strain ratio. No significant differences were found in radial thickening or in the circumferential whole wall or endocardial strain (Table 4). Mutation load and clinical status There was a significant correlation between the urinary mutation load (mean: 62 ± 20%, range: 22–90%) and disease burden (NMDAS mean score: 18 ± 11, range: 2–42) among patients (r = 0.59, P < 0.02). Both these clinical markers displayed significant correlations with the LVMI (respectively, r = 0.71 and r = 0.79, both P < 0.001), and peak endocardial circumferential strain (r = −0.59 and r = −0.57, both P < 0.03, Figure 2). Figure 2 LVMI and clinical markers. LVMI correlated positively with both (A) urinary mutation load and (B) NMDAS score, while peak endocardial circumferential strain correlated negatively with (C) urinary mutation load and (D) NMDAS score. LVMI, left ventricular mass index; NMDAS, Newcastle Mitochondrial Disease Adult Scale.

2 LVMI and clinical markers. LVMI correlated positively with both (A) urinary mutation load and (B) NMDAS score, while peak endocardial circumferential strain correlated negatively with (C) urinary mutation load and (D) NMDAS score. LVMI, left ventricular mass index; NMDAS, Newcastle Mitochondrial Disease Adult Scale. Myocardial bioenergetics The PCr/ATP ratio was decreased (Figure 3, mean decrease 21%, P < 0.001) in patients (1.51 ± 0.34) compared with controls (1.92 ± 0.20). There were no significant correlations between the PCr/ATP ratio and markers of disease burden or myocardial mass or function. Thirteen patients (59%) had an abnormal PCr/ATP ratio (<1.6) but there were no significant differences in markers of disease burden, cardiac morphology or the function between patients with the PCr/ATP ratio >1.6 and those <1.6.18 Figure 3 31P magnetic resonance spectroscopy. Representative spectra from (A) a patient carrying m.3243A>G, with the PCr/ATP ratio of 1.23, and (B) a matched control subject, with the PCr/ATP ratio of 2.10, showing a difference in the PCr concentration. Spectra are presented as acquired before correction for heart rate, flip angle and blood content. (C) A box plot of range and quartiles of the PCr/ATP ratio in patient and control groups. 2,3DPG, 2,3-disphosphoglycerate; PDE, phosphodiesters; PCr, phosphocreatine; ATP, adenosine triphosphate; ppm, parts per million; *P < 0.001 compared with controls.

d before correction for heart rate, flip angle and blood content. (C) A box plot of range and quartiles of the PCr/ATP ratio in patient and control groups. 2,3DPG, 2,3-disphosphoglycerate; PDE, phosphodiesters; PCr, phosphocreatine; ATP, adenosine triphosphate; ppm, parts per million; *P < 0.001 compared with controls. Discussion This study used a combined approach of comprehensive cardiac MRI and 31P MRS to examine myocardial morphology, function, and bioenergetics in 22 patients harbouring the m.3243A>G mutation without clinical cardiac involvement. The major findings in these patients compared with age- and gender-matched controls were: (i) LVMI was greater and was a more sensitive indicator of subtle cardiac hypertrophy than the LV mass; (ii) concentric remodelling occurred in the absence of hypertension or diabetes mellitus; (iii) altered systolic myocardial strains occurred, with reduced longitudinal shortening and increased peak torsion, in the absence of global systolic or diastolic dysfunction; (iv) early changes in the cardiac morphology and strains were associated with increased mtDNA mutation load and NMDAS score; and (v) PCr/ATP ratio was reduced, but did not correlate with structural or functional cardiac indices or markers of disease burden.

n, in the absence of global systolic or diastolic dysfunction; (iv) early changes in the cardiac morphology and strains were associated with increased mtDNA mutation load and NMDAS score; and (v) PCr/ATP ratio was reduced, but did not correlate with structural or functional cardiac indices or markers of disease burden. Cardiac morphology and function Patients in this study displayed concentric hypertrophic remodelling, as evidenced by increased M/V ratio and wall thicknesses, which were independent of diabetic or hypertensive status. Reductions in the end-systolic and end-diastolic blood pool volumes, consistent with concentric remodelling, resulted in decreased stroke volume and index in this study. The finding of significantly elevated heart rate, that ensured no difference in the cardiac output or index was however intriguing and has not been previously reported in this patient group. No likely culprit medications were identified in the patient group as a whole and the possibility of a relationship to the underlying disease process remains to be explored. The pattern of concentric hypertrophic remodelling observed in this study is similar to that in normal ageing where reduced ventricular volumes and increased M/V ratio, with a minimal change in the LV mass, have been linked to adverse cardiac outcomes, particularly when present in those <65 years of age.25

be explored. The pattern of concentric hypertrophic remodelling observed in this study is similar to that in normal ageing where reduced ventricular volumes and increased M/V ratio, with a minimal change in the LV mass, have been linked to adverse cardiac outcomes, particularly when present in those <65 years of age.25 Small studies have suggested differing estimates of the prevalence of LVH in m.3243A>G mutation carriers.4,6 This study, performed in patients without preceding evidence of cardiac involvement, demonstrates the critical importance of referencing LV mass to the BSA. Although not reaching statistical significance, patients had smaller body mass, height, BMI and BSA compared with controls. The LV mass was significantly increased in patients, but only when indexed to BSA. Taken together, these findings suggest that previous cohort studies that used absolute measures of the LV mass and did not employ indexation may have underestimated the number of patients with an increased indexed LV mass. While standard definitions of LVH should be used, this implies that the frequency of LVH in patients, which is potentially amenable to treatment, may be higher than previously indicated.4 A larger cross-sectional study would be required to investigate this issue.

he number of patients with an increased indexed LV mass. While standard definitions of LVH should be used, this implies that the frequency of LVH in patients, which is potentially amenable to treatment, may be higher than previously indicated.4 A larger cross-sectional study would be required to investigate this issue. Myocardial strains and torsion Consistent with LVH in other clinical contexts,15,28 concentric remodelling in this study was associated with reduced longitudinal shortening. In healthy ageing and patients with diabetes mellitus or neuromuscular diseases, circumferential strain is reduced, and associated with reduced longitudinal shortening.25,28,29 In this study, endocardial and whole wall circumferential strains tended to be reduced in patients without reaching statistical significance. The higher mutation load and NMDAS score, indicative of greater disease burden, did however correlate with increased LVMI and reduced endocardial circumferential strain.

25,28,29 In this study, endocardial and whole wall circumferential strains tended to be reduced in patients without reaching statistical significance. The higher mutation load and NMDAS score, indicative of greater disease burden, did however correlate with increased LVMI and reduced endocardial circumferential strain. Increased torsion and/or TSR, often with reduced longitudinal shortening, have been reported in patients with LVH secondary to HCM,15,16 or conditions of increased afterload including aortic stenosis and hypertension.14,30 In all cases, such changes in torsion, TSR and longitudinal shortening are believed to be due to a reduction in the contractile function in the subendocardium compared with subepicardium.14 Recently Chung et al.31 demonstrated increased torsion in the absence of morphological cardiac disease in diabetic patients. Although the pathophysiology of diabetes in mitochondrial disease is distinct,32 our results of increased torsion and TSR without significant change in circumferential strains concur with these findings and those reported in HCM mutation carriers without LVH.16 Additionally, as in our cohort, both these studies reported isolated systolic abnormalities of torsion with no difference in controls in the rate of torsion dissipation during diastole, or basic measures of diastolic function. Myocardial perfusion defects could contribute to this increased torsion, with abnormalities of subendocardial arterioles in HCM hearts,33 and small vessel disease in diabetes. Such abnormalities could similarly be responsible for the differences in myocardial deformation in our study. However, our patients also demonstrated significant reductions in end-diastolic and end-systolic volumes, yielding smaller radii for the myocardium. Increased torsion could have resulted from the additional dominance this gave to the subepicardium.34 To distinguish between these explanations would require a measure of perfusion at the subendocardium.

strated significant reductions in end-diastolic and end-systolic volumes, yielding smaller radii for the myocardium. Increased torsion could have resulted from the additional dominance this gave to the subepicardium.34 To distinguish between these explanations would require a measure of perfusion at the subendocardium. Disease burden Cardiac involvement is an important prognostic factor in mitochondrial disease since complications of cardiomyopathy are a frequent cause of premature death. Yet the pathophysiological mechanisms linking the m.3243A>G mutation to LVH and cardiomyopathy remain unknown. Our group has previously shown that urinary mutation load is the best predictor of overall clinical outcome in the m.3243A>G mutation carriers.21 In this study, we report for the first time a correlation between the urinary mutation load and cardiac involvement specifically, as evidenced by increased LVMI. The NMDAS score correlated strongly with both the urinary mutation load and the LVMI. These important findings support the primary importance of mtDNA mutations in the changes observed in cardiac morphology, and may support more intensive cardiac evaluation of patients with higher mutation loads and/or NMDAS scores.

I. The NMDAS score correlated strongly with both the urinary mutation load and the LVMI. These important findings support the primary importance of mtDNA mutations in the changes observed in cardiac morphology, and may support more intensive cardiac evaluation of patients with higher mutation loads and/or NMDAS scores. Cardiac bioenergetics Reductions in the PCr/ATP ratio, assessed non-invasively using 31P MRS, have prognostic importance in diverse forms of cardiomyopathy.18 Indeed myocardial energy depletion has been proposed as a critical mechanism linking sarcomeric defects to hypertrophy in the HCM.17 Our study confirms the findings of a study in m.3243A>G mutation carriers, which found that the PCr/ATP ratio was significantly reduced.19 However, we were unable to detect any correlation between the cardiac bioenergetic defect and MRI-based parameters of myocardial structure or function, or markers of disease burden in patients without clinical cardiac disease. In different forms of inherited cardiomyopathy, several groups have suggested the primacy of bioenergetic defects by detection of abnormalities in mutation carriers without evidence of LVH.17,35 However, we detected significant differences in cardiac remodelling, known itself to cause a reduction in the PCr/ATP ratio and potentially explaining the lack of correlation with other parameters. The PCr/ATP ratio does not in isolation appear to have a prognostic value in the detection of early cardiac remodelling in patients harbouring the m.3243A>G mutation, but a larger natural history study would be required to confirm this suggestion.

nd potentially explaining the lack of correlation with other parameters. The PCr/ATP ratio does not in isolation appear to have a prognostic value in the detection of early cardiac remodelling in patients harbouring the m.3243A>G mutation, but a larger natural history study would be required to confirm this suggestion. Clinical implications There are several clinical implications from our findings. First, the indexing of measures of the LV mass to BSA or the end-diastolic volume is essential to detect early concentric remodelling in patients harbouring the m.3243A>G mutation. Cardiac involvement may be more prevalent than previously suspected. Secondly, patients with a higher urinary mutation load, or NMDAS score, may be at an increased risk of developing cardiomyopathy, supporting more frequent cardiac screening. Finally, natural history studies of pathogenesis and eventual clinical therapeutic trials are dependent on an ability to identify the earliest biomechanical changes attributable to the m.3243A>G mutation. We have shown for the first time increased torsion and abnormal myocardial strains in this cohort, and suggest that the measurement of LV mechanics may be useful in assessing disease progression and response to intervention.

ability to identify the earliest biomechanical changes attributable to the m.3243A>G mutation. We have shown for the first time increased torsion and abnormal myocardial strains in this cohort, and suggest that the measurement of LV mechanics may be useful in assessing disease progression and response to intervention. Limitations Although this study is the largest cardiac MRI-based investigation performed to date in this patient group, it remains limited in the sample size and was not designed to investigate pathogenetic mechanisms or disease progression. Cardiac involvement in mitochondrial disease is linked to clinical outcomes, yet we acknowledge that the prognostic importance of the changes we describe must be determined through longitudinal studies. We studied a relatively homogenous cohort of patients, harbouring the single commonest mtDNA point mutation, without known cardiac involvement; such patients account for ∼25% of our specialist clinic attendees yet we recognize that our findings may not be generalizable to all patients with mtDNA point mutations. We did not perform echocardiography in controls and so cannot use age- and gender-matching to compare echocardiographic and MRI-derived parameters, particularly with regard to the diastolic dysfunction. Longitudinal shortening was used to provide a global measure of the long-axis function rather than the longitudinal strain, which would require additional tagged long-axis slices. Similarly, in an already extensive MRI protocol, we did not study flow-based analyses of the diastolic function or first-pass perfusion. Finally, we did not perform LGE imaging in controls and cannot exclude the presence of focal fibrosis in these individuals, although the probability of this is very low.

-axis slices. Similarly, in an already extensive MRI protocol, we did not study flow-based analyses of the diastolic function or first-pass perfusion. Finally, we did not perform LGE imaging in controls and cannot exclude the presence of focal fibrosis in these individuals, although the probability of this is very low. Conclusions Concentric remodelling is prevalent in patients harbouring the m.3243A>G mutation and occurs in association with characteristic changes in systolic intramyocardial strains and torsion. These findings, which are closely related to urinary mutation load and disease burden, occur in patients without existing evidence of cardiac involvement, and may provide an early marker of myocardial pathology, enabling future studies of pathogenesis and intervention. Supplementary data Supplementary data are available at European Heart Journal – Cardiovascular Imaging online.

Conclusions Concentric remodelling is prevalent in patients harbouring the m.3243A>G mutation and occurs in association with characteristic changes in systolic intramyocardial strains and torsion. These findings, which are closely related to urinary mutation load and disease burden, occur in patients without existing evidence of cardiac involvement, and may provide an early marker of myocardial pathology, enabling future studies of pathogenesis and intervention. Supplementary data Supplementary data are available at European Heart Journal – Cardiovascular Imaging online. Funding This work was supported by the Wellcome Trust [BH092142 to M.G.D.B., 096919Z/11/Z to P.F.C., D.M.T. and R.W.T., 074454/Z/04/Z to D.M.T. and R.W.T.]; the Medical Research Council [G1100160 to K.G.H,. G0601943 to D.M.T. and G0800674 to P.F.C., D.M.T. and R.W.T.]; the UK National Institute for Health Research Biomedical Research Centre for Ageing and Age-related Diseases award to Newcastle upon Tyne Hospitals NHS Foundation Trust [for P.F.C., D.M.T., G.S.G., control data and cardiac tagging software]; the British Heart Foundation [CH/07/001 to B.D.K.]; and the UK NHS Specialized Services and Newcastle upon Tyne Hospitals NHS Foundation Trust that support the ‘Rare Mitochondrial Disorders of Adults and Children’ Diagnostic Service [http://www.mitochondrialncg.nhs.uk]. Magnetic resonance imaging data from four patients in this study were collected during the screening phase of a commercially sponsored study funded by Penwest Pharmaceuticals.

Foundation Trust that support the ‘Rare Mitochondrial Disorders of Adults and Children’ Diagnostic Service [http://www.mitochondrialncg.nhs.uk]. Magnetic resonance imaging data from four patients in this study were collected during the screening phase of a commercially sponsored study funded by Penwest Pharmaceuticals. Supplementary Material Supplementary Data Acknowledgements We thank patients and volunteers involved in this study. We are grateful to Louise Morris, Tamsin Gaudie and Tim Hodgson, Research Radiographers, and to Ben Dixon and Rajiv Das for assistance with image analysis. Conflict of interest: None declared.

Introduction Heart failure (HF) represents the end stage of the continuum of cardiovascular diseases. This disorder is common, and creates an enormous health-care burden. The prognosis of HF patients is dismal, with 5-year survival rates worse than many of the most common cancers.1 Cardiac remodelling is a key component of HF that progresses from adaptive to maladaptive as the disease worsens, and is associated with increased risks of symptoms and mortality.2 Patients with early or mild HF are often clinically compensated and their left ventricular (LV) stroke volume (SV) may be within the normal range. Despite normal LVSV, the prognosis and clinical outcomes of such patients are impaired compared with healthy comparators. Blood flow through the heart and vessels is a fundamental aspect of cardiovascular function.3,4 In HF, altered intra-cardiac flow patterns have been recognized.5,6 Three-dimensional, time-resolved flow encoded MRI provides comprehensive velocity data of the intrinsically three-dimensional and time-varying (3D+time = 4D) flow patterns within the beating heart.7–15 Recently, the four-dimensional (4D) flow within the LV has been quantified and separated into different functional flow components.7,9,10 These components have been used to demonstrate component-specific routes and energetics in healthy hearts that may represent important aspects of normal ventricular function, including an improved preservation of LV inflow kinetic energy (KE) for effective and rapid systolic ejection.

functional flow components.7,9,10 These components have been used to demonstrate component-specific routes and energetics in healthy hearts that may represent important aspects of normal ventricular function, including an improved preservation of LV inflow kinetic energy (KE) for effective and rapid systolic ejection. We hypothesized that compensated HF patients, despite having comparable LVSVs to healthy subjects, would manifest their functional LV impairment as less LV diastolic inflow KE preservation and a diminished percentage of the flow component that transits the LV in a single cardiac cycle. Such discrepancies in LV diastolic–systolic coupling may represent important pathophysiological aspects of occult LV dysfunction. Accordingly, we measured LV 4D flow-specific measures in patients with compensated HF and compared them with those in healthy subjects.

ponent that transits the LV in a single cardiac cycle. Such discrepancies in LV diastolic–systolic coupling may represent important pathophysiological aspects of occult LV dysfunction. Accordingly, we measured LV 4D flow-specific measures in patients with compensated HF and compared them with those in healthy subjects. Methods Study population Ten patients with clinically compensated dilated cardiomyopathy (DCM) and ten healthy subjects were included in the study (Table 1). At the time of diagnosis, DCM was defined as the presence of symptoms and signs of HF in the presence of echocardiographic findings of ventricular enlargement and systolic myocardial dysfunction by parameters of depressed ejection fraction and mitral annular descent. The patients were recruited from outpatients seen at the Department of Cardiology, Linköping University Hospital. Seven of the patients and seven of the controls were matched for age and gender. All subjects were in sinus rhythm. Inclusion criteria for (i) healthy subjects: normal electrocardiographic and echocardiographic examinations; and (ii) DCM patients: ≤65 years of age. Exclusion criteria for (i) all subjects: contraindications to MRI examination; (ii) normal subjects: a history of prior or current heart disease or the use of cardiac medication, and (iii) DCM patients: significantly irregular cardiac rhythm, a history of myocardial infarction, ≥moderate arterial hypertension, as well as ≥moderate valvular disorder, <mild LV dilatation, and <mild LV systolic dysfunction defined by echocardiography. Three of the DCM patients were included in a prior method validation study.9 All subjects gave written informed consent before participation and the study was approved by the regional ethical review board. The study complies with the declaration of Helsinki. Table 1 Demographical and clinical data of the study population

ree of the DCM patients were included in a prior method validation study.9 All subjects gave written informed consent before participation and the study was approved by the regional ethical review board. The study complies with the declaration of Helsinki. Table 1 Demographical and clinical data of the study population Parameter DCM (n = 10) Healthy (n = 10) P-value Gender (female:male) 6:4 4:6 — Age (years) 49 ± 14 44 ± 17 0.54 Weight (kg) 82 ± 18 72 ± 8 0.172 Heart rate (bpm) 61 ± 11 68 ± 10 0.129 Systolic BP (mmHg) 122 ± 14 129 ± 11 0.269 Diastolic BP (mmHg) 77 ± 9 80 ± 4 0.439 LV end-diastolic volume (mL) 179 ± 33 147 ± 22 0.021 LV sphericity index 0.75 ± 0.12 0.57 ± 0.06 0.001 LV ejection fraction 42 ± 5 54 ± 6 0.000 LV diastolic function according to echo Doppler indices Normal diastolic function 5 10 — Relaxation abnormality 2 — — Pseudonormal filling 1 — — Restrictive filling 2 — — Medication ARB/ACE-I 10 0 — Beta-blocker 10 0 — Diuretic 3 0 — Aldosterone I 2 0 — Statin 3 0 — Warfarin 4 0 — Aspirin 2 0 — Digoxin 1 0 — All values are mean ± 1 SD unless otherwise stated. P-values <0.05 are significant. LV, left ventricle; BP, blood pressure; ARB, angiotensin II receptor blocker; ACE-I, angiotensin-converting enzyme inhibitor.

