Browse the corpus

ounts. The dose of cisplatin in subsequent cycles was adjusted based on creatinine clearance. Gemcitabine administration was suspended in patients with white blood cell counts less than 2000/mm3, thrombocyte counts less than 5 × 104/mm3 or other non-tolerable grade 3 or higher non-hematological toxicity on day 8 or 15. In patients in the neoadjuvant setting, before starting chemotherapy, the bladder tumor was resected as completely as possible by TUR-BT. Interim analyses were performed after two cycles of chemotherapy by abdominal and pelvic computed tomography (CT) scan or magnetic resonance imaging (MRI). If no tumor response was identified, the neoadjuvant chemotherapy was stopped and then patients received a salvage cystectomy immediately. The clinical response after neoadjuvant chemotherapy was judged by the CT scan or MRI with the use of the response evaluation criteria in solid tumor (RECIST).

Introduction Radical cystectomy with pelvic lymph-node dissection is a standard treatment option in patients with clinically localized muscle-invasive bladder cancer.1,2 However, muscle-invasive bladder cancer has a high potential for systemic disease recurrence, which has been reported to occur in approximately 50% of patients during their clinical course, and then cause their death in almost all of them.3,4 Based on these observations, it is assumed that micrometastases already exist at the time of radical cystectomy. Traditionally, in order to avoid disease recurrence and death, many studies have tested the efficacy of perioperative systemic chemotherapy.5–8

al course, and then cause their death in almost all of them.3,4 Based on these observations, it is assumed that micrometastases already exist at the time of radical cystectomy. Traditionally, in order to avoid disease recurrence and death, many studies have tested the efficacy of perioperative systemic chemotherapy.5–8 The main rationale for early systemic chemotherapy (neoadjuvant) is to eradicate a micrometastasis and reduce the primary bladder tumor volume in order to facilitate the subsequent surgical procedure. For almost 20 years many platinum-based combinations of neoadjuvant chemotherapy have been explored.2,5,9–13 Several clinical trials have demonstrated improved survival benefits in patients who received neoadjuvant chemotherapy.5,7,11 A recent meta-analysis concluded that the addition of neoadjuvant platinum-based chemotherapy to a radical cystectomy provided a 5-year survival advantage of 5% on an additive scale.14,15 Based on the large body of this type of evidence, platinum-based neoadjuvant chemotherapy for patients with muscle-invasive bladder cancer is widely used in daily practice. However, neoadjuvant chemotherapy also includes definite potential disadvantages. It may, for example, delay the radical cystectomy and then cause some patients who cannot achieve a tumor response to become ineligible for receiving a radical cystectomy.

h muscle-invasive bladder cancer is widely used in daily practice. However, neoadjuvant chemotherapy also includes definite potential disadvantages. It may, for example, delay the radical cystectomy and then cause some patients who cannot achieve a tumor response to become ineligible for receiving a radical cystectomy. On the other hand, compared with neoadjuvant chemotherapy, adjuvant chemotherapy for bladder cancer has not provided as much strong evidence of a therapeutic response because of a lack of results from large randomized prospective trials. But adjuvant chemotherapy has already been accepted in daily practice to a certain degree. The main rationale for adjuvant chemotherapy administration is to reduce local and/or metastatic recurrence. The potential disadvantage of adjuvant chemotherapy is that it may cause a delay in micrometastasis eradication and that it may be impossible to administer a full dose of the chemotherapeutic agent due to surgical complications. A few small prospective and many retrospective studies have revealed a significant prolongation of relapse-free survival and the survival benefit in adjuvant chemotherapy.16–18 Based on the present weak evidence, adjuvant chemotherapy cannot be considered as a standard treatment option. Nevertheless, in daily practice, several patients, such as those with severe pollakiuria or hematuria, with renal functions that were inadequate for chemotherapy, have been treated with a radical cystectomy followed by adjuvant chemotherapy. In such cases, adjuvant chemotherapy is frequently utilized in daily practice. However, in order to become a standard treatment option, evidence that supports the equal efficacy of neoadjuvant and adjuvant chemotherapy might be needed. The equal efficacy of neoadjuvant and adjuvant chemotherapy has already been established in other malignancies such as breast cancer.19 However, in muscle-invasive bladder cancer very little evidence is available. No randomized trials have directly compared neoadjuvant with adjuvant chemotherapy. A few retrospective studies have been published but in those the chemotherapy regimen was not gemcitabine plus cisplatin (GC) but almost always methotrexate, vinblastine, doxorubicin and cisplatin.20

r very little evidence is available. No randomized trials have directly compared neoadjuvant with adjuvant chemotherapy. A few retrospective studies have been published but in those the chemotherapy regimen was not gemcitabine plus cisplatin (GC) but almost always methotrexate, vinblastine, doxorubicin and cisplatin.20 We hypothesized that the efficacies of neoadjuvant and adjuvant chemotherapy in patients undergoing a radical cystectomy might be equivalent. In order to test this hypothesis in this retrospective analysis, we reviewed and compared the outcomes, such as recurrence-free survival (RFS), in patients who received neoadjuvant or adjuvant GC chemotherapy at our institution.

and adjuvant chemotherapy in patients undergoing a radical cystectomy might be equivalent. In order to test this hypothesis in this retrospective analysis, we reviewed and compared the outcomes, such as recurrence-free survival (RFS), in patients who received neoadjuvant or adjuvant GC chemotherapy at our institution. Methods We retrospectively reviewed and analyzed the records of patients who were scheduled to undergo a radical cystectomy plus perioperative chemotherapy with GC at the National Cancer Center Hospital East (Kashiwa, Japan). The eligible patients were diagnosed as clinical stage T2–4, N0–2, M0 bladder cancer from April 2005 to December 2010. All patients were confirmed as having pathological muscle-invasive bladder cancer by transurethral resection. In this analysis, the pathological component was not limited to urothelial carcinoma; non-urothelial variants were allowed. Patients receiving other chemotherapy regimens and those with clinical stage < T2, with distant metastasis or with upper tract urothelial carcinoma, were excluded from this analysis. In principle, muscle-invasive or node positive disease was treated with neoadjuvant chemotherapy followed by a radical cystectomy in our institution. The patients who received preceding radical cystectomy followed by adjuvant chemotherapy had a reason for choosing this sequence, such as symptoms with a hematuria necessitating a cystectomy prior to chemotherapy or muscle-invasion of bladder discovered in the cystectomy specimen.

radical cystectomy in our institution. The patients who received preceding radical cystectomy followed by adjuvant chemotherapy had a reason for choosing this sequence, such as symptoms with a hematuria necessitating a cystectomy prior to chemotherapy or muscle-invasion of bladder discovered in the cystectomy specimen. A radical cystectomy with pelvic lymph node dissection was performed by a urological surgeon at our institution. The pelvic lymph node dissection included the hypogastric, external iliac, obturator and distal common iliac lymph nodes. The patients who received a partial cystectomy (organ-sparing surgery) were also excluded from this analysis. In principle, neoadjuvant or adjuvant chemotherapy was administered as four cycles of GC. Patients received gemcitabine 1000 mg/m2 on days 1, 8 and 15 plus cisplatin 70 mg/m2 on day 1 by i.v. infusion. The four cycles of chemotherapy were administered with 4-week intervals between the cycles. Prior to every chemotherapeutic agent infusion, the patients' laboratory data were assessed, such as kidney function by creatinine clearance and bone marrow function by white blood cell and thrombocyte counts. The dose of cisplatin in subsequent cycles was adjusted based on creatinine clearance. Gemcitabine administration was suspended in patients with white blood cell counts less than 2000/mm3, thrombocyte counts less than 5 × 104/mm3 or other non-tolerable grade 3 or higher non-hematological toxicity on day 8 or 15.

onance imaging (MRI). If no tumor response was identified, the neoadjuvant chemotherapy was stopped and then patients received a salvage cystectomy immediately. The clinical response after neoadjuvant chemotherapy was judged by the CT scan or MRI with the use of the response evaluation criteria in solid tumor (RECIST). The primary outcome measure was RFS. The secondary outcome measures were a pathological complete response (pCR) rate and relative-dose intensity (RDI). All toxicity profiles were graded according to the National Cancer Institute common toxicity criteria for adverse events (CTC-AE) version 4.0. Treatment duration was determined from the start date of neoadjuvant chemotherapy administration to the date of the cystectomy in neoadjuvant patients and from the date of the cystectomy to the date of the final adjuvant chemotherapy administration in adjuvant patients. The length of follow up and start of survival analysis was determined from the start date of treatment to the date of recurrence confirmation or last follow up. Survival analysis was performed using the Kaplan–Meier method with patients stratified into neoadjuvant and adjuvant chemotherapy groups. A log–rank test was used to estimate and compare RFS among patients. Proportions were analyzed with the χ2 test. Continuous variables were analyzed with the Mann–Whitney U-test. All tests were two-sided. P < 0.05 was considered to be statistically significant. All analyses were performed according to the intention-to-treat principle. All data analyses were calculated using SPSW statistics 18.0 (SPSS, Chicago, IL, USA).

e χ2 test. Continuous variables were analyzed with the Mann–Whitney U-test. All tests were two-sided. P < 0.05 was considered to be statistically significant. All analyses were performed according to the intention-to-treat principle. All data analyses were calculated using SPSW statistics 18.0 (SPSS, Chicago, IL, USA). Results During the study period, 25 patients received neoadjuvant GC chemotherapy followed by a cystectomy and 17 patients received a cystectomy followed by adjuvant GC. In the adjuvant group, the reason for choosing adjuvant chemotherapy was as follow. Eight patients had symptoms with severe hematuria, seven patients had severe pollakiuria and muscle-invasion was discovered in the cystectomy specimen in two patients. Baseline clinical and pathological characteristics of these groups are summarized in Table 1. Comparing baseline characteristics of the two groups, significant differences were observed in sex (P = 0.016) and clinical node stage (P = 0.026). The neoadjuvant group was slightly older and at a lower clinical T stage than the adjuvant group, but this difference was not significant. No significant difference was observed among their other characteristics. Table 1 Patients' characteristics

Results During the study period, 25 patients received neoadjuvant GC chemotherapy followed by a cystectomy and 17 patients received a cystectomy followed by adjuvant GC. In the adjuvant group, the reason for choosing adjuvant chemotherapy was as follow. Eight patients had symptoms with severe hematuria, seven patients had severe pollakiuria and muscle-invasion was discovered in the cystectomy specimen in two patients. Baseline clinical and pathological characteristics of these groups are summarized in Table 1. Comparing baseline characteristics of the two groups, significant differences were observed in sex (P = 0.016) and clinical node stage (P = 0.026). The neoadjuvant group was slightly older and at a lower clinical T stage than the adjuvant group, but this difference was not significant. No significant difference was observed among their other characteristics. Table 1 Patients' characteristics Neoadjuvant Adjuvant P-value Number of patients 25(%) 17(%) Age, years Mean 65 65 0.849 Median 67 64 0.877 Range 47–79 50–76 Sex(n, %) 0.016 Male 15(60) 16(94) Female 10(40) 1(6) Clinical T stage(n, %) 0.391 ≤cT2 9(36) 4(24) >cT2 16(64) 13(77) Clinical N stage(n, %) 0.026 cN0 16(64) 16(94) cN1 or 2 9(36) 1(6) Histology 0.298 Pure urothelial carcinoma 23(84) 13(77) Non-urothelial carcinoma or mixed component 4(16) 4(24) In principle, four cycles of GC chemotherapy were planned in both neoadjuvant and adjuvant settings. The mean and median numbers of cycles of GC between the two groups were not significantly different (neoadjuvant vs adjuvant; median 4.0 vs 4.0, P = 0.166; mean 3.80 vs 3.53, P = 0.073). Only one and two patients in the neoadjuvant and adjuvant groups, respectively, dropped out of chemotherapy due to non-tolerable toxicity. The hematological toxicity profiles of GC chemotherapy are listed in Table 2. The incidence rates of grade 3 or 4 thrombocytopenia, anemia and neutropenia were not significantly different between the two groups (data not shown). The RDI of gemcitabine and cisplatin showed a tendency to be higher in the neoadjuvant than in the adjuvant group but these differences were not statistically significant (neoadjuvant vs adjuvant; gemcitabine; 82 vs 78%; P = 0.460, cisplatin; 95 vs 82%; P = 0.073). The median treatment duration was 134 days for patients in the neoadjuvant group and 150 days for those in the adjuvant group. This difference in treatment duration between the two groups was statistically significant (P = 0.016). More details of the treatment procedure are shown in Table 3.

tin; 95 vs 82%; P = 0.073). The median treatment duration was 134 days for patients in the neoadjuvant group and 150 days for those in the adjuvant group. This difference in treatment duration between the two groups was statistically significant (P = 0.016). More details of the treatment procedure are shown in Table 3. Tabel 2 Hematological toxicity of perioperative gemcitabine plus cisplatin regimen (n, %) Neoadjuvant (n = 25) Adjuvant (n = 17) Grade 1–2 Grade 3 Grade 4 Grade 1–2 Grade 3 Grade 4 Number of patients (%) Anemia 17(68) 8(32) 0 15(88) 2(12) 0 Thrombocytopenia 14(56) 7(28) 3(12) 9(53) 3(17) 2(12) Neutropenia 13(52) 7(28) 3(12) 8(47) 5(29) 1(5.8) Febrile neutropenia – 1(4) 1(4) – 1(5.8) 0 Table 3 Treatment procedure

juvant (n = 25) Adjuvant (n = 17) Grade 1–2 Grade 3 Grade 4 Grade 1–2 Grade 3 Grade 4 Number of patients (%) Anemia 17(68) 8(32) 0 15(88) 2(12) 0 Thrombocytopenia 14(56) 7(28) 3(12) 9(53) 3(17) 2(12) Neutropenia 13(52) 7(28) 3(12) 8(47) 5(29) 1(5.8) Febrile neutropenia – 1(4) 1(4) – 1(5.8) 0 Table 3 Treatment procedure Neoadjuvant Adjuvant P-value Number of gemcitabine plus cisplatin cycles Median 4.0 4.0 0.166 Mean 3.80 3.53 0.073 Range 2–4 1–4 Relative dose intensity (%) Gemcitabine 82 78 0.460 Cisplatin 95 82 0.073 Treatment duration, days Median 134.0 150.0 0.016 Mean 131.7 153.4 0.023 Range 62–195 94–246 Follow-up, months 0.377 Median 25.1 33.8 Range 3.4–58.4 6.8–64.9 The median follow-up period was 28.6 months in all patients (range; 3.4–64.9 months). The difference of median follow-up period between the two groups was not statistically significant (P = 0.377). During the follow-up periods, 12 patients (29%) experienced metastatic recurrences. Recurrence was observed in nine and three patients in the neoadjuvant and adjuvant groups, respectively. Kaplan–Meier curves of RFS stratified by treatment group are shown in Figure 1. The RFS rate at median follow up was 66.7 and 76%, in the neoadjuvant and adjuvant groups, respectively. No significant difference in RFS at median follow up was detected between the two groups (P = 0.124). Figure 1 Disease-free survival curves stratified by patients on (----) neoadjuvant and (____) adjuvant chemotherapy. Log–rank P = 0.124.

Neoadjuvant Adjuvant P-value Number of gemcitabine plus cisplatin cycles Median 4.0 4.0 0.166 Mean 3.80 3.53 0.073 Range 2–4 1–4 Relative dose intensity (%) Gemcitabine 82 78 0.460 Cisplatin 95 82 0.073 Treatment duration, days Median 134.0 150.0 0.016 Mean 131.7 153.4 0.023 Range 62–195 94–246 Follow-up, months 0.377 Median 25.1 33.8 Range 3.4–58.4 6.8–64.9 The median follow-up period was 28.6 months in all patients (range; 3.4–64.9 months). The difference of median follow-up period between the two groups was not statistically significant (P = 0.377). During the follow-up periods, 12 patients (29%) experienced metastatic recurrences. Recurrence was observed in nine and three patients in the neoadjuvant and adjuvant groups, respectively. Kaplan–Meier curves of RFS stratified by treatment group are shown in Figure 1. The RFS rate at median follow up was 66.7 and 76%, in the neoadjuvant and adjuvant groups, respectively. No significant difference in RFS at median follow up was detected between the two groups (P = 0.124). Figure 1 Disease-free survival curves stratified by patients on (----) neoadjuvant and (____) adjuvant chemotherapy. Log–rank P = 0.124. The clinical and pathological responses in patients receiving neoadjuvant chemotherapy are shown in Table 4. The clinical response and complete response rates after neoadjuvant chemotherapy were 60% (15/25) and 44% (11/25), respectively. Three patients in the neoadjuvant group (12%) showed no tumor response after two cycles of neoadjuvant GC chemotherapy by radiological assessment and were then defined as having progressive disease (PD). The three patients who had experienced PD stopped neoadjuvant chemotherapy and underwent salvage cystectomy immediately. The pathological complete response (pT0) rate and downstaging (<pT2) rate after neoadjuvant chemotherapy were 40% (10/25) and 44% (11/25), respectively. None of the patients who achieved pT0 showed disease recurrence. Kaplan–Meier survival curves stratified by pT0 and non-pT0 in patients receiving neoadjuvant GC chemotherapy are shown in Figure 2. The RFS rates at median follow up were 100 and 50%, in patients with pT0 and non-pT0, respectively. The difference in RFS between these groups at median follow up was significant (P = 0.018). On the other hand, in the adjuvant group, no patient achieved pT0 in the cystectomy specimens.

emotherapy are shown in Figure 2. The RFS rates at median follow up were 100 and 50%, in patients with pT0 and non-pT0, respectively. The difference in RFS between these groups at median follow up was significant (P = 0.018). On the other hand, in the adjuvant group, no patient achieved pT0 in the cystectomy specimens. Figure 2 Recurrence-free survival curves stratified by (____) pathological complete response (pT0) and (----) non-pT0 in patients in neoadjuvant group. Log–rank P = 0.018. Table 4 Clinical and pathological outcomes of neoadjuvant chemotherapy (N = 25) Number of patients (%) Clinical outcome Complete response 11(44) Partial response 4(16) Stable disease 7(28) Progressive disease 3(12) Pathological outcome pT0 10(40) pT1 1(4) ≥pT2 14(56) pT, pathological complete response. Positive pelvic node metastasis, which was analyzed in surgical specimens, was present in six patients in each of the neoadjuvant and adjuvant groups. The pathological node metastasis rates were not significantly different between the two groups (25 vs 35%; P = 0.49).