Beta-blocker 10 0 — Diuretic 3 0 — Aldosterone I 2 0 — Statin 3 0 — Warfarin 4 0 — Aspirin 2 0 — Digoxin 1 0 — All values are mean ± 1 SD unless otherwise stated. P-values <0.05 are significant. LV, left ventricle; BP, blood pressure; ARB, angiotensin II receptor blocker; ACE-I, angiotensin-converting enzyme inhibitor. Data acquisition All subjects underwent MRI examination to acquire 4D flow data and morphological images. The 4D flow data were acquired during free-breathing, using an ECG-triggered, retrospectively navigator-gated, three-dimensional, three-directional, time-resolved phase contrast MRI sequence on a clinical 1.5 T Philips Achieva scanner (Philips Healthcare, Best, the Netherlands). The acquisition and post-processing of the 4D flow data were performed as described previously.9 Scan parameters included: velocity encoding (VENC) = 100 cm/s, flip angle = 8°, echo time = 3.7 ms, repetition time = 6.3 ms, parallel imaging (SENSE) speed-up factor = 2, and k-space segmentation factor = 2. These settings gave a temporal resolution of 50.4 ms. The spatial resolution was 3 × 3 × 3 mm3 and the field-of-view (FOV) was adjusted for each subject to cover the left heart. The mean scan time, including the navigator efficiency and arrhythmia rejection for the Cartesian 3DcinePC sequence, was 31 min (range 16–57, median: 30, std: 8). The mean time required to acquire all k-space lines required for the 4D flow data, excluding navigator efficiency and arrhythmia rejection, was ∼12 min (std: 2, median: 12, and range: 9–18). After post-processing, where corrections were made for concomitant gradient fields, phase-wraps, and background phase errors, the first of a set of data quality control steps was applied as in Eriksson et al.9 This consists of emitting pathlines continuously throughout the cardiac cycle from an emitter grid positioned in the left atrium. A pathline shows the path that an imaginary mass-less particle would take through the time resolved velocity field. If pathlines leave the blood pool at locations incompatible with the actual blood flow, e.g. leaving the LV through the ventricular apex, this is an indication of data imperfections. All data sets passed this quality control.

path that an imaginary mass-less particle would take through the time resolved velocity field. If pathlines leave the blood pool at locations incompatible with the actual blood flow, e.g. leaving the LV through the ventricular apex, this is an indication of data imperfections. All data sets passed this quality control. Morphological long-axis and two stacks of short-axis images were acquired in 30 time frames during end-expiratory breath holds, using cine-balanced steady-state free-precession. Slice thickness was 8 mm. The acquired and reconstructed pixel size for the short-axis images was 2.19 × 1.78 and 1.37 × 1.37 mm2, respectively. Echocardiography was performed using a Vivid 7 scanner and a 2.0 MHz probe (GE, Vingmed Ultrasound, Horten, Norway), and used for enrollment criteria. In addition, LV diastolic function at rest was characterized according to conventional echo Doppler indices.16

Morphological long-axis and two stacks of short-axis images were acquired in 30 time frames during end-expiratory breath holds, using cine-balanced steady-state free-precession. Slice thickness was 8 mm. The acquired and reconstructed pixel size for the short-axis images was 2.19 × 1.78 and 1.37 × 1.37 mm2, respectively. Echocardiography was performed using a Vivid 7 scanner and a 2.0 MHz probe (GE, Vingmed Ultrasound, Horten, Norway), and used for enrollment criteria. In addition, LV diastolic function at rest was characterized according to conventional echo Doppler indices.16 Data analysis All data sets were analysed using a previously presented and validated method.7,9 In short, this consists of a manual segmentation of the LV from short-axis images at time of end-diastole (ED) and end-systole (ES) using freely available software (Segment).17 The segmentation at ED is resampled to give a volume with isotropic voxels of the same size as the flow data. From the centre of each voxel in the LV segmentation, a pathline is emitted. Pathlines are created backwards and forwards in time until the preceding or subsequent ES, respectively. In combination, these forwards and backwards pathlines represent the entire LV end-diastolic volume (EDV) tracked over one complete cardiac cycle. The positions of all pathlines at the time of ES relative to the cardiac chambers defined by the ES segmentation are then used to separate them into four different flow components: direct flow, retained inflow, delayed ejection flow, and residual volume, per definitions listed in Table 2 and Figure 1.9,10 Further, the inflow components direct flow and retained inflow were divided into early and late inflow as in Eriksson et al.7 The early and late diastolic phases were defined as the interval from onset diastole until mid-diastasis and the interval from mid-diastasis until ED, respectively. The time of mid-diastasis was defined visually as the time frame at which the lowest number of pathlines crossed the mitral valve plane. As each pathline represents a volume of blood with a calculable mass (by using the density of blood, ρblood = 1060 kg/m3) and a known velocity at every point in time, the KE for each pathline can be calculated at every point in time by , where Vpathline is the volume that one pathline represents and v­pathline is the velocity of the pathline at a given point in time. Each component's KE is the sum of the KE for each of its pathlines.7 The KE for each component was then divided by its volume. The KE of each component's pathlines at the time of ED were compared. Table 2 Definition of LV blood flow components

nts and v­pathline is the velocity of the pathline at a given point in time. Each component's KE is the sum of the KE for each of its pathlines.7 The KE for each component was then divided by its volume. The KE of each component's pathlines at the time of ED were compared. Table 2 Definition of LV blood flow components Component Definition Direct flow Blood that enters the LV during diastole and leaves the LV during systole in the analysed heart beat; component of both inflow and ejected volume Retained inflow Blood that enters the LV during diastole but does not leave during systole in the analysed heart beat; component of inflow volume only Delayed ejection flow Blood that starts and resides inside the LV during diastole and leaves during systole in the analysed heart beat; component of ejected volume only Residual volume Blood that resides within the LV for at least two cardiac cycles; not a component of inflow or ejected volume Figure 1 Blood flow component definitions: illustration showing the components defined in Table 2. Direct flow, green; retained inflow, yellow, delayed ejection flow, blue; and residual volume, red. LA, left atrium; LV, left ventricle.

Component Definition Direct flow Blood that enters the LV during diastole and leaves the LV during systole in the analysed heart beat; component of both inflow and ejected volume Retained inflow Blood that enters the LV during diastole but does not leave during systole in the analysed heart beat; component of inflow volume only Delayed ejection flow Blood that starts and resides inside the LV during diastole and leaves during systole in the analysed heart beat; component of ejected volume only Residual volume Blood that resides within the LV for at least two cardiac cycles; not a component of inflow or ejected volume Figure 1 Blood flow component definitions: illustration showing the components defined in Table 2. Direct flow, green; retained inflow, yellow, delayed ejection flow, blue; and residual volume, red. LA, left atrium; LV, left ventricle. Statistical analysis Inter-group comparisons of inflow and ejected volumes and of parameters in Table 1 were made using two-sample t-tests; a P-value <0.05 was considered significant. Inter-group comparisons of component volumes (Figure 3) and end-diastolic KE (Figure 5) were made using two-sample t-tests, Bonferroni corrected for multiple comparisons, and hence a P-value <0.0125 was considered significant. For intra-group comparison of components, a two-way analysis of variance with the Tukey post hoc test was used, and a P-value <0.05 was considered significant. For comparisons between early (E) and late (A) inflow (Figure 6), the following was done: for intra-group comparison, i.e. E vs. A in the same subject, Bonferroni-corrected paired t-tests were used and a P-value <0.025 was considered significant. For inter-group comparison, Bonferroni-corrected two-sample t-tests were used, and a P-value <0.0125 was considered significant. Minitab 16.2 was used for statistical calculations.

omparison, i.e. E vs. A in the same subject, Bonferroni-corrected paired t-tests were used and a P-value <0.025 was considered significant. For inter-group comparison, Bonferroni-corrected two-sample t-tests were used, and a P-value <0.0125 was considered significant. Minitab 16.2 was used for statistical calculations. Results The analysis of clinical data at rest (Table 1) shows that there was no statistically significant difference between the groups for age, heart rate, weight, or blood pressure. There were differences in LV morphological characteristics and global systolic function: the EDV and the sphericity index were higher and the ejection fraction was lower in the DCM patients compared with healthy subjects. All healthy subjects and half of the patients showed normal LV diastolic function. All patients were in NYHA class II. Two of the patients had a left bundle branch block.

stolic function: the EDV and the sphericity index were higher and the ejection fraction was lower in the DCM patients compared with healthy subjects. All healthy subjects and half of the patients showed normal LV diastolic function. All patients were in NYHA class II. Two of the patients had a left bundle branch block. Visualizations were created of all flow data sets. These were inspected as an initial data quality assurance check. Confirmation of adequate data quality required pathline constraint to anatomical structures via visual inspection (Figure 2 and Supplementary data online, Movies S1–S3). A second data quality check was performed by comparison of quantitative inflow and outflow LV volumes to assure equivalence. No data sets were rejected based on visualization, and there was no significant difference in inflow volume vs. ejected volume in either the DCM patients or the healthy subjects (P = 0.99 and P = 0.73, respectively). Figure 2 Blood flow visualization: pathline visualization of the four flow components (direct flow, retained inflow, delayed ejection flow, and residual volume). (A–C) a healthy 50-year-old woman with normal LV diastolic function; (D–F) a 62-year-old male with DCM and LV relaxation abnormality; (G–I) a 61-year-old female with DCM and restrictive LV filling. Semi-transparent three-chamber images provide anatomical orientation. Ao, aorta; LA, left atrium; LV, left ventricle.

me). (A–C) a healthy 50-year-old woman with normal LV diastolic function; (D–F) a 62-year-old male with DCM and LV relaxation abnormality; (G–I) a 61-year-old female with DCM and restrictive LV filling. Semi-transparent three-chamber images provide anatomical orientation. Ao, aorta; LA, left atrium; LV, left ventricle. LV stroke volumes While the LVSVs in the DCM group and the healthy group were equal (inflow volume: 77 ± 18 mL in DCM vs. 80 ± 17 mL in healthy; ejected volume: 77 ± 19 mL in DCM vs. 79 ± 16 mL in the healthy, P = 0.73 and P = 0.81, respectively) the direct flow was significantly smaller in DCM compared with healthy (35 ± 11 vs. 56 ± 14 mL, P = 0.002), resulting in a smaller direct flow to total inflow ratio in DCM compared with healthy subjects (46 ± 9 vs. 70 ± 6%, P = 0.000). The KE was calculated over time for all components (Figure 4). At ED there was no significant difference between the KE/mL of the direct flow in DCM patients compared with healthy subjects. The KE/mL for the other three components was significantly larger in the DCM group than in normals (Figure 5). Figure 3 Blood flow components: blood flow components presented as a percentage of end-diastolic volume (mean ± SD). Top panel: DCM patients (n = 10); lower panel: healthy subjects (n = 10). *Components with a P-value <0.0125 compared with the corresponding component in the healthy group. Intra-group comparison: †P-value ≤0.0002 among DCM patients vs. residual volume. ‡P-value ≤0.005 vs. direct flow in healthy subjects. §P-value =0.000 vs. residual volume in healthy subjects.

el: healthy subjects (n = 10). *Components with a P-value <0.0125 compared with the corresponding component in the healthy group. Intra-group comparison: †P-value ≤0.0002 among DCM patients vs. residual volume. ‡P-value ≤0.005 vs. direct flow in healthy subjects. §P-value =0.000 vs. residual volume in healthy subjects. Figure 4 Kinetic energy over diastole: for each individual in the study kinetic energy per mL of blood (J/mL) is shown over the diastolic interval, normalized by the length of diastole, left panels: DCM patients; and right panels: healthy subjects. Direct flow (A and B), retained inflow (C and D), delayed ejection flow (E and F), and residual volume (G and H). The direct flow volume was divided into early and late inflow in order to study the impact of atrial contraction.7 The ratio of early diastolic direct flow/total inflow volume was equal in the groups, whereas the late diastolic direct flow volume/total inflow volume ratio was smaller in the DCM group compared with healthy (P < 0.001) (Figure 6A). The direct flow KE/mL at ED was smaller for the early diastolic sub-volume in healthy subjects (P = 0.003), whereas there was no difference between the diastolic phases in DCM (NS) (Figure 6B). Figure 5 Pre-systolic kinetic energy: the kinetic energy (KE) at end-diastole, in mJ/mL, for each flow component in DCM patients (left) and healthy subjects (right). Bars show group mean and standard deviation. *P-value <0.0125 vs. the corresponding component in healthy subjects. Intra-group comparison: †P-value ≤0.0001 vs. direct flow. ‡P-value ≤0.0001 vs. residual volume.

(KE) at end-diastole, in mJ/mL, for each flow component in DCM patients (left) and healthy subjects (right). Bars show group mean and standard deviation. *P-value <0.0125 vs. the corresponding component in healthy subjects. Intra-group comparison: †P-value ≤0.0001 vs. direct flow. ‡P-value ≤0.0001 vs. residual volume. LV residual volume Although the EDV was significantly larger in the DCM group (P = 0.021) (Table 1), the amount of the residual volume relative to the EDV did not differ between the DCM group and the healthy subjects (NS) (Figure 3). The residual volume possessed a significantly larger KE/mL at ED in DCM patients compared with healthy subjects (P = 0.002, Figure 5). Figure 6 Early vs. late inflow: characteristics of direct flow during early (E) and late (A) diastole. At end-diastole (A) the direct flow volume as a percentage of the total inflow for E and A. The sum of the values at E and A, for each group, gives the direct flow to total inflow ratio for the entire diastolic filling phase (DCM: 46 ± 9% vs. healthy: 70 ± 6%, P = 0.000). (B) The kinetic energy at E and A. Data are presented as mean and standard deviation. (A) *P-value =0.000 for direct flow A in DCM vs. direct flow A in healthy subjects. (B) †P-value =0.003 for direct flow E in healthy subjects vs. direct flow A in healthy subjects.

phase (DCM: 46 ± 9% vs. healthy: 70 ± 6%, P = 0.000). (B) The kinetic energy at E and A. Data are presented as mean and standard deviation. (A) *P-value =0.000 for direct flow A in DCM vs. direct flow A in healthy subjects. (B) †P-value =0.003 for direct flow E in healthy subjects vs. direct flow A in healthy subjects. Discussion It is straightforward to assume that ventricular remodelling and dysfunction in patients with HF would be associated with alterations in flows through the heart chambers. In this study, we apply a previously validated 4D flow method9 in patients with compensated HF and an equal number of healthy subjects. This novel method allows the separation of the total LV volume into functionally distinct components. Comparison of the volume, timing, and KE of those components in these two states clarifies those differences, and yields some surprises.

method9 in patients with compensated HF and an equal number of healthy subjects. This novel method allows the separation of the total LV volume into functionally distinct components. Comparison of the volume, timing, and KE of those components in these two states clarifies those differences, and yields some surprises. The present study shows that, despite mild LV remodelling associated with DCM, the SV in these compensated HF patients is similar to normals. The composition of the SV is different, however: a smaller proportion of the flow transits the LV in a single cardiac cycle (direct flow). Although the volume of the direct flow is lower in these HF patients, on a per mL basis its KE at the time of ED does not differ from normal subjects. The percentage of the total EDV that consists of a re-circulating residual volume was also similar between DCM patients and normals, but in contrast, on a per mL basis the residual volume had a greater end-diastolic KE compared with normals. Finally, the late diastolic inflow in normal LVs had a larger proportion of the direct flow compared with DCM. These observed differences may be relevant to the function and remodelling of the LV. The normal interaction between flowing blood and the developing heart in utero stimulates a continuous positive remodelling process that creates an optimal geometry for efficient flow.3,4,18,19 The same responses to flow-induced forces may also play a role in the pathophysiology of HF.2,18,20

and remodelling of the LV. The normal interaction between flowing blood and the developing heart in utero stimulates a continuous positive remodelling process that creates an optimal geometry for efficient flow.3,4,18,19 The same responses to flow-induced forces may also play a role in the pathophysiology of HF.2,18,20 The KE of inflowing blood has several possible consequences in the receiving ventricle. The KE of the entering blood may be (i) translated into motion of the blood already residing in the LV, (ii) converted into potential energy stored in the myocardium, (iii) dissipated as heat, or (iv) decelerated with the elevation of LV diastolic pressure. Less deceleration of flow in the LV may be an important component of normal diastolic function. A computational fluid dynamics study of the vortical flow in the right ventricle during filling, for example, suggested that the early diastolic vortices' KE prevented inflow-impeding pressure rise within the RV.21 Higher KE of intra-ventricular blood resulting from altered LV myocardial properties, such as impaired active relaxation or increased stiffness, will likely lead to elevated filling pressures as elevated velocities decelerate.

hat the early diastolic vortices' KE prevented inflow-impeding pressure rise within the RV.21 Higher KE of intra-ventricular blood resulting from altered LV myocardial properties, such as impaired active relaxation or increased stiffness, will likely lead to elevated filling pressures as elevated velocities decelerate. The differences in the volumes and diastolic energetics of the four functional flow components observed between myopathic and normal LVs can be considered in the light of the diastolic impact of the KE. The volume of the direct flow component is significantly less in DCM than normals, so that while the end-diastolic KE/mL is equivalent in both groups, the total direct flow KE at ED is smaller in the myopathic LVs. As this is volume destined for ejection during the subsequent systole, a lower KE at ED may indicate that incremental contribution from systolic contraction will be required for its ejection. In contrast, the retained inflow has a larger KE/mL at ED in DCM compared with normal LVs. This suggests a shift in the specific effects of inflow KE on the myopathic ventricle, with increased interaction between the inflowing blood and the resident volume (delayed ejection flow and residual volume) and surrounding myocardium. Separation of inflow into early and late diastolic phases demonstrates that the end-diastolic KE/mL of the direct flow volume was smaller for the early diastolic portion compared with the late diastolic portion in healthy subjects, whereas there was no difference between the diastolic phases in DCM. This may reflect impaired active relaxation of the myocardium in myopathic LVs. Late diastolic differences were also observed between the two groups: the proportion of the direct flow following atrial contraction was smaller in DCM compared with normal LVs. This shift might correlate with impaired diastolic–systolic coupling since the direct flow contributed by atrial contraction in normal hearts preserves inflow KE particularly well. The loss of organized atrial contraction, which is clinically related to increased symptoms in HF patients, may eliminate this advantageous KE preservation. The degree to which late diastolic inflow augments LV outflow might be a parameter that could influence target heart rates and optimal pacing strategies also in patients with HF.

anized atrial contraction, which is clinically related to increased symptoms in HF patients, may eliminate this advantageous KE preservation. The degree to which late diastolic inflow augments LV outflow might be a parameter that could influence target heart rates and optimal pacing strategies also in patients with HF. The residual volume outlines the periphery of the LV chamber and appears to provide a fluid–fluid interface that impacts the flow paths of the retained inflow and delayed ejection flow. Although there was no significant difference in the relative amount of the non-exchanging residual volume between the groups, the exchanging components (retained inflow and delayed ejection flow) were relatively larger in the DCM patients. This is in agreement with preliminary data suggesting that decompensated HF patients have a larger residual volume than compensated patients.12 In the current study, the KE/mL of the residual volume at ED was higher in patients compared with healthy subjects, which may reflect the impact of stiffer and less compliant myocardium. Moreover, converting motion of inflowing blood to motion of the residual volume may offer advantages in terms of less diastolic pressure rise, as well as in avoidance of thrombosis. A large amount of the residual volume with low velocities may create conditions promoting intra-ventricular thrombus formation. This is supported by an echo-Doppler study that demonstrated lower apical blood velocities in DCM patients with LV thrombus compared with DCM patients without thrombus.22

idance of thrombosis. A large amount of the residual volume with low velocities may create conditions promoting intra-ventricular thrombus formation. This is supported by an echo-Doppler study that demonstrated lower apical blood velocities in DCM patients with LV thrombus compared with DCM patients without thrombus.22 Study limitations Meticulous quality control of the post-processed data was fundamental in this study and evidence of aberrant pathlines indicating data imperfections was carefully sought.9 Accuracy of pathline computation is dependent on a large number of factors, such as temporal and spatial resolution, VENC, field strength, MRI hardware, and intra-voxel dephasing due to turbulence. Errors in pathlines computation will propagate from one time frame to the next and accumulate over time. Eddy currents and concomitant gradient fields may also hamper the accuracy of 4D PC-MRI flow data. The potential effects of these factors were anticipated and minimized by the acquisition design and by tailored post-processing. The data acquisition was optimized for the blood flow through the left ventricle, with VENC targeted to the range of diastolic inflow velocities. As systolic velocities downstream from the aortic valve can exceed the upper limit of the velocity scale used, aliasing artefacts may occur there and data from that region were not studied.

ition was optimized for the blood flow through the left ventricle, with VENC targeted to the range of diastolic inflow velocities. As systolic velocities downstream from the aortic valve can exceed the upper limit of the velocity scale used, aliasing artefacts may occur there and data from that region were not studied. The scan times in this study were relatively long. However, new MRI sequences23 as well as improvements in MRI hardware continue to decrease scan times. The morphologic images and the 3DcinePC velocity data were acquired in two separate acquisitions; hence, there is a risk of mismatch due to patient movement. Mismatch would be detectable as a large difference in inflow and outflow volumes, and visible offsets between the morphologic images and the contrast poor magnitude data taken from the flow data acquisition. This problem was addressed by acquiring morphological images both before and after the 3DcinePC velocity data. In the future, registration algorithms might improve matching of the 3DcinePC velocity data to the underlying morphological data and further reduction in the potential for this type of error.

data acquisition. This problem was addressed by acquiring morphological images both before and after the 3DcinePC velocity data. In the future, registration algorithms might improve matching of the 3DcinePC velocity data to the underlying morphological data and further reduction in the potential for this type of error. The present findings relate only to a relatively small number of patients and healthy subjects in the supine position at rest and in sinus rhythm with no significant valvular disorder. It is reasonable to assume that different flow patterns and energetics would be seen with altered rhythm, valve disease, and at stress with higher heart rates and different loading conditions. The residual volume is likely to be the most inaccurately assessed component, due to partial volume effects at the chamber boundaries and imperfect segmentation, which may inadvertently include some myocardium. This fact may partly explain the relatively low flow-based LV ejection fraction values observed in the healthy subjects. The same methodology was used in both groups, however.

component, due to partial volume effects at the chamber boundaries and imperfect segmentation, which may inadvertently include some myocardium. This fact may partly explain the relatively low flow-based LV ejection fraction values observed in the healthy subjects. The same methodology was used in both groups, however. Conclusion Although total LVSV is equal in healthy subjects and HF patients with mild LV remodelling, the SV's transventricular flow paths to ejection and diastolic energetics are significantly different in the two states. Late filling facilitates a direct transit of blood through the normal LV, but this aspect of diastolic–systolic coupling is less pronounced in myopathic LVs. These multi-dimensional flow-based abnormalities are detectable despite clinical compensation, and may prove useful as subclinical markers of impaired LV function. Supplementary data Supplementary data are available at European Journal of Echocardiography online. Funding This study was funded by The Swedish Research Council, The Swedish Heart and Lung Foundation, and The Emil and Wera Cornell foundation. Conflict of interest: none declared. Supplementary Material Supplementary Data Acknowledgements The authors appreciate the assistance provided by Jan Engvall and Johan Kihlberg.

A 67-year-old woman initially presented with dyspnoea. Echocardiography demonstrated an inter-atrial mass. Cardiovascular magnetic resonance (CMR) showed a well-marginated mass (35 mm × 45 mm) in the inter-atrial septum (IAS) sparing the fossa ovalis, extending to the posterior right atrial wall. The mass had high signal intensity on T1-weighted images (Panels A and B, arrows) consistent with fat, which attenuated on fat-suppression imaging (Panels C and D, arrows). Native T1-maps (Panels E and F, arrows) showed homogenous and significantly lower T1 values (230–350 ms) compared with the normal myocardium (green) and blood (yellow/red), but similar to subcutaneous and epicardial fat (blue). The mass had no late gadolinium enhancement and was stable in the size compared with 1-year prior. Findings were consistent with lipomatous hypertrophy of the IAS.

ntly lower T1 values (230–350 ms) compared with the normal myocardium (green) and blood (yellow/red), but similar to subcutaneous and epicardial fat (blue). The mass had no late gadolinium enhancement and was stable in the size compared with 1-year prior. Findings were consistent with lipomatous hypertrophy of the IAS. This is a first demonstration of using native T1-mapping for immediate visual and quantitative tissue characterization of an intra-cardiac mass to help confirm its fat content. T1-maps provide pixel-wise estimates of the T1 relaxation times of the underlying tissue, and each tissue type exhibits a characteristic normal range of T1 values. This allows the visual differentiation of tissues using colour scales as shown. Lipomatous hypertrophy of the IAS is an unencapsulated proliferation of adipose tissue ≥20 mm within the atrial septum that is in continuity with epicardial fat. It is associated with older age, increased fat elsewhere in the body, atrial arrhythmias and in some cases, obstruction of the superior vena cava. The correct differentiation of this benign cardiac mass from other atrial tumours using multi-parametric tissue characterization techniques on CMR can prevent unnecessary invasive interventions. Funding Funding to pay the Open Access publication charges for this article was provided by the Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford.

Echocardiography plays an essential role in guiding complex cardiovascular interventions: atrial septal defect (ASD) closure, left atrial appendage (LAA) occlusion, and transcatheter aortic valve implantation (TAVI). Most procedures are guided by three-dimensional (3D) transoesophageal echocardiography. Intracardiac echocardiography (ICE) has the advantage of not requiring general sedation, does not interfere with fluoroscopy, and provides simultaneous guidance of the entire procedure. However, until now, ICE could only provide 2D images. We report the first worldwide clinical use of 3D intracardiac echocardiography. Images were acquired using an Acuson SC 2000™ ultrasound unit (Siemens Medical Solutions, Erlangen, Germany) and a 10 F AcuNav V™, able to acquire a 60° × 15° volume, at a volume rate of 20 vps. Images in Panels A–C and E–F were acquired with the catheter in the right atrium. Panel A shows the right and left atrium (RA, LA) and an ASD (see Supplementary data online, video S1). This view was used in this patient to guide the closure of the ASD: Panels B and C show 3D ICE images during the procedure. Panels E and F show a long- and short-axis view of the aortic valve (Ao), which can be used to guide TAVI. The image in Panel D was acquired with the catheter advanced into the left atrium, showing the LAA opening (see Supplementary data online, video S2), which can be used to guide LAA occlusion procedures. The circumflex artery (Cx) can also be viewed. Finally, Panel G shows the pulmonary valve (PV), right ventricle outflow tract (RVOT) and pulmonary artery (PA).

The image in Panel D was acquired with the catheter advanced into the left atrium, showing the LAA opening (see Supplementary data online, video S2), which can be used to guide LAA occlusion procedures. The circumflex artery (Cx) can also be viewed. Finally, Panel G shows the pulmonary valve (PV), right ventricle outflow tract (RVOT) and pulmonary artery (PA). Supplementary data are available at European Heart Journal – Cardiovascular Imaging online. Supplementary Material Supplementary Data

Introduction The use of optical coherence tomography (OCT) in clinical practice has revealed that various types of abnormal vessel reactions associated with stent implantation (stent malapposition, thrombus, prolapse, and edge dissection) develop more frequently than expected.1,2 Given the high incidence of these abnormalities after treatment with the currently available metallic stents and the possible link between such phenomena and future clinical events, elucidation of the natural course of these findings is important regarding the safety and cost-effectiveness of percutaneous coronary intervention (PCI), because such information might help avoid unnecessary additional procedures frequently performed after stent implantation to achieve stent optimization. In this context, several investigators including our group have reported that high-resolution OCT allows for visualization of microstructural abnormal findings before and after PCI when compared with the current gold standard, intravascular ultrasound (IVUS).2 In the present study, we performed a serial OCT examination to clarify the natural course of minor abnormalities, such as stent malapposition, thrombus, prolapse, and edge dissection.

f microstructural abnormal findings before and after PCI when compared with the current gold standard, intravascular ultrasound (IVUS).2 In the present study, we performed a serial OCT examination to clarify the natural course of minor abnormalities, such as stent malapposition, thrombus, prolapse, and edge dissection. Methods Study population and methods Study subjects were collected from among patients who provided written informed consent for OCT-guided PCI and follow-up OCT examination. We retrospectively selected patients implanted with sirolimus-eluting stents (SES: CypherTM, Cordis Corp., Miami Lakes, FL, USA) or paclitaxel-eluting stents (PES: Taxus LiberteTM, Boston Scientific Corp., Natick, MA, USA) having excellent OCT images immediately after stent implantation and at the 8-month follow-up. A total of 35 stable or unstable angina patients treated with SES or PES were enrolled into this study from December 2007 to September 2010. Inclusion criteria for this study were as follows: (i) native coronary arteries with diameter stenosis ≥75% and (ii) reference vessel diameter between 2.5 and 3.5 mm based on visual estimation. Exclusion criteria were as follows: (i) acute myocardial infarction, (ii) apparent congestive heart failure, (iii) contraindication to dual antiplatelet therapy, and (iv) lesions unsuitable for OCT (severe tortuous lesions and ostial lesions). All the patients were taking aspirin (100 mg/day). Ticlopidine (200 mg/day), or clopidogrel (75 mg/day) was also given for at least 8 months after stent implantation. The study was approved by the Ethics Committee of Kobe University and all enrolled study patients provided their written informed consent.

tial lesions). All the patients were taking aspirin (100 mg/day). Ticlopidine (200 mg/day), or clopidogrel (75 mg/day) was also given for at least 8 months after stent implantation. The study was approved by the Ethics Committee of Kobe University and all enrolled study patients provided their written informed consent. OCT examination In this study, as frequency-domain OCT had not been approved for clinical use in Japan during the study period, time-domain OCT with coronary artery occlusion was used as previously reported.3 OCT analysis All images were analysed by an independent observer who was blinded to the clinical presentations and lesion characteristics. Cross-sectional OCT images were analysed at every frame. Bifurcation lesions with major side branches were excluded from the analysis. To obtain identical analysis of segments at baseline and at follow-up OCT, we displayed the baseline and follow-up OCT images side-by-side and performed serial OCT analysis using information about the motorized pullback speed and landmarks such as the presence of calcium deposits, side branches, and plaque shape.

ysis. To obtain identical analysis of segments at baseline and at follow-up OCT, we displayed the baseline and follow-up OCT images side-by-side and performed serial OCT analysis using information about the motorized pullback speed and landmarks such as the presence of calcium deposits, side branches, and plaque shape. For qualitative analysis, we evaluated the presence of stent malapposition, thrombus, tissue prolapse, and stent edge dissection. Stent malapposition was defined as the distance between the centre reflection of the strut and the vessel wall greater than the actual stent thickness plus the OCT resolution limit. This resulted in stent malapposition thresholds of ≥170 µm for SES and ≥140 µm for PES.3 The malapposition length was defined as the number of consecutive frames with malapposed struts. We evaluated the maximum malapposition length for every stent. Malapposition identified after stent implantation but absent at follow-up was defined as resolved; otherwise, it was persistent. If both resolved and persistent malapposition were existed within the same stent on follow-up OCT images, it was defined as partially resolved stents. Late-acquired malapposition was defined as that not present immediately after the procedure but observed at the follow-up.4

ed as resolved; otherwise, it was persistent. If both resolved and persistent malapposition were existed within the same stent on follow-up OCT images, it was defined as partially resolved stents. Late-acquired malapposition was defined as that not present immediately after the procedure but observed at the follow-up.4 Intracoronary thrombus was defined as a mass protruding beyond the stent strut into the lumen with significant attenuation behind the mass.5,6 Resolved thrombus was defined as that observed after stent implantation, but absent at the follow-up. Persistent thrombus was defined as that still present during the follow-up period. Late-acquired thrombus was defined as that not present after the procedure, but newly appearing at the follow-up. Tissue prolapse was defined as a protrusion of tissue between the stent struts, extending inside a circular arc connecting adjacent struts in both OCT images.7 Stent edge dissection was defined as disruption of the luminal vessel surface in the edge segments.7 Representative serial OCT images of these figures are shown in Figure 1. Figure 1 Representative OCT images: (A) classification of malapposed struts: (a) resolved malapposition, (b) persistent malapposition, (c) late-acquired malapposition, (B) classification of thrombus: (a) resolved thrombus, (b) persistent thrombus, (c) late-acquired thrombus, (C) tissue prolapse, (D) edge dissection.

1. Figure 1 Representative OCT images: (A) classification of malapposed struts: (a) resolved malapposition, (b) persistent malapposition, (c) late-acquired malapposition, (B) classification of thrombus: (a) resolved thrombus, (b) persistent thrombus, (c) late-acquired thrombus, (C) tissue prolapse, (D) edge dissection. For quantitative analysis, the neointimal thickness inside each stent strut was measured for all malapposed struts. Also, we calculated the incidence of frames with malapposed struts, thrombus, and tissue prolapse on baseline and follow-up OCT images (number of frames with malapposed struts, thrombus, prolapse/total number of frames) to evaluate the changes in the longitudinal length of these abnormalities during the follow-up. Clinical follow-up The incidence of death, myocardial infarction, target lesion revascularization (TLR), and stent thrombosis was evaluated 8 months after the index stent procedure. TLR was defined as any reintervention (surgical or percutaneous) to treat restenosis of the analysed segment.

For quantitative analysis, the neointimal thickness inside each stent strut was measured for all malapposed struts. Also, we calculated the incidence of frames with malapposed struts, thrombus, and tissue prolapse on baseline and follow-up OCT images (number of frames with malapposed struts, thrombus, prolapse/total number of frames) to evaluate the changes in the longitudinal length of these abnormalities during the follow-up. Clinical follow-up The incidence of death, myocardial infarction, target lesion revascularization (TLR), and stent thrombosis was evaluated 8 months after the index stent procedure. TLR was defined as any reintervention (surgical or percutaneous) to treat restenosis of the analysed segment. Statistical analysis All statistical analyses were performed using the StatView 5.0 software (SAS Institute, Inc., Cary, NC, USA). Qualitative data are expressed as frequencies, and quantitative data are shown as mean values ± SD. Continuous variables were compared using an unpaired or paired Student's t-test or Mann–Whitney U test. Differences in categorical variables were assessed using the χ2 test and Fisher's exact tests. We performed a ROC analysis of the S–V distance to identify the optimal cut-off value for predicting the natural course of malapposed struts during the follow-up (resolved or persistent) using Medcalc (MedCalc Software, Mariakerke, Belgium). For all comparisons, a P-value <0.05 was considered statistically significant.

We performed a ROC analysis of the S–V distance to identify the optimal cut-off value for predicting the natural course of malapposed struts during the follow-up (resolved or persistent) using Medcalc (MedCalc Software, Mariakerke, Belgium). For all comparisons, a P-value <0.05 was considered statistically significant. Results Baseline patient demographic, lesion, and procedural characteristics are shown in Tables 1 and 2. In all stents, angiographic optimization was accomplished after PCI. Table 1 Baseline patient characteristics Variable 35 patients (40 stents) Age 65.3 ± 10.3 Men (%) 28 (80) Hypertension (%) 24 (68) Dyslipidaemia (%) 26 (74) Cigarette smoking (%) 17 (49) Diabetes (%) 19 (54) Renal dysfunction (%) 7 (20) Prior myocardial infarction (%) 3 (9) Prior PCI (%) 8 (23) Prior CABG (%) 0 (0) Clinical presentation (%) Stable angina pectoris 33 (94) Unstable angina pectoris 2 (6) Statin 24 (69) ACE-I, ARB 22 (63) Values are presented as number of patients (%). PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting, ACE, angiotensin-converting enzyme, ARB, angiotensin II receptor blocker. Table 2 Procedural characteristics

Variable 35 patients (40 stents) Age 65.3 ± 10.3 Men (%) 28 (80) Hypertension (%) 24 (68) Dyslipidaemia (%) 26 (74) Cigarette smoking (%) 17 (49) Diabetes (%) 19 (54) Renal dysfunction (%) 7 (20) Prior myocardial infarction (%) 3 (9) Prior PCI (%) 8 (23) Prior CABG (%) 0 (0) Clinical presentation (%) Stable angina pectoris 33 (94) Unstable angina pectoris 2 (6) Statin 24 (69) ACE-I, ARB 22 (63) Values are presented as number of patients (%). PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting, ACE, angiotensin-converting enzyme, ARB, angiotensin II receptor blocker. Table 2 Procedural characteristics Variable 35 patients (40 stents) Location, proximal/mid/distal 8/27/5 ACC/AHA classification, type A/type B/type C 3/31/6 Stent type Cypher™ 19 Taxus Liberte™ 21 Average stent diameter (mm) 3.0 ± 0.4 Average stent length (mm) 21.9 ± 7.0 Multiple overlapping stents 2 Multivessel stenting 3 Stent implantation pressure (atm) 10.8 ± 2.1 Post-dilatation 23 Largest balloon size for dilatation (mm) 3.2 ± 0.4 Maximum inflation pressure (atm) 13.8 ± 3.2 Angiographically detected dissection 0 Follow-up duration (month) 8.1 ± 1.5 Values are presented as mean ± SD or number of stents. ACC, American College of Cardiology; AHA, American Heart Association.

Variable 35 patients (40 stents) Location, proximal/mid/distal 8/27/5 ACC/AHA classification, type A/type B/type C 3/31/6 Stent type Cypher™ 19 Taxus Liberte™ 21 Average stent diameter (mm) 3.0 ± 0.4 Average stent length (mm) 21.9 ± 7.0 Multiple overlapping stents 2 Multivessel stenting 3 Stent implantation pressure (atm) 10.8 ± 2.1 Post-dilatation 23 Largest balloon size for dilatation (mm) 3.2 ± 0.4 Maximum inflation pressure (atm) 13.8 ± 3.2 Angiographically detected dissection 0 Follow-up duration (month) 8.1 ± 1.5 Values are presented as mean ± SD or number of stents. ACC, American College of Cardiology; AHA, American Heart Association. Incidence of persistent, resolved, and late-acquired stent malapposition Stent malapposition was observed in 65.0% of enrolled stents (26/40 stents) immediately after stenting, and in 32.5% (13/40 stents) at the follow-up (Figure 2A). Among a total of 26 stents with post-procedural malapposition, serial OCT analysis revealed that 6 stents (6/26: 23.1%) were partially resolved with at least one persistent malapposed strut at the follow-up, and 20 stents (20/26: 76.9%) were completely resolved without any malapposed struts observed on the follow-up OCT images (Figure 2A). Late-acquired malapposition was observed in seven stents (7/40; 17.5%). According to strut-based analysis, 431 struts among a total of 73 929 struts (0.58%) were malapposed on post-procedural OCT images. Among these, serial OCT analysis revealed persistent malapposition of 49 struts, but the other 382 struts had resolved during the follow-up (resolved malapposition), and there were 108 newly appearing malapposed struts at the follow-up (late-acquired malapposition). Figure 2 Number of abnormal findings after stenting: (A) stent malapposition, (B) thrombus, (C) tissue prolapse, (D) edge dissection.

but the other 382 struts had resolved during the follow-up (resolved malapposition), and there were 108 newly appearing malapposed struts at the follow-up (late-acquired malapposition). Figure 2 Number of abnormal findings after stenting: (A) stent malapposition, (B) thrombus, (C) tissue prolapse, (D) edge dissection. In most of the stents with malapposed struts after PCI, the maximum malapposition length was equal to or less than five cross-sections (Table 3). The incidences of cross-sections with malapposed struts decreased from post-procedure to the follow-up (Table 4). Table 3 Malapposition length analysis Maximum malapposition length After PCI At follow-up ≤5 cross-sections 19 stents 9 stents 6 ≤ 15 cross-sections 5 stents 3 stents ≥16 cross-sections 2 stents 1 stent Table 4 Cross-sectional OCT analysis Variable Immediately after PCI Follow-up P-value Stent malapposition 231 cross-sections 2.04% 117 cross-sections 0.97% 0.054 Thrombus 488 cross-sections 5.10% 108 cross-sections 0.89% 0.021 Tissue prolapse 3141 cross-sections 28.17% 0 cross-section 0.00% <0.01

Maximum malapposition length After PCI At follow-up ≤5 cross-sections 19 stents 9 stents 6 ≤ 15 cross-sections 5 stents 3 stents ≥16 cross-sections 2 stents 1 stent Table 4 Cross-sectional OCT analysis Variable Immediately after PCI Follow-up P-value Stent malapposition 231 cross-sections 2.04% 117 cross-sections 0.97% 0.054 Thrombus 488 cross-sections 5.10% 108 cross-sections 0.89% 0.021 Tissue prolapse 3141 cross-sections 28.17% 0 cross-section 0.00% <0.01 Relation between neointimal coverage and natural course of malapposed struts The extent of the strut coverage among the different types of malapposition differed significantly (Figure 3). The incidence of uncovered struts was significantly higher among late-acquired and persistent malapposed struts compared with resolved malapposed struts. Additionally, the mean neointimal thickness of resolved malapposed struts was significantly greater than that of the persistent and late-acquired malapposed struts. Figure 3 The frequency of neointimal coverage and the measurement of neointimal thickness between resolved malapposed struts, persistent malapposed struts, and late-acquired malapposed struts: (A) the incidence of struts without neointima was significantly higher in late-acquired and persistent malapposed struts compared with resolved malapposed struts. (B) The mean neointimal thickness of resolved malapposed struts was significantly thicker than that of the persistent and late-acquired malapposed struts.