Number of patients (%) Clinical outcome Complete response 11(44) Partial response 4(16) Stable disease 7(28) Progressive disease 3(12) Pathological outcome pT0 10(40) pT1 1(4) ≥pT2 14(56) pT, pathological complete response. Positive pelvic node metastasis, which was analyzed in surgical specimens, was present in six patients in each of the neoadjuvant and adjuvant groups. The pathological node metastasis rates were not significantly different between the two groups (25 vs 35%; P = 0.49). Discussion Perioperative systemic chemotherapy, especially neoadjuvant chemotherapy added to radical cystectomy, has been demonstrated in many randomized trials to provide a significant survival benefit.5,7 Based on this level I evidence, at our institution neoadjuvant chemotherapy followed by radical cystectomy has been increasingly used in patients with muscle-invasive bladder cancer. However, in almost all previous trials the chemotherapy regimen was not GC but other platinum-based combinations. To the best of our knowledge, studies with the GC regimen as neoadjuvant chemotherapy were performed only as minimal retrospective analyses.21,22 In addition, information about the usefulness of adjuvant chemotherapy is limited because of a lack of results from large randomized trials. No prospective studies or small retrospective analyses have directly compared neoadjuvant with adjuvant GC chemotherapy in patients with muscle-invasive bladder cancer. Despite the limited evidences of the benefit of adjuvant chemotherapy, especially the GC regimen, adjuvant chemotherapy already is widely used in daily practice.

tive studies or small retrospective analyses have directly compared neoadjuvant with adjuvant GC chemotherapy in patients with muscle-invasive bladder cancer. Despite the limited evidences of the benefit of adjuvant chemotherapy, especially the GC regimen, adjuvant chemotherapy already is widely used in daily practice. This aim of this investigation was to compare the efficacy of the GC regimen in neoadjuvant and adjuvant settings. The results revealed no significant difference in RFS between the two modes of administration in patients with muscle-invasive bladder cancer. Our finding thus confirms two hypotheses. First, the efficacy of chemotherapy in neoadjuvant and adjuvant settings might be equivalent. Second, the GC regimen might be a candidate for perioperative standard chemotherapy. The result of the non-inferior RFS comparison between neoadjuvant and adjuvant administrations in this investigation is similar to that of the previous retrospective study.20 However, in the previous study the chemotherapy regimen was not limited to GC. All platinum-based chemotherapy regimens were allowed and the proportion of patients who received GC was only 35%. In addition, a subgroup analysis of the GC regimen in the previous study showed that the disease-specific survival was significantly worse in the adjuvant than in the neoadjuvant setting (hazard ratio 10.6; P = 0.049). This result was different from that of our study. Moreover, PFS curves were not significantly different but clearly separated in our study. There was some possibility that the P-value was not significant simply due to under-powering or a small sample size. There might be a number of confounding features and/or biases between the previous study and ours, such as patient selection, treatment decisions and RDI. Therefore, we cannot directly compare these results.

e was some possibility that the P-value was not significant simply due to under-powering or a small sample size. There might be a number of confounding features and/or biases between the previous study and ours, such as patient selection, treatment decisions and RDI. Therefore, we cannot directly compare these results. Other previous prospective trials have also supported the hypothesis of equal efficacy of the two modes of administration in perioperative chemotherapy. A trial at the University of Texas MD Anderson Cancer Center evaluated 140 patients who were randomized to a radical cystectomy with either five cycles of adjuvant MVAC or two cycles of neoadjuvant MVAC with three cycles of adjuvant MVAC. The result from this study demonstrated that the difference in disease-specific survival and overall survival was not significantly different between the two arms.8 However, the chemotherapy regimen in the MD Anderson trial also was not GC. In our investigation there was no significant difference in the RDI of neoadjuvant and adjuvant GC administrations. According to this result, a potential disadvantage of adjuvant administration is eliminated by using neoadjuvant administration, such as surgical complications preventing administration of a full dose of the chemotherapeutic agent. Based on these results, including those of our investigation, the timing of perioperative chemotherapy is less important than whether or not patients actually receive it. Future prospective randomized trials (neoadjuvant vs adjuvant) of GC chemotherapy are thus warranted.

n of a full dose of the chemotherapeutic agent. Based on these results, including those of our investigation, the timing of perioperative chemotherapy is less important than whether or not patients actually receive it. Future prospective randomized trials (neoadjuvant vs adjuvant) of GC chemotherapy are thus warranted. The efficacy equivalence of GC and MVAC has been generally accepted and GC has been adopted due to its being less toxic but the previous study was powered only to show superiority of GC and the results showed a trend towards equivalence.23 Better tolerability is an important factor in considering whether to introduce perioperative chemotherapy in daily practice. However, there have been no prospective trials reported of GC in neoadjuvant and/or adjuvant settings, but only a few retrospective studies.21,22 The tolerability gleaned from the RDI and the interruption rate seen in our investigation are satisfactory and similar to those seen in previous analyses. The concepts of RFS or disease-free survival are in general used as efficacy measurements in perioperative chemotherapy. The pCR may be relied on as a surrogate marker for efficacy measurement in a neoadjuvant setting for muscle-invasive bladder cancer. According to the subset analysis of the Southwest Oncology Group 8710 trial, the median overall survival in patients who achieved pT0 was 13.6 years.5 This duration is surprisingly long and may be regarded as a cure. From our results, no disease recurrence was observed in patients who achieved pT0. In addition, pCR appears to be a reasonable and suitable end-point for assessing the efficacy of various neoadjuvant chemotherapy regimens. The pCR rates were 38 and 32%, respectively, with MVAC in the SWOG8710 trial and with cisplatin, methotrexate and vinblastine in the Medical Research Council NC trial.9 In our investigation with GC, 40 and 44% of patients in the neoadjuvant setting achieved pCR and <pT2 at cystectomy, respectively. These results are almost the same as those of the previous trial with non-GC. Thus, the GC regimen seems to be suitable for neoadjuvant administration. However, various outcomes have been reported for the neoadjuvant GC, but the results are confusing. For example, Dash et al.

d pCR and <pT2 at cystectomy, respectively. These results are almost the same as those of the previous trial with non-GC. Thus, the GC regimen seems to be suitable for neoadjuvant administration. However, various outcomes have been reported for the neoadjuvant GC, but the results are confusing. For example, Dash et al. used a 21-day schedule for four cycles of GC and reported that 26 and 36% of patients achieved pT0 and <pT2, respectively.24 The Cleveland Clinic used a standard administration of three cycles of GC and reported that only 7 and 31% of patients achieved pT0 and <pT2, respectively.25 There are several possible explanations for these varying and confusing results associated with the GC regimen. Major reasons may be that the planned neoadjuvant GC cycles were different in the studies and that the bladder tumor was resected by TUR-BT as completely as possible before neoadjuvant chemotherapy, which is the practice in some institutions including ours. These procedural differences might affect the pCR rate. In addition, there might be many minor reasons for the differences, such as baseline characteristics and RDI. However, a high pCR rate with four cycles of GC is impressive, and further studies are warranted to confirm this result.

some institutions including ours. These procedural differences might affect the pCR rate. In addition, there might be many minor reasons for the differences, such as baseline characteristics and RDI. However, a high pCR rate with four cycles of GC is impressive, and further studies are warranted to confirm this result. The strength of our analysis may be summarized as follows: first, direct access to all medical records allowed for accurate documentation of all information. Second, none of our patients were lost to follow up. Third, a consistent chemotherapy drug delivery was assured due to patients being treated at single institution analysis, even though it was an off-protocol study. On the other hand, our study also has several limitations. The most important of these is that it is a retrospective study. Other limitations would be its small sample size and biases such as different baseline characteristics. However, we believe that our results might be useful in daily practice and for planning further prospective trials.

study also has several limitations. The most important of these is that it is a retrospective study. Other limitations would be its small sample size and biases such as different baseline characteristics. However, we believe that our results might be useful in daily practice and for planning further prospective trials. A review from the National Cancer Database with over 7000 patients demonstrated that adjuvant chemotherapy was used in 10% of patients, compared with 2% for neoadjuvant chemotherapy from 1998 to 2003.26 The reason why neoadjuvant chemotherapy is not generally used may be that the optimal chemotherapy cycles have not been definitely established. Further trials are needed to determine the optimal cycles for both neoadjuvant and adjuvant chemotherapy. In fact, further prospective randomized trial of neoadjuvant and adjuvant chemotherapy are expected, but they seem to be impossible for number of reasons, such as need for a huge sample size and the unethical deviation from level 1 evidence of treatment not with neoadjuvant chemotherapy but with adjuvant chemotherapy for muscle-invasive bladder cancer. I believe a large sample size and many retrospective studies are the best we can get. In addition, new methodological approaches, such as determining the p53 status and gene expression profile to establish chemo-sensitivity, may be expected to be introduced in daily practice in the near future.

invasive bladder cancer. I believe a large sample size and many retrospective studies are the best we can get. In addition, new methodological approaches, such as determining the p53 status and gene expression profile to establish chemo-sensitivity, may be expected to be introduced in daily practice in the near future. In conclusion, our results demonstrated there was no significant difference in RFS between neoadjuvant and adjuvant GC chemotherapy. We expect to validate these findings in a prospective randomized trial.

Introduction Gastric cancer has one of the highest incidences of cancer worldwide. Invasion is a major cause of recurrence and patient death in gastric cancer. Ras-related C3 botulinum toxin substrate 1 (Rac1) is an important member of the small molecule G-protein Rho family (Ras homologue) and is an important class of intracellular signaling molecules. It affects tumor growth, invasion and metastasis, and tumor angiogenesis.1,2 P21-activated kinase 1 (Pakl) is a conserved serine/threonine protein kinase that is an important downstream target protein of Rho-GTPase Cdc42 and Rac1, which are involved in many important cellular activities and play an important role in cytoskeletal reorganization, cell migration, apoptosis and survival, cell cycle, gene transcription regulation and cell transformation.3,4 The Rock1 gene is highly expressed abnormally in a variety of tumor cells and plays a role in tumor cell invasion and metastasis.5 At present, little research has focused on the relationship between the expression of Rac1, Pak1 and Rock1 in gastric carcinoma and clinical pathology. In the present study, immunohistochemistry and a tissue microarray were used in the detection of Rac1, Pak1 and Rock1 protein expression levels in gastric cancer cells, intraepithelial neoplasia and normal tissues. The correlation between lymph node metastasis and TNM stages was analyzed.

cinoma and clinical pathology. In the present study, immunohistochemistry and a tissue microarray were used in the detection of Rac1, Pak1 and Rock1 protein expression levels in gastric cancer cells, intraepithelial neoplasia and normal tissues. The correlation between lymph node metastasis and TNM stages was analyzed. Methods Clinical data The specimens (resection specimens of 158 cases of gastric cancer) were recruited from the Department of Pathology, Xiangtan Affiliated Clinical College of Nanhua University from 2004 to 2010. All specimens were sorted according to the World Health Organization classification of digestive system cancer in 2010,6 and confirmed using hematoxylin–eosin (HE) slice biopsy. The patients had not received radiotherapy and chemotherapy before surgery. Tissue microarray The primary tumor tissue was taken. Based on the slice determined using H–E staining, the representative lesion distribution was determined to construct the tissue microarray. Immunohistochemistry Substance P (SP) immunohistochemistry was carried out according to the manufacturer's instructions. Phosphate buffered saline instead of primary antibodies were used as the negative control. A known cancer-positive biopsy was used as the positive control.

Tissue microarray The primary tumor tissue was taken. Based on the slice determined using H–E staining, the representative lesion distribution was determined to construct the tissue microarray. Immunohistochemistry Substance P (SP) immunohistochemistry was carried out according to the manufacturer's instructions. Phosphate buffered saline instead of primary antibodies were used as the negative control. A known cancer-positive biopsy was used as the positive control. Results determination The results were determined according to the method described by Wang et al.7 The positive rate was determined by three pathology experts. We checked 10 high power fields from each slice randomly, and made a positive cell count score and staining intensity score. Rac1, Pak1 and Rock1 protein positive expression consisted of tan or brown granules located in the cytoplasm and/or cell membrane. Based on the degree of positive staining and the percentage of stained cells, the specimens were scored as follows: 0 corresponds to unstained, 1 point corresponds to brown and 2 points correspond to dark brown; 0 for stained cells <5%, 1 for 5–25%, 2 for 26–50% and 3 for above 50%. Two kinds of scoring were employed: ≥2 points was considered positive, <2 points was considered negative. Statistical analysis The results were analyzed via a χ2 test using SPSS 17.0 (SPSS, Chicago, IL, USA) statistical software, and differences of P < 0.05 were considered statistically significant.

Results determination The results were determined according to the method described by Wang et al.7 The positive rate was determined by three pathology experts. We checked 10 high power fields from each slice randomly, and made a positive cell count score and staining intensity score. Rac1, Pak1 and Rock1 protein positive expression consisted of tan or brown granules located in the cytoplasm and/or cell membrane. Based on the degree of positive staining and the percentage of stained cells, the specimens were scored as follows: 0 corresponds to unstained, 1 point corresponds to brown and 2 points correspond to dark brown; 0 for stained cells <5%, 1 for 5–25%, 2 for 26–50% and 3 for above 50%. Two kinds of scoring were employed: ≥2 points was considered positive, <2 points was considered negative. Statistical analysis The results were analyzed via a χ2 test using SPSS 17.0 (SPSS, Chicago, IL, USA) statistical software, and differences of P < 0.05 were considered statistically significant. Results Clinical data The subjects included 112 men and 46 women with a mean age of 56.25 years (28–83-years old). Gastric cancer with lymph node metastasis was found in 109 cases, and 49 cases did not exhibit lymph node metastasis. A total of 67 cases were classified into TNM stages I to II and 91 were diagnosed with stages III–IV. Intraepithelial neoplastic tissue was collected in 54 cases and normal gastric mucosa (gastric resection specimens from the foci of the cancer more than 10 cm of normal gastric mucosa as the control group) was collected from 64 cases.

ses were classified into TNM stages I to II and 91 were diagnosed with stages III–IV. Intraepithelial neoplastic tissue was collected in 54 cases and normal gastric mucosa (gastric resection specimens from the foci of the cancer more than 10 cm of normal gastric mucosa as the control group) was collected from 64 cases. Expression of Rac1, Pak1 and Rock1 protein The expression of Rac1, Pak1 and Rock1 in normal epithelium and intraepithelial neoplastic epithelium was weakly positive or positive; however, a large number of positively stained cells were heterogeneously distributed in the gastric carcinoma tissues. Positive staining was tan-yellow, with bulky granules that were heterogeneously distributed in the cell membrane and cytoplasm. The Rac1 expression rates in the normal gastric tissue, intraepithelial neoplastic tissue and gastric carcinoma were 27, 43 and 68 percent, respectively. The Pak1 expression rates in the three groups were 20, 35 and 60%, respectively, whereas those of Rock1 in the three groups were 16, 28, and 58%, respectively. The differences among the groups were statistically significant (P < 0.05) (Fig. 1). Figure 1 (a) Positive Rac1 in gastric tube adenocarcinoma (substance P (SP) × 400); (b) positive Pak1 in mixed gastric adenocarcinoma (SP × 400); (c) positive Rock1 in stomach cancer of poor adhesion (SP 400).

Expression of Rac1, Pak1 and Rock1 protein The expression of Rac1, Pak1 and Rock1 in normal epithelium and intraepithelial neoplastic epithelium was weakly positive or positive; however, a large number of positively stained cells were heterogeneously distributed in the gastric carcinoma tissues. Positive staining was tan-yellow, with bulky granules that were heterogeneously distributed in the cell membrane and cytoplasm. The Rac1 expression rates in the normal gastric tissue, intraepithelial neoplastic tissue and gastric carcinoma were 27, 43 and 68 percent, respectively. The Pak1 expression rates in the three groups were 20, 35 and 60%, respectively, whereas those of Rock1 in the three groups were 16, 28, and 58%, respectively. The differences among the groups were statistically significant (P < 0.05) (Fig. 1). Figure 1 (a) Positive Rac1 in gastric tube adenocarcinoma (substance P (SP) × 400); (b) positive Pak1 in mixed gastric adenocarcinoma (SP × 400); (c) positive Rock1 in stomach cancer of poor adhesion (SP 400). Relationship between Rac1, Pak1 and Rock1 protein expression and clinicopathological indicators As shown in Table 1, Rac1, Pak1 and Rock1 expression levels and patients' sex were not statistically significant (P > 0.05). Rac1, Pak1 and Rock1 expression in the lymph node metastasis group (75, 71 and 66%) were significantly higher than in the group without metastasis (51, 43 and 41%). Rac1 expression in stages I and II was 48 percent, which was significantly lower than that in stages III and IV (82%). Pak1 expression in stages I and II was 46 percent, which was significantly lower than that in stages III and IV (74%). Rock1 expression in stages I and II was 42 percent, which was significantly lower than that of stages III and IV (70%) (P < 0.05).

t, which was significantly lower than that in stages III and IV (82%). Pak1 expression in stages I and II was 46 percent, which was significantly lower than that in stages III and IV (74%). Rock1 expression in stages I and II was 42 percent, which was significantly lower than that of stages III and IV (70%) (P < 0.05). Table 1 Relationship between the expression of Rac1, Pak1 Rock1 and gastric cancer clinical and pathological features Rac1 Pak1 Rock1 Clinical pathological features Positive rate (%) P value Positive rate (%) P value Positive rate (%) P value Sex Male 112 73 (65) 0.350 67 (60) 0.586 63 (56) 0.481 Female 46 34 (74) 31 (68) 29 (63) Group Normal epithelial 64 17 (27) 0.000 13 (20) 0.000 10 (16) 0.000 Epithelial neoplasia 54 23 (43) 19 (35) 15 (28) Gastric cancer 158 107 (68) 98 (62) 92 (58) Lymph node metastasis No 49 25 (51) 0.003 21 (44) 0.001 20 (41) 0.005 Yes 109 82 (75) 77 (71) 72 (66) TNM classification I, II stage 67 32 (48) 0.000 31 (46) 0.001 28 (49) 0.001 III, IV stage 91 75 (82) 67 (74) 64 (70) [Correction added on 19 March 2013, after first online publication: Positive rate (%) of Rac1 for Male was amended to be 65.] The relationship between Rac1, Pak1, and Rock1 expression and clinical pathology The correlation among the expression of Rac1, Pak1 and Rock1 in gastric cancer was analyzed using Spearman's rank correlation. Rac1, Pak1 and Rock1 expression in gastric cancer was positively correlated (r = 0.555, P < 0.05).