s: (A) the incidence of struts without neointima was significantly higher in late-acquired and persistent malapposed struts compared with resolved malapposed struts. (B) The mean neointimal thickness of resolved malapposed struts was significantly thicker than that of the persistent and late-acquired malapposed struts. Relation between S–V distance and natural course of malapposed struts The mean S–V distance for all malapposed struts on post-procedural OCT images was 224 ± 71 μm. The mean post-procedural S–V distance of persistent malapposed struts was significantly greater than that of resolved malapposed struts (342 ± 99 μm vs. 210 ± 49 μm; P <0.01). Based on the ROC analysis, an S–V distance ≤260 µm was the corresponding cut-off point for a resolved malapposed strut with a maximum sensitivity of 89.3% and a specificity of 83.7% [area under the curve (AUC) = 0.884]. Indeed, most malapposed struts with an S–V distance ≤260 μm (341/349 struts; 97.7%) were observed to be well-apposed on follow-up OCT images (Figure 4). Only eight struts with an S–V distance ≤260 μm (8/349 struts; 2.3%) persisted, and all were observed in the same stent with suspicious plaque regression or thrombus dissolution (Figure 5). Figure 4 ROC curve analysis and distribution of S–V distance of post-procedural malapposed struts. (A) An S–V distance ≤260 µm was the corresponding cut-off point for a resolved malapposed strut with a maximum sensitivity of 89.3% and a specificity of 83.7% (AUC = 0.884, P = 0.001). (B) Only eight struts with an S–V distance ≤260 μm persisted.

is and distribution of S–V distance of post-procedural malapposed struts. (A) An S–V distance ≤260 µm was the corresponding cut-off point for a resolved malapposed strut with a maximum sensitivity of 89.3% and a specificity of 83.7% (AUC = 0.884, P = 0.001). (B) Only eight struts with an S–V distance ≤260 μm persisted. Figure 5 A case of persistent malapposed struts with a baseline S–V distance ≤260 μm: left panel shows the OCT image immediately after Taxus Liberte™ implantation. The S–V distance is 180 μm, which is less than the cut-off value of the S–V distance of 260 μm. These malapposed struts persisted on follow-up OCT images (right panel), probably due to plaque regression or thrombus dissolution. Serial OCT analysis of thrombus, tissue prolapse, and edge dissection Thrombus was observed in 15 stents (15/40: 37.5%) post-PCI and in 8 stents (8/40: 20%) at the follow-up. Serial OCT analysis revealed persistent thrombus in 1 stent (1/15: 6.7%), resolved thrombus in 14 stents (14/15: 93.3%), and late-acquired thrombus in 8 stents (8/40: 20%) (Figure 2B). Tissue prolapse was observed in 95% cases (38/40 stents) immediately after PCI, and was not observed in any of the cases at the 8-month follow-up examination (Figure 2C). There were eight stents with edge dissections immediately after PCI, all of which were repaired at the follow-up (Figure 2D). The incidences of cross-sections with thrombus and tissue prolapse decreased from post-procedure to the follow-up (Table 4).

in any of the cases at the 8-month follow-up examination (Figure 2C). There were eight stents with edge dissections immediately after PCI, all of which were repaired at the follow-up (Figure 2D). The incidences of cross-sections with thrombus and tissue prolapse decreased from post-procedure to the follow-up (Table 4). Comparison between sirolimus-eluting stents and paclitaxel-eluting stents The percentage of malapposed struts after PCI was 0.45% (128/27843 struts) in SES, and 0.66% (303/46086 struts) in PES. On the basis of the ROC curve analysis according to the stent type, we identified an S–V distance ≤280 µm as the best cut-off point for a resolved malapposed strut with a maximum sensitivity of 95.0% and a specificity of 100% for SES (AUC: 0.991), and an S–V distance ≤260 µm for PES (sensitivity 87.8%, specificity 80.0%, AUC = 0.865; Figure 6). The AUC for SES was numerically higher than that of PES. The incidence of thrombus, edge dissection, and tissue prolapse did not differ between SES and PES. Figure 6 Comparison of ROC curve analysis and distribution of S–V distance between SES and PES: (A) SES; an S–V distance of ≤280 was the best cut-off point for a resolved malapposed strut with a maximum sensitivity of 95.0% and a specificity of 100% (AUC: 0.991). All the malapposed struts with an S–V distance ≤280 μm changed to be well-apposed. (B) PES; an S–V distance of ≤260 was the best cut-off point for a resolved malapposed strut (sensitivity 87.8%, specificity 80.0%, AUC = 0.865). Eight malapposed struts with S–V distance ≤260 μm persisted.

ecificity of 100% (AUC: 0.991). All the malapposed struts with an S–V distance ≤280 μm changed to be well-apposed. (B) PES; an S–V distance of ≤260 was the best cut-off point for a resolved malapposed strut (sensitivity 87.8%, specificity 80.0%, AUC = 0.865). Eight malapposed struts with S–V distance ≤260 μm persisted. Clinical outcome There were no deaths at the 8-month follow-up. One patient had myocardial infarction at another vessel occlusion. TLR was necessary in six patients with seven stents. Malapposition, thrombus, tissue prolapse, and edge dissection on post-stenting OCT images were not associated with the rate of TLR at the 8-month clinical follow-up. Stent thrombosis did not occur during the follow-up (Table 5). Table 5 Clinical Outcome at 8-month Variable n = 35 Death (%) 0 (0) Myocardial infarction (%) 1 (2.9) Target lesion revascularization (%) 6 (17.1) Stent thrombosis (%) 0 (0)

Clinical outcome There were no deaths at the 8-month follow-up. One patient had myocardial infarction at another vessel occlusion. TLR was necessary in six patients with seven stents. Malapposition, thrombus, tissue prolapse, and edge dissection on post-stenting OCT images were not associated with the rate of TLR at the 8-month clinical follow-up. Stent thrombosis did not occur during the follow-up (Table 5). Table 5 Clinical Outcome at 8-month Variable n = 35 Death (%) 0 (0) Myocardial infarction (%) 1 (2.9) Target lesion revascularization (%) 6 (17.1) Stent thrombosis (%) 0 (0) Discussion The findings of the present study demonstrated the following: (i) despite a relatively high incidence of post-procedural stent malapposition (65%), >70% of such malapposed stents had completely resolved during the follow-up period; (ii) resolved malapposed struts had a significantly thicker neointima together with fewer incidences of uncovered struts than persistent and late-acquired malapposed struts; (iii) an S–V distance ≤260 µm was the corresponding cut-off value for a resolved malapposed strut after first-generation DES with a sensitivity of 89.3% and a specificity of 83.7% (SES: S–V distance ≤280 µm; PES: S–V distance ≤260 µm); (iv) the incidence of late-acquired malapposition and thrombus were relatively high (late-acquired malapposition: 7/40 stents; 17.5%; thrombus 8/40 stents: 20%); (v) most cases of thrombus, tissue prolapse, and minor dissection detected immediately after PCI were resolved during the follow-up period.

distance ≤260 µm); (iv) the incidence of late-acquired malapposition and thrombus were relatively high (late-acquired malapposition: 7/40 stents; 17.5%; thrombus 8/40 stents: 20%); (v) most cases of thrombus, tissue prolapse, and minor dissection detected immediately after PCI were resolved during the follow-up period. Stent malapposition Ako et al.8 reported that the incidence of post-procedural stent malapposition was 16.3% in the IVUS analysis of SES in De Novo Coronary Lesions (SIRIUS) trial. In the IVUS analysis of the Taxus II study, the incidence of post-procedural stent malapposition was reported to be 11.5% for slow-release and 2.6% for moderate-release PES.9 In the present study, malapposed struts were detected by post-procedural OCT in >50% of enrolled stents, which was more frequent than expected based on these previous IVUS studies. Also, although still controversial, previous IVUS studies implied a possible link between late stent malapposition (persistent or late-acquired malapposition) and an increased risk for stent thrombosis.10–12 Therefore, it is important to identify the natural course of post-procedural malapposed struts to maximize the potential benefit and cost-effectiveness of performing additional procedures for stent optimization by avoiding unnecessary post-stent balloon dilation.

sition) and an increased risk for stent thrombosis.10–12 Therefore, it is important to identify the natural course of post-procedural malapposed struts to maximize the potential benefit and cost-effectiveness of performing additional procedures for stent optimization by avoiding unnecessary post-stent balloon dilation. Despite a relatively high incidence of stents with at least one malapposed struts, the frequency and extent of malapposed struts appeared to be low. The average incidence of malapposed struts per stent was <1% (0.6 ± 1.2%) and the average S–V distance was only 224 ± 71 μm. Additionally, among a total of 431 malapposed struts observed on post-procedural OCT images, >85% of such struts (382 struts) resolved spontaneously during the follow-up period. Therefore, we speculate that additional interventional treatment for stent optimization might not be necessary for stents with minor malapposed struts with an S–V distance ≤260 μm after first-generation DES.

post-procedural OCT images, >85% of such struts (382 struts) resolved spontaneously during the follow-up period. Therefore, we speculate that additional interventional treatment for stent optimization might not be necessary for stents with minor malapposed struts with an S–V distance ≤260 μm after first-generation DES. Although various factors may be associated with the healing process of malapposed struts (e.g. patient and lesion characteristics, procedural factors, type of stents used, etc), we speculate that the S–V distance at the index procedure and the extent of subsequent neointima proliferation are the most relevant factors for the healing process of malapposed struts. Indeed, the S–V distance of persistent malapposed struts was significantly greater than that of resolved malapposed struts. Additionally, the mean neointimal thickness of resolved malapposed struts was significantly greater than that of persistent and late-acquired malapposed struts, suggesting that increased neointimal proliferation might decrease the incidence of persistent and/or late-acquired malapposed struts through its proliferative healing process, enough to ‘fill in’ the space between the stents and the vessel wall. Our findings might support this speculation. Gutiérrez-Chico et al.13 suggested a different healing mechanism in which the neointima progresses over time to cover the remaining malapposed struts. In our study, neointimal coverage represented as a bridge or partial bridge pattern by Gutiérrez-Chico et al.13 may have been misclassified as thrombus formation. Also, a previous report demonstrated that stent design and drug release kinetics are important factors for stent strut coverage.14 Everolimus-eluting and zotarolimus-eluting stents were used these previous studies, and the progression pattern of neointima may differ depending on the stent type.

d as thrombus formation. Also, a previous report demonstrated that stent design and drug release kinetics are important factors for stent strut coverage.14 Everolimus-eluting and zotarolimus-eluting stents were used these previous studies, and the progression pattern of neointima may differ depending on the stent type. A recent report suggested that larger acute stent malapposition might be responsible for persistent malapposition and delayed neointimal coverage.13 Our study demonstrated the cut-off value of the S–V distance between resolved and persistent malapposition, and the high rate of incomplete neointimal coverage in cases with persistent malapposition (>80%). The size of acute stent malapposition after PCI can be estimated by various parameters. Among them, S–V distance can be measured directly and easily by OCT. We suggest that S–V distance is an easy-to-use parameter for decision-making at the time of PCI.

neointimal coverage in cases with persistent malapposition (>80%). The size of acute stent malapposition after PCI can be estimated by various parameters. Among them, S–V distance can be measured directly and easily by OCT. We suggest that S–V distance is an easy-to-use parameter for decision-making at the time of PCI. In our study, the relationship between malapposition and late stent thrombosis was not clear. Guagliumi et al.15 reported an association between late stent thrombosis and OCT evidence of an increased frequency and length of uncovered and malapposed strut. Moreover, they suggested that late-acquired stent malapposition and extensive vessel remodelling are markers of underlying vascular toxicity and inflammation, the actual likely cause of stent thrombosis. We also found a high rate of incomplete neointimal coverage on the late-acquired malapposed struts and on the persistent malapposed struts. The uncovered and malapposed struts at the mid-term follow-up might persist for several years and thus cause late stent thrombosis. Although our study excluded malapposed side-branch struts from OCT analysis, the first human in vivo study addressing the coverage of these struts was reported by Gutiérrez-Chico et al.16 They demonstrated coverage of malapposed side-branch struts is delayed with respect to well-apposed struts in drug-eluting stents.

In our study, the relationship between malapposition and late stent thrombosis was not clear. Guagliumi et al.15 reported an association between late stent thrombosis and OCT evidence of an increased frequency and length of uncovered and malapposed strut. Moreover, they suggested that late-acquired stent malapposition and extensive vessel remodelling are markers of underlying vascular toxicity and inflammation, the actual likely cause of stent thrombosis. We also found a high rate of incomplete neointimal coverage on the late-acquired malapposed struts and on the persistent malapposed struts. The uncovered and malapposed struts at the mid-term follow-up might persist for several years and thus cause late stent thrombosis. Although our study excluded malapposed side-branch struts from OCT analysis, the first human in vivo study addressing the coverage of these struts was reported by Gutiérrez-Chico et al.16 They demonstrated coverage of malapposed side-branch struts is delayed with respect to well-apposed struts in drug-eluting stents. Comparison of S–V distance between SES and PES We identified different cut-off values for predicting the natural course of malapposed struts between SES and PES in this study. Although previous reports have consistently demonstrated greater neointimal proliferation with PES than with SES,17 our results indicated a longer cut-off S–V distance for SES than for PES. We speculate that this is mainly due to the difference in strut thickness between SES and PES (Cypher™ SES: 140 µm, Taxus Liberte™ PES: 97 µm). Even in cases with the same S–V distance, a thicker SES strut might lead to a shorter distance from abluminal side of the strut to the vessel, so that post-procedural malapposition is more likely to become well apposed during the follow-up.

thickness between SES and PES (Cypher™ SES: 140 µm, Taxus Liberte™ PES: 97 µm). Even in cases with the same S–V distance, a thicker SES strut might lead to a shorter distance from abluminal side of the strut to the vessel, so that post-procedural malapposition is more likely to become well apposed during the follow-up. In addition to the different cut-off S–V distance, a different predicting accuracy of S–V distance was observed between the two stent types. SES showed a greater AUC than PES in the ROC analysis, suggesting that S–V distance is more accurately predictive of the time course of post-procedural malapposition for SES than for PES. Although speculative, we consider that this can be partially explained by non-uniform vessel healing with PES when compared with SES.17,18 In PES, even malapposed struts with a long S–V distance (e.g. >260 μm) could resolve due to unexpectedly greater neointimal proliferation. On the other hand, malapposed struts with a short S–V distance (e.g. ≤260 μm) might persist due to extremely suppressed neointimal proliferation. The relatively uniform neointimal proliferation of SES might explain the more accurate prediction of the S–V distance for the time course of post-procedural malapposed struts.

n the other hand, malapposed struts with a short S–V distance (e.g. ≤260 μm) might persist due to extremely suppressed neointimal proliferation. The relatively uniform neointimal proliferation of SES might explain the more accurate prediction of the S–V distance for the time course of post-procedural malapposed struts. Thrombus In the present study, a considerable incidence of thrombus attachment (15/40 stents: 37.5%) was observed immediately after stenting. This might be due to procedural problems, such as a longer time required to locate the stent or insufficient heparinization during PCI. Interestingly, serial OCT examination demonstrated that most cases of such thrombus, however, had disappeared at the 8-month follow-up examination (14/15 stents: 93%). On the other hand, despite dual antiplatelet therapy, late-acquired thrombus was observed in eight stents (8/40 stents: 20.0%) at the 8-month follow-up. Subclinical thrombus attachment after DES implantation has been reported in previous OCT and angioscopic reports.5,19 According to such reports, the incidence of thrombus at mid-term follow-up OCT was 20–30% after first-generation DES, which is consistent with our results. Although the clinical impact of such thrombus remains unclear, there is concern regarding a possible link between subclinical thrombus attachment and DES restenosis or stent thrombosis.

orts, the incidence of thrombus at mid-term follow-up OCT was 20–30% after first-generation DES, which is consistent with our results. Although the clinical impact of such thrombus remains unclear, there is concern regarding a possible link between subclinical thrombus attachment and DES restenosis or stent thrombosis. A recent OCT study reported a significant association between late stent malapposition and the development of OCT-detected thrombus at the follow-up.20 The present study also revealed a case in which a thrombus was present on malapposed struts at the follow-up (Figure 7). Figure 7 A representative case of thrombus with late stent malapposition. In a previous OCT study, we reported a possible association between stent eccentricity and thrombus formation after SES implantation. Additionally, the cytochrome P450 2C19*2 polymorphism is associated with subclinical OCT-detectable thrombus in patients treated with SES.21 On the basis of these previous findings, we speculate that the mechanism of thrombus formation involves multiple factors, including patient, lesion, and procedural factors. A larger study is necessary to confirm our speculation.

hism is associated with subclinical OCT-detectable thrombus in patients treated with SES.21 On the basis of these previous findings, we speculate that the mechanism of thrombus formation involves multiple factors, including patient, lesion, and procedural factors. A larger study is necessary to confirm our speculation. Tissue prolapse and stent edge dissection In a post-mortem study, compression of the coronary plaque after stent implantation with the protrusion of tissue between the struts was observed in 94% of the patients.22 This finding is in agreement with our study in which tissue prolapse between the struts was visible in the vast majority of enrolled stents (95%: 38/40 stents). In a previous IVUS study, minor plaque prolapse was not associated with late angiographic in-stent restenosis.23 In our study, OCT-detectable tissue prolapse immediately after PCI was not correlated with restenosis at the follow-up. With regard to stent edge dissection, Hong et al.24 reported that non-flow-limiting edge dissections detected by IVUS are not associated with an increase in acute or long-term clinical events. In our study, none of the cases with stent edge dissection on OCT images had restenosis and all of such dissections healed spontaneously during the follow-up. On the basis of these findings, we suggest that non-flow-limiting tissue prolapse and edge dissection might not require further intervention.

rm clinical events. In our study, none of the cases with stent edge dissection on OCT images had restenosis and all of such dissections healed spontaneously during the follow-up. On the basis of these findings, we suggest that non-flow-limiting tissue prolapse and edge dissection might not require further intervention. Clinical significance of minor abnormalities The present OCT study investigated the natural course of minor abnormalities after stenting during the 8-month follow-up periods. These findings were not associated with an adverse clinical outcome. Our study suggested that these abnormalities might be benign phenomenon. Previous reports provided this information, and our result also supported their studies.20,25

l course of minor abnormalities after stenting during the 8-month follow-up periods. These findings were not associated with an adverse clinical outcome. Our study suggested that these abnormalities might be benign phenomenon. Previous reports provided this information, and our result also supported their studies.20,25 Limitation Our study has a number of limitations. First, this is a non-randomized retrospective study based on a relatively limited sample size, raising the possibility of selection bias. Secondly, although we carefully reviewed both post-stenting and follow-up OCT images, it is sometimes difficult to completely match the lesions on the post-stenting and follow-up OCT images. Thirdly, although OCT is a high-resolution modality, it may not be able to differentiate neointima from an organized white thrombus, as both show high backscatter on the image. The smooth shape thrombus might thus be misdiagnosed as neointima. Fourthly, the cut-off value obtained by ROC curve analysis may be specific for the parameters of our study design, such as follow-up duration, stent design, and strut thickness with a limited number of lesions. Fifthly, strut-based analysis was performed at every frame interval. Stent-based analysis by our methodology equally evaluates stents with a single-malapposed strut as well as stents with large sections of malapposed struts. This methodology is different from that of previous IVUS studies, and the results of the present study might not be directly comparable. Finally, although we investigated the association between the minor abnormalities and the mid-term clinical event, this study is not sufficiently powered to detect possible association between them. A larger study with a longer-term follow-up will be warranted.

lts of the present study might not be directly comparable. Finally, although we investigated the association between the minor abnormalities and the mid-term clinical event, this study is not sufficiently powered to detect possible association between them. A larger study with a longer-term follow-up will be warranted. Conclusion Most cases of minor stent malapposition, thrombus, tissue prolapse, and edge dissection after stenting improved spontaneously during the follow-up period. Thus, these minor abnormalities observed immediately after PCI may not require additional interventional treatment. OCT has a potential to provide information for a specific recommendation regarding the optimal endpoint for stent implantation in daily clinical practice. Funding Fundings to pay the Open Access publication charges for this article was provided by Dr Junya Shite. Conflict of interest: Junya Shite is a consultant of St Jude Company and Goodman Company. Toshiro Shinke is a consultant of Goodman Company. The remaining authors report no conflicts of interests.

A 65-year-old man presented in our hospital complaining of continuous chest discomfort and massive peripheral oedema. Past medical history was negative for heart diseases. Transthoracic echocardiography (TTE) revealed a septum apparently dividing the left atrium longitudinally into two chambers connected by a small opening. An abnormal continuous flow, through the abnormal septum, crossing the whole atrium, was directed towards the right pulmonary veins (Panel A; Supplementary data online, Video S1). The transoesophageal echocardiography (TEE) confirmed the presence of a partition of the left atrium (Panel B), but no connections between pulmonary veins and the lateral chamber. Coronary arteriography (Panel C; Supplementary data online, Video S2) and coronary computed tomography (CT) (Panel D) showed a complex coronary malformation with cirsoid abnormal vessels and with a big aneurysm, merging with the left atrium, formed by the left circumflex artery (LCx) and a superior abnormal branch of the right coronary artery (RCA). A cardiac CT scan showed a thick pericardium. Equal diastolic pressures (19 mmHg) in the right and left chambers were found at catheterization indicating constriction. This was the reason for the presenting symptoms of peripheral oedema. The patient was referred for pericardiectomy to the surgeon and went, throughout the procedure, without any problem. The surgical procedure consisted in removal of the pericardium from phrenic to phrenic with a good immediate result.

A cardiac CT scan showed a thick pericardium. Equal diastolic pressures (19 mmHg) in the right and left chambers were found at catheterization indicating constriction. This was the reason for the presenting symptoms of peripheral oedema. The patient was referred for pericardiectomy to the surgeon and went, throughout the procedure, without any problem. The surgical procedure consisted in removal of the pericardium from phrenic to phrenic with a good immediate result. Panel A. TTE: an abnormal continuous flow in the left atrium coming from a lateral chamber (LC) going towards the right pulmonary veins. Panel B. TEE showing a chamber on the left part of the left atrium. A: coronary aneurysm. Panel C. Right coronary arteriography demonstrating a cirsoid coronary artery with an abnormal superior branch. Panel D. Coronary CT: coronary computed tomography; RCA: right coronary artery; AN: aneurysm formed by LCx and RCA. Supplementary data are available at European Heart Journal – Cardiovascular Imaging online. Supplementary Material Supplementary Data

Introduction Coronary vasodilators are typically used to diagnose obstructive coronary artery disease (CAD) and to risk stratify patients. Currently, three vasodilator stress agents are approved by the European Medicines Agency (EMA) and the United States Food and Drug Administration (FDA) for radionuclide myocardial perfusion imaging: regadenoson, adenosine, and dipyridamole. Use of these agents in perfusion cardiac magnetic resonance (CMR) imaging is considered an off-label indication. Vasodilator stress agents bind to adenosine receptors (A1, A2A, A2B, and A3), which are located in multiple tissue types.1 Activation of A2A results in coronary vasodilation as well as partial peripheral vasodilation; whereas, activation of A1, A2B, and A3 results in side-effects such as bronchospasm and high-grade atrioventricular (AV) block. An ideal vasodilator stress agent is one that binds preferentially to the A2A receptor to cause coronary vasodilation with minimal activation of other receptor subtypes. Regadenoson has higher selectively for A2A activation while adenosine binds non-selectively to A1, A2A, A2B, and A3. Dipyridamole decreases the degradation of adenosine and thus indirectly affects all adenosine receptors.

he A2A receptor to cause coronary vasodilation with minimal activation of other receptor subtypes. Regadenoson has higher selectively for A2A activation while adenosine binds non-selectively to A1, A2A, A2B, and A3. Dipyridamole decreases the degradation of adenosine and thus indirectly affects all adenosine receptors. Regadenoson has been shown to be safe, non-inferior to adenosine, and has fewer side-effects in nuclear imaging trials.2–4 Regadenoson is also safe in patients with stage 3–4 renal failure,5,6 end-stage liver disease,7 post-cardiac transplant,8 chronic obstructive pulmonary disease (COPD) and mild-to-moderate asthma.9,10 However, there is a paucity of data to address the safety and tolerability of regadenoson in perfusion CMR,11,12 where ECG monitoring is less reliable due to magnetohydrodynamic effects13 and resuscitation necessitates prompt removal of the patient from the scanner. Knowledge of the adverse events associated with regadenoson perfusion CMR has implications for patient safety and staff training. Thus, we sought to prospectively assess the safety and tolerability of regadenoson in perfusion CMR.

effects13 and resuscitation necessitates prompt removal of the patient from the scanner. Knowledge of the adverse events associated with regadenoson perfusion CMR has implications for patient safety and staff training. Thus, we sought to prospectively assess the safety and tolerability of regadenoson in perfusion CMR. Methods Subject recruitment Patients (age ≥18 years) with indications for vasodilator stress testing were prospectively enrolled from August 2009 to March 2012. Exclusion criteria included active wheezing, active symptoms of myocardial ischaemia or myocardial infarction (MI) within 24 h, estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m2, or contraindications for regadenoson perfusion CMR. Pregnant and lactating females who were not willing to discard their breast milk for 24 h following the CMR exam were also excluded. Twenty-five normal volunteers (defined as non-smoking subjects without chest pain within 6 months and without known risk factors for coronary disease) were recruited as a control group. The study was approved by the Institutional Review Board and was compliant with the Health Insurance Portability and Accountability Act.

enty-five normal volunteers (defined as non-smoking subjects without chest pain within 6 months and without known risk factors for coronary disease) were recruited as a control group. The study was approved by the Institutional Review Board and was compliant with the Health Insurance Portability and Accountability Act. Imaging protocol CMR imaging was performed using a 1.5 Tesla imaging system (Siemens Medical Solution, Erlangen, Germany). First-pass stress and rest perfusion images were obtained using a steady-state-free precession sequence (SSFP) (n = 706) (TR 2.5 ms, TE 1.04 ms, flip angle 50°, voxel size 3 × 3 × 8 mm, bandwidth 1085 Hz/pixel) or a gradient spoiled echo sequence (n = 22) (TR 2.17 ms, TE 1.03 ms, flip angle 12°, voxel size 3 × 3 × 8 mm, bandwidth 651 Hz/pixel). Gadolinium (Magnevist©, Gadopentetate Dimeglumine, Bayer Healthcare, Wayne, NJ, USA) 0.05 mmol/kg body weight was given at 5 mL/s for both stress and rest image acquisition. Depending on the heart rate (HR), either three or four left ventricular short-axis slices (base, mid-ventricle, and apex) were obtained. SSFP cine images were obtained during the 20-min post-stress period (TR 2.90 ms, TE 1.19 ms, flip angle 50°, voxel size 1 × 1 × 6 mm, bandwidth 930 Hz/ pixel). Late gadolinium enhancement images were acquired using a phase sensitive inversion recovery fast gradient echo sequence (TR 8.3 ms, TE 3.25 ms, TI individualized to null the myocardium, flip angle 25°, voxel size 1 × 1 × 6 mm, bandwidth 140 Hz/ pixel) (Figure 1). Figure 1 Regadenoson perfusion imaging protocol. sec (s), second (s); min (s), minute (s).

were acquired using a phase sensitive inversion recovery fast gradient echo sequence (TR 8.3 ms, TE 3.25 ms, TI individualized to null the myocardium, flip angle 25°, voxel size 1 × 1 × 6 mm, bandwidth 140 Hz/ pixel) (Figure 1). Figure 1 Regadenoson perfusion imaging protocol. sec (s), second (s); min (s), minute (s). Stress protocol and assessment of symptoms, adverse events, and heart rate response Patients were asked to abstain from caffeine intake and to refrain from taking anti-anginal medications including beta-blockers 24 h prior to the exam. Fixed-dose (0.4 mg) regadenoson (Astellas, Northbrook, IL, USA) was given as an iv bolus over 10 s. Within 5 min after acquisition of first-pass perfusion images, aminophylline 100 mg iv was given to reverse the effects of regadenoson (Figure 1). Sublingual nitroglycerine and iv metoprolol were available for severe and persistent chest pain. A 12-lead ECG was performed before and after the exam. Owing to magnetohydrodynamic effects causing ECG signal distortion, ECG tracing during examination was used only for gating purposes. Oxygen saturation, blood pressure (BP), and HR were monitored throughout the exam. Emergency medical supplies including a defibrillator were available in the immediate vicinity. One physician, one nurse, and one technologist were present during the exam.

ECG tracing during examination was used only for gating purposes. Oxygen saturation, blood pressure (BP), and HR were monitored throughout the exam. Emergency medical supplies including a defibrillator were available in the immediate vicinity. One physician, one nurse, and one technologist were present during the exam. Patients were queried about their symptoms before and after regadenoson and aminophylline administration. Stress-related adverse events including death, MI, ventricular tachycardia (VT)/ventricular fibrillation (VF), hospitalization, bronchospasm, and non-life-threatening arrhythmias were noted. Other adverse events, including nephrogenic systemic fibrosis, contrast extravasation, reaction to gadolinium, and thrombophlebitis, were also assessed. Baseline HR and BP were obtained at rest prior to stress exam in the supine position. Peak HR was defined as the highest HR during the stress perfusion scan and prior to administration of aminophylline. Peak BP was defined as the BP prior to reversal with aminophylline. Heart rate response (HRR) and blood pressure response (BPR) were calculated as previously described14 (HR response = [(HRpeak − HRbaseline)/HRbaseline] × 100; BP response = [(BPpeak − BPbaseline)/BPbaseline] × 100).

to administration of aminophylline. Peak BP was defined as the BP prior to reversal with aminophylline. Heart rate response (HRR) and blood pressure response (BPR) were calculated as previously described14 (HR response = [(HRpeak − HRbaseline)/HRbaseline] × 100; BP response = [(BPpeak − BPbaseline)/BPbaseline] × 100). Statistical analysis Continuous variables are reported as median [inter-quartile range (IQR)] and compared using the Mann–Whitney U test. Categorical data are reported as discrete values and percentages and compared using the Chi square test. Nine variables [age ≥64 years, BMI ≥30 kg/m2, diabetes (DM), left ventricular ejection fraction (LVEF) ≤40%, abnormal perfusion, eGFR 30–44.9 mL/min/1.73 cm2, eGFR 45–60 mL/min/1.73 cm2, eGFR >60 mL/min/1.73 cm2, and beta-blocker use) were chosen based on their potential association with cardiac autonomic function and HRR and evaluated using univariable logistic regression analysis. Significant predictors were then entered into a multivariable logistic regression model to predict HRR in the lowest quartile. Interactions among significant predictors were assessed and adjusted in the best-fit model. Model sensitivity and specificity were assessed via area under the curve (ROC) analysis and goodness of fit was assessed by the Hosmer–Lemeshow test. Two-tailed P-values were used for all statistical assessment and a P-value <0.05 was considered significant. Analyses were performed using MedCalc Version 12.0.1.0 (Mariakerke, Belgium).

y and specificity were assessed via area under the curve (ROC) analysis and goodness of fit was assessed by the Hosmer–Lemeshow test. Two-tailed P-values were used for all statistical assessment and a P-value <0.05 was considered significant. Analyses were performed using MedCalc Version 12.0.1.0 (Mariakerke, Belgium). Results Study population Seven hundred and eighty consecutive subjects were evaluated over a period of 2.6 years, but 27 patients were excluded because they did not receive regadenoson for various reasons (Figure 2). Thus, a total of 753 subjects [728 patients (median age 58 (IQR: 49–64, range 19–86), 44% female, 33% BMI ≥30 kg/m2, 20% DM and 25 normal volunteers (median age 21 (IQR: 20–23, range 18–48), 24% female)] were included in the final analysis. Two per cent of subjects (17 of 780) developed claustrophobia during the initial stages of the CMR exam and did not receive regadenoson nor complete the CMR exam—thereby accounting for 63% (17 of 27) of those excluded from the final analysis. Patient characteristics are summarized in Table 1. Table 1 Baseline patient characteristics

less marked than with standard-resolution (1.6 ± 0.2 vs. 3.2 ± 0.8 mm; P = 0.004). Seven high-resolution data sets (10%) were affected by k-t reconstruction artefacts at stress and/or rest due to respiratory motion, but this did not affect myocardial contrast passage and generally occurred at the end of a breath hold. Detection of 3VD pattern In patients with angiographic 3VD (n = 35), perfusion defects in all three territories were detected in 29% of patients (10 of 35) by standard-resolution and in 57% of patients (20 of 35) by high-resolution imaging (P = 0.04) (Figure 2A). Overall, there was poor agreement between the two techniques in determining the pattern of ischaemia in patients with angiographic 3VD [kappa = 0.09, 95% confidence interval (CI): −0.10–0.29] (Figure 2B). Figure 2 Distribution of ischaemia detected by perfusion CMR. (A) In patients with 3VD (n = 35), hypoperfusion in all three territories was detected in 57% using high-resolution imaging but in only 29% using standard-resolution (P = 0.04). (B) There was also poor agreement between high-resolution and standard-resolution imaging in determining the distribution of ischaemia in patients with 3VD (kappa = 0.09, 95% CI: −0.10–0.29). IT = ischaemic territories.

f subjects (17 of 780) developed claustrophobia during the initial stages of the CMR exam and did not receive regadenoson nor complete the CMR exam—thereby accounting for 63% (17 of 27) of those excluded from the final analysis. Patient characteristics are summarized in Table 1. Table 1 Baseline patient characteristics Patient group (n = 728) Age, y 58 (49–64) Female 322 (44%) BMI (kg/m2) 28 (25–31) Creatinine 0.90 (0.80–1.1) eGFR >60 mL/min/1.73 cm2 647 (89%) eGFR 45–60 mL/min/1.73 cm2 72 (10%) eGFR 30–44.9 mL/min/1.73 cm2 9 (1%) Ethnicity (%) Hispanic 82 (11) White 76 (92) Non-Hispanic 646 (89) White 365 (57) Black 124 (19) Asian 147 (23) Other 10 (2) Medications (%) ACE inhibitors 232 (32) ARB 69 (9) Aspirin 394 (54) Beta-blocker 337 (46) CCB 108 (15) Clopidogrel/prasugrel 82 (11) Statin 384 (53) CAD risk factors (%) Family history of CAD 164 (23) HTN 410 (56) Dyslipidaemia 381 (52) Smoking 201 (28) CAD equivalent (%) Diabetes 147 (20) Known CAD 175 (24) Prior MI 93 (13) Prior PCI 91 (13) CABG 46 (6) Atrial fibrillation 8 (1) CVA/TIA 24 (3) COPD/asthma 10 (1) MRI LV function and morphology n = 684a LV EF, % 63 (57–68) LV ESVI, mL/m2 28 (23–36) LV EDVI, mL/m2 77 (67–88) LV mass index, g/m2 49 (42–57) LV SVI, mL/m2 47 (42–52) *Continuous variables are reported as median (IQR) and compared using the Mann–Whitney U test. Categorical variables are reported as absolute values and percentages and compared using the Chi square test. aForty-four studies had real-time cine images and thus volumetric measurements were not calculated.

Patient group (n = 728) Age, y 58 (49–64) Female 322 (44%) BMI (kg/m2) 28 (25–31) Creatinine 0.90 (0.80–1.1) eGFR >60 mL/min/1.73 cm2 647 (89%) eGFR 45–60 mL/min/1.73 cm2 72 (10%) eGFR 30–44.9 mL/min/1.73 cm2 9 (1%) Ethnicity (%) Hispanic 82 (11) White 76 (92) Non-Hispanic 646 (89) White 365 (57) Black 124 (19) Asian 147 (23) Other 10 (2) Medications (%) ACE inhibitors 232 (32) ARB 69 (9) Aspirin 394 (54) Beta-blocker 337 (46) CCB 108 (15) Clopidogrel/prasugrel 82 (11) Statin 384 (53) CAD risk factors (%) Family history of CAD 164 (23) HTN 410 (56) Dyslipidaemia 381 (52) Smoking 201 (28) CAD equivalent (%) Diabetes 147 (20) Known CAD 175 (24) Prior MI 93 (13) Prior PCI 91 (13) CABG 46 (6) Atrial fibrillation 8 (1) CVA/TIA 24 (3) COPD/asthma 10 (1) MRI LV function and morphology n = 684a LV EF, % 63 (57–68) LV ESVI, mL/m2 28 (23–36) LV EDVI, mL/m2 77 (67–88) LV mass index, g/m2 49 (42–57) LV SVI, mL/m2 47 (42–52) *Continuous variables are reported as median (IQR) and compared using the Mann–Whitney U test. Categorical variables are reported as absolute values and percentages and compared using the Chi square test. aForty-four studies had real-time cine images and thus volumetric measurements were not calculated. ACE, angiotensin; ARB, angiotensin receptor blocker; ASA, aspirin; BMI, body mass index; CAD, coronary artery disease; CABG, coronary artery bypass graft; CCB, calcium channel blockers; COPD, chronic obstructive pulmonary disease; CVA, cerebral vascular accident; EF, ejection fraction; eGFR, glomerular filtration index; HTN, hypertension; LV ESVI, left ventricular end-systolic volume index; LV EDVI, left ventricular end-diastolic volume index; LV SVI, left ventricular stroke volume index; MI, myocardial infarction; PCI, percutaneous intervention; TIA, transient ischaemic attack; y, year (s).

raction; eGFR, glomerular filtration index; HTN, hypertension; LV ESVI, left ventricular end-systolic volume index; LV EDVI, left ventricular end-diastolic volume index; LV SVI, left ventricular stroke volume index; MI, myocardial infarction; PCI, percutaneous intervention; TIA, transient ischaemic attack; y, year (s). Figure 2 Recruitment of subjects. AV, atrioventricular; CMR, cardiac magnetic resonance imaging. Adverse events Overall, there were few adverse events (Table 2). There were no deaths, MIs, VT/VF, high-grade AV block, regadenoson-induced atrial fibrillation, or nephrogenic systemic fibrosis. There was one hospitalization related to acute exacerbation of chronic heart failure and one episode of bronchospasm requiring observation in the emergency department despite reversal with aminophylline. Six per cent (46 of 728) of patients had minor stress-induced dysrhythmias (premature atrial and/or ventricular contractions). Two patients experienced transient symptomatic hypotension (one was secondary to transient bigeminy; one was secondary to transient narrow complex bradycardia with difficult to distinguish P-wave morphology). Two patients had contrast extravasation. Rash or hives occurred in one subject and may be related to gadolinium or regadenoson. Nine patients required sublingual nitroglycerine for chest pain; whereas six patients required additional iv metoprolol for symptom resolution. Table 2 Frequency of adverse events associated with regadenoson CMR

ntrast extravasation. Rash or hives occurred in one subject and may be related to gadolinium or regadenoson. Nine patients required sublingual nitroglycerine for chest pain; whereas six patients required additional iv metoprolol for symptom resolution. Table 2 Frequency of adverse events associated with regadenoson CMR Adverse events Patient cohort (n = 728) Death 0 VT/VF 0 Myocardial infarction 0 Hospitalization 1 Bronchospasm 1 High-grade AV block 0 Stress-induced atrial fibrillation 0 Nephrogenic systemic fibrosis 0 Stress-induced ectopies (PACs/PVCs) 46 (6%) Bigeminy 2 (<1%) Symptomatic hypotension 2 (<1%) Contrast extravasation 2 (<1%) Minor reaction to gadolinium (rash/hives) 1 (<1%) Thrombophlebitis 0 Chest pain requiring NTG 9 (1%) Chest pain requiring iv metoprolol 6 (<1%) AV, atrioventricular; iv, intravenous; NTG, nitroglycerine; PACs, premature atrial contractions; PVCs, premature ventricular contractions; VF, ventricular fibrillation; VT, ventricular tachycardia. Frequency of symptoms Dyspnoea, chest pain, and headache were the three symptoms most frequently reported by patients (Figure 3). More normal volunteers experienced palpitations when compared with the patient cohort (60 vs. 8%; P = 0.652), while dyspnoea was experienced at a similar frequency (P = 0.525). Figure 3 Frequency of symptoms reported by patients and normal volunteers. Abd, abdominal; CP nitro, chest pain requiring nitroglycerine; CP MTP, chest pain requiring metoprolol.

palpitations when compared with the patient cohort (60 vs. 8%; P = 0.652), while dyspnoea was experienced at a similar frequency (P = 0.525). Figure 3 Frequency of symptoms reported by patients and normal volunteers. Abd, abdominal; CP nitro, chest pain requiring nitroglycerine; CP MTP, chest pain requiring metoprolol. Haemodynamic response to regadenoson Systolic and diastolic BPR among patient subgroups and normal volunteers was not statistically significant (P > 0.05, Figure 4). In the patient cohort, median systolic and diastolic BPR were −2% (IQR: −10 to 5) and −5% (IQR: −14 to 3), respectively. In normal volunteers, median systolic and diastolic BPR were −3% (IQR: −6 to 6) and −10% (IQR: −17 to 1), respectively. Despite relatively similar baseline median HR between normal volunteers [65 bpm (IQR: 53–71)] and patient cohort [66 bpm (IQR: 58–76), P = 0.066], normal volunteers had a higher median HRR [71% (IQR: 58–97)] when compared with the patient cohort [48% (IQR: 35–63), P < 0.001] (Figure 4). The higher HRR by normal volunteers likely represent a robust sympathetic response as one would expect in a younger cohort of normal healthy volunteers. A statistically significant blunted HRR was noted in those with BMI ≥30 kg/m2 and DM (Figure 5). Patients with BMI ≥30 kg/m2 had a higher median baseline HR [68 bpm (IQR: 62–77)] when compared with those with BMI <30 kg/m2 [64 bpm (IQR: 57–65 bpm), P = 0.001]. A higher resting HR was also present in patients with DM [69 bpm (IQR: 62–80)] compared with those without DM [65 bpm (IQR: 58–75), P = 0.001]. Figure 4 Haemodynamic response with regadenoson. Values reported are medians. Error bars represent the inter-quartile range. Systolic and diastolic BP response among patient subgroups was not statistically significant (P > 0.05). BMI, body mass index; BP, blood pressure.