Clinical pathological features Positive rate (%) P value Positive rate (%) P value Positive rate (%) P value Sex Male 112 73 (65) 0.350 67 (60) 0.586 63 (56) 0.481 Female 46 34 (74) 31 (68) 29 (63) Group Normal epithelial 64 17 (27) 0.000 13 (20) 0.000 10 (16) 0.000 Epithelial neoplasia 54 23 (43) 19 (35) 15 (28) Gastric cancer 158 107 (68) 98 (62) 92 (58) Lymph node metastasis No 49 25 (51) 0.003 21 (44) 0.001 20 (41) 0.005 Yes 109 82 (75) 77 (71) 72 (66) TNM classification I, II stage 67 32 (48) 0.000 31 (46) 0.001 28 (49) 0.001 III, IV stage 91 75 (82) 67 (74) 64 (70) [Correction added on 19 March 2013, after first online publication: Positive rate (%) of Rac1 for Male was amended to be 65.] The relationship between Rac1, Pak1, and Rock1 expression and clinical pathology The correlation among the expression of Rac1, Pak1 and Rock1 in gastric cancer was analyzed using Spearman's rank correlation. Rac1, Pak1 and Rock1 expression in gastric cancer was positively correlated (r = 0.555, P < 0.05). Relationship between Rac1, Pak1 and Rock1 expression and survival of gastric cancer patients Kaplan–Meier survival curves showed that the median survival time of the 158 patients with follow up was 26 months. The 3-year and 5-year overall survival rates of the whole group were 47 and 41 percent, respectively. Median survival was significantly shorter in Rac1, Pak1 and Rock-positive patients than in Rac1, Pak1 and Rock-negative patients (19 months vs 78 months, χ2 = 6.857, P = 0.009) (Fig. 2). Univariate survival analysis showed that Rac1, Pak1 and Rock1 expression was a risk factor affecting the survival of the patients.

y. Median survival was significantly shorter in Rac1, Pak1 and Rock-positive patients than in Rac1, Pak1 and Rock-negative patients (19 months vs 78 months, χ2 = 6.857, P = 0.009) (Fig. 2). Univariate survival analysis showed that Rac1, Pak1 and Rock1 expression was a risk factor affecting the survival of the patients. Figure 2 Relationship between Rac1, Pak1 and Rock1 expression and survival of gastric cancer patients showing () positive, () negative, () positive deletion and () negative deletion groups. The results of multivariate analysis by Cox regression for all patients among the four prognostic factors (tumor differentiation, lymph node metastasis, TNM stage and expression of the markers) showed that the expression of the markers may be recognized as the significant independent factor related to disease-free survival (χ2 = 17.594, P < 0.001). Discussion The characteristics of the cytoskeletal structure result in different exercise capacities, which are related to the genetic diversity of tumor cells and normal cells and different metastatic potentials of the tumor cells.8 The cytoskeleton is the intracellular structure mainly composed of protein fiber. It plays an important role in maintaining cell morphology and the internal structure of cells, as well as cell movement, material transport, energy conversion, information transfer and cell differentiation.9

otentials of the tumor cells.8 The cytoskeleton is the intracellular structure mainly composed of protein fiber. It plays an important role in maintaining cell morphology and the internal structure of cells, as well as cell movement, material transport, energy conversion, information transfer and cell differentiation.9 The Rac1 gene, located on the short arm of human chromosome 7 (7p22), encodes a small G-protein that is an important member of the Rho (Ras homologue) family. The Rac1 protein has two states, Rac1-GDP (inactive) and Rac1-GTP (activated).2 The biological functions of Rac1 depends on its conversion between the two states. When Rac1 is activated it participates in the formation of actin stress fibers and adhesion plaques, promotes cytoskeleton reorganization, regulates sheet pseudopodia and filopodia extension, affects the structure and polarization of the cell, promotes cell motility and migration and inhibits apoptosis.10 Studies in recent years have shown that Rac1 expression in colon, breast and lung cancer, among others, was significantly increased. Moreover, Rac1 expression is closely related to invasion and metastasis.11,12 Rac1 cell motility signaling, which promotes cancer cell invasion and metastasis, is quite complex. This process may be achieved through the following ways: (i) activated Rac1 promotes the assembly of cell surface integrin protein molecules in the surface of the head of the cells and passes regulative signals to the actin cytoskeleton, thus inducing actin filaments to aggregate in the plasma membrane and form sheet pseudopodia, leading to cell membrane multi-polarization, which eventually affects the movement of cell migration'2 (ii) Rac1, through the activation of type IV collagenase type 2 matrix metalloproteinase to increase collagenase type I expression, promotes extracellular matrix degradation and enhances the penetration ability of tumor cells.13 (iii) Rac1 regulates nuclear factor kappa-light-chain-enhancer of activated B cells activity and increases intracellular superoxide anion concentration to suppress apoptosis.14

ase to increase collagenase type I expression, promotes extracellular matrix degradation and enhances the penetration ability of tumor cells.13 (iii) Rac1 regulates nuclear factor kappa-light-chain-enhancer of activated B cells activity and increases intracellular superoxide anion concentration to suppress apoptosis.14 Pakl is a class of evolutionarily conserved serine/threonine protein kinase that is widely expressed in many tissues as downstream target proteins of the small molecule G-protein Rho family Cdc42 and Rac l. Pakl can be activated by growth factors and other extracellular signals, either through the GTPase-dependent signaling pathway or not. It has a variety of biological effects. PAK, as an important biological regulator, plays an important role in a series of cell functions of mammals, such as cell motility, cell survival, cell cycle, angiogenesis and the regulation of gene transcription. Recent research indicates that Pakl activation of lysophosphatidic acid and toxins from the body induces cell motility in melanoma cells.15 Head and neck cancer was found to have higher Pakl activity than normal tissue.16 In Kaposi's sarcoma the activity of Pakl is enhanced.4 Combined analysis using gene hybridization array and tissue microarray confirmed that Pakl is an upregulated key cancer target gene and it is positively correlated with cyclin D1 expression.17 Studies have shown that with increasing Pakl expression, the malignant evolution of colorectal cancer is increased.18 Moreover, 55 percent of breast cancer had a high expression of Pakl.19

icroarray confirmed that Pakl is an upregulated key cancer target gene and it is positively correlated with cyclin D1 expression.17 Studies have shown that with increasing Pakl expression, the malignant evolution of colorectal cancer is increased.18 Moreover, 55 percent of breast cancer had a high expression of Pakl.19 Rock (Rho-associated kinases) are direct downstream target proteins of RhoA, which are involved in a variety of cell functions, such as smooth muscle contraction, cytoskeleton construction, cell adhesion and movement and gene expression. The Rock gene has two subtypes, Rock1 and Rock2. Rock1 is located on chromosome 18 and encodes a 1354 amino acid protein. Rock2 is located on chromosome 12 and encodes a 1388 amino acid protein. Rocks include an amino acid kinase domain, a carboxyl terminal cysteine-rich region, are coiled coil domains in the middle, which includes the Rho-binding domain. Rock1 and Rock2 have 65 percent amino acid homology and the kinase domain has 92 percent homology. Rock1 is a GTP-dependent serine/threonine protein kinase that interacts with the Rho G-protein through its Rho-binding domain, thereby mediating Rho signaling. Rock1 overexpression or activation stimulates Rho activity. In addition, Rock1 can independently stimulate Rho and directly regulate cell biological behavior. Activated Rock1 induces a variety of proteins that can regulate cytoskeleton phosphorylation, thus producing corresponding biological effects, such as the reliance of Rock1 on MLC kinase or direct phosphorylation of serine 19 of MLC. Through the phosphorylation of LIMK-1 of Section 508 threonine and LIMK-2 of Section 505 threonine and the ezrin-radixin-moesin family proteins and adducin to promote cell cortex actin network formation and actin filament contact with the cell membrane.17,20

on MLC kinase or direct phosphorylation of serine 19 of MLC. Through the phosphorylation of LIMK-1 of Section 508 threonine and LIMK-2 of Section 505 threonine and the ezrin-radixin-moesin family proteins and adducin to promote cell cortex actin network formation and actin filament contact with the cell membrane.17,20 Some studies have shown that GTP enzymes, such as Rho, Rac and Cdc42, through the downstream effectors Pak1, Pak4 and Rock activate LIMK1.17,21 Previous studies have shown that LIMK1 is closely related to the differentiation of gastric cancer, lymph node metastasis and TNM stage, and plays an important role in invasion and metastasis in gastric cancer. Rac1, Pak1 and Rock1, through the phosphorylation activation of the threonine residue within the LIMK1 ring, regulates the activity of LIMK1 and plays a role in cancer invasion and metastasis.22 LIMK is regulated by a variety of upstream signals, where the main upstream signal involved in the migration and invasion is the Rho GTP enzyme family. The Rho GTP enzyme family, including Rho, Rac and Cdc42, are activated by different transmembrane receptors and transmit signals to downstream effector proteins Rock1 and Pak1. Rock1, a Rho-associated protein kinase 1, can activate the protein function, and Rac can indirectly activate LIMK1 by Pak1 (P21-activated protein kinase1). Conformational changes of Rock1 and Pak1 caused by connecting to the active GTP enzyme leads to the first 508 threonine phosphorylation of LIMK1, thereby causing the third serine phosphorylation of cofilin1, ultimately causing actin dynamics,23–28 caused the formation of a signaling pathway regulation of cell migration and invasion of Rho-Rac1-ROCK1/PAK1-LIMK1-Cofilins. (actin-depolymerizing factor)/cofilins belongs to actin depolymerization factor, is a key factor regulating the actin cytoskeleton. It includes three members: destrin (ADF), cofilin1 and cofilin2, low concentration monomers G-actin, maintaining the actin monomers pool; high concentration by nucleation effect, promoting the formation of pseudopodia, drive tumor cell migration. The results indicate that Rac1 expression in normal gastric tissue, intraepithelial neoplastic tissues and gastric carcinoma were 27, 43 and 68 percent, respectively.

-actin, maintaining the actin monomers pool; high concentration by nucleation effect, promoting the formation of pseudopodia, drive tumor cell migration. The results indicate that Rac1 expression in normal gastric tissue, intraepithelial neoplastic tissues and gastric carcinoma were 27, 43 and 68 percent, respectively. Pak1 expression levels in normal gastric tissue, intraepithelial neoplastic tissues and gastric carcinoma were 20, 35 and 60 percent, respectively, whereas Rock1 expression was 16, 28 and 58 percent, respectively. The difference between the groups was significant (P < 0.05). The Rac1 expression rate in lymph node metastasis was 75 percent, which is significantly higher than in the group without metastasis (51%). The Pak1 expression rate in lymph node metastasis was 71 percent, which is significantly higher than in the group without metastasis (43%). The Rock1 expression rate in lymph node metastasis was 66 percent, which is significantly higher than in the group without metastasis (41%) (P < 0.05). The Rac1 expression rate in stage II was 47.8 percent, which was significantly lower than that in stage III (82%). The Pak1 expression rate in stage II was 46 percent, which is significantly lower than that in stage III (74%). The Rock1 expression rate in stage II was 42 percent, which is significantly lower than that in stage III (70%) (P < 0.05). The correlation of Rac1, Pak1 and Rock1 expression in gastric cancer was analyzed using Spearman's rank correlation, and the expression of the three groups in gastric cancer was positively correlated (r = 0.555, P < 0.05). Using a Kaplan–Meier curve to analyze the survival of 158 patients with follow up, it was found that Rac1, Pak1 and Rock expression was closely associated with survival and was a risk factor for survival.

relation, and the expression of the three groups in gastric cancer was positively correlated (r = 0.555, P < 0.05). Using a Kaplan–Meier curve to analyze the survival of 158 patients with follow up, it was found that Rac1, Pak1 and Rock expression was closely associated with survival and was a risk factor for survival. We found that there were significant differences of Rac1, Pak1 and Rock expression in gastric cancer, epithelial neoplasia and normal epithelial tissue, which was an early molecular event in gastric cancer; Rac1, Pak1 and Rock expression, closely related to lymph node metastasis, depth of invasion and degree of differentiation, were valuable indicators in evaluating the degree of malignancy in gastric cancer and thus could contribute to the predication of invasion and metastasis of gastric cancer. Rac1, Pak1 and Rock expression was closely related to the survival of patients and could help in predicting the prognosis of patients. The possible mechanism of interaction between Rac1, Pak1 and Rock and other genes in gastric cancer deserves further study.

the predication of invasion and metastasis of gastric cancer. Rac1, Pak1 and Rock expression was closely related to the survival of patients and could help in predicting the prognosis of patients. The possible mechanism of interaction between Rac1, Pak1 and Rock and other genes in gastric cancer deserves further study. Rac1, Pak1 and Rock1 expression in gastric cancer is closely related with the degree of gastric cancer lymph node metastasis and TNM stage and they play an important role in the invasion and metastasis of gastric cancer and might be key biological markers for invasion and metastasis. The expression of these genes might be valuable indicators for evaluating the degree of malignancy of gastric cancer, perhaps as new markers of the biological behavior of gastric cancer. They could contribute to predicting gastric cancer invasion, metastasis and prognosis of patients. The actin cytoskeleton dynamics mechanism in tumor biology behavior, as the regulation of the actin cytoskeleton in cancer prevention and control in the sense Rac1, Pak1 and Rock1 is worth studying. Drugs including paclitaxel and cytochalasin B in making the cytoskeleton stable have been used for the treatment of cancer, but the wide range of their toxic effects make people worry. The development of drugs regulating the expression of Rac1, Rock1 and Pak1 may more accurately regulate actin activity which may be more beneficial in the prevention and treatment of diseases such as cancer.

stable have been used for the treatment of cancer, but the wide range of their toxic effects make people worry. The development of drugs regulating the expression of Rac1, Rock1 and Pak1 may more accurately regulate actin activity which may be more beneficial in the prevention and treatment of diseases such as cancer. This study was supported by the Scientific and Technological Project of Hunan Province (2008SK3010); and the Science and Technology Planning Project of Xiangtan Science and Technology Bureau (SF20081003).

1 INTRODUCTION Cervical cancer is one of the most common cancers in China. It was estimated that there were 98.9 thousand new cervical cancer cases and 30.5 thousands deaths in 2015.1 With the changes of treatment pattern and development of treatment technique, the local control (LC) of cervical cancer improved significantly. After definitive treatment, there was more distant failure than local failure.2, 3 Pulmonary metastasis is one of the most common sites of distant failure. It was reported that the incidences of pulmonary metastasis were 1.8–2.1%.4, 5

reatment technique, the local control (LC) of cervical cancer improved significantly. After definitive treatment, there was more distant failure than local failure.2, 3 Pulmonary metastasis is one of the most common sites of distant failure. It was reported that the incidences of pulmonary metastasis were 1.8–2.1%.4, 5 For pulmonary metastases from cervical cancer after initial treatment (radical surgery or definitive radiotherapy), pulmonary resection is an efficacy approach, with a 5‐year survival of more than 30%.6, 7 However, only patients with few metastatic lesions limited to the lungs and good general condition could receive pulmonary resection and only a small number of patients meets these criteria. Most patients with pulmonary metastases from cervical cancer were treated with chemotherapy or radiotherapy. As a palliative treatment, the survival after single chemotherapy was poor. Panek et al. reported that, after platinum‐5‐fluorouracil chemotherapy, only 12% of patients with pulmonary metastases from cervical cancer achieved complete response (CR). The overall survival (OS) and progression‐free survival (PFS) at 3 years were 17.6% and 14.3%, respectively.8 Patients with oligo‐pulmonary metastases from cervical cancer still have opportunity to have long‐term survival. Therefore, just giving these patients single chemotherapy is unreasonable.

omplete response (CR). The overall survival (OS) and progression‐free survival (PFS) at 3 years were 17.6% and 14.3%, respectively.8 Patients with oligo‐pulmonary metastases from cervical cancer still have opportunity to have long‐term survival. Therefore, just giving these patients single chemotherapy is unreasonable. In the past, radiotherapy was often used as a palliative approach for pulmonary metastases. As a new technique, stereotactic body radiation therapy (SBRT) could accurately deliver high dose to the lesions in few fractions. In pulmonary metastasis from rectum and soft tissue sarcoma, SBRT had been proved to be a safe and effective approach9, 10, 11 and became an important alternative to resection. For the use of SBRT in pulmonary metastases from cervical cancer, researches are limited. In this article, we retrospectively analyzed the efficacy and toxicity of SBRT for patients with pulmonary metastases from cervical cancer after definitive treatment in our institute.

In the past, radiotherapy was often used as a palliative approach for pulmonary metastases. As a new technique, stereotactic body radiation therapy (SBRT) could accurately deliver high dose to the lesions in few fractions. In pulmonary metastasis from rectum and soft tissue sarcoma, SBRT had been proved to be a safe and effective approach9, 10, 11 and became an important alternative to resection. For the use of SBRT in pulmonary metastases from cervical cancer, researches are limited. In this article, we retrospectively analyzed the efficacy and toxicity of SBRT for patients with pulmonary metastases from cervical cancer after definitive treatment in our institute. 2 MATERIALS AND METHODS 2.1 Patients We retrospectively analyzed patients with pulmonary metastases from cervical cancer treated with SBRT from January 2011 to July 2016 in Peking Union Medical College Hospital. The inclusive criteria were as follows: biopsy‐diagnosed cervical cancer; received definitive treatment for cervical cancer (surgery or radiotherapy) and achieved clinical CR after treatment; diagnosed of pulmonary metastases after a period of disease‐free survival (DFS); pulmonary metastases were treated with SBRT; KPS score ≥60. The exclusive criteria included: with extra‐pulmonary lesions before or simultaneous with pulmonary metastases; pulmonary metastases treated with radiotherapy or surgery before SBRT. There were 19 patients eligible for this research.

sease‐free survival (DFS); pulmonary metastases were treated with SBRT; KPS score ≥60. The exclusive criteria included: with extra‐pulmonary lesions before or simultaneous with pulmonary metastases; pulmonary metastases treated with radiotherapy or surgery before SBRT. There were 19 patients eligible for this research. The patients, tumor and treatment characteristics were shown in Table 1. The most common histology was squamous cell carcinoma (17 patients, 89.5%). When diagnosed of cervical cancer, the most common stage was stage IB (six patients, 31.6%) and stage IIB (six patients, 31.6%). Fifteen patients (78.9%) received definitive concurrent chemoradiotherapy (CCRT) or radiotherapy as primary treatment and four patients (21.1%) received surgery. Table 1 Patient, tumor and treatment characteristics

The patients, tumor and treatment characteristics were shown in Table 1. The most common histology was squamous cell carcinoma (17 patients, 89.5%). When diagnosed of cervical cancer, the most common stage was stage IB (six patients, 31.6%) and stage IIB (six patients, 31.6%). Fifteen patients (78.9%) received definitive concurrent chemoradiotherapy (CCRT) or radiotherapy as primary treatment and four patients (21.1%) received surgery. Table 1 Patient, tumor and treatment characteristics Characteristics Number Percentage (%) Histology SCC 17 89.5 Adenocarcinoma 1 5.3 Neuroendocrine carcinoma 1 5.3 Primary FIGO stage IB 6 31.6 IIA 3 15.8 IIB 6 31.6 IIIB 4 21.1 Primary treatment approach Concurrent chemoradiotherapy 14 73.7 Radiotherapy 1 5.3 surgery 2 10.5 Surgery+radiotherapy 2 10.5 DFI ≤12 months 12 63.2 >12 months 7 36.8 SCC Antigen <1.5 ng/mL 9 47.4 1.5–10 ng/mL 5 26.3 >10 ng/mL 5 26.3 No. of pulmonary metastases 1 13 68.4 2 3 15.8 3 2 10.5 4 1 5.3 Unilateral or bilateral lung metastases Unilateral lung metastases 16 84.2 Bilateral lung metastases 3 15.8 Diameter of lung metastases <1 cm 4 13.8 1–1.9 cm 10 34.5 2–2.9 cm 8 27.6 ≥3 cm 7 24.1 Location of lung metastases Central 8 27.6 Peripheral 21 72.4 Fractionated dose (BED) 64 Gy in 8 fractions (115.2 Gy) 8 27.6 56 Gy in 7 fractions (100.8 Gy) 7 24.1 48 Gy in 6 fractions (86.4 Gy) 4 13.8 60 Gy in 10 fractions (96.0 Gy) 2 6.9 50 Gy in 10 fractions (75.0 Gy) 2 6.9 60 Gy in 6 fractions (120.0 Gy) 1 3.4 60 Gy in 15 fractions (84.0 Gy) 1 3.4 50 Gy in 5 fractions (100.0 Gy) 1 3.4 48 Gy in 4 fractions (105.6 Gy) 1 3.4 45 Gy in 3 fractions (112.5 Gy) 1 3.4 40 Gy in 10 fractions (56.0 Gy) 1 3.4 Chemotherapy Yes 9 47.4 No 10 52.6 Abbreviations: BED, biological effective dose; SCC, squamous cell carcinoma.

s (120.0 Gy) 1 3.4 60 Gy in 15 fractions (84.0 Gy) 1 3.4 50 Gy in 5 fractions (100.0 Gy) 1 3.4 48 Gy in 4 fractions (105.6 Gy) 1 3.4 45 Gy in 3 fractions (112.5 Gy) 1 3.4 40 Gy in 10 fractions (56.0 Gy) 1 3.4 Chemotherapy Yes 9 47.4 No 10 52.6 Abbreviations: BED, biological effective dose; SCC, squamous cell carcinoma. John Wiley & Sons, Ltd.At the time of SBRT, the median age was 47 years old (range, 26–67). The time interval from CR of primary tumor to the appearance of lung metastases was disease‐free interval (DFI). The median DFI was 10.7 months. There were 29 lesions for these 19 patients in total. The mean number of lung metastases was 1.5 (range, 1–4). Thirteen patients (68.4%) had single metastases, three patients (15.8%) suffered with multiple lung metastases in single lung (two patients with two lesions and one patient with three lesions) and three patients had bilateral lung lesions (one patient with two lesions, one patient with three lesions, one patient with four lesions). The median diameter of the lesions was 2 cm (range, 0.7–5.6 cm). 2.2 Treatment A CT simulation was performed with 16‐slice Philips Brilliance Big Bore CT at a slice thickness of 5 mm, and patients were immobilized with thermoplastic. The gross tumor volume (GTV) was contoured on the CT image for each patient. The planning target volume (PTV) was defined as the GTV plus a 5–10 mm margin. The prescribed dose was decided considering the size and location of the tumors.