ose without DM [65 bpm (IQR: 58–75), P = 0.001]. Figure 4 Haemodynamic response with regadenoson. Values reported are medians. Error bars represent the inter-quartile range. Systolic and diastolic BP response among patient subgroups was not statistically significant (P > 0.05). BMI, body mass index; BP, blood pressure. Figure 5 Box-and-Whisker plot of median heart rate response in patient subgroups. Differences between patient subgroups were evaluated using the Mann–Whitney U test. The height of the box represents the inter-quartile range. The middle horizontal line in the box represents the median. Whiskers (error bars) extending from the box represent minimum–maximum values. Circles represent data points. BMI, body mass index. Using a multivariable logistic regression model, the following variables predicted the lowest quartile of HRR (Hosmer–Lemeshow test for goodness of fit χ2 = 8, P = 0.37): age ≥64, BMI ≥30 kg/m2, DM, LVEF ≤40%, and abnormal perfusion (Table 3). Of the significant predictors of HRR in the lowest quartile, abnormal perfusion was the weakest (P = 0.042). Interactions were found between age*BMI*DM (P = 0.023) and abnormal perfusion*DM (P = 0.037). Table 3 Univariable and multivariable logistic analysis of predictors for the lowest quartile of heart rate response

Of the significant predictors of HRR in the lowest quartile, abnormal perfusion was the weakest (P = 0.042). Interactions were found between age*BMI*DM (P = 0.023) and abnormal perfusion*DM (P = 0.037). Table 3 Univariable and multivariable logistic analysis of predictors for the lowest quartile of heart rate response Variables Coefficient (β) Standard error P-value Odds ratio 95% CI Multivariable analysisa Age ≥64 years 1.010 (0.884) 0.310 (0.204) 0.001 (<0.001) 2.745 (2.421) 1.495–5.040 (1.623–3.611) BMI ≥30 kg/m2 0.784 (0.737) 0.212 (0.192) 0.001 (0.001) 2.190 (2.089) 1.445–3.320 (1.433–3.045) Diabetes 1.009 (0.552) 0.277 (0.212) 0.001 (0.009) 2.743 (1.736) 1.595–4.718 (1.146–2.631) LVEF ≤40% 0.944 (0.945) 0.391 (0.391) 0.016 (0.016) 2.569 (2.573) 1.195–5.522 (1.195–5.537) Abnormal perfusion 0.472 (0.264) 0.232 (0.203) 0.042 (0.194) 1.603 (1.302) 1.018–2.524 (0.875–1.937) Univariable analysis Age ≥64 years 0.783 0.184 <0.001 2.189 1.527–3.138 BMI ≥30 kg/m2 0.494 0.173 0.004 1.639 1.168–2.301 Diabetes 0.624 0.197 0.002 1.867 1.269–2.746 LVEF ≤40% 0.987 0.371 0.008 2.683 1.298–5.547 Abnormal perfusion 0.371 0.176 0.035 1.449 1.027–2.044 eGFR 30–44.9 1.209 0.676 0.074 3.348 0.890–12.600 eGFR 45–60 −0.307 0.289 0.288 0.736 0.418–1.296 eGFR >60 0.096 0.263 0.716 1.100 0.657–1.843 Gender 0.053 0.168 0.753 1.054 0.759–1.465 Beta-blocker use 0.436 0.168 0.010 1.547 1.113–2.149 BMI, body mass index; CI, confidence interval; LVEF, left ventricular ejection fraction; eGFR, estimated glomerular filtration rate (mL/min/1.73 cm2).

.289 0.288 0.736 0.418–1.296 eGFR >60 0.096 0.263 0.716 1.100 0.657–1.843 Gender 0.053 0.168 0.753 1.054 0.759–1.465 Beta-blocker use 0.436 0.168 0.010 1.547 1.113–2.149 BMI, body mass index; CI, confidence interval; LVEF, left ventricular ejection fraction; eGFR, estimated glomerular filtration rate (mL/min/1.73 cm2). aBeta-blocker use was removed from the best-fit multivariable model for P > 0.05 after entry into the model. The best-fit model was also adjusted for interactions between age*BMI*DM (P = 0.023) and abnormal perfusion*DM (P = 0.037). Values without adjustment for interactions are in parentheses. Goodness of fit for the best-fit model using the Hosmer–Lemeshow test: model adjusted for interactions χ2 = 8, P = 0.374, area under ROC curve 0.694 (95% CI: 0.658–0.729); Model unadjusted for interactions χ2 = 17, P = 0.020, area under ROC curve 0.686 (95% CI: 0.650–0.721). Discussion Our study demonstrates that regadenoson perfusion CMR can be performed in a clinical setting with few adverse events and that regadenoson is well tolerated. There were no occurrences of death, MI, VT/VF, high-grade AV block, or stress-induced atrial fibrillation. A blunted HRR was noted in patients with a BMI ≥30 kg/m2 and diabetes.

nstrates that regadenoson perfusion CMR can be performed in a clinical setting with few adverse events and that regadenoson is well tolerated. There were no occurrences of death, MI, VT/VF, high-grade AV block, or stress-induced atrial fibrillation. A blunted HRR was noted in patients with a BMI ≥30 kg/m2 and diabetes. Several studies have reported on the safety and tolerability of adenosine and dobutamine stress CMR.15–18 Large-scale trials have also established the safety of regadenoson stress testing with single photon emission computed tomography (SPECT).2–4,19 However, there are no published large-scale, prospective studies assessing the safety and tolerability of regadenoson perfusion CMR. The CMR environment represents a confined space with a strong magnetic field, where the ECG signalsmay be distorted,13 and resuscitation requires prompt patient removal. Thus, knowledge of adverse events relating to the safety and tolerability of regadenoson CMR is important for patient safety and for staff training.

. The CMR environment represents a confined space with a strong magnetic field, where the ECG signalsmay be distorted,13 and resuscitation requires prompt patient removal. Thus, knowledge of adverse events relating to the safety and tolerability of regadenoson CMR is important for patient safety and for staff training. The EMA defines adverse events as ‘any untoward medical occurrence in a patient or clinical investigation subject administered a pharmaceutical product and which does not necessarily have to have a causal relationship with the treatment’ while the United States FDA defines adverse events as ‘an undesirable experience associated with the use of a medical product or device’. Both define serious events as those resulting in death, life-threatening conditions, hospitalization, disability or permanent damage, or congenital anomaly or birth defect. Based on the above definition, two adverse events in our study merit further discussion.

e associated with the use of a medical product or device’. Both define serious events as those resulting in death, life-threatening conditions, hospitalization, disability or permanent damage, or congenital anomaly or birth defect. Based on the above definition, two adverse events in our study merit further discussion. Event #1 involved regadenoson-induced bronchospasm in a patient with known CAD but no history of COPD or asthma. He developed bronchospasm with active wheezing following regadenoson injection. After reversal with aminophylline, albuterol, and methylprednisolone were administered. He was admitted to the emergency department for further monitoring. No intubation or hospitalization was required. According to the package inserts, bronchoconstrictive or bronchospastic conditions such as asthma are contraindications for adenosine.20,21 However, these conditions are not listed as contraindications for regadenoson.22,23 The regadenoson package insert contains a warning for potential bronchoconstriction and suggests that bronchodilator therapy and resuscitative measures be available. However, multiple studies evaluated the specific safety of regadenoson in patients with COPD and asthma and found no increase in acute COPD or asthma exacerbation.9,10,24

package insert contains a warning for potential bronchoconstriction and suggests that bronchodilator therapy and resuscitative measures be available. However, multiple studies evaluated the specific safety of regadenoson in patients with COPD and asthma and found no increase in acute COPD or asthma exacerbation.9,10,24 Event #2 involved exacerbation of chronic heart failure requiring hospitalization. The patient had known multivessel CAD and declined bypass surgery 2 years prior to his presentation. Because of recurrent heart failure, he was referred for assessment of his ischaemia and scar burden. On presentation, he reported stable dyspnoea and lower extremity oedema. He was haemodynamically stable before, during, and after the perfusion CMR. However, his exam showed multiple moderate to severe perfusion defects with viable myocardium. After discharge, he had worsening of dyspnoea and presented to the hospital where he did not have ischaemic ECG changes, but did have a troponin-I of 0.15 μg/L and a pro-BNP of 2550 pg/mL. He was admitted for three-vessel revascularization and heart failure management. In reviewing the case, we could not delineate whether exacerbation of his symptoms was secondary to a stress-induced increase in left ventricular end-diastolic and wedge pressure25 or whether this was a natural progression of his disease. Although his heart failure medications were held the morning of the exam, a 3- to 4-h lapse in the usual timing of his medications would unlikely lead to his decompensation. To our knowledge, there are no reports of regadenoson- or aminophylline-induced heart failure in the literature or on the package insert.

disease. Although his heart failure medications were held the morning of the exam, a 3- to 4-h lapse in the usual timing of his medications would unlikely lead to his decompensation. To our knowledge, there are no reports of regadenoson- or aminophylline-induced heart failure in the literature or on the package insert. The mechanism of adenosine-induced tachycardia has been attributed to a baroreflex-mediated activation of the sympathetic nervous system.26 However, a recent study using regadenoson suggests that activation of the A2A receptor causes direct activation of the sympathetic nervous system.27 Because regadenoson has greater selectivity for the A2A receptor, the effect of regadenoson-mediated tachycardia is exaggerated. Abidov et al.28 first reported on the prognostic significance of HRR following adenosine infusion in 2003, thereby spurring an interest in HRR in vasodilator testing. Recently, Hage et al.29 hypothesized that a blunted HRR may reflect the health of the sympathetic system and therefore, be prognostically useful. In their recent work, they demonstrated that a blunted HRR in both regadenoson and adenosine perfusion SPECT is an independent predictor of poor outcome.30,31 A blunted HRR in regadenoson SPECT was noted in those with DM14 and metabolic syndrome.32 In this study, we report a blunted HRR in those with BMI ≥30 kg/m2 and DM. Further, DiBella et al.12 found that fixed-dose regadenoson was sufficient in obese subjects. Taken together, these data suggest that blunted HRR observed in our obese subjects is unlikely due to a simple dose effect, but additional studies are warranted. Interestingly, analysis of individual absolute HRR via Box–Whisker plots showed great overlap between-patient subgroups thereby suggesting that individual data points have limited diagnostic value in individual patients.

n our obese subjects is unlikely due to a simple dose effect, but additional studies are warranted. Interestingly, analysis of individual absolute HRR via Box–Whisker plots showed great overlap between-patient subgroups thereby suggesting that individual data points have limited diagnostic value in individual patients. Perfusion CMR imaging has progressed in recent years.33,34 Despite its superiority to SPECT in the diagnosis of CAD,35 its high sensitivity and specificity,35,36 potential overall cost reduction in diagnosing chest pain,37 and lack of radiation, the incorporation of perfusion CMR into daily clinical routine has been slow. Requirements for MRI compatible infusion pumps and weight-based vasodilator infusions complicate the workflow. Regadenoson can simplify current stress protocols in perfusion CMR. Fixed-dose bolus administration obviates the need for infusion pumps and shortens exam time.

n CMR into daily clinical routine has been slow. Requirements for MRI compatible infusion pumps and weight-based vasodilator infusions complicate the workflow. Regadenoson can simplify current stress protocols in perfusion CMR. Fixed-dose bolus administration obviates the need for infusion pumps and shortens exam time. Several limitations in our study merit discussion. Our sample size is modest compared with prior clinical trials evaluating the safety of regadenoson SPECT. Secondly, our study reflects a single-centre experience, but the patient demographics are representative of the general population being referred for stress testing. We note that the prevalence of patients with COPD/asthma is low. Many referred patients have been pre-screened by other cardiologists followed by a second round of screening by our nursing staff. Additionally, we excluded one patient with acute exacerbation of asthma, which may represent a potential limitation of this study. Thirdly, because of ECG distortion by a strong magnetic field, the diagnosis of heart block is limited. With regadenoson, the fixed rapid bolus administration does not allow for dose-modification based on ECG findings and thus is one reason why our study primarily focused on haemodynamically significant adverse events, which were infrequent. Despite recent work to reduce ECG signal distortion during stress testing, ECG monitoring during stress CMR perfusion exams remains suboptimal. Lastly, recent work by Bhave et al.38 demonstrated partial reversal of regadenoson-induced hyperaemia despite aminophylline. This finding has implications for the quantification of myocardial perfusion reserve and diagnostic accuracy. Although these issues are of clinical significance, they are beyond the scope of this study.

recent work by Bhave et al.38 demonstrated partial reversal of regadenoson-induced hyperaemia despite aminophylline. This finding has implications for the quantification of myocardial perfusion reserve and diagnostic accuracy. Although these issues are of clinical significance, they are beyond the scope of this study. In conclusion, our findings demonstrate that regadenoson perfusion CMR is safe and the frequency of adverse events is low. Regadenoson perfusion CMR is well-tolerated and symptoms are comparable with those reported in the nuclear literature. Funding This research was supported by the Division of Intramural Research of the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (1ZIAHL004607, 1ZIAHL006137, 1ZIEHL006139, 1ZIDHL006140). Funding to pay the Open Access publication charges for this article was provided by the Division of Intramural Research of the National Heart, Lung and Blood Institute, National Institutes of Health. Acknowledgements The authors thank Dr Xin Tian, PhD (Office of Biostatistics Division, NHLBI) for her statistical guidance on logistic regression modelling. The authors also thank all patients, referring physicians, research coordinators (particularly Tracy Lowrey and Luis Rosario), clinical fellows within the Advanced Cardiovascular Imaging Fellowship Program, and technologists at NHLBI for their assistance. Conflict of interest: A.E.A. is the principal investigator on a U.S. Government Cooperative Research and Development Agreement (CRADA) with Siemens.

Introduction Three-vessel coronary artery disease (3VD) is found in ∼9% of patients undergoing elective coronary angiography and these patients have a considerably poorer prognosis than those with less extensive disease.1 Detection of 3VD with non-invasive imaging can be challenging due to the effects of balanced ischaemia leading to false-negative results in up to 20% of cases.2,3 This limitation has been well documented with single-photon emission computed tomography (SPECT), and although its overall sensitivity for detecting coronary artery disease (CAD) in multi-vessel disease is 80–95%, it often only detects perfusion defects in one territory.2,4,5 In one SPECT study, inducible perfusion abnormalities in all three territories were identified in only 12% of patients with known angiographic 3VD.6

hough its overall sensitivity for detecting coronary artery disease (CAD) in multi-vessel disease is 80–95%, it often only detects perfusion defects in one territory.2,4,5 In one SPECT study, inducible perfusion abnormalities in all three territories were identified in only 12% of patients with known angiographic 3VD.6 Myocardial perfusion cardiovascular magnetic resonance (perfusion CMR) imaging is a highly accurate method for the detection of significant CAD.7–9 One of the major advantages of perfusion CMR compared with SPECT is its higher spatial resolution (typically 2–3 vs. 8–10 mm). Balanced ischaemia can lead to diffuse subendocardial hypoperfusion and although there are few comparisons between perfusion CMR and SPECT in 3VD, it is expected that the higher resolution of CMR can better resolve the transmural perfusion gradient in balanced ischaemia and thereby potentially improve the detection of 3VD.10 With recently developed spatio-temporal undersampling methods such as k-t broad-use linear acquisition speed-up technique (k-t BLAST), the spatial resolution of perfusion CMR can be improved further to <2 mm.11

e transmural perfusion gradient in balanced ischaemia and thereby potentially improve the detection of 3VD.10 With recently developed spatio-temporal undersampling methods such as k-t broad-use linear acquisition speed-up technique (k-t BLAST), the spatial resolution of perfusion CMR can be improved further to <2 mm.11 Several studies have demonstrated the feasibility and accuracy of high-resolution perfusion CMR.12–16 In a direct comparison, we have previously shown that high-resolution perfusion CMR has a higher overall diagnostic accuracy compared with standard-resolution imaging in patients with suspected CAD. This previous study included a small subset of patients with multi-vessel disease.16 The present study aims to compare the distribution of and extent of ischaemia in patients with 3VD detected by both techniques and tests the hypothesis that improved performance of high-resolution CMR is due to better detection of subendocardial ischaemia in 3VD.

a small subset of patients with multi-vessel disease.16 The present study aims to compare the distribution of and extent of ischaemia in patients with 3VD detected by both techniques and tests the hypothesis that improved performance of high-resolution CMR is due to better detection of subendocardial ischaemia in 3VD. Methods Study population A total of 105 patients were included in this analysis. All had undergone coronary angiography for suspected angina within the last 30 days. Seventy patients were prospectively recruited: 35 had 3VD on quantitative coronary angiography and 35 with normal coronary arteries served as a control group. Data from 24 of the patients with angiographic 3VD have been previously reported with different endpoints (diagnostic accuracy rather than pattern of ischaemia or ischaemic burden).16 Additionally, we selected 35 consecutive patients with angiographic 1VD or 2VD from the same previous study to prevent a spectrum bias for the secondary analyses relating to myocardial ischaemic burden (MIB).16 Exclusion criteria for all patients were contra-indications to CMR, adenosine, or gadolinium contrast agent; or a history of recent (within 6 months) myocardial infarction (MI), unstable angina, or revascularization. Additionally, patients with angiographic 1VD or 2VD and co-existing moderate coronary artery stenoses (i.e. 40–69%) in other territories were not included. All patients gave written consent and the study was approved by the regional ethics committee.

s) myocardial infarction (MI), unstable angina, or revascularization. Additionally, patients with angiographic 1VD or 2VD and co-existing moderate coronary artery stenoses (i.e. 40–69%) in other territories were not included. All patients gave written consent and the study was approved by the regional ethics committee. CMR protocol All patients underwent a standard-resolution and a high-resolution perfusion scan on separate days (within 4 weeks) using a 1.5-T scanner (Intera, Philips Healthcare, Best, The Netherlands). The standard perfusion pulse sequence was a saturation recovery gradient-echo method accelerated with sensitivity encoding (SENSE) (SENSE acceleration factor 2, repetition time (TR) 2.7 ms, echo time (TE) 1.0 ms, flip-angle = 15°, acquisition time per slice = 136 ms, single-saturation pre-pulse per R–R interval shared over three slices, matrix = 144 × 144, median field-of-view (FOV) = 360 mm, in-plane spatial resolution = 2.5 mm). The high-resolution perfusion pulse sequence used a similar saturation recovery gradient-echo method, but was accelerated with k-t BLAST (acceleration factor 8 with 11 training profiles, TR = 3.4 ms, TE = 1.7 ms, flip-angle = 15°, one saturation pre-pulse per slice, acquisition time per slice = 103 ms, matrix = 192 × 192, median FOV = 310 mm, in-plane spatial resolution = 1.6 mm). For both techniques, perfusion data were acquired in three short-axis slices in each R–R interval.

th 11 training profiles, TR = 3.4 ms, TE = 1.7 ms, flip-angle = 15°, one saturation pre-pulse per slice, acquisition time per slice = 103 ms, matrix = 192 × 192, median FOV = 310 mm, in-plane spatial resolution = 1.6 mm). For both techniques, perfusion data were acquired in three short-axis slices in each R–R interval. For both studies, stress perfusion started after 4 min of an intravenous adenosine infusion (140 mcg/kg/min) during an intravenous bolus injection of dimeglumine gadopentetate (Magnevist; Schering AG, West Sussex, UK) and a 15-mL saline flush delivered at 5 mL/s. Contrast dose and administration protocols for both studies were chosen to optimize their visual analysis performance based on their use in previous studies. For the standard-resolution method, a contrast dose of 0.05 mmol/kg body weight was used during perfusion acquisition, identical to previous studies with this pulse sequence.9,16 To compensate for the lower signal-to-noise ratio (SNR) associated with smaller voxel size, a contrast dose of 0.1 mmol/kg body weight was used for the high-resolution method, consistent with previous reports.13,14,16,17

was used during perfusion acquisition, identical to previous studies with this pulse sequence.9,16 To compensate for the lower signal-to-noise ratio (SNR) associated with smaller voxel size, a contrast dose of 0.1 mmol/kg body weight was used for the high-resolution method, consistent with previous reports.13,14,16,17 Rest perfusion imaging was performed 15 min later. Late gadolinium-enhancement (LGE) imaging was performed in all patients on their first visit using conventional methods (1.6 mm in-plane spatial resolution) and a cumulative contrast dose of 0.2 mmol/kg body weight (the same for both protocols). During standard-resolution perfusion CMR scans, this cumulative dose was achieved by administration of an additional contrast bolus of 0.1 mmol/kg body weight immediately after rest perfusion.

s (1.6 mm in-plane spatial resolution) and a cumulative contrast dose of 0.2 mmol/kg body weight (the same for both protocols). During standard-resolution perfusion CMR scans, this cumulative dose was achieved by administration of an additional contrast bolus of 0.1 mmol/kg body weight immediately after rest perfusion. CMR analysis CMR images were anonymized, randomly ordered and visually reported by two observers (S.P. and M.M., 10 and 2 years experience, respectively) acting in consensus and blinded to all clinical and angiographic data (QMASS 6.1.6, Medis, Leiden, The Netherlands). In case of disagreement, arbitration from a third observer was sought (J.P.G., 10 years experience).Using a 16-segment model, perfusion in a segment was considered abnormal if signal intensity was reduced compared with remote myocardium or an endocardial-to-epicardial perfusion gradient was present.12,18 Additionally, any perfusion defect was required to persist beyond the peak myocardial signal enhancement to distinguish it from artefact. Corresponding LGE images were reviewed side-by-side with the perfusion data. Perfusion defects present at stress but not rest and occurring outside any hyperenhanced myocardial tissue on LGE images were considered as inducible defects. Perfusion in each segment was graded on a four-point scale (transmural ischaemia index: 0 = normal, 1 = inconclusive, 2 = subendocardial defect, 3 = transmural defect). A typical example of perfusion images is seen in Figure 1. All segmental scores were summed to produce a perfusion score (0–48) for each patient. MIB as a percentage of the total myocardium (MIB%) was estimated by dividing the perfusion score by 48 and multiplying by 100.19 In patients with 3VD, perfusion scores were also calculated for the left anterior descending (LAD), left circumflex (LCX) and right coronary artery (RCA) territories according to american heart association (AHA) segmentation adjusted for arterial dominance.18 Figure 1 Case example. Standard and high-resolution stress perfusion CMR in a patient with three-vessel coronary artery disease. Standard-resolution shows perfusion defects (arrows) in the basal-inferior (A), mid-inferior, mid-inferoseptal (B), apical-anterior and apical-inferior segments (C). High-resolution shows a similar distribution of perfusion defects but demonstrates additional ischaemia in the basal-lateral (D), mid-anterior, and mid-anterolateral segments (E) with a circumferential defect in the apical slice (F).

mid-inferior, mid-inferoseptal (B), apical-anterior and apical-inferior segments (C). High-resolution shows a similar distribution of perfusion defects but demonstrates additional ischaemia in the basal-lateral (D), mid-anterior, and mid-anterolateral segments (E) with a circumferential defect in the apical slice (F). Perfusion defects are also better delineated at high-resolution and the transmural extent of ischaemia more clearly seen. Image quality was graded 1–4 (1 = unusable, 2 = poor, 3 = adequate, 4 = excellent). Occurrence of artefacts related to k-t reconstruction, respiratory motion, electrocardiographic gating, and endocardial dark-rim was scored 0–3 (0 = none, 1 = minor, 2 = moderate, 3 = severe). Where present, the width of dark-rim artefact (DRA) (a frequent finding in perfusion CMR at the myocardial–blood pool interface relating to cardiac motion, Gibb's ringing, susceptibility, and partial volume cancellation) was measured with electronic callipers.16,20

d 0–3 (0 = none, 1 = minor, 2 = moderate, 3 = severe). Where present, the width of dark-rim artefact (DRA) (a frequent finding in perfusion CMR at the myocardial–blood pool interface relating to cardiac motion, Gibb's ringing, susceptibility, and partial volume cancellation) was measured with electronic callipers.16,20 Quantitative coronary angiography Quantitative coronary angiography (QCA) was performed (QCAPlus, Sanders Data Systems, Palo Alto, CA, USA) by an experienced observer blinded to CMR data (M.M., 7years of experience in coronary angiography). Stenoses were assigned to the appropriate myocardial segments of an AHA 16-segment model using standard criteria adjusted for arterial dominance and lesion location.18,21 As per convention, significant CAD was defined as luminal stenosis ≥70% in any of the major epicardial coronary arteries or first-order branches ≥2 mm. Angiographic 3VD was defined as stenosis ≥70% in all three coronary arteries; or the presence of ≥50% stenosis in the left main stem with ≥70% in the RCA. Normal coronary arteries were defined as an absence of any stenosis ≥40%. Collateral circulation was graded according to the Rentrop classification (RC) depending on the angiographic findings of the occluded artery using the best injection: 0 = no collateral circulation; 1 = collateral filling of side branches without visualization of any epicardial segments; 2 = collaterals partially filling the epicardial segment; 3 = collaterals completely filling the epicardial segment.22

ng on the angiographic findings of the occluded artery using the best injection: 0 = no collateral circulation; 1 = collateral filling of side branches without visualization of any epicardial segments; 2 = collaterals partially filling the epicardial segment; 3 = collaterals completely filling the epicardial segment.22 Statistical analysis Analysis was performed using SPSS 17.0 (SPSS, Chicago, IL, USA). Mean values were compared using paired Student t-tests. Ordinal data were compared using χ2 or Wilcoxon signed-rank tests as appropriate. Paired proportions were compared using McNemar's exact test. The pattern of ischaemia determined by both techniques for patients with angiographic 3VD was compared using Cohen's kappa statistic. MIB% was compared across 1VD, 2VD, and 3VD groups using one-way analysis of variance and Tukey's post hoc test. Receiver-operator characteristic (ROC) curve analyses were performed on summed perfusion scores for individual territories and on MIB% per patient. Area-under-the-curve (AUC) for both imaging techniques were compared using methods described by DeLong and DeLong. All tests were two-tailed and P < 0.05 was considered statistically significant.

cteristic (ROC) curve analyses were performed on summed perfusion scores for individual territories and on MIB% per patient. Area-under-the-curve (AUC) for both imaging techniques were compared using methods described by DeLong and DeLong. All tests were two-tailed and P < 0.05 was considered statistically significant. Results Study population A total of 105 patients were enrolled into the study, including 35 with 3VD by QCA. Of these 35 patients, 32 qualified as angiographic 3VD on the basis of significant stenoses in the proximal coronary segments and no patients qualified as angiographic 3VD on the basis of distal coronary segment disease. Further demographics are given in Table 1. All 14 patients with a clinical history of MI (but no additional patients) had evidence of hyperenhancement on LGE imaging. A chronic total occlusion (CTO) was seen in six patients (all in the group with prior MI) and two of these patients were in the 3VD group. There were no patients with more than one CTO. In two cases of CTO (neither in the 3VD group) there was mild collateral flow (RC = 2), but in the remaining four cases there was no or minimal collateralization (RC ≤ 1). Table 1 Patient clinical characteristics (n = 105)

) and two of these patients were in the 3VD group. There were no patients with more than one CTO. In two cases of CTO (neither in the 3VD group) there was mild collateral flow (RC = 2), but in the remaining four cases there was no or minimal collateralization (RC ≤ 1). Table 1 Patient clinical characteristics (n = 105) Age, years 68 ± 10 Males 75 (71) Medical history Hypertension 65 (62) Hypercholesterolaemia 62 (59) Diabetes mellitus 18 (17) Smoking 43 (41) Family history of CAD 39 (37) Previous MI 14 (13) Previous PCI 10 (9) Atrial fibrillation 1 (1) LV ejection fraction, % 55 ± 11 Angiography findingsa No significant disease 35 (33) One-vessel disease 25 (24) Two-vessel disease 10 (9) Three-vessel disease 35 (33) LAD disease 51 (49) LCX disease 50 (48) RCA disease 49 (47) Significant lesions per patient 1.4 ± 1.2 Total significant lesions 150 70–90% stenoses 109 (73) 90–99% stenoses 35 (23) Chronic total occlusions 6 (4) Values are mean ± SD or n (%). MI, myocardial infarction; PCI, percutaneous coronary intervention; LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; RCA, right coronary artery. aSignificant coronary artery disease (CAD) defined as coronary stenosis ≥70% on quantitative coronary analysis.

Age, years 68 ± 10 Males 75 (71) Medical history Hypertension 65 (62) Hypercholesterolaemia 62 (59) Diabetes mellitus 18 (17) Smoking 43 (41) Family history of CAD 39 (37) Previous MI 14 (13) Previous PCI 10 (9) Atrial fibrillation 1 (1) LV ejection fraction, % 55 ± 11 Angiography findingsa No significant disease 35 (33) One-vessel disease 25 (24) Two-vessel disease 10 (9) Three-vessel disease 35 (33) LAD disease 51 (49) LCX disease 50 (48) RCA disease 49 (47) Significant lesions per patient 1.4 ± 1.2 Total significant lesions 150 70–90% stenoses 109 (73) 90–99% stenoses 35 (23) Chronic total occlusions 6 (4) Values are mean ± SD or n (%). MI, myocardial infarction; PCI, percutaneous coronary intervention; LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; RCA, right coronary artery. aSignificant coronary artery disease (CAD) defined as coronary stenosis ≥70% on quantitative coronary analysis. Image quality and artefacts There was no significant difference in the haemodynamic stress response during standard- and high-resolution imaging (rate–pressure product, mmHg × beats/min: 10 251 ± 2321 vs. 10 201 ± 2109; P = 0.92). No images were graded as unusable and therefore there were no exclusions from the image analysis for either technique. Image quality (median score = 3 for both; P = 0.67) and artefact scores (median = 0 for both; P = 0.06) were similar for both standard- and high-resolution imaging across the full spectrum of patients (n = 105). DRA was significantly less frequent with high-resolution (7% vs. 26%; P = 0.03) and when it did occur, it was less marked than with standard-resolution (1.6 ± 0.2 vs. 3.2 ± 0.8 mm; P = 0.004). Seven high-resolution data sets (10%) were affected by k-t reconstruction artefacts at stress and/or rest due to respiratory motion, but this did not affect myocardial contrast passage and generally occurred at the end of a breath hold.

detected in 57% using high-resolution imaging but in only 29% using standard-resolution (P = 0.04). (B) There was also poor agreement between high-resolution and standard-resolution imaging in determining the distribution of ischaemia in patients with 3VD (kappa = 0.09, 95% CI: −0.10–0.29). IT = ischaemic territories. Detection of CAD in each territory In patients with angiographic 3VD (n = 35), separate ROC analyses of perfusion scores for each of the three coronary territories were performed. The AUC for each territory in patients with angiographic 3VD was greater with high-resolution than with standard-resolution imaging, but reached statistical significance only for the LCX territory (Table 2). Table 2 Diagnostic performance of perfusion CMR in each territory in patients with 3VD AUC (95% CI) Standard-resolution High-resolution P-value LAD 0.82 (0.69–0.95) 0.85 (0.73–0.97) 0.62 LCX 0.62 (0.46–0.79) 0.83 (0.70–0.95) 0.02 RCA 0.83 (0.70–0.95) 0.90 (0.80–1.00) 0.27 With standard-resolution imaging, diagnostic accuracy was significantly lower for the LCX than for the LAD (0.62 vs. 0.82; P < 0.01) or RCA territory (0.62 vs. 0.83; P = 0.02). With high-resolution, diagnostic accuracies were more homogenous between territories (although still lowest in the LCX territory) with no statistical difference between them (LAD: 0.85 vs. LCX: 0.83 vs. RCA: 0.90; all P-values >0.05).

han for the LAD (0.62 vs. 0.82; P < 0.01) or RCA territory (0.62 vs. 0.83; P = 0.02). With high-resolution, diagnostic accuracies were more homogenous between territories (although still lowest in the LCX territory) with no statistical difference between them (LAD: 0.85 vs. LCX: 0.83 vs. RCA: 0.90; all P-values >0.05). Detection of subendocardial ischaemia Five hundred and sixty myocardial segments were available from the 35 patients with angiographic 3VD. Of these, 420 were determined as angiographically hypoperfused and used for further analysis. With high-resolution acquisition, significantly more of these hypoperfused segments were determined as having subendocardial ischaemia than with standard-resolution (195 vs. 101; P < 0.0001); and there was a significant reduction in the number of segments determined as being normal (135 vs. 212; P < 0.001) or inconclusive (15 vs. 40; P < 0.001). The number of segments assessed as having transmural ischaemia was similar with both techniques (75 vs. 67; P = 0.25) (Figure 3). Figure 3 Distribution of transmural ischaemia index. In patients with three-vessel disease (n = 35), high-resolution perfusion CMR determined significantly more segments as having subendocardial ischaemia and fewer as normal or inconclusive compared with standard-resolution imaging.

es (75 vs. 67; P = 0.25) (Figure 3). Figure 3 Distribution of transmural ischaemia index. In patients with three-vessel disease (n = 35), high-resolution perfusion CMR determined significantly more segments as having subendocardial ischaemia and fewer as normal or inconclusive compared with standard-resolution imaging. Detection of ischaemic burden In patients with angiographic 3VD (n = 35), the overall extent of myocardial ischaemia detected was significantly greater with high-resolution than standard-resolution imaging, with more abnormal segments per patient (7.2 ± 3.8 vs. 5.3 ± 4.0; P = 0.004), more abnormal territories per patient (2.4 ± 0.9 vs. 1.6 ± 1.1; P = 0.002), a higher perfusion score per territory (5.9 ± 4.3 vs. 4.7 ± 5.0; P = 0.01) and a higher overall perfusion score per patient (20.1 ± 7.7 vs. 11.9 ± 9.4; P < 0.0001) (Figure 1 and Table 3). Table 3 Detection of ischaemic burden with perfusion CMR in patients with 3VD Standard-resolution High-resolution P-value Mean abnormal segments per patient 5.3 ± 4.0 7.2 ± 3.8 0.004 Mean abnormal territories per patient 1.6 ± 1.0 2.4 ± 0.9 0.002 Mean perfusion score per patient 11.9 ± 9.4 20.1 ± 7.7 <0.0001 Mean perfusion score per territory 4.7 ± 5.0 5.9 ± 4.3 0.01

Detection of ischaemic burden In patients with angiographic 3VD (n = 35), the overall extent of myocardial ischaemia detected was significantly greater with high-resolution than standard-resolution imaging, with more abnormal segments per patient (7.2 ± 3.8 vs. 5.3 ± 4.0; P = 0.004), more abnormal territories per patient (2.4 ± 0.9 vs. 1.6 ± 1.1; P = 0.002), a higher perfusion score per territory (5.9 ± 4.3 vs. 4.7 ± 5.0; P = 0.01) and a higher overall perfusion score per patient (20.1 ± 7.7 vs. 11.9 ± 9.4; P < 0.0001) (Figure 1 and Table 3). Table 3 Detection of ischaemic burden with perfusion CMR in patients with 3VD Standard-resolution High-resolution P-value Mean abnormal segments per patient 5.3 ± 4.0 7.2 ± 3.8 0.004 Mean abnormal territories per patient 1.6 ± 1.0 2.4 ± 0.9 0.002 Mean perfusion score per patient 11.9 ± 9.4 20.1 ± 7.7 <0.0001 Mean perfusion score per territory 4.7 ± 5.0 5.9 ± 4.3 0.01 When the full spectrum of 105 patients were assessed, high-resolution imaging found a significant upward trend in estimated MIB% in those with significant CAD (n = 70) across advancing disease groups (1VD: 12 ± 7% vs. 2VD: 39 ± 15% vs. 3VD: 42 ± 16%; P < 0.0001).16 However, with standard-resolution, there was no discriminate difference in estimated MIB% across the disease groups (1VD: 21 ± 11% vs. 2VD: 25 ± 6% vs. 3VD: 25 ± 19%; P = 0.53) (Figure 4A). Figure 4 Myocardial ischaemic burden (MIB). (A) In patients with significant coronary artery disease (n = 70), high-resolution perfusion CMR was able to detect significant differences in MIB across disease categories unlike standard-resolution imaging (error bars = SEM). (B) The AUC for detecting 3VD using MIB among patients with suspected angina (n = 105) was significantly greater with high-resolution than standard-resolution (AUC = 0.90 vs. 0.69; P < 0.0001). VD = vessel disease.

icant differences in MIB across disease categories unlike standard-resolution imaging (error bars = SEM). (B) The AUC for detecting 3VD using MIB among patients with suspected angina (n = 105) was significantly greater with high-resolution than standard-resolution (AUC = 0.90 vs. 0.69; P < 0.0001). VD = vessel disease. Accordingly, on ROC analysis, the AUC for detecting angiographic 3VD using the estimated MIB% was significantly greater with high-resolution (AUC = 0.90, 95% CI: 0.84–0.96) than standard-resolution (AUC = 0.69, 95% CI: 0.62–0.76; P < 0.0001) (Figure 4B) imaging. For high-resolution imaging the optimal MIB% threshold to detect angiographic 3VD was 31% which resulted in a sensitivity and specificity of 80% (95% CI: 64–90%) and 90% (95% CI: 80–95%), respectively. For standard-resolution imaging, the optimal threshold was 23%, which resulted in a sensitivity and specificity of 51% (95% CI: 36–67%) and 78% (95% CI: 66–86%), respectively (Figure 4B). Discussion This study shows that in patients with angiographic 3VD, high-resolution perfusion CMR detects a greater ischaemic burden than standard-resolution and more frequently identifies a 3VD pattern of ischaemia due to a higher detection rate of subendocardial ischaemia.

Accordingly, on ROC analysis, the AUC for detecting angiographic 3VD using the estimated MIB% was significantly greater with high-resolution (AUC = 0.90, 95% CI: 0.84–0.96) than standard-resolution (AUC = 0.69, 95% CI: 0.62–0.76; P < 0.0001) (Figure 4B) imaging. For high-resolution imaging the optimal MIB% threshold to detect angiographic 3VD was 31% which resulted in a sensitivity and specificity of 80% (95% CI: 64–90%) and 90% (95% CI: 80–95%), respectively. For standard-resolution imaging, the optimal threshold was 23%, which resulted in a sensitivity and specificity of 51% (95% CI: 36–67%) and 78% (95% CI: 66–86%), respectively (Figure 4B). Discussion This study shows that in patients with angiographic 3VD, high-resolution perfusion CMR detects a greater ischaemic burden than standard-resolution and more frequently identifies a 3VD pattern of ischaemia due to a higher detection rate of subendocardial ischaemia. The evidence for using perfusion CMR in patients with 3VD is limited. Although previous studies evaluating perfusion CMR have included patients with 3VD, they have rarely reported separate results for this specific patient population. In the published data, the incidence of patients with 3VD was only 2% (2 patients) in the study by Sakuma et al., 15% (8 patients) in the study by Ishida et al., and 19% (7 patients) in the study by Schwitter et al.8,23,24 Although other studies may contain larger numbers, patients with 3VD are rarely analysed separately and they are usually grouped together with patients found to have 2VD.7,9,25 Only one study to date has been specifically designed to evaluate standard-resolution perfusion CMR in patients with 3VD (n = 78) and it demonstrated a sensitivity of 85% for CAD detection and superiority of perfusion CMR compared with SPECT.6 Data on the use of high-resolution perfusion CMR in the specific 3VD population are even scarcer. However, in a recent related analysis, we reported on 38 patients with angiographic multi-vessel disease (24 had 3VD) and found a greater diagnostic accuracy for the detection of CAD (any perfusion defect) with high-resolution perfusion CMR compared with standard-resolution imaging (AUC, 0.98 vs.0.91; P < 0.002), but the pattern of perfusion defects detected and the ischaemic burden were not assessed.16

lti-vessel disease (24 had 3VD) and found a greater diagnostic accuracy for the detection of CAD (any perfusion defect) with high-resolution perfusion CMR compared with standard-resolution imaging (AUC, 0.98 vs.0.91; P < 0.002), but the pattern of perfusion defects detected and the ischaemic burden were not assessed.16 Detection of subendocardial ischaemia The finding in the present study indicates that high-resolution acquisition identified significantly more ischaemic segments and in particular more segments with subendocardial ischaemia in angiographic 3VD is consistent with the expected improvement in subendocardial definition with higher spatial resolution. Although we have previously demonstrated this finding across the full spectrum of CAD, it was important to confirm that this advantage is maintained in the 3VD population against the competing challenge of balanced ischaemia.16 It means that one of the major limitations of myocardial perfusion imaging and visual analysis in 3VD, i.e. its dependence on a reference area of normal perfusion, can be overcome to some extent with high-resolution techniques that are able to resolve subendocardial ischaemia and transmural perfusion gradients, reducing the need for intra-patient comparison. An alternative approach is quantitative analysis of standard perfusion CMR and although this has been shown to identify patients with 3VD better than visual analysis, it remains a time-consuming research tool until technical developments lead to greater automation.21 The significant reduction of DRA with high-resolution may also partially account for its improved detection of subendocardial ischaemia and notably there were fewer inconclusive segments with high-resolution acquisition compared with standard-resolution (Figure 3). Previous work investigating DRA has shown the prominent role of spatial resolution on the occurrence and extent of this artefact.17,20