CT at a slice thickness of 5 mm, and patients were immobilized with thermoplastic. The gross tumor volume (GTV) was contoured on the CT image for each patient. The planning target volume (PTV) was defined as the GTV plus a 5–10 mm margin. The prescribed dose was decided considering the size and location of the tumors. Coplanar fixed‐field intensity modulated radiation therapy plan was made and treatment was delivered with a Varian Trilogy linear accelerator (Varian Medical Systems, USA). A cone beam CT (CBCT) was acquired and registered to the planning CT before every treatment. The target position error was corrected by shifting the treatment couch. The prescription dose was modulated according to size, number and location of the lesions. The most common prescribed doses included 64 Gy in eight fractions (eight lesions) and 56 Gy in seven fractions (seven lesions). The median fractions of SBRT was seven fractions (range 3–15 fractions). The fractionated dose ranged from 4 to 15 Gy and the most common one was 8 Gy (19 lesions, 65.5%). We calculate the biological effective dose (BED, α/β = 10 for tumors) for every tumor. The median BED was 100.8 Gy (range 56–120). Of the 29 pulmonary metastases, the BED of 18 lesions (62%) was higher than 100 Gy and the BED of 26 lesions (89.7%) was higher than 80 Gy.

the most common one was 8 Gy (19 lesions, 65.5%). We calculate the biological effective dose (BED, α/β = 10 for tumors) for every tumor. The median BED was 100.8 Gy (range 56–120). Of the 29 pulmonary metastases, the BED of 18 lesions (62%) was higher than 100 Gy and the BED of 26 lesions (89.7%) was higher than 80 Gy. Nine patients (47.4%) received chemotherapy. The chemotherapy regime included paclitaxel and carboplatin (five patients), paclitaxel and cisplatin (two patients), cisplatin and 5‐Fu (one patient), cisplatin alone (one patient). The median courses of chemotherapy were four courses (range 3–6). Chemotherapy was performed before radiotherapy in three patients and after radiotherapy in six patients. The median interval between radiotherapy and chemotherapy was 1.5 months (range, 0.4–3.1 months). The detailed treatment characteristics were shown in Table 1. 2.3 Outcome evaluation and statistical analysis Response Evaluation Criteria in Solid Tumors criteria was used to assess the treatment efficacy. Treatment‐related adverse effects were graded according to Common Terminology Criteria for Adverse Effects version 3.0. Statistical analysis was performed using the SPSS version 19.0. OS, PFS and LC were calculated using Kaplan–Meier analysis from the time of SBRT completion. LC was defined as no progression in their radiation field during follow up.

s were graded according to Common Terminology Criteria for Adverse Effects version 3.0. Statistical analysis was performed using the SPSS version 19.0. OS, PFS and LC were calculated using Kaplan–Meier analysis from the time of SBRT completion. LC was defined as no progression in their radiation field during follow up. 3 RESULTS The median follow‐up was 9.5 months (range, 3.0–62.4). The median follow‐up for patients alive was 18.9 months (range, 3.6–62.4). The 1‐year OS, PFS and LC were 76.8%, 55.8% and 75.6%, respectively. The estimated 2‐year OS, PFS and LC were 76.8%, 37.2% and 75.6%. The median PFS was 12.7 months. By the end of follow‐up, 11 patients (57.9%) experienced progressive disease. Seven of them had progression in the lung, including two patients with progression in the irradiation field and five patients with new lesions outside the irradiation volume. Four patients had extrapulmonary metastases, including two patients with http://mediastinum lymph nodes metastases, one patient with bone metastases and one patient with liver metastases.

including two patients with progression in the irradiation field and five patients with new lesions outside the irradiation volume. Four patients had extrapulmonary metastases, including two patients with http://mediastinum lymph nodes metastases, one patient with bone metastases and one patient with liver metastases. Of the eight patients (42.1%) without progression during follow‐up, six patients (31.6%) appeared to be disease free. Five of these six disease‐free patients had solitary lesion in the lung and the other one patient had three lesions in the bilateral lungs. The median diameter of the lesions was 2.3 cm (range, 1–3.5 cm). The dose fractionations delivered to the tumor included 64 Gy in eight fractions (three lesions), 56 Gy in seven fractions (two lesions), 60 Gy in six fractions (one lesion), 50 Gy in five fractions (one lesion) and 45 Gy in three fractions (one lesion). The median follow‐up of theses six patients was 40.4 months (range, 20.5–62.4). Figure 1 was the PET/CT images before treatment, CBCT image before the first delivery and the CT image 40 months after treatment of a patient with long‐time survival. She was a 47‐year‐old women who diagnosed with cervical squamous cell carcinoma of stage FIGO IIB in November 2009. She was treated with CCRT between December 2009 and February 2010 and acquired clinical CR. In November 2012, the squamous cell carcinoma antigen raised to 3.9 ng/mL. PET/CT scan showed a 2.6‐cm hypermetabolic lesion in the right lung and no evidence of extrapulmonary metastases. The lesion in the lung was treated with SBRT, 56 Gy in seven fractions. The patient acquired CR and DFS until the last follow up (April 2016).

2012, the squamous cell carcinoma antigen raised to 3.9 ng/mL. PET/CT scan showed a 2.6‐cm hypermetabolic lesion in the right lung and no evidence of extrapulmonary metastases. The lesion in the lung was treated with SBRT, 56 Gy in seven fractions. The patient acquired CR and DFS until the last follow up (April 2016). Figure 1 The pretreatment PET/CT images (A), the cone beam CT image before first delivery (B) and the chest CT 40 months after treatment (C) [Color figure can be viewed at http://wileyonlinelibrary.com] Symptomatic radiation pneumonitis was observed in one patient (5.3%). This patient received 64 Gy in eight fractions irradiation to 1.5 cm lesion in the left lung and suffered with Grade 2 radiation pneumonitis 2 months later. The symptoms included cough, expectoration and dyspnea. These symptoms limited instrumental activities of daily living but she did not need oxygen uptake. The chest CT images shown large patchy shadow in the left lung. After treated with glucocorticoid, the symptom relieved. The chest CT images before SBRT, at the time of radiation pneumonitis and after the treatment of glucocorticoid were shown in Figure 2. Figure 2 The chest CT images of the patient with grade 2 radiation induced pneumonitis (A: pretreatment image; B: 2 months after radiotherapy; C: 8 months after radiotherapy; D: 22 months after treatment) [Color figure can be viewed at http://wileyonlinelibrary.com]

Symptomatic radiation pneumonitis was observed in one patient (5.3%). This patient received 64 Gy in eight fractions irradiation to 1.5 cm lesion in the left lung and suffered with Grade 2 radiation pneumonitis 2 months later. The symptoms included cough, expectoration and dyspnea. These symptoms limited instrumental activities of daily living but she did not need oxygen uptake. The chest CT images shown large patchy shadow in the left lung. After treated with glucocorticoid, the symptom relieved. The chest CT images before SBRT, at the time of radiation pneumonitis and after the treatment of glucocorticoid were shown in Figure 2. Figure 2 The chest CT images of the patient with grade 2 radiation induced pneumonitis (A: pretreatment image; B: 2 months after radiotherapy; C: 8 months after radiotherapy; D: 22 months after treatment) [Color figure can be viewed at http://wileyonlinelibrary.com] 4 DISCUSSION For oligometastatic pulmonary tumors from cervical cancer, surgery is an important treatment approach. In the study of Anderson et al., six patients with pulmonary metastases from cervical cancer were treated with pulmonary resection. The median survival was 36 months.5 Yamamoto et al. reported 29 patients with pulmonary metastases from cervical cancer, which were detected after a disease‐free period after initial treatment and were resected with the intention to cure. A solitary lesion was found in 17 patients, 2 lesions were found in 6 patients, 3 lesions in 3 patients and 4 lesions in 3 patients. The 5‐year DFS rate was 32.9% for all patients. For 23 patients with one to two pulmonary lesions, the 5‐year DFS was 42.2%.6 Anraku et al. treated 76 patients with pulmonary metastases from cervical cancer with surgical resection, the 5‐year OS was 45.7%.7 For patients with pulmonary metastases from cervical cancer after radical treatment, surgical resection is an efficacy approach. The 5‐year survival was higher than 30%.

ar DFS was 42.2%.6 Anraku et al. treated 76 patients with pulmonary metastases from cervical cancer with surgical resection, the 5‐year OS was 45.7%.7 For patients with pulmonary metastases from cervical cancer after radical treatment, surgical resection is an efficacy approach. The 5‐year survival was higher than 30%. For patients with pulmonary metastases, SBRT was an effective treatment option. The 2 or 3‐year LC rate was higher than 70%.9, 10, 11 In the pooled analysis of Germen working group, 700 patients with medically inoperable lung metastases in 20 centers were treated with SBRT. The most common primary tumors included non–small cell lung cancer (NSCLC, n = 210), colorectal cancer (n = 153), sarcoma (n = 51) and so on. Solitary pulmonary metastases were found in 246 patients (42.4%). The median diameter of the lesions was 2.2 cm (range 0.4–9.4 cm). The median single fraction dose of PTV was 12.5 Gy (range 3–33 Gy) and the median number of SBRT fractions was 3 (range, 1–13). The median follow‐up was 14.3 months. The 2‐year LC and OS rates were 81.2% and 54.4%, respectively.9 A study from South Korea involved 50 patients (79 pulmonary lesions) with one to three lung metastases from colorectal cancer. A total dose of 40–60 Gy (median, 48 Gy) in three or four fractions was prescribed to the pulmonary lesions with SBRT. The 3‐year OS, PFS and LC rates were 64.0%, 24.0% and 70.6%, respectively.10 In the study of Baumann et al., 30 sarcoma patients with 39 pulmonary metastases received 50 Gy in four to five fractions with SBRT. Two‐year OS and LC rates were 43% and 86%, respectively.11 In our study, the 2‐year OS, PFS and LC rates were 76.8%, 37.2% and 75.6%, respectively. To our knowledge, this was the first study on SBRT for pulmonary metastases from cervical cancer.

9 pulmonary metastases received 50 Gy in four to five fractions with SBRT. Two‐year OS and LC rates were 43% and 86%, respectively.11 In our study, the 2‐year OS, PFS and LC rates were 76.8%, 37.2% and 75.6%, respectively. To our knowledge, this was the first study on SBRT for pulmonary metastases from cervical cancer. SBRT or stereotactic ablative radiotherapy (SABR) have become an important treatment option for patients with inoperable stage I NSCLC. A pooled analysis indicated that, for operable stage I NSCLC, patients treated with SABR could get better OS compared with lobectomy.12 For patients with single or oligo‐pulmonary metastases, SBRT was often used as an treatment option for inoperable tumors or patients refusing surgery, and there was no randomized control trial to compare the efficacy of SBRT and surgery. Yu et al. retrospectively analyzed 58 patients with pulmonary metastases from osteosarcoma. Of them, 27 patients were treated with stereotactic radiosurgery (SRS). A dose of 50 Gy in 10 fractions was delivered to PTV and 70 Gy in 10 fractions was prescribed to the GTV. The other 31 patients were treated by surgical resection. For SRS group and surgical group, the 2‐year OS rates were 40.7% and 48.3% (P > 0.05), the 2‐year PFS rates were 33.3% and 38.7% (P > 0.05).13 In this study, the 2‐year PFS rate was 37.2%. Six of 19 patients (31.6%) appeared to be disease free for more than 20 months. The outcome was close to the data of pulmonary resection.6, 7 This indicated that, for operable patients, SBRT was also a potential alternative to resection for oligometastatic pulmonary tumors from cervical cancer. Considering the small number of patients and short follow‐up period, we need more studies on this.

nths. The outcome was close to the data of pulmonary resection.6, 7 This indicated that, for operable patients, SBRT was also a potential alternative to resection for oligometastatic pulmonary tumors from cervical cancer. Considering the small number of patients and short follow‐up period, we need more studies on this. According to National Comprehensive Cancer Network (NCCN) guidelines of NSCLC, the dose of SBRT delivered to stage I NSCLC was decided based on the size, location of the tumor. The fractionated dose range from 25 to 34 Gy in one fraction to 60–70 Gy in 8–10 fractions.14 For pulmonary metastases, studies on the dose of SBRT were limited. In the previous studies, the dose fractionation delivered included 48 Gy in four fractions or 50–60 Gy in five to eight fractions,15 50 Gy in four to five fractions,11 40–60 Gy(median 48 Gy) in three to four fractions,10 median PTV single dose of 12.5 Gy (range 3.0–33.0 Gy) in a median number of three fractions (range 1–13).9 The radiation sensitivity of lung metastases varies with different primary tumors. The radiation sensitivity of cervical cancer was comparatively higher than sarcoma, colorectal cancer, renal cell carcinoma and NSCLC et al. In this study, the most common dose fractionations were 64 Gy in 8 fractions (8 lesions), 56 Gy in 7 fractions (7 lesions). The median BED was 100.8 Gy (56–120 Gy). With these dose fractionations, SBRT achieved excellent LC and promising OS. This may provide an acceptable dose fractionation option for pulmonary metastases from cervical cancer.

ommon dose fractionations were 64 Gy in 8 fractions (8 lesions), 56 Gy in 7 fractions (7 lesions). The median BED was 100.8 Gy (56–120 Gy). With these dose fractionations, SBRT achieved excellent LC and promising OS. This may provide an acceptable dose fractionation option for pulmonary metastases from cervical cancer. It was reported that the incidence rates of grade 2 or greater radiation pneumonitis were 0–6.5% for patients with pulmonary metastases when they treated with SBRT.9, 10, 11, 15, 16 In this study, only 1 (5.3%) patients had grade 2 radiation pneumonitis, which was similar with previous reports. There were some limitations in this study. The number of patients with pulmonary metastases from cervical cancer was small and the follow‐up period was short. As pulmonary metastases from cervical cancer are comparatively rare,4 multi‐institute studies are needed to collect more patients in the future. SBRT was an efficacy and safe approach for patents with oligometastatic pulmonary tumor from cervical cancer. SBRT should be considered as a potential alternative to resection for these patients. CONFLICTS OF INTEREST The authors declare that they have no conflict of interest. FUNDING This study is supported by the Ministry of Science and Technology of the People's Republic of China (Grant No 2016YFC0105207).

INTRODUCTION Following the publication of the Trastuzumab for Gastric Cancer (ToGA) trial results1 and the approval of trastuzumab for the treatment of human epidermal growth factor receptor 2 (HER2)-positive advanced gastric cancer, the quality of HER2 testing in gastric cancer emerged as an area of interest in the Asia-Pacific region, which reports a high incidence of gastric cancer. A multidisciplinary panel of experts from the Asia-Pacific region convened to discuss the applicability of current HER2 testing recommendations in gastric cancer. Experts were selected based on scientific merit, on recognition as regional key opinion leaders in gastric cancer management, and on advocacy of HER2 testing quality through programs such as the Scientific Partnership for HER2 testing Excellence (SPHERE). SPHERE is an ongoing educational initiative that engages with pathology laboratories in the Asia-Pacific region to achieve the highest standards in HER2 testing. The program also runs advocacy activities to engage members of HER2 testing teams to coordinate diagnosis, treatment plans and overall patient care. These expert discussions assessed latest research and real-world practice in the region to supplement Rüschoff et al.'s recommendations on HER2 testing in gastric cancer, which have been developed primarily based on experience from Europe.2 Where necessary, Rüschoff et al.'s recommendations have been modified or expanded upon to reflect the experience of Asia-Pacific laboratories in routine testing. It should be noted that the following recommendations are not binding and should be adapted to reflect national level reimbursement plans.

ience from Europe.2 Where necessary, Rüschoff et al.'s recommendations have been modified or expanded upon to reflect the experience of Asia-Pacific laboratories in routine testing. It should be noted that the following recommendations are not binding and should be adapted to reflect national level reimbursement plans. EPIDEMIOLOGY Worldwide, gastric cancer is the fourth most commonly diagnosed cancer3,4 and the second leading cause of cancer death in men and women.4 The Asia-Pacific region records among the highest disease rates globally, with 50% of all cases reported originating in East Asian countries.4 The estimated mortality rates in East Asia, the highest in the world, are 28.1 per 100 000 in men and 13.0 per 100 000 in women.4 Gastric cancer is associated with a poor prognosis.1 Patients with stage III and IV advanced or metastatic gastric cancer have a 5-year survival rate between 5.0 and 26.9%; the median overall survival is typically less than 1 year.1,5 These poor survival rates are more often reported in countries that do not have government-supported nationwide screening programs, such as the United States (<25.0%).3 In countries with screening programs, such as Korea and Japan, gastric cancer is more likely to be detected at the early stages of the disease, thus resulting in more favorable 5-year survival rates of 55.6–66.0%6 and 50.0%, respectively.7

nt-supported nationwide screening programs, such as the United States (<25.0%).3 In countries with screening programs, such as Korea and Japan, gastric cancer is more likely to be detected at the early stages of the disease, thus resulting in more favorable 5-year survival rates of 55.6–66.0%6 and 50.0%, respectively.7 THE CLINICAL SIGNIFICANCE OF HER2 IN GASTRIC CANCER Across Asian populations, HER2 overexpression in gastric cancer is reported in approximately 9.8–23.0% of cases,8–15 with higher rates generally observed in gastroesophageal junction (GEJ) cancer.14,15 HER2 is a key driver of tumorigenesis through its association with tumor cell proliferation, apoptosis, adhesion, migration, and differentiation, and is suggested to be associated with aggressive disease.1 In gastric cancer, HER2 overexpression is predictive of response to HER2-targeted therapy;1 however, its prognostic significance remains controversial. While some studies failed to establish the prognostic value of HER2 positivity,12 other studies suggested that there were correlations between HER2 positivity and poorer outcomes, as evidenced by higher HER2 positivity rates in advanced gastric carcinomas and prognostic significance found in early gastric carcinomas.2,9,16 A large cohort study by Cho et al. suggested that frequent HER2 positivity observed in advanced gastric carcinomas is indicative that HER2 may be involved in tumor progression and poor prognosis.17

HER2 positivity rates in advanced gastric carcinomas and prognostic significance found in early gastric carcinomas.2,9,16 A large cohort study by Cho et al. suggested that frequent HER2 positivity observed in advanced gastric carcinomas is indicative that HER2 may be involved in tumor progression and poor prognosis.17 Further analyses found that poorer prognosis in HER2-positive cases was significantly correlated with the patient's age and histopathologic type,9,18 whereby intestinal-type carcinomas are more likely to demonstrate HER2 overexpression compared with diffuse- or mixed-type carcinomas.9,16,19 TOGA'S SIGNIFICANCE ON THE TREATMENT OF HER2-POSITIVE GASTRIC CARCINOMAS ToGA was an open-label, multicenter, randomized controlled trial that investigated trastuzumab in combination with chemotherapy for the first-line treatment of patients with HER2-positive disease with advanced gastric or GEJ cancer.1

Further analyses found that poorer prognosis in HER2-positive cases was significantly correlated with the patient's age and histopathologic type,9,18 whereby intestinal-type carcinomas are more likely to demonstrate HER2 overexpression compared with diffuse- or mixed-type carcinomas.9,16,19 TOGA'S SIGNIFICANCE ON THE TREATMENT OF HER2-POSITIVE GASTRIC CARCINOMAS ToGA was an open-label, multicenter, randomized controlled trial that investigated trastuzumab in combination with chemotherapy for the first-line treatment of patients with HER2-positive disease with advanced gastric or GEJ cancer.1 Results showed that the addition of trastuzumab to chemotherapy provided the greatest survival benefit to patients with high HER2 protein expression (defined as immunohistochemistry [IHC] modified HercepTest [Dako Corp, Carpinteria, CA, USA] score of 2+ and fluorescence in situ hybridization [FISH]-determined HER2 gene amplification, or IHC modified HercepTest score of 3+). The median overall survival was 16.0 months for patients treated with combination chemotherapy and trastuzumab versus 11.8 months for those treated with chemotherapy alone.1 To date, ToGA is the only trial to evidence a treatment strategy that extends overall survival beyond the 12-month median mark in advanced gastric/GEJ cancer patients who are HER2 positive.1 Trastuzumab has approved indications for the treatment of HER2-positive gastric cancer in five East Asian countries and areas, as shown in Table 1. Table 1 Approved trastuzumab indications for HER2-positive gastric cancer in East Asian countries

Results showed that the addition of trastuzumab to chemotherapy provided the greatest survival benefit to patients with high HER2 protein expression (defined as immunohistochemistry [IHC] modified HercepTest [Dako Corp, Carpinteria, CA, USA] score of 2+ and fluorescence in situ hybridization [FISH]-determined HER2 gene amplification, or IHC modified HercepTest score of 3+). The median overall survival was 16.0 months for patients treated with combination chemotherapy and trastuzumab versus 11.8 months for those treated with chemotherapy alone.1 To date, ToGA is the only trial to evidence a treatment strategy that extends overall survival beyond the 12-month median mark in advanced gastric/GEJ cancer patients who are HER2 positive.1 Trastuzumab has approved indications for the treatment of HER2-positive gastric cancer in five East Asian countries and areas, as shown in Table 1. Table 1 Approved trastuzumab indications for HER2-positive gastric cancer in East Asian countries Country Indication Japan Advanced or recurrent gastric cancer overexpressing HER2, not amenable to curative resection Hong Kong Combination with cisplatin and capecitabine or 5-fluorouracil for HER2-positive metastatic gastric adenocarcinoma in treatment-naive patients South Korea and China Combination with capecitabine or 5-fluorouracil and cisplatin for the treatment of patients with HER2-positive metastatic or unresectable adenocarcinoma of the stomach or GEJ, who have not received prior anticancer treatment for their metastatic or unresectable disease Taiwan Combination with capecitabine or 5-fluorouracil and cisplatin is indicated for the treatment of patients with HER2-positive (IHC2+/ISH+ or IHC3+) metastatic adenocarcinoma of the stomach (or GEJ) who have not received prior chemotherapy for their metastatic disease GEJ, gastroesophageal junction; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization.

he treatment of patients with HER2-positive (IHC2+/ISH+ or IHC3+) metastatic adenocarcinoma of the stomach (or GEJ) who have not received prior chemotherapy for their metastatic disease GEJ, gastroesophageal junction; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization. THE HER2 TESTING ALGORITHM IN GASTRIC CANCER HER2 status in gastric cancer is determined through IHC testing for HER2 protein overexpression and in situ hybridization (ISH) testing for HER2 gene amplification. The ToGA trial demonstrated that trastuzumab provides the greatest benefit for patients with high HER2 protein overexpression, thereby establishing IHC as the primary testing method for gastric cancer.1 ISH testing is reserved for HER2 status confirmation of cases with equivocal HER2 protein expression (IHC 2+) (Fig. 1).2,20 Bright-field ISH methodologies are now preferred to effectively identify regions of HER2 amplification; however, selection of ISH methodology should adhere to national level guidelines and reimbursement policies. Figure 1 Recommended HER2 testing algorithm in gastric cancer. IHC, immunohistochemistry; GEJ, gastroesophageal junction; HER2, human epidermal growth factor receptor 2; ISH, in situ hybridization.

The ToGA trial demonstrated that trastuzumab provides the greatest benefit for patients with high HER2 protein overexpression, thereby establishing IHC as the primary testing method for gastric cancer.1 ISH testing is reserved for HER2 status confirmation of cases with equivocal HER2 protein expression (IHC 2+) (Fig. 1).2,20 Bright-field ISH methodologies are now preferred to effectively identify regions of HER2 amplification; however, selection of ISH methodology should adhere to national level guidelines and reimbursement policies. Figure 1 Recommended HER2 testing algorithm in gastric cancer. IHC, immunohistochemistry; GEJ, gastroesophageal junction; HER2, human epidermal growth factor receptor 2; ISH, in situ hybridization. While IHC is the primary test for all gastric cancers, there is a growing interest in the role of ISH techniques. The Australian Gastric HER2 Testing Study (GaTHER), which examined expanding the application of ISH to determine HER2 positivity in gastric cancer, concluded that an IHC 3+ score strongly predicts a positive ISH result, and that IHC triage before ISH testing was the most cost-effective strategy.21 Given the unique features of gastric cancers and the difficulty of ensuring consistency of HER2 staining in the community setting, when the laboratory initiates a testing program for HER2 in advanced gastric/GEJ cancer, it should test IHC and ISH in parallel until the results are consistent.21 Gómez-Martin et al. considered hybridization techniques in general, and dual-color silver-enhanced ISH (dc-SISH) in particular, may optimize the selection of patients with HER2-positive gastric cancer who will respond to HER2-targeted therapy.22 However, due to limited resources, ISH's role in the HER2 testing algorithm within the Asia-Pacific context will most likely be applied for the confirmation of HER2 status in all HER2 IHC 2+ cases.

lar, may optimize the selection of patients with HER2-positive gastric cancer who will respond to HER2-targeted therapy.22 However, due to limited resources, ISH's role in the HER2 testing algorithm within the Asia-Pacific context will most likely be applied for the confirmation of HER2 status in all HER2 IHC 2+ cases. PRINCIPLES FOR A MULTIDISCIPLINARY HER2 TESTING TEAM HER2 testing in gastric cancer requires multidisciplinary coordination to ensure accurate and timely diagnosis of patients' HER2 status. The multidisciplinary team, which involves core members of the surgical, oncology and pathology teams, should collectively promote several key principles toward achieving the goal of quality testing. First, all inoperable, locally advanced, recurrent and metastatic gastric and GEJ cancers should be routinely tested for HER2 overexpression at diagnosis. While the Task Force accepts that access to HER2-targeted therapy is an important factor in clinical decision making, it should not influence decisions to test patients. Second, turnaround time from histological diagnosis of gastric cancer to determination of HER2 status should be as short as possible. Appropriate early treatment is important given the rapid progression of disease and high relapse rates in gastric or GEJ cancers. Third, treatment plans must be formulated with full knowledge of a patient's HER2 status. HER2 test reports should therefore provide a conclusive diagnosis; an equivocal result should not be submitted or accepted.

Second, turnaround time from histological diagnosis of gastric cancer to determination of HER2 status should be as short as possible. Appropriate early treatment is important given the rapid progression of disease and high relapse rates in gastric or GEJ cancers. Third, treatment plans must be formulated with full knowledge of a patient's HER2 status. HER2 test reports should therefore provide a conclusive diagnosis; an equivocal result should not be submitted or accepted. Finally, HER2 testing should be automated if possible to minimize the risks of testing variations and to produce more reliable and accurate results. RECOMMENDATIONS FOR GASTRIC HER2 TESTING Appropriate specimen type Sampling is a crucial early determinant in the accuracy of HER2 testing. Surgical resection and biopsies are both accepted. However, biopsies are more commonly provided because advanced and metastatic gastric cancers are generally inoperable.

Finally, HER2 testing should be automated if possible to minimize the risks of testing variations and to produce more reliable and accurate results. RECOMMENDATIONS FOR GASTRIC HER2 TESTING Appropriate specimen type Sampling is a crucial early determinant in the accuracy of HER2 testing. Surgical resection and biopsies are both accepted. However, biopsies are more commonly provided because advanced and metastatic gastric cancers are generally inoperable. Current recommendations state that six to eight biopsies should be taken to account for intratumoral heterogeneity and to provide sufficient tissue for testing.2 This range may not necessarily be achievable in clinical settings; the Task Force considers four to six biopsies as acceptable. Endoscopic teams should be aware that insufficient tumor tissue cannot account for heterogeneity and increases the risk of false-negative results.23,24 To reduce sampling errors and necrotic tissue, which limit the available volume of tissue for testing, endoscopists should be careful to acquire sufficient viable tumor samples and routinely record the number of biopsy fragments submitted. Tissue microarrays are not recommended for routine specimen assessment because they cannot fully represent heterogeneity. Handling and fixation Gastric biopsy specimens are susceptible to rapid dehydration. Stringent procedures should be in place to transfer biopsy samples as quickly as possible after removal into the recommended fixative 10% neutral buffered formalin (NBF), preferably within 20 min of excision, as proposed by Rüschoff et al.2

ing and fixation Gastric biopsy specimens are susceptible to rapid dehydration. Stringent procedures should be in place to transfer biopsy samples as quickly as possible after removal into the recommended fixative 10% neutral buffered formalin (NBF), preferably within 20 min of excision, as proposed by Rüschoff et al.2 Surgical excisions should be handled under similar stringent procedures with the transfer of the specimen into 10% NBF within 1 h of resection. Pre-analytic cold ischemia time, the time between tissue removal and initiation of formalin fixation, can affect accurate measurement of protein expression patterns in tissues, leading to inaccurate HER2 test interpretation and false-negative results.25,26 In some clinical situations, a longer cold ischemia time may be acceptable. For example, in one laboratory's practical experiments with gastric specimens showing HER2 protein overexpression, delayed time to fixation not exceeding 4 h provided surprisingly acceptable IHC and ISH results (Professor Kyoung-Mee Kim, pers. comm., 2013). This is however not recommended as a guideline without revalidation in the original laboratory as well as further validation in other laboratories. The Task Force recommends that laboratories validate the cold ischemia time for biopsies and surgical excisions separately to provide evidence for appropriate tissue handling protocols. Validations must reflect whether specimens are to be transferred fresh to the laboratory or whether fixation is to be initiated in the operating suite.

commends that laboratories validate the cold ischemia time for biopsies and surgical excisions separately to provide evidence for appropriate tissue handling protocols. Validations must reflect whether specimens are to be transferred fresh to the laboratory or whether fixation is to be initiated in the operating suite. If initiated in the operating suite, the time that specimens are removed and placed into 10% NBF should be noted in the testing request form. Specimens should be placed in sturdy, appropriate-sized containers and submerged in 10% NBF; surgical specimens should be opened and oriented prior to submersion (Fig. 2). While laboratory staff should ideally lead in process validation, surgical and other operating suite staff should confirm adherence to validated tissue handling procedures. Figure 2 An example of a custom-designed container provided to an operating theater in Taiwan. The container allows surgical staff to open, orientate and pin resected specimens to a corkboard prior to submersion in 10% neutral buffered formalin. (Image courtesy of J. Wang).

If initiated in the operating suite, the time that specimens are removed and placed into 10% NBF should be noted in the testing request form. Specimens should be placed in sturdy, appropriate-sized containers and submerged in 10% NBF; surgical specimens should be opened and oriented prior to submersion (Fig. 2). While laboratory staff should ideally lead in process validation, surgical and other operating suite staff should confirm adherence to validated tissue handling procedures. Figure 2 An example of a custom-designed container provided to an operating theater in Taiwan. The container allows surgical staff to open, orientate and pin resected specimens to a corkboard prior to submersion in 10% neutral buffered formalin. (Image courtesy of J. Wang). RECOMMENDATIONS FOR LABORATORIES: PARAMETERS COMMONLY ASSOCIATED WITH TESTING VARIATIONS Fixation Task Force members have noted that variable and non-validated fixation procedures are commonly used in the Asia-Pacific region. This has caused testing variations, as noted in confidential feedback provided to individual laboratories by the United Kingdom National External Quality Assessment Service (UK NEQAS), the quality assurance program to which many laboratories subscribe. This section highlights the accepted standards that should benchmark the validation of fixation procedures.

oted in confidential feedback provided to individual laboratories by the United Kingdom National External Quality Assessment Service (UK NEQAS), the quality assurance program to which many laboratories subscribe. This section highlights the accepted standards that should benchmark the validation of fixation procedures. Fixation is influenced by a number of factors: temperature, penetration rate, specimen dimensions (preferably no more than 4 mm thick), fixative-to-tissue volume ratio, pH and duration of fixation. Optimizing these parameters will ensure that cross-linkages are appropriately formed between tissue elements and fixative, and that specimens are stable for processing. Fresh supplies of 10% NBF, the only fixative associated with consistent testing results, should be used where possible in a fixative-to-tissue volume ratio of 10:1. Even when fixation is initiated in operating theaters, the Task Force suggests that laboratory staff monitor the process to ensure compliance with validated procedures, and to ensure appropriate fixative type and volume upon receipt in the laboratory.

be used where possible in a fixative-to-tissue volume ratio of 10:1. Even when fixation is initiated in operating theaters, the Task Force suggests that laboratory staff monitor the process to ensure compliance with validated procedures, and to ensure appropriate fixative type and volume upon receipt in the laboratory. The duration of fixation for biopsies and surgical excisions should be validated separately within the recommended range of 8–72 h (Fig. 3). Access to HER2 testing centers is an important issue in the Asia-Pacific region. The Task Force recognizes that tissue sampling may be performed in remote settings and some distance away from HER2 testing centers, which are more commonly located in urban areas. HER2 testing centers should work closely with local laboratories in these remote settings to appropriately prepare tissues for transportation. If specimens will reach the HER2 testing laboratory within the validated fixation duration, local laboratories should initiate fixation by submerging specimens in sufficient volume of 10% NBF and sealing them within containers that are appropriate for long-distance transportation. If specimens will reach the laboratory after the validated fixation duration, specimens should be processed and embedded in paraffin blocks prior to transportation, in accordance with validated protocols.

ficient volume of 10% NBF and sealing them within containers that are appropriate for long-distance transportation. If specimens will reach the laboratory after the validated fixation duration, specimens should be processed and embedded in paraffin blocks prior to transportation, in accordance with validated protocols. Figure 3 Staining of human epidermal growth factor receptor 2 (HER2) immunohistochemistry (IHC) 3 + specimen fixed for 8 h (a), 48 h (b), 72 h (c), and 96 h (d). The HER2 IHC 3+ specimen that was fixed for 96 h shows relatively weak staining compared with the other specimens. (Image courtesy of K-M. Kim). Antigen retrieval Automated, temperature-controlled antigen retrieval, optimized and validated based on the instrument manufacturer's recommendations, is recommended. Poor retrieval techniques are a common cause of false results. Over-retrieval, for instance, may activate endogenous biotin, leading to high levels of background staining with avidin-biotin reaction-based detection and may produce false-positive results.27

the instrument manufacturer's recommendations, is recommended. Poor retrieval techniques are a common cause of false results. Over-retrieval, for instance, may activate endogenous biotin, leading to high levels of background staining with avidin-biotin reaction-based detection and may produce false-positive results.27 Assays Standardized, commercially available IHC and ISH assays are preferred over in-house assays to ensure reliability and reproducibility of HER2 testing results. Assays should ideally be validated and approved by regional regulatory agencies for testing specifically in gastric cancer specimens. In-house validation of assays is critical, as based on the Task Force's collective experiences, different IHC antibodies may produce different staining intensities, and positivity rates may vary.13 The validation process should produce reference images for no, low and high levels of HER2 protein expression to standardize day-to-day practice in scoring in laboratories. Controls Wherever possible, controls with defined HER2 expression levels should be present on each specimen slide to detect false results. For IHC testing, controls should include sections with no (IHC 0) and strong IHC staining (IHC 3+), and if possible, equivocal staining (IHC 2+). Controls, preferably those recommended and/or provided in test kits, should be prepared using similar fixation, processing and paraffin embedding techniques as specimens.

ts. For IHC testing, controls should include sections with no (IHC 0) and strong IHC staining (IHC 3+), and if possible, equivocal staining (IHC 2+). Controls, preferably those recommended and/or provided in test kits, should be prepared using similar fixation, processing and paraffin embedding techniques as specimens. RECOMMENDATIONS FOR HER2 TEST INTERPRETATION AND SCORING IN GASTRIC CANCER The unique features of gastric carcinomas Reflecting the unique features of gastric carcinomas, gastric interpretation and scoring guidelines differ from those for breast cancer.2 Unlike breast carcinomas, gastric carcinomas are gland-forming, mucin-producing carcinomas that demonstrate incomplete, basolateral/lateral staining patterns (Fig. 4). When strongly stained, this U-shaped pattern is classified as HER2 positive.20,28 If breast guidelines, which require complete membrane staining, are applied to gastric carcinomas, there would be a significant number of gastric specimens that would be incorrectly scored as HER2 negative. Figure 4 Incomplete lateral/basolateral membrane staining produces the characteristic U-shaped patterns (arrowheads) associated with gastric cancer cells. (Image courtesy of J. Wang).

RECOMMENDATIONS FOR HER2 TEST INTERPRETATION AND SCORING IN GASTRIC CANCER The unique features of gastric carcinomas Reflecting the unique features of gastric carcinomas, gastric interpretation and scoring guidelines differ from those for breast cancer.2 Unlike breast carcinomas, gastric carcinomas are gland-forming, mucin-producing carcinomas that demonstrate incomplete, basolateral/lateral staining patterns (Fig. 4). When strongly stained, this U-shaped pattern is classified as HER2 positive.20,28 If breast guidelines, which require complete membrane staining, are applied to gastric carcinomas, there would be a significant number of gastric specimens that would be incorrectly scored as HER2 negative. Figure 4 Incomplete lateral/basolateral membrane staining produces the characteristic U-shaped patterns (arrowheads) associated with gastric cancer cells. (Image courtesy of J. Wang). Heterogeneous staining is a true biological feature that is far more common in gastric and GEJ carcinomas than in breast carcinomas20,28 and is estimated to be present in up to 30% of HER2-positive gastric cases.2 Heterogeneity is more often detected in IHC 2+ cases or in mixed histologic type,17 and overexpression of HER2 protein is associated with histologic types (differentiated or intestinal) and aggressive biological behaviors (lymphovascular invasion and lymph node metastasis).12,13,17,28

% of HER2-positive gastric cases.2 Heterogeneity is more often detected in IHC 2+ cases or in mixed histologic type,17 and overexpression of HER2 protein is associated with histologic types (differentiated or intestinal) and aggressive biological behaviors (lymphovascular invasion and lymph node metastasis).12,13,17,28 Heterogeneity impacts each phase of testing, from sampling to scoring, and may cause HER2-positive foci to be missed in gastric specimens (Fig. 5). The heterogeneity issue has raised questions on whether tumors with <10% positively stained cells would respond to HER2-targeted therapy, despite classification as HER2 negative with current scoring criteria. In tumors in which strong complete/basolateral or lateral membrane staining is seen in <10% of the cells, IHC staining should be repeated on a different paraffin block section, and if still inconclusive, an ISH test should be performed to determine HER2 gene amplification and final HER2 status. Figure 5 Heterogeneous human epidermal growth factor receptor 2 (HER2) protein expression in a gastric specimen. (a) Strong staining characterizing a HER2-positive foci under low (4×) magnification; (b) weak-to-moderate staining intensity under 10× magnification, classified as immunohistochemistry (IHC) 2+; (c) barely perceptible staining under higher (20×) magnification, which, if seen in isolation, would give the specimen a score of IHC 1+. (Images courtesy of J. Ying).

ositive foci under low (4×) magnification; (b) weak-to-moderate staining intensity under 10× magnification, classified as immunohistochemistry (IHC) 2+; (c) barely perceptible staining under higher (20×) magnification, which, if seen in isolation, would give the specimen a score of IHC 1+. (Images courtesy of J. Ying). Besides intratumoral heterogeneity, discordant HER2 status between primary and metastatic gastric tumors is not uncommon in gastric carcinoma13,24 and could also impact patients' eligibility for HER2-targeted therapy. Therefore, in stage IV disease, HER2 testing should ideally be performed on samples from both primary and distant metastatic sites, subject to the availability of specimens. Evidence of discordance between the primary and metastatic sites should prompt a discussion between the pathologist and oncologist, prior to making a final treatment decision. ASSESSMENT OF IHC HER2 STAINING When assessing a specimen following IHC staining, care should be taken to interpret an invasive area of the carcinoma. Nonspecific staining and artifacts should not be assessed. Tests should be repeated on different samples when faced with poorly preserved tissue.

Besides intratumoral heterogeneity, discordant HER2 status between primary and metastatic gastric tumors is not uncommon in gastric carcinoma13,24 and could also impact patients' eligibility for HER2-targeted therapy. Therefore, in stage IV disease, HER2 testing should ideally be performed on samples from both primary and distant metastatic sites, subject to the availability of specimens. Evidence of discordance between the primary and metastatic sites should prompt a discussion between the pathologist and oncologist, prior to making a final treatment decision. ASSESSMENT OF IHC HER2 STAINING When assessing a specimen following IHC staining, care should be taken to interpret an invasive area of the carcinoma. Nonspecific staining and artifacts should not be assessed. Tests should be repeated on different samples when faced with poorly preserved tissue. Table 2 summarizes the widely accepted IHC scoring criteria for gastric carcinomas that were developed from validation studies, including the preclinical validation phase of the ToGA trial.1,2,20,30 Rüschoff et al. further suggest that the microscopic magnification rule can complement the interpretation process.2 The rule correlates staining intensity and microscopic magnification with IHC scores as outlined below.IHC 3+: any membranous staining visible at low (×2.5–5) magnification. IHC 2+: membranous staining visible at medium (×10–20) magnification. IHC 1+: staining visible only at high (×40) magnification. Table 2 HER2 IHC and ISH scoring criteria in gastric cancer

Table 2 summarizes the widely accepted IHC scoring criteria for gastric carcinomas that were developed from validation studies, including the preclinical validation phase of the ToGA trial.1,2,20,30 Rüschoff et al. further suggest that the microscopic magnification rule can complement the interpretation process.2 The rule correlates staining intensity and microscopic magnification with IHC scores as outlined below.IHC 3+: any membranous staining visible at low (×2.5–5) magnification. IHC 2+: membranous staining visible at medium (×10–20) magnification. IHC 1+: staining visible only at high (×40) magnification. Table 2 HER2 IHC and ISH scoring criteria in gastric cancer IHC scoring criteria† ISH scoring criteria‡ Score Surgical specimen staining pattern Biopsy specimen staining pattern Diagnosis Scoring criteria IHC 0 (negative) No reactivity or membranous reactivity in <10% of tumor cells No reactivity or no membranous reactivity in any tumor cell ISH positive HER2:CEP17 ratio ≥2.0; if present, record polysomic amplified IHC 1+ (negative) Faint/barely perceptible membranous reactivity in ≥10% of tumor cells; cells are reactive only in part of their membrane Tumor cell cluster§ with a faint/barely perceptible membranous reactivity, irrespective of percentage of tumor cells stained ISH negative HER2:CEP17 ratio <2.0; if present, record polysomic not amplified IHC 2+ (equivocal) Weak-to-moderate, complete, basolateral or lateral membranous reactivity in ≥10% of tumor cells Tumor cell cluster§ with a weak-to-moderate, complete, basolateral or lateral membranous reactivity, irrespective of percentage of tumor cells stained ISH borderline amplification (ratio close to 2.0) HER2:CEP17 ratio 1.8–2.2Where the HER2:CEP17 ratio suggests borderline amplification, count the overall number of HER2 gene copies

cluster§ with a weak-to-moderate, complete, basolateral or lateral membranous reactivity, irrespective of percentage of tumor cells stained ISH borderline amplification (ratio close to 2.0) HER2:CEP17 ratio 1.8–2.2Where the HER2:CEP17 ratio suggests borderline amplification, count the overall number of HER2 gene copies If the HER2 gene count is four to six copies, count another 20 cells to reconfirm the HER2:CEP17 ratio If the HER2/CEP17 ratio remains 1.80–1.99, report as borderline not amplified, HER2 negative29 If the HER2/CEP17 ratio remains 2.00–2.20, report as borderline amplified and HER2 positive29 IHC 3+ (positive) Strong, complete, basolateral or lateral membranous reactivity in ≥10% of tumor cells Tumor cell cluster§ with a strong, complete, basolateral or lateral membranous reactivity, irrespective of percentage of tumor cells stained † Adapted from Bang et al.1 ‡ Adapted from Rüschoff et al.2 § For biopsies, there is no percentage cutoff; however, a cluster of at least five positive cells is required. CEP17, chromosome enumeration probe 17; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization.

IHC 3+ (positive) Strong, complete, basolateral or lateral membranous reactivity in ≥10% of tumor cells Tumor cell cluster§ with a strong, complete, basolateral or lateral membranous reactivity, irrespective of percentage of tumor cells stained † Adapted from Bang et al.1 ‡ Adapted from Rüschoff et al.2 § For biopsies, there is no percentage cutoff; however, a cluster of at least five positive cells is required. CEP17, chromosome enumeration probe 17; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization. Using this magnification rule in conjunction with IHC scoring criteria may improve concordance between IHC and ISH test results.2 In the ToGA trial, as many as 18.6% of IHC 1+ gastric carcinomas were HER2 gene amplified.1 However, in the opinion of the Task Force, this rule should not be used in isolation without the scoring criteria listed in Table 2, as fluctuations in temperature, microscope settings and environmental brightness could affect staining visibility. If a surgical excision specimen shows <10% strongly stained tumor cells, a different section of the specimen that contains intestinal-type carcinoma by the Lauren classification system should be re-stained. The Task Force further recommends that if a specimen's HER2 status is in doubt, an ISH test should be performed for confirmation.12,17,31 If the ISH test confirms the discordance, a HER2-positive diagnosis should be reported in the presence of HER2-positive foci.

y the Lauren classification system should be re-stained. The Task Force further recommends that if a specimen's HER2 status is in doubt, an ISH test should be performed for confirmation.12,17,31 If the ISH test confirms the discordance, a HER2-positive diagnosis should be reported in the presence of HER2-positive foci. The final HER2 status is that of the invasive carcinoma only. HER2 positivity confined to the dysplasia of Barrett's esophagus or in situ gastric carcinoma can be recorded in the report notes. Assessment of HER2 gene amplification Due to the heterogeneous nature of gastric carcinomas, hematoxylin and eosin and IHC stains should guide screening and selection of possible HER2-amplified areas for assessment. The entire specimen is screened to confirm successful hybridization, detection and visualization, and a selection of normal cells is formally counted before interpreting tumor cells.29 Approximately 20 adjacent or continuous cells from IHC 2+ areas should be counted based on the principles outlined in Table 2. It should be noted that polysomy of chromosome 17 (detected using chromosome enumeration probe 17 [CEP17]), defined as ≥3 CEP17 signals per nuclei, is often associated with IHC 2+ cases.17 While its occurrence should be noted in the report, it should not affect accuracy of the diagnosis with assessment of the HER2/CEP17 ratio.

d be noted that polysomy of chromosome 17 (detected using chromosome enumeration probe 17 [CEP17]), defined as ≥3 CEP17 signals per nuclei, is often associated with IHC 2+ cases.17 While its occurrence should be noted in the report, it should not affect accuracy of the diagnosis with assessment of the HER2/CEP17 ratio. Caution should be exercised when counting overlapping nuclei, because the signals cannot be accurately counted or assigned to a particular cell in the region of the overlap. Nuclei that are not intact and contain central unstained areas should not be counted, since the missing nuclear area may have contained signals that cannot be assessed. Cells with weak signal intensity should also be excluded from assessment. With dark-field methodologies, cells within regions of nonspecific or high-background staining should not be counted. When using dual probes, only cells with at least one ISH signal for each probe should be counted. FISH-stained slides will fade with time. It is recommended that representative images are taken and that slides are stored at −20°C for a minimum of 12 months.16 A test should be rejected in the presence of the following factors, which could impact the accuracy of assessment.Pre-analytic parameters, particularly fixation, are not in accordance with validated procedures. Analytic parameters are not as expected due to microtome artifacts, intensity of hybridization signals and/or presence of both signals in the dual-probe test. Unsatisfactory results in the controls. Lack of tumor component in the histological section.

A test should be rejected in the presence of the following factors, which could impact the accuracy of assessment.Pre-analytic parameters, particularly fixation, are not in accordance with validated procedures. Analytic parameters are not as expected due to microtome artifacts, intensity of hybridization signals and/or presence of both signals in the dual-probe test. Unsatisfactory results in the controls. Lack of tumor component in the histological section. Negative results in sections prepared during a period in excess of 4–6 weeks (which have not been paraffinized and/or stored at 4°C). The GaTHER study concluded that good agreement across laboratories for HER2 copy number, as determined by single-probe chromogenic ISH or silver ISH (SISH), suggests that this is the optimal HER2 ISH testing method in gastric/GEJ cancer.21 This recommendation differs from United Kingdom guidelines, which no longer recommend modalities that do not include CEP17 probes.29 While the Task Force considers all ISH tests to provide accurate test results following validation, there is a preference for bright-field, dual-probe tests, which produce signals that are easier to score and interpret.

United Kingdom guidelines, which no longer recommend modalities that do not include CEP17 probes.29 While the Task Force considers all ISH tests to provide accurate test results following validation, there is a preference for bright-field, dual-probe tests, which produce signals that are easier to score and interpret. QUALITY CONTROL AND QUALITY ASSURANCE Quality control and quality assurance measures implemented at each phase of testing will ensure reliable and consistent results. The Task Force recognizes that experience contributes significantly to accuracy of testing, and recommends that HER2 testing laboratories in the region should routinely perform at least 10 IHC tests per day for any antigen, including HER2. Where laboratories are unable to comply with the recommendations, they should consider referring cases to central testing laboratories for HER2 testing. For laboratories with both IHC and ISH testing capabilities that are performing HER2 testing in gastric cancer for the first time, 25–50 gastric cancer cases should be analyzed in parallel, using IHC and ISH, with the aim of achieving concordance of greater than 90%. At least five HER2-positive cases (10–20% validation cases) should be included.

IHC and ISH testing capabilities that are performing HER2 testing in gastric cancer for the first time, 25–50 gastric cancer cases should be analyzed in parallel, using IHC and ISH, with the aim of achieving concordance of greater than 90%. At least five HER2-positive cases (10–20% validation cases) should be included. HER2 testing in gastric cancer must be validated independently of breast cancer protocols, and the optimized procedures documented in in-house HER2 testing protocols that are specific for gastric cancer specimens. If procedures vary from these protocols, records must be made detailing these variations. Revalidation must be performed if there is a change to testing conditions, reagents, materials and methods. Testing in the laboratory should be performed by a designated team specifically trained in gastric HER2 testing; members should be reevaluated once a year. Besides competency in laboratory procedures, staff involved in testing and scoring should learn to recognize histological features of gastric cells, and identify and assess the invasive component of the tumor. Prior HER2 testing experience in breast cancer should not be considered an indicator of competency in gastric cancer testing. The team's pathologists should score a minimum of 50 gastric cancer cases per year to build expertise and accuracy of scoring and interpretation, verified by participation in ring studies.

umor. Prior HER2 testing experience in breast cancer should not be considered an indicator of competency in gastric cancer testing. The team's pathologists should score a minimum of 50 gastric cancer cases per year to build expertise and accuracy of scoring and interpretation, verified by participation in ring studies. Ongoing quality assurance programs can also help laboratories to ensure standards are consistently achieved in day-to-day testing. Internally, laboratories can maintain an audit of HER2 test results to detect changes in testing trends, such as positivity rates and scoring distributions, and to provide evidence when investigating sources of variations. Externally, laboratories may consider participation in proficiency testing and external quality assurance programs, like UK NEQAS. These external programs allow laboratories to continuously monitor HER2 testing quality against peers/published positivity rates.

provide evidence when investigating sources of variations. Externally, laboratories may consider participation in proficiency testing and external quality assurance programs, like UK NEQAS. These external programs allow laboratories to continuously monitor HER2 testing quality against peers/published positivity rates. CONCLUSION HER2 testing laboratories in Asia-Pacific countries see a high incidence of gastric cancer and as such perform thousands of tests on unselected patients outside the context of clinical trials. The experience of a multidisciplinary task force from Asia-Pacific countries has been utilized to provide HER2 testing recommendations that are of particular relevance to the Asia-Pacific region. Multidisciplinary collaborations are the cornerstone of best practice to ensure patients receive the best possible treatment for their disease. Laboratories should make every effort to comply with best practice guidelines in order to improve the accuracy and reliability of HER2 testing in gastric cancer. Support for third-party writing assistance for this manuscript was provided by F. Hoffmann-La Roche Ltd.

CONCLUSION HER2 testing laboratories in Asia-Pacific countries see a high incidence of gastric cancer and as such perform thousands of tests on unselected patients outside the context of clinical trials. The experience of a multidisciplinary task force from Asia-Pacific countries has been utilized to provide HER2 testing recommendations that are of particular relevance to the Asia-Pacific region. Multidisciplinary collaborations are the cornerstone of best practice to ensure patients receive the best possible treatment for their disease. Laboratories should make every effort to comply with best practice guidelines in order to improve the accuracy and reliability of HER2 testing in gastric cancer. Support for third-party writing assistance for this manuscript was provided by F. Hoffmann-La Roche Ltd. Conflicts of interest: KMK, MB, BSK, YSP and MHR have received speaker's and/or consultancy fees from Roche. KMC has received speaker's and/or consultancy fees from Roche and has had clinical trial involvement with Roche, Bristol Myers Squibb, ImClone Systems, Merck Serono and Novartis. WHK and JW have received speaker's and/or consultancy fees and travel grants from Roche. WS, YC, JMY and SZ have no conflicting interests to disclose. KMK, MB, BSK, YSP, MHR, KMC, WHK and JW received speaker's and/or consultancy fees or travel grants as part of SPHERE. SPHERE is an educational initiative that promotes and facilitates excellence in HER2 testing in breast and gastric cancer across the Asia-Pacific region. It aims to achieve this by conducting regional pathology workshops and by highlighting the importance of multidisciplinary collaboration between surgeons, oncologists, pathologists and laboratory scientists in ensuring accurate test results. This program is funded by F. Hoffmann-La Roche Ltd.

1 INTRODUCTION Nanoparticle albumin‐bound paclitaxel (nab‐PTX) showed a significantly higher overall response rate (ORR) than solvent‐based paclitaxel (sb‐PTX) in combination with carboplatin for patients with non‐small cell lung cancer (NSCLC) as first‐line treatment (33% vs 25%, P = 0.001). The frequency of some serious adverse events (AEs) in the nab‐PTX arm, such as peripheral neuropathy and neutropenia, was less than that in the sb‐PTX arm.1 A 4‐week cycle treatment (days 1, 8 and 15) of 125 mg/m2 of nab‐PTX for patients with naïve advanced NSCLC demonstrated efficacy and tolerability in phase I/II of the study; ORR and median time to progression were 30% and 5 months, respectively.2 In addition, a phase II study of weekly nab‐PTX in patients previously treated for advanced NSCLC demonstrated acceptable toxicity and promising activity; ORR and median progression‐free survival (mPFS) were 31.7% (95% confidence interval [CI] of 19.3–44.1%) and 4.9 months (95% CI, 2.4–7.4 months), respectively.3

ely.2 In addition, a phase II study of weekly nab‐PTX in patients previously treated for advanced NSCLC demonstrated acceptable toxicity and promising activity; ORR and median progression‐free survival (mPFS) were 31.7% (95% confidence interval [CI] of 19.3–44.1%) and 4.9 months (95% CI, 2.4–7.4 months), respectively.3 Epidermal growth factor receptor (EGFR) mutations may be predictive biomarkers for the effects of cytotoxic chemotherapy, according to some phase III randomized studies comparing the efficacy of EGFR‐tyrosine kinase inhibitors with cytotoxic chemotherapies in NSCLC patients. In the INTEREST study, the ORR of docetaxel (DOC) was 21.1% for the EGFR mutant, whereas it was 9.8% for wild‐type patients.4 In the V‐15‐32 study, the ORR of DOC was 46% for the EGFR mutant, whereas it was 13% for wild‐type patients.5 In the TRIBUTE trial, the progressive disease rate of carboplatin and paclitaxel chemotherapy was 21% for the EGFR mutant, whereas it was 37% for wild‐type patients.6 However, in the DELTA study, comparing erlotinib with DOC in previously treated NSCLC patients, the ORRs of DOC were similar: 17.9% for the EGFR‐unselected patients and 20.0% for wild‐type patients.7, 8

e of carboplatin and paclitaxel chemotherapy was 21% for the EGFR mutant, whereas it was 37% for wild‐type patients.6 However, in the DELTA study, comparing erlotinib with DOC in previously treated NSCLC patients, the ORRs of DOC were similar: 17.9% for the EGFR‐unselected patients and 20.0% for wild‐type patients.7, 8 As the importance of individualized medicine increases, it is necessary to establish a treatment strategy that takes into consideration the presence or absence of driver mutations, such as EGFR or anaplastic lymphoma kinase (ALK). To date, no clinical trial prospectively examined the efficacy and safety of nab‐PTX for previously treated NSCLC patients without EGFR or ALK mutations. We believed that the development of a superior second‐line treatment was important for such cases. With this background, we determined the recommended schedule of weekly nab‐PTX in phase I of the study. Subsequently, in phase II, we evaluated the efficacy and safety of nab‐PTX with the recommended schedule as second‐ or third‐line treatment for patients with advanced NSCLC without any driver mutation.

As the importance of individualized medicine increases, it is necessary to establish a treatment strategy that takes into consideration the presence or absence of driver mutations, such as EGFR or anaplastic lymphoma kinase (ALK). To date, no clinical trial prospectively examined the efficacy and safety of nab‐PTX for previously treated NSCLC patients without EGFR or ALK mutations. We believed that the development of a superior second‐line treatment was important for such cases. With this background, we determined the recommended schedule of weekly nab‐PTX in phase I of the study. Subsequently, in phase II, we evaluated the efficacy and safety of nab‐PTX with the recommended schedule as second‐ or third‐line treatment for patients with advanced NSCLC without any driver mutation. 2 PATIENTS AND METHODS 2.1 Patients Patients aged ≥20 years were enrolled in the study. The inclusion criteria are as follows: histologically or cytologically confirmed advanced NSCLC; an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0–2; measurable lesions documented by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 and adequate organ functions; and progressed after one or two chemotherapy regimens including platinum‐doublet chemotherapy. The driver mutation status for each patient was also confirmed; EGFR mutation was negative and ALK fusion status was negative or unknown. Patients who had a history of sb‐PTX or nab‐PTX and had symptomatic brain or meningeal metastasis were excluded.

or two chemotherapy regimens including platinum‐doublet chemotherapy. The driver mutation status for each patient was also confirmed; EGFR mutation was negative and ALK fusion status was negative or unknown. Patients who had a history of sb‐PTX or nab‐PTX and had symptomatic brain or meningeal metastasis were excluded. This study protocol was developed according to the Declaration of Helsinki, and ethical guidelines for clinical research were approved by the ethics review boards of Shikoku Cancer Center, Matsuyama, Japan [H25‐71] and each participating institution. The unique ID issued by UMIN was UMIN000012404. Written informed consent was obtained from each patient before enrollment. All patients provided written informed consent before enrollment. 2.2 Treatment and assessment in phase I This study was an open‐label, multicenter, single‐arm prospective study.

This study protocol was developed according to the Declaration of Helsinki, and ethical guidelines for clinical research were approved by the ethics review boards of Shikoku Cancer Center, Matsuyama, Japan [H25‐71] and each participating institution. The unique ID issued by UMIN was UMIN000012404. Written informed consent was obtained from each patient before enrollment. All patients provided written informed consent before enrollment. 2.2 Treatment and assessment in phase I This study was an open‐label, multicenter, single‐arm prospective study. In phase I, a 4‐week cycle (dose level 0: days 1, 8, 15 and 22 or dose level –1: days 1, 8 and 15) of 100 mg/m2 nab‐PTX was administered until disease progression or unacceptable AEs were observed. The study treatment was started at dose level 0 and six patients were initially enrolled. If the predefined dose‐limiting toxicity (DLT) was observed in zero or one patient, we determined the recommended schedule to be dose level 0, and subsequently proceeded to phase II. If DLT was observed in two patients, six additional patients were enrolled and evaluated at dose level 0. If DLT was observed in ≥3 out of six patients, we determined the recommended schedule to be dose level –1. If DLT was observed in ≤5 out of 12 patients, we determined the recommended schedule to be dose level 0. If DLT was observed in ≥6 out of 12 patients, we determined the recommended schedule to be dose level –1. The definition of DLTs is defined as follows; nab‐PTX was not administered on days 8, 15 and 22 in the first cycle because of neutrophil count of < 1000/#x000B5;L, platelet count of <50 000/#x000B5;L, infection, peripheral neuropathy, aspartate aminotransferase (AST) increased, alanine aminotransferase (ALT) increased, blood bilirubin increased, creatinine increased, mucositis or diarrhea of ≥grade 2 or the other nonhematological toxicity of ≥ grade 3, the administration of nab‐PTX on day 1 in the second cycle was late for 8 or more days because of neutrophil count of <1500/#x000B5;L, platelet count of <100 000/#x000B5;L, hemoglobin of <8.0 g/dL, blood bilirubin of ≥ 1.5 mg/dL or creatinine of >1.5 mg/dL, peripheral neuropathy or infection of ≥grade 2, pneumonitis of any grade, the other nonhematological toxicities of ≥grade 3 or ECOG PS 3 or more.

ays because of neutrophil count of <1500/#x000B5;L, platelet count of <100 000/#x000B5;L, hemoglobin of <8.0 g/dL, blood bilirubin of ≥ 1.5 mg/dL or creatinine of >1.5 mg/dL, peripheral neuropathy or infection of ≥grade 2, pneumonitis of any grade, the other nonhematological toxicities of ≥grade 3 or ECOG PS 3 or more. The primary endpoints of phase I were feasibility and determination of the recommended nab‐PTX schedule. 2.3 Treatment and assessment in phase II In phase II, nab‐PTX was administered according to the recommended schedule determined in phase I until disease progression and unacceptable AEs were observed. The primary endpoint of phase II was ORR and the secondary endpoints were overall survival (OS), PFS and safety. Tumor response was assessed according to the RECIST v1.1. AEs were graded in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI‐CTCAE), version 4.0.

e observed. The primary endpoint of phase II was ORR and the secondary endpoints were overall survival (OS), PFS and safety. Tumor response was assessed according to the RECIST v1.1. AEs were graded in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI‐CTCAE), version 4.0. 2.4 Statistical analysis The sample size was calculated according to Simon's optimal and minimax two‐stage sequential design. The expected ORR and threshold ORR for this study were assumed to be 15% and 5%, respectively.9 Given this assumption, calculation of the required number of subjects with α = 0.05 (one‐sided) and β = 0.2 yielded 52 patients; considering that some patients may be ineligible, the planned enrollment number for phase II was set at 55 patients. These 55 patients included patients enrolled in phase I with the recommended schedule. We considered weekly nab‐PTX to be an effective treatment regimen, if complete response (CR) or partial response (PR) was confirmed in ≥9 patients from the total of 55. 3 RESULTS 3.1 Phase I A total of five patients were enrolled in phase I. The patient characteristics in phase I are presented in Table 1. The recommended schedule of weekly nab‐PTX was determined as level –1, because DLT was observed in four of five patients. The contents of DLTs were grade 3/4 neutropenia in three patients (60%) and grade 2 pneumonitis in one patient (20%). Table 1 Characteristics of patients in the study (n = 60)

3 RESULTS 3.1 Phase I A total of five patients were enrolled in phase I. The patient characteristics in phase I are presented in Table 1. The recommended schedule of weekly nab‐PTX was determined as level –1, because DLT was observed in four of five patients. The contents of DLTs were grade 3/4 neutropenia in three patients (60%) and grade 2 pneumonitis in one patient (20%). Table 1 Characteristics of patients in the study (n = 60) Phase I (n = 5) Phase II (n = 55) Characteristics n % n % Age (years) Median (range) 67 (61–71) 66 (41–90) Sex Male 5 100.0 40 72.7 Female 0 0.0 15 27.3 Smoking status Nonsmoker 2 40.0 12 21.8 Ex‐smoker 1 20.0 39 70.9 Current smoker 2 40.0 4 7.3 ECOG performance status 0 1 20.0 12 21.8 1 4 80.0 39 70.9 2 0 0.0 4 7.3 Histology Adenocarcinoma 4 80.0 34 61.8 Squamous cell carcinoma 1 20.0 17 30.9 Large cell carcinoma 0 0.0 2 3.6 Others 0 0.0 2 3.6 Disease stage IIIB 1 20.0 6 10.9 IV 4 80.0 37 67.3 Postoperative recurrence 0 0.0 12 21.8 Treatment line Second line 5 100.0 34 61.8 Third line 0 0.0 21 38.2 Previous treatment with docetaxel 1 20.0 20 36.4 Abbreviation: ECOG, Eastern Cooperative Oncology Group.

30.9 Large cell carcinoma 0 0.0 2 3.6 Others 0 0.0 2 3.6 Disease stage IIIB 1 20.0 6 10.9 IV 4 80.0 37 67.3 Postoperative recurrence 0 0.0 12 21.8 Treatment line Second line 5 100.0 34 61.8 Third line 0 0.0 21 38.2 Previous treatment with docetaxel 1 20.0 20 36.4 Abbreviation: ECOG, Eastern Cooperative Oncology Group. John Wiley & Sons, Ltd.3.2 Phase II A total of 55 patients were enrolled in phase II between April 2014 and July 2016. The patient characteristics in phase II are presented in Table 1. The median age was 66 years (range, 41–90 years). The proportion of male patients was 72.7% and the PS 0/1/2 was 12/39/4, respectively. Thirty‐four (61.8%) patients were administered second‐line therapy. Twenty patients (36.3%) previously received DOC treatment. In 21 patients, (38.2%) who previously received both first‐ and second‐line treatments, DOC was most frequently administered as the second‐line treatment (52.4%). The proportion of patients with adenocarcinoma and squamous cell carcinoma (Sq) was 34 (61.8%) and 17 (30.9%) patients, respectively.

received DOC treatment. In 21 patients, (38.2%) who previously received both first‐ and second‐line treatments, DOC was most frequently administered as the second‐line treatment (52.4%). The proportion of patients with adenocarcinoma and squamous cell carcinoma (Sq) was 34 (61.8%) and 17 (30.9%) patients, respectively. 3.3 Efficacy Treatment efficacy is summarized in Table 2. All 55 patients were eligible for efficacy analysis. Based on the investigator's assessment, four patients had a PR, and none demonstrated CR, yielding an ORR of 7.3% (95% CI, 2.0–17.6%). Twenty‐six patients had stable disease (SD), yielding a disease control rate (DCR: CR + PR + SD) of 54.5% (95% CI, 40.6–68.0%). In addition, subanalysis of clinical concerned factors predicting ORR and DCR was conducted (Table 3). The patients with non‐Sq histology tended to have better ORR and DCR. At the median follow‐up time of 9.6 months (range, 2.1–34.8 months) for all patients, mPFS was 3.4 months (95% CI, 1.9–4.0 months) and median survival time was 10.6 months (95% CI, 6.9–17.8 months) (Figure 1). The median number of treatment cycles was three. Among all treated patients, nine (16.4%) reduced the dose of nab‐PTX. The median dose intensity (DI) was 62.5 mg/m2 per week and the median relative DI (RDI) was 83.3%. Table 2 Objective response

3.3 Efficacy Treatment efficacy is summarized in Table 2. All 55 patients were eligible for efficacy analysis. Based on the investigator's assessment, four patients had a PR, and none demonstrated CR, yielding an ORR of 7.3% (95% CI, 2.0–17.6%). Twenty‐six patients had stable disease (SD), yielding a disease control rate (DCR: CR + PR + SD) of 54.5% (95% CI, 40.6–68.0%). In addition, subanalysis of clinical concerned factors predicting ORR and DCR was conducted (Table 3). The patients with non‐Sq histology tended to have better ORR and DCR. At the median follow‐up time of 9.6 months (range, 2.1–34.8 months) for all patients, mPFS was 3.4 months (95% CI, 1.9–4.0 months) and median survival time was 10.6 months (95% CI, 6.9–17.8 months) (Figure 1). The median number of treatment cycles was three. Among all treated patients, nine (16.4%) reduced the dose of nab‐PTX. The median dose intensity (DI) was 62.5 mg/m2 per week and the median relative DI (RDI) was 83.3%. Table 2 Objective response Best response n % CR 0 0 PR 4 7.3 SD 26 47.3 PD 24 43.6 NE 1 1.8 Response rate % [95% CI] ORR 7.3 [2.0–17.6] DCR 54.5 [40.6–68.0] Abbreviations: CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not evaluated; ORR, overall response rate; DCR, disease control rate; CI, confidence interval. John Wiley & Sons, Ltd.Table 3 Subanalysis of ORR and DCR

Best response n % CR 0 0 PR 4 7.3 SD 26 47.3 PD 24 43.6 NE 1 1.8 Response rate % [95% CI] ORR 7.3 [2.0–17.6] DCR 54.5 [40.6–68.0] Abbreviations: CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not evaluated; ORR, overall response rate; DCR, disease control rate; CI, confidence interval. John Wiley & Sons, Ltd.Table 3 Subanalysis of ORR and DCR No. of pts (n = 55) ORR ［95% CI］ DCR ［95% CI］ Previous docetaxel treatment Yes 20 5.0 ［0.1–24.8］ 45.0 ［23.1–68.4］ No 35 8.5 ［1.8–23.1］ 60.0 ［42.1–76.1］ Treatment line Second line 34 8.8 ［1.9–23.7］ 61.8 ［43.6–77.8］ Third line 21 4.7 ［0.1–23.8］ 42.9 ［21.8–66.0］ Histology Nonsquamous 38 10.5 ［2.9–24.8］ 63.2 ［46.0–78.2］ Squamous 17 0 ［0–19.5］ 35.3 ［14.2–61.7］ Abbreviations: ORR, overall response rate; DCR, disease control rate; CI, confidence interval. John Wiley & Sons, Ltd.Figure 1 Progression‐free survival and overall survival Progression‐free survival curve (A) and overall survival curve (B) Abbreviations: PFS, progression‐free survival; CI, confidence interval; OS, overall survival; MST, median survival time

No. of pts (n = 55) ORR ［95% CI］ DCR ［95% CI］ Previous docetaxel treatment Yes 20 5.0 ［0.1–24.8］ 45.0 ［23.1–68.4］ No 35 8.5 ［1.8–23.1］ 60.0 ［42.1–76.1］ Treatment line Second line 34 8.8 ［1.9–23.7］ 61.8 ［43.6–77.8］ Third line 21 4.7 ［0.1–23.8］ 42.9 ［21.8–66.0］ Histology Nonsquamous 38 10.5 ［2.9–24.8］ 63.2 ［46.0–78.2］ Squamous 17 0 ［0–19.5］ 35.3 ［14.2–61.7］ Abbreviations: ORR, overall response rate; DCR, disease control rate; CI, confidence interval. John Wiley & Sons, Ltd.Figure 1 Progression‐free survival and overall survival Progression‐free survival curve (A) and overall survival curve (B) Abbreviations: PFS, progression‐free survival; CI, confidence interval; OS, overall survival; MST, median survival time 3.4 Safety All 55 patients enrolled in the study treatment were eligible for safety analysis. The major treatment‐related toxicities are presented in Tables 4 and 5. The major nonhematologic toxicities (total/grade 3 or more) were peripheral sensory neuropathy (49.1%/1.8%), fatigue (27.3%/0%) and anorexia (27.3%/1.8%). The most frequent treatment‐related grade 3 or 4 toxicities were neutropenia (36%), febrile neutropenia (5.5%) and pulmonary infection (3.6%). Five patients (9.1%) and one patient received granulocyte‐colony‐stimulating factor support and erythrocyte transfusion, respectively. Seven patients (13%) discontinued the study treatment because of grade 2 pneumonitis (n = 3), grade 3 AST/ALT elevation (n = 1), grade 4 sepsis (n = 1), grade 3 anorexia (n = 1) and grade 5 adult respiratory distress syndrome (ARDS) (n = 1).

e‐colony‐stimulating factor support and erythrocyte transfusion, respectively. Seven patients (13%) discontinued the study treatment because of grade 2 pneumonitis (n = 3), grade 3 AST/ALT elevation (n = 1), grade 4 sepsis (n = 1), grade 3 anorexia (n = 1) and grade 5 adult respiratory distress syndrome (ARDS) (n = 1). Table 4 Treatment‐related adverse events in phase II of the trial Adverse event Total (%) Grade 3 or above (%) General Peripheral sensory neuropathy 49.1 1.8 Fatigue 27.3 0.0 Anorexia 27.3 1.8 Nausea 12.7 0.0 Myalgia 12.7 0.0 Peripheral motor neuropathy 9.1 1.8 Arthralgia 7.3 0.0 Vomiting 5.5 1.8 Pneumonitis 5.5 0.0 Febrile neutropenia 5.5 5.5 Pulmonary infection 3.6 3.6 Diarrhea 3.6 1.8 Sepsis 1.8 1.8 ARDS 1.8 1.8 Abbreviation: ARDS, adult respiratory distress syndrome. John Wiley & Sons, Ltd.Table 5 Treatment‐related adverse events in phase II of the trial Adverse event Grade 3 or above (%) Hematologic Neutropenia 36.4 Anemia 1.8 Thrombocytopenia 0.0 Biochemical Blood bilirubin increased 1.8 AST/ALT elevation 1.8 Hyponatremia 1.8 Hypokalemia 1.8 Abbreviations: ALT, alanine amino transferase; AST, aspartate amino transferase.

John Wiley & Sons, Ltd.Table 5 Treatment‐related adverse events in phase II of the trial Adverse event Grade 3 or above (%) Hematologic Neutropenia 36.4 Anemia 1.8 Thrombocytopenia 0.0 Biochemical Blood bilirubin increased 1.8 AST/ALT elevation 1.8 Hyponatremia 1.8 Hypokalemia 1.8 Abbreviations: ALT, alanine amino transferase; AST, aspartate amino transferase. John Wiley & Sons, Ltd.Two deaths (3.6%) were observed during the protocol study. One patient developed grade 5 ARDS on day 26 of the second cycle. Considering that the patient had increasing bilateral pleural effusion and exacerbation of dyspnea on day 1 of the second cycle and home oxygen therapy was introduced, his cancer may have progressed after the first cycle. Nevertheless, the patient received 2 mg oral dexamethasone to treat dyspnea, which was an inhibited agent in this protocol and the second treatment cycle was started. This case was considered to deviate from the study protocol. This deviation may have caused the patient's death. Thus, we considered this patient's demise to be probably a nab‐PTX treatment‐related death. The other patient died on day 10 of the fourth cycle, two days after day 8 of nab‐PTX administration. Loss of consciousness occurred as the patient was going to the toilet at his home on day 9. He vomited, and subsequently experienced a cardiac arrest. In this case, AEs of nausea, vomiting and loss of appetite had not yet been reported. Despite emergency cardiopulmonary resuscitation, the patient died the next day. Computed tomography imaging of the head revealed no evidence of cerebral bleeding or infarction. Although grade 4 vomiting in this case was considered to be possibly related to nab‐PTX treatment, the causal relationship between death and the treatment protocol is unknown.

y resuscitation, the patient died the next day. Computed tomography imaging of the head revealed no evidence of cerebral bleeding or infarction. Although grade 4 vomiting in this case was considered to be possibly related to nab‐PTX treatment, the causal relationship between death and the treatment protocol is unknown. 3.5 Poststudy chemotherapy Overall, 70.9% (39/55) patients received subsequent systemic chemotherapy after poststudy treatment. The median number of poststudy treatment lines was one (range: 0–5). Drugs administered to the patients were as follows: nivolumab (30.9%), DOC (18.2%), S1 (12.7%), pemetrexed (10.9%) and vinorelbine (10.9%).

emotherapy Overall, 70.9% (39/55) patients received subsequent systemic chemotherapy after poststudy treatment. The median number of poststudy treatment lines was one (range: 0–5). Drugs administered to the patients were as follows: nivolumab (30.9%), DOC (18.2%), S1 (12.7%), pemetrexed (10.9%) and vinorelbine (10.9%). 4 DISCUSSION We performed a prospective, multicenter phase I/II study of weekly nab‐PTX therapy in patients with advanced NSCLC without EGFR mutations or ALK rearrangement who were previously treated with platinum‐doublet chemotherapy. Although the level –1 schedule was not investigated in phase I, the proportion of AEs as the cause of discontinuation of therapy and dose reduction in phase II was 9/55 (16.4%) and 7/55 (13%) patients, respectively. The incidence of nonhematologic toxicities of grade ≥3 with the level –1 schedule was < 5%, except for febrile neutropenia (5.5%). Hematological toxicity was mild, with an incidence of 36.4% for grade ≥3 neutropenia. In contrast, the incidence of grade ≥3 neutropenia was 90.1% and that of febrile neutropenia was 19.1% in the DOC arm of phase III studies in previously treated Japanese patients with advanced NSCLC.10 Considering these factors, the level –1 schedule was feasible and appropriate in this setting.

r grade ≥3 neutropenia. In contrast, the incidence of grade ≥3 neutropenia was 90.1% and that of febrile neutropenia was 19.1% in the DOC arm of phase III studies in previously treated Japanese patients with advanced NSCLC.10 Considering these factors, the level –1 schedule was feasible and appropriate in this setting. Although patients with histology of non‐Sq tended to have better ORR in subanalysis, the experimental regimen yielded an ORR of 7.3% (95% CI, 2.0–17.6%) in total, which did not meet the primary endpoint of the study. In other clinical trials involving EGFR‐mutated NSCLC patients, ORR was 14.5–31.7%, indicating good efficacy.3, 11, 12 In the KTOSG trial 1301, indicating the proportion of cases with EGFR mutations, 56.1% patients had wild‐type EGFR mutations, but the effectiveness analysis with or without EGFR mutation has not yet been investigated.3 Although this may not be the only reason for the low ORR obtained in this study, targeting only wild‐type EGFR patients may have resulted in this outcome. In this study, the median DI was 62.5 mg/m2/week. This value was 89.1 mg/m2/week in the KTOSG trial 1301. The low DI may be one of the reasons for low ORR in our study. Although the prescribed DI was 75 mg/m2/week, skipping treatment or dose reduction of nab‐PTX may have resulted in the reduction of the RDI to 83.3%. The proportion of patients who skipped nab‐PTX treatment on day 15 was 62/194 cycles (31.9%), which may be the main reason for the reduced RDI.

for low ORR in our study. Although the prescribed DI was 75 mg/m2/week, skipping treatment or dose reduction of nab‐PTX may have resulted in the reduction of the RDI to 83.3%. The proportion of patients who skipped nab‐PTX treatment on day 15 was 62/194 cycles (31.9%), which may be the main reason for the reduced RDI. However, DCR, PFS and OS were not lower than those in other trials, where patients with EGFR mutation were included (Table 6).3, 11, 12 The argument whether ORR is appropriate to determine the efficiency criteria in second‐ or third‐line treatment is ongoing. Because PFS and OS of patients ≥70 years administered nab‐PTX in the CA031 trial tended to be superior to those administered sb‐PTX, a phase III trial (ABOUND .70+ trial) was conducted to investigate the efficacy and safety of weekly nab‐PTX either continuously or with a 1‐week interval, both in combination with carboplatin for patients aged ≥70 years.13 The study showed that the 1‐week break between treatment cycles significantly improved PFS (mPFS was 3.6 and 7.0 months [Hazard Ratio 0.48, P < 0.0019]) and ORR (23.9% and 40.3%; P = 0.0376). These findings support the safety and efficacy of first‐line nab‐paclitaxel/carboplatin in elderly patients with advanced NSCLC. Table 6 The treatment response and survival benefit

However, DCR, PFS and OS were not lower than those in other trials, where patients with EGFR mutation were included (Table 6).3, 11, 12 The argument whether ORR is appropriate to determine the efficiency criteria in second‐ or third‐line treatment is ongoing. Because PFS and OS of patients ≥70 years administered nab‐PTX in the CA031 trial tended to be superior to those administered sb‐PTX, a phase III trial (ABOUND .70+ trial) was conducted to investigate the efficacy and safety of weekly nab‐PTX either continuously or with a 1‐week interval, both in combination with carboplatin for patients aged ≥70 years.13 The study showed that the 1‐week break between treatment cycles significantly improved PFS (mPFS was 3.6 and 7.0 months [Hazard Ratio 0.48, P < 0.0019]) and ORR (23.9% and 40.3%; P = 0.0376). These findings support the safety and efficacy of first‐line nab‐paclitaxel/carboplatin in elderly patients with advanced NSCLC. Table 6 The treatment response and survival benefit Age, median ORR DCR mPFS MST Author n (range) Dose (%) (%) [95% CI] [95% CI] Sakata et al.3 41 68 100 mg/m2 31.7 65.9 4.9 11 (43–77) Q3w, days 1, 8 and 15 [2.4–7.4] Hu et al.12 56 59.6 100 mg/m2 16.1 51.7 3.5 6.8 (32–83) Q4w, days 1, 8 and 15 [1.9–5.8] [4.7–9.3] Liu et al.11 55 52.5 150 mg/m2 14.5 65.5 4.9 11 (29–74) Q3w, days 1 and 8 [2.4–7.4] Present study 55 66 100 mg/m2 7.3 54.5 3.4 10.6 (41–90) Q4w, days 1, 8 and 15 [1.9–4.0] [6.9–17.8] Abbreviations: ORR, objective response rate; DCR, disease control rate; MST, median survival time; mPFS, median progression‐free survival; CI, confidence interval.

g/m2 14.5 65.5 4.9 11 (29–74) Q3w, days 1 and 8 [2.4–7.4] Present study 55 66 100 mg/m2 7.3 54.5 3.4 10.6 (41–90) Q4w, days 1, 8 and 15 [1.9–4.0] [6.9–17.8] Abbreviations: ORR, objective response rate; DCR, disease control rate; MST, median survival time; mPFS, median progression‐free survival; CI, confidence interval. John Wiley & Sons, Ltd.In preclinical data, it was shown that combination therapy with PTX/nab‐PTX and angiogenesis inhibitor bevacizumab increased the antitumor effect as compared with PTX/nab‐PTX monotherapy.14 Actually, in advanced gastric cancer, angiogenesis inhibitor ramucirumab + PTX has been statistically significantly prolonged OS compared with PTX alone as a secondary treatment in patients who progressed after platinum‐containing chemotherapy, and it is regarded as standard treatment.15 In the future, it might be necessary to verify the effectiveness and safety of combination therapy of nab‐PTX and angiogenesis inhibitor as a second‐line chemotherapy in NSCLC. This study failed to meet predefined primary endpoints for patients with advanced NSCLC, although the PFS and OS were comparable with those in previous reports and toxicity was acceptable. The weekly nab‐PTX was not a promising treatment for NSCLC patients without EGFR or ALK mutations as a second‐ or third‐line treatment setting.

ailed to meet predefined primary endpoints for patients with advanced NSCLC, although the PFS and OS were comparable with those in previous reports and toxicity was acceptable. The weekly nab‐PTX was not a promising treatment for NSCLC patients without EGFR or ALK mutations as a second‐ or third‐line treatment setting. ACKNOWLEDGMENTS For the study, there are no financial grants and other funding. We disclose the conflict of interests as follows; Daijiro Harada has been paid lecture fees to participate in a speakers’ bureau by Ono Pharmaceutical Co., Ltd., Bristol‐Myers Squibb Company, Yakult Honsha Co., Ltd., Kyowa Hakko Kirin Co., Ltd., AstraZeneca K.K., Nippon Boehringer Ingelheim Co., Ltd. and Eli Lilly Japan. Toshiyuki Kozuki has received honoraria outside the current work from Chugai Pharmaceutical, AstraZeneca, Eli Lilly Japan, Boehringer–Ingelheim, Ono Pharmaceutical, Bristol‐Myers Squibb Company, Taiho Pharmaceutical, MSD K.K., Pfizer Inc. Japan and Kyowa Hakko Kirin. Naoyuki Nogami has received honoraria from Astellas Pharma, AstraZeneca, Ono Pharmaceutical, Taiho Pharmaceutical, Chugai Pharmaceutical, Eli Lilly, Boehringer Ingelheim and Pfizer. Akihiro Bessho has received a speaker's fee from Taiho Pharmaceutical, Ono Pharmaceutical, Eli Lilly Japan, Bristol‐Myers Squibb and MSD K.K. Katsuyuki Hotta has received grants and personal fees from AstraZeneca, Ono Pharmaceutical, Boehringer‐Ingelheim, Chugai Pharmaceutical, Novartis, BMS, Eli Lilly Japan and MSD, personal fees from Nihon Kayaku and Taiho Pharmaceutical. Hiroshige Yoshioka has received honoraria outside the current work from Boehringer–Ingelheim, AstraZeneca, Eli Lilly Japan, Chugai Pharmaceutical, Bristol‐Myers Squibb Company, Ono Pharmaceutical, Taiho Pharmaceutical and Takeda Pharmaceutical Co. Ltd. Toshihide Yokoyama has been paid lecture fees to participate in a speakers’ bureau by Ono Pharmaceutical Co., Ltd., AstraZeneca K.K., Nippon Boehringer Ingelheim Co., Ltd., Chugai Pharmaceutical, MSD, Taiho Pharmaceutical and Eli Lilly Japan. Nagio Takigawa has received honoraria outside the current work from Eli Lilly Japan, AstraZeneca, Daiichi–Sankyo Pharmaceutical, Chugai Pharmaceutical, Taiho Pharmaceutical, Pfizer Inc. Japan, Boehringer–Ingelheim and Ono Pharmaceutical. Katsuyuki Kiura has received honoraria from Eli Lilly Japan, Nihon Kayaku, AstraZeneca, Daiichi‐Sankyo Pharmaceutical, Chugai Pharmaceutical, Taiho Pharmaceutical and Sanofi‐Aventis.

iichi–Sankyo Pharmaceutical, Chugai Pharmaceutical, Taiho Pharmaceutical, Pfizer Inc. Japan, Boehringer–Ingelheim and Ono Pharmaceutical. Katsuyuki Kiura has received honoraria from Eli Lilly Japan, Nihon Kayaku, AstraZeneca, Daiichi‐Sankyo Pharmaceutical, Chugai Pharmaceutical, Taiho Pharmaceutical and Sanofi‐Aventis. The remaining authors declare that they have no competing interests. We would like to thank the patients and their family members, all participating physicians and medical assistants from cooperating hospitals and the independent review committee members for their involvement in the study, and Editage (http://www.editage.jp) for English language editing.

1 INTRODUCTION Prostate cancer is the second most common cancer affecting men, with an estimated 1.1 million new cases in the world in 2012.1, 2 Primary therapy consists of radical prostatectomy or radiotherapy. However, patients with a positive margin, extraprostatic extension, lymph node involvement, high prostate‐specific antigen (PSA), or high Gleason Score (GS) are at high risk of prostate cancer recurrence following primary therapy. In these patients, androgen deprivation therapy (ADT) can be given as neoadjuvant therapy prior to primary therapy to shrink the tumor and reduce margin positivity. Radiotherapy, ADT, or a combination of the two can also be given as adjuvant treatment following primary therapy to reduce the risk of recurrence. There are several different ADT modalities, which aim to deplete androgen levels by suppressing testicular androgen secretion or by inhibiting circulating androgens through targeting the androgen receptor. Consequently, ADT can be delivered by medical or surgical castration, antiandrogen therapy, and combined androgen blockade (CAB).3, 4

ifferent ADT modalities, which aim to deplete androgen levels by suppressing testicular androgen secretion or by inhibiting circulating androgens through targeting the androgen receptor. Consequently, ADT can be delivered by medical or surgical castration, antiandrogen therapy, and combined androgen blockade (CAB).3, 4 Although ADT monotherapy is not appropriate for clinically localized prostate cancer, the addition of ADT to primary therapy has been shown to improve outcomes significantly for certain men with intermediate‐ or high‐risk prostate cancer.4 However, many questions still remain unanswered, including the ideal patient population and optimal timing and duration of therapy. Exploring these questions is complicated by long survival and observation times, leading to fewer opportunities to conduct ideal randomized clinical trials. Furthermore, diverse study endpoints make comparisons difficult and a standard comparator is lacking. A need to address this challenge in prostate cancer patients has been exemplified by the Intermediate Clinical Endpoints in Cancer of the Prostate (ICECap) working group in the development of an intermediate clinical endpoint to serve as a robust surrogate for overall survival (OS).5 This review, therefore, discusses the current findings on the role of ADT in the treatment of castrate‐sensitive nonmetastatic prostate cancer. We also propose a treatment roadmap for ADT in this setting based on the available evidence.

ediate clinical endpoint to serve as a robust surrogate for overall survival (OS).5 This review, therefore, discusses the current findings on the role of ADT in the treatment of castrate‐sensitive nonmetastatic prostate cancer. We also propose a treatment roadmap for ADT in this setting based on the available evidence. 2 METHODS A PubMed search of all prospective and retrospective studies or meta‐analyses evaluating the outcomes of men treated with ADT for nonmetastatic prostate cancer published since 2000 was conducted. Findings on the use of neoadjuvant and adjuvant therapy in combination with radical prostatectomy or radiotherapy, and ADT at the time of biochemical recurrence were reviewed. Based on existing publications, long‐term ADT treatment has been defined as treatment duration ≥18 months.6 3 RESULTS 3.1 ADT for patients who received radical prostatectomy as a primary treatment Radical prostatectomy is typically used for patients with localized disease who have an estimated life expectancy of over 10 years, and in patients with locally advanced disease. 3.1.1 Neoadjuvant ADT plus radical prostatectomy Men with early‐stage prostate cancer with intermediate or high risk of recurrence may be considered for neoadjuvant ADT prior to primary treatment. Neoadjuvant ADT before prostatectomy has been shown to provide long‐term progression‐free survival (PFS)2 and to significantly reduce the risk of recurrence (Table 1)7; however, it has generally not been shown to extend OS.7

te or high risk of recurrence may be considered for neoadjuvant ADT prior to primary treatment. Neoadjuvant ADT before prostatectomy has been shown to provide long‐term progression‐free survival (PFS)2 and to significantly reduce the risk of recurrence (Table 1)7; however, it has generally not been shown to extend OS.7 Table 1 Summary of results from key clinical trials that investigated ADT as neoadjuvant, adjuvant therapy, or treatment at biochemical recurrence in patients with prostate cancer who received radical prostatectomy Study (years) Level of evidencea Study type N Patient characteristics Treatment and duration Outcomes (1966–2006)7 1 Meta‐analysis 11 149 Localized or locally advanced PC with or without lymph node involvement (T1‐4, N1, M0) Neoadjuvant ADT + RP versus RP alone OS: OR, 1.11 (0.67–1.85); P = 0.69 Disease recurrence: OR, 0.74 (0.55–1.0); P = 0.05 Positive surgical margin rate: OR, 0.34 (0.27–0.42); P < 0.00001 (1966–2006)7 1 Meta‐analysis 11 149 Localized or locally advanced PC with or without lymph node involvement (T1‐4, N1, M0) Adjuvant ADT following RP versus RP alone 5‐y OS: OR, 1.50 (0.79–2.85); P = 0.2 5‐y DFS: OR, 3.73 (2.3–6.03); P < 0.00001 10‐y DFS: OR, 2.06 (1.34–3.15); P = 0.0009 Timing Of Antigen Deprivation (TOAD) therapy in patients with prostate cancer27 (2004–2012) 2 Prospective, randomized, phase 3 293 PSA relapse after curative treatment (RP or RT), or ineligible for curative treatment Immediate salvage ADT or delayed salvage ADT (recommended interval ≥ 2 y, unless clinically indicated) 5‐y OS: 91.2% (84.2–95.2) versus 86.4% (78.5–91.5); P = 0.047

prostate cancer27 (2004–2012) 2 Prospective, randomized, phase 3 293 PSA relapse after curative treatment (RP or RT), or ineligible for curative treatment Immediate salvage ADT or delayed salvage ADT (recommended interval ≥ 2 y, unless clinically indicated) 5‐y OS: 91.2% (84.2–95.2) versus 86.4% (78.5–91.5); P = 0.047 Among men with PSA relapse, 5‐y OS: 84.3% (73.9–90.8) versus 78.2% (67.2–85.8); P = 0.10 French Genito‐Urinary Group and the French Association of Urology (GETUG‐AFU) 1628 (2006–2010) 2 Prospective, randomized, phase 3 743 pT2‐4a PC with rising PSA of 0.2–2.0 ng/mL following RP without evidence of clinical disease Salvage RT (66 Gy in 33 fractions 5 d/wk for 7 wk) + 6 mo ADT (goserelin) versus salvage RT alone 5‐y PFS: 80% (75–84) versus 62% (57–67); HR, 0.50 (0.38–0.66); P < 0.0001 Radiation Therapy Oncology Group (RTOG) 960129 (1998–2003) 2 Prospective, randomized, phase 3 760 pT3pN0 or pT2pN0 and positive margins; rising PSA (0.2–4.0 ng/mL) following RP ADT (bicalutamide 150 mg daily for 2 y) during and after salvage RT (64.8 Gy in 36 fractions of 1.8 Gy) versus salvage RT alone 12‐y OS: 76.3% versus 71.3%; HR, 0.77 (0.59–0.99); P = 0.04 12‐y PC: 5.8% versus 13.4%; P < 0.001 10‐y PC deaths: 4.5% versus 10.1%; P < 0.001 Eastern Cooperative Oncology Group (ECOG) 388616 (1988–1993) 2 Prospective, randomized 98 Clinically localized PC (T1b or T2) and had previously undergone RP + PLND Immediate adjuvant ADT (goserelin monthly or bilateral orchiectomy) versus RP + salvage ADT Median OS: 13.9 y versus 11.3 y; HR, 1.84 (1.01–3.35); P = 0.04

Eastern Cooperative Oncology Group (ECOG) 388616 (1988–1993) 2 Prospective, randomized 98 Clinically localized PC (T1b or T2) and had previously undergone RP + PLND Immediate