for its improved detection of subendocardial ischaemia and notably there were fewer inconclusive segments with high-resolution acquisition compared with standard-resolution (Figure 3). Previous work investigating DRA has shown the prominent role of spatial resolution on the occurrence and extent of this artefact.17,20 Detection of 3VD pattern Standard-resolution perfusion CMR identified defects in all three perfusion territories in only 29% of patients with angiographic 3VD (10 of 35) (Figure 2B). In the published CMR literature, there is only one study that investigates the same question and it found a significantly greater figure of over 50% using standard-resolution techniques.6 However, this previous retrospective study only analysed patients with 3VD, without a control group; and additionally did not use LGE imaging to exclude areas of infarction from inducible ischaemia, both of which are likely to have led to a positive bias. In the comparative SPECT literature, we see the ability to detect a 3VD pattern is similarly only 29% in one series and as low as 12% in another.2,6

a control group; and additionally did not use LGE imaging to exclude areas of infarction from inducible ischaemia, both of which are likely to have led to a positive bias. In the comparative SPECT literature, we see the ability to detect a 3VD pattern is similarly only 29% in one series and as low as 12% in another.2,6 With the high-resolution perfusion CMR method, twice as many patients with angiographic 3VD, i.e. 57% (20 of 35) were correctly classified as having a 3VD pattern of ischaemia (Figure 2B). The improved detection of a 3VD pattern of ischaemia with high-resolution compared with standard-resolution acquisition was due to better detection of ischaemia in the LCX territory and more subendocardial ischaemia detection. Disease of the LCX territory can be difficult to detect because this territory is farthest from the radiofrequency coil and because visual, unlike quantitative assessment analysis, cannot correct signal intensity for distance from the coil.6,26 In keeping with previous studies, high-resolution acquisition used a higher contrast dose (in order to compensate for the lower SNR associated with smaller voxel size) and arguably this may have contributed to its superior performance in the LCX territory.13,14,16,17

ect signal intensity for distance from the coil.6,26 In keeping with previous studies, high-resolution acquisition used a higher contrast dose (in order to compensate for the lower SNR associated with smaller voxel size) and arguably this may have contributed to its superior performance in the LCX territory.13,14,16,17 Rather than inadequacy of the perfusion analysis, the low detection rate of ischaemia in multiple territories by different imaging modalities may reflect at least in part the inadequacy of the purely anatomical angiographic endpoint in these studies including our own. In a sub-analysis of the FAME study, only 14% of patients with angiographic 3VD (n = 115) had concordant three-vessel functional disease determined by fractional flow reserve (FFR) (i.e. FFR < 0.80 in all three vessels).27

the inadequacy of the purely anatomical angiographic endpoint in these studies including our own. In a sub-analysis of the FAME study, only 14% of patients with angiographic 3VD (n = 115) had concordant three-vessel functional disease determined by fractional flow reserve (FFR) (i.e. FFR < 0.80 in all three vessels).27 Estimation of ischaemic burden It can be argued that in clinical practice, it is less relevant whether a functional scan depicts a typical three-vessel pattern if there is myocardial ischaemia involving a significant proportion of total myocardium. An accurate assessment of ischaemic burden is important because the extent of ischaemia is a marker of patient prognosis—and a large ischaemic burden supports aggressive medical treatment and angiography with a view to revascularization regardless of the territorial pattern of perfusion defect.19,28–30 When MIB% was estimated using high-resolution CMR, it was found to reliably discriminate angiographic 3VD from less extensive disease and normal controls (AUC = 0.89) (Figure 4). This was not the case for standard-resolution CMR, for which MIB% had a significantly poorer diagnostic accuracy (AUC = 0.69; P < 0.0001) and was not able to reliably differentiate between patients with 1VD, 2VD, or 3VD (Figure 4). The latter observation regarding standard-resolution perfusion CMR has been previously noted by Patel et al.21 who found similar estimates of ischaemic burden in patients with angiographic single-vessel disease and 3VD with visual assessment (21% vs. 31%; P = 0.26), but a significant difference if quantified with myocardial perfusion reserve analysis (25% vs. 60%; P = 0.02). A similar phenomenon has been described with positron emission tomography.31 High-resolution perfusion CMR acquisition appears to overcome this limitation seen with lower spatial resolution imaging methods.

), but a significant difference if quantified with myocardial perfusion reserve analysis (25% vs. 60%; P = 0.02). A similar phenomenon has been described with positron emission tomography.31 High-resolution perfusion CMR acquisition appears to overcome this limitation seen with lower spatial resolution imaging methods. Although there is no agreed reference standard for MIB%, the current data suggest that standard-resolution imaging underestimates MIB%—and this should be considered in the interpretation of future perfusion CMR studies that may use a particular threshold of ischaemic burden as a defined end-point or inclusion criteria. In the nuclear sub-study of COURAGE, patients with a MIB% >10% had a lower risk of death or MI if they underwent revascularization rather than optimal medical therapy alone.30 Notably, there has been no direct comparison of CMR and SPECT for MIB% assessment—but if the threshold defined for SPECT is applied to the CMR data in this study, 24 of 35 patients with angiographic 3VD had an MIB% >10% using standard-resolution perfusion CMR, compared with 33 patients with the high-resolution technique. Thus, high-resolution perfusion CMR may offer an improved non-invasive assessment of ischaemic burden and help identify the optimal therapeutic approach.

study, 24 of 35 patients with angiographic 3VD had an MIB% >10% using standard-resolution perfusion CMR, compared with 33 patients with the high-resolution technique. Thus, high-resolution perfusion CMR may offer an improved non-invasive assessment of ischaemic burden and help identify the optimal therapeutic approach. Study limitations Our findings are mainly technical, and further studies with clinical outcome data would be required to support the proposed incremental value of high-resolution perfusion CMR. We also accept that although we hypothesize the superior performance of high-resolution perfusion CMR relates to greater spatial resolution and better detection of subendocardial ischaemia, we cannot exclude the influence of other factors such as differences in SNR between scans and the difference in contrast protocols used as compensation. A functional endpoint such as FFR would have been preferable—but this is not easily achievable given the logistics of performing multiple FFR assessments on serial and complex stenoses in three diffusely diseased arteries to define each subject. However, our findings predominantly relate to comparative differences in ischaemic burden assessment between the two techniques rather than their absolute ability to detect any ischaemia. Because the visual CMR analysis was performed by two observers acting in consensus, inter-observer and intra-observer variability for perfusion scoring was not tested; however, arbitration from a third reader was only required in 5 out of the 210 analyses (3 with standard-and 2 with high-resolution).

A functional endpoint such as FFR would have been preferable—but this is not easily achievable given the logistics of performing multiple FFR assessments on serial and complex stenoses in three diffusely diseased arteries to define each subject. However, our findings predominantly relate to comparative differences in ischaemic burden assessment between the two techniques rather than their absolute ability to detect any ischaemia. Because the visual CMR analysis was performed by two observers acting in consensus, inter-observer and intra-observer variability for perfusion scoring was not tested; however, arbitration from a third reader was only required in 5 out of the 210 analyses (3 with standard-and 2 with high-resolution). Finally, quantitative methods for the estimation of myocardial blood flow (MBF) based on perfusion CMR data have been validated in animal models and applied to clinical studies.21 However, in this study, contrast agent dose and administration (single-bolus technique) were optimized for visual analysis and therefore quantitative analysis was not performed. Although high-resolution perfusion CMR offers further intriguing opportunities for quantitative analysis the algorithms applied for the reconstruction of high-resolution perfusion CMR data acquired with temporospatial undersampling methods give rise to a degree of low-pass temporal filtering, posing additional challenges to quantitation of MBF including an underestimation bias.32,33 Recent developments such as k-t principal component analysis are likely to overcome some of these challenges but will require evaluation in future studies.34

sampling methods give rise to a degree of low-pass temporal filtering, posing additional challenges to quantitation of MBF including an underestimation bias.32,33 Recent developments such as k-t principal component analysis are likely to overcome some of these challenges but will require evaluation in future studies.34 Conclusions High-resolution perfusion CMR increases the observed ischaemic burden and distribution of ischaemia detected in angiographic 3VD. The incremental value of high-resolution acquisition for correctly identifying, stratifying, and managing this high-risk group has to be determined in further clinical studies. Funding S.P. is funded by British Heart Foundation fellowship (FS/10/62/28409). S.P and J.P.G received a research grant from Philips Healthcare. Funding to pay the Open Access publication charges for this article was provided by The British Heart Foundation. Acknowledgements The authors thank Margaret Saysell, Caroline Richmond, and Gavin Bainbridge (radiographers) for their technical assistance; and Petra Bijsterveld and Fiona Richards (research nurses) for their assistance with patient recruitment. Conflicts of interest: none declared.

Introduction Myocardial fibrosis is a common pathological finding in a wide range of cardiovascular diseases and has been associated with an adverse prognosis.1–3 Using cardiovascular magnetic resonance (CMR), the late gadolinium enhancement (LGE) technique has become widely used to evaluate focal myocardial fibrosis. However in many conditions, including aortic stenosis, a more diffuse form of fibrosis predominates, which crucially is reversible and therefore a potential target for novel therapeutic strategies.4–7 LGE imaging has inherent limitations in assessing diffuse fibrosis because it relies upon detecting a difference in signal intensity between normal and fibrotic regions.8 Consequently, it has difficulty in discriminating areas of diffuse myocardial fibrosis, which tend to have an even distribution.

eutic strategies.4–7 LGE imaging has inherent limitations in assessing diffuse fibrosis because it relies upon detecting a difference in signal intensity between normal and fibrotic regions.8 Consequently, it has difficulty in discriminating areas of diffuse myocardial fibrosis, which tend to have an even distribution. Recently, several T1 mapping approaches have been developed to quantify diffuse fibrosis. The first approach measures intrinsic myocardial T1 on the basis that T1 relaxation times are longer in regions of fibrosis (pre-contrast T1).9–11 Alternatively, myocardial T1 can be measured following gadolinium administration, which accumulates in fibrotic areas on account of the increased extracellular volume (post-contrast T1).12,13 However, post-contrast T1 is potentially confounded by individual variations in gadolinium kinetics and by the precise timing of imaging.8 As a result, investigators have proposed methods to correct for these factors using either blood-pool T1 values to derive the partition coefficient (λ),14 or plasma volume to calculate the contrast volume of distribution in the myocardium. The latter is commonly referred to as the myocardial extracellular volume fraction (ECV).15–20 Each of these approaches have been validated against the extent of myocardial fibrosis on histology.12,13,15,17,18,20 However, the optimal technique remains uncertain due to a lack of consistent acquisition sequences and disease states studied, whilst direct comparative studies are relatively lacking.20 In addition, there is insufficient reproducibility data (particularly scan–rescan) and few studies have been performed at 3T,21–24 which may offer potential improvements compared with 1.5T.25

lack of consistent acquisition sequences and disease states studied, whilst direct comparative studies are relatively lacking.20 In addition, there is insufficient reproducibility data (particularly scan–rescan) and few studies have been performed at 3T,21–24 which may offer potential improvements compared with 1.5T.25 Therefore, the purpose of this study was to perform a systematic and comprehensive assessment to determine the optimal T1 approach at 3T. In particular, we aimed to characterize the temporal and regional T1 profiles of the myocardium and to identify the optimal technique based upon its reproducibility and ability to differentiate asymptomatic patients with aortic stenosis from healthy volunteers. Patients with advanced symptoms and focal scarring were excluded so as to focus on patients in whom diffuse myocardial fibrosis is most likely to be of clinical interest. Methods Study participants Twenty asymptomatic patients with mild-to-severe aortic stenosis were recruited from outpatient clinics at the Edinburgh Heart Centre (see Supplementary data online). Twenty healthy volunteers were recruited from the community and the University of Edinburgh. All individuals had normal renal function and a left ventricular ejection fraction within the normal range.

e aortic stenosis were recruited from outpatient clinics at the Edinburgh Heart Centre (see Supplementary data online). Twenty healthy volunteers were recruited from the community and the University of Edinburgh. All individuals had normal renal function and a left ventricular ejection fraction within the normal range. The exclusion criteria for patients with aortic stenosis were as follows: (i) other significant valvular heart disease (moderate to severe in nature), (ii) acquired or inherited cardiomyopathies, (iii) previous myocarditis and (iv) the presence of focal LGE. The exclusion criteria for healthy volunteers were as follows: (i) hypertension, (ii) diabetes mellitus, (iii) coronary artery disease (previous myocardial infarction, evidence of myocardial ischaemia, or >50% luminal stenosis in a major epicardial vessel) (iv) valvular heart disease, (v) cardiomyopathy or previous myocarditis, and (vi) the presence of focal LGE. All clinical assessments and imaging studies were carried out at the Wellcome Trust Clinical Research Facility and the Clinical Research Imaging Centre, Edinburgh. Studies were performed with the approval of the local research ethics committee, and with the written informed consent from each participant.

The exclusion criteria for patients with aortic stenosis were as follows: (i) other significant valvular heart disease (moderate to severe in nature), (ii) acquired or inherited cardiomyopathies, (iii) previous myocarditis and (iv) the presence of focal LGE. The exclusion criteria for healthy volunteers were as follows: (i) hypertension, (ii) diabetes mellitus, (iii) coronary artery disease (previous myocardial infarction, evidence of myocardial ischaemia, or >50% luminal stenosis in a major epicardial vessel) (iv) valvular heart disease, (v) cardiomyopathy or previous myocarditis, and (vi) the presence of focal LGE. All clinical assessments and imaging studies were carried out at the Wellcome Trust Clinical Research Facility and the Clinical Research Imaging Centre, Edinburgh. Studies were performed with the approval of the local research ethics committee, and with the written informed consent from each participant. Imaging protocols Echocardiography Transthoracic echocardiography was performed in all participants (iE33, Philips Medical Systems, The Netherlands) by two independent operators. The severity of aortic stenosis was classified using aortic valve jet velocity, mean pressure gradient, and the aortic valve area, according to the American Heart Association/American College of Cardiology guidelines.26 Diastolic function was assessed using pulse-wave Doppler [early (E) and late (A) mitral inflow velocities] and tissue Doppler imaging (average of medial and lateral annulus velocities of the mitral valve, e′).

he aortic valve area, according to the American Heart Association/American College of Cardiology guidelines.26 Diastolic function was assessed using pulse-wave Doppler [early (E) and late (A) mitral inflow velocities] and tissue Doppler imaging (average of medial and lateral annulus velocities of the mitral valve, e′). CMR imaging CMR was performed at 3T (MAGNETOM Verio, Siemens AG, Healthcare Sector, Erlangen, Germany) according to the study protocol (Supplementary data online). Short-axis cine images were obtained using a balanced steady-state free precession sequence (8-mm parallel slices with 2-mm spacing) for the assessment of left ventricular function and volumes.

MAGNETOM Verio, Siemens AG, Healthcare Sector, Erlangen, Germany) according to the study protocol (Supplementary data online). Short-axis cine images were obtained using a balanced steady-state free precession sequence (8-mm parallel slices with 2-mm spacing) for the assessment of left ventricular function and volumes. T1 mapping was performed using the MOdified Look-Locker Inversion recovery (MOLLI; flip angle 35°; minimum TI 100 ms; TI increment of 80 ms; time delay of 150 ms with a heart beat acquisition scheme of 3-3-5) with built-in motion correction.27 A gradient echo field map and associated shim were performed to minimize off-frequency artefact. Short-axis T1 maps of the mid-cavity slice were acquired in diastole before and at 2, 5, 10, 15, 20, and 30 min following the administration of 0.1 mmol/kg of gadobutrol (Gadovist/Gadavist, Bayer Pharma AG, Germany). Additional basal and apical T1 maps were obtained in diastole at 0, 15, 20, and 30 min (see Supplementary data online). The basal slice was defined as the first complete ring of myocardium below the aortic outflow tract, and the mid-cavity slice as the most basal slice to include both papillary muscles. The apical slice was selected between the apex and the mid-cavity on the image least affected by trabeculations and partial volume averaging. Figure 1 Methodology for measuring myocardial T1 at multiple time points and in multiple segments of the left ventricle (A) Measurement of myocardial T1 at multiple time points. ROI were drawn within the borders on the pre-contrast myocardial T1 maps and then copied onto the corresponding post-contrast images at all time points. Minor adjustments were made to avoid artefact and blood pool. An ROI was also drawn in the left ventricular blood pool in order to calculate the partition coefficient (λ) and extracellular volume fraction (ECV) at each time point. This approach demonstrated excellent intra- and inter-observer reproducibility. (B) Assessment of regional variation in T1 measures. Using the anterior and inferior ventricular insertion points as well as the mid-point of the ventricular cavity as reference points, three intersecting lines were drawn to divide the left ventricle into 16 segments. ROI were drawn onto the basal (six segments), mid-cavity (six segments), and apical (four segments) pre-contrast T1 maps with the standardized approach described above. Subsequently, the ROI were copied onto the 20-min post-contrast T1 maps.

e intersecting lines were drawn to divide the left ventricle into 16 segments. ROI were drawn onto the basal (six segments), mid-cavity (six segments), and apical (four segments) pre-contrast T1 maps with the standardized approach described above. Subsequently, the ROI were copied onto the 20-min post-contrast T1 maps. Pre- and post-contrast T1, λ, and ECV values were assessed in each segment LGE imaging was performed between 8 and 15 min using two approaches: an inversion-recovery fast gradient-echo sequence and a phase-sensitive inversion recovery sequence, performed in two phase-encoding directions to differentiate true enhancement from artefact.28,29 The inversion time was optimized for each slice to achieve satisfactory nulling of the myocardium. Imaging analysis Assessment of the left ventricle and myocardial enhancement The quantification of left ventricular function, volumes, and mass was performed using the Argus Ventricular Function software (Siemens AG Healthcare Sector). All volumes and mass were indexed to body surface area. Qualitative assessment of myocardial enhancement was assessed independently by two experienced operators (M.D. and S.M.). Assessing the quality of T1 maps The quality of T1 map data was examined using the individual inversion recovery images. All segments affected by off-resonance, excessive breathing motion artefacts not corrected by the inline motion correction, and mistriggering were excluded from the analysis.

Imaging analysis Assessment of the left ventricle and myocardial enhancement The quantification of left ventricular function, volumes, and mass was performed using the Argus Ventricular Function software (Siemens AG Healthcare Sector). All volumes and mass were indexed to body surface area. Qualitative assessment of myocardial enhancement was assessed independently by two experienced operators (M.D. and S.M.). Assessing the quality of T1 maps The quality of T1 map data was examined using the individual inversion recovery images. All segments affected by off-resonance, excessive breathing motion artefacts not corrected by the inline motion correction, and mistriggering were excluded from the analysis. Assessing myocardial T1 at multiple time points Myocardial T1 values were assessed at multiple time points to establish the optimal time point for post-contrast T1 mapping. This was defined at the flattest point on the T1 relaxation curve at which variation in T1 values with time was at a minimum. Mid-cavity motion-corrected T1 maps were analysed using a dedicated workstation (OsiriX version 4.1.1, Geneva, Switzerland). To minimize partial volume effects from surrounding tissues and blood pool, we standardized the windowing and placement of regions of interest (ROI) around the mid-cavity myocardium using a pre-defined protocol (Figure 1). The ROI were first drawn on the pre-contrast T1 maps and then copied onto each of the corresponding post-contrast T1 maps with stringent adjustments applied to avoid blood pool and artefact (Figure 1A). This approach ensured consistency in the placement of ROI across the different time points, allowing us to investigate the temporal variation in our T1 measures.

T1 maps and then copied onto each of the corresponding post-contrast T1 maps with stringent adjustments applied to avoid blood pool and artefact (Figure 1A). This approach ensured consistency in the placement of ROI across the different time points, allowing us to investigate the temporal variation in our T1 measures. Derivation of the partition coefficient and ECV Myocardial λ and ECV were also calculated at all time points. These measures were derived from pre- and post-contrast myocardial T1 values corrected for blood-pool T1 (measured at the mid-cavity, Figure 1) and haematocrit (sampled at the time of CMR), according to: (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$$\lambda = \Delta\hbox{R1}_{{\rm myocardium}} /\Delta\hbox{R1}_{{\rm blood}\,{\rm pool}} ,\hbox{where R1} = \hbox{1}/\hbox{T1}$$\end{document} (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$$\hbox{ECV} = ({\hbox{1} - \hbox{hematocrit}} )\times \lambda$$\end{document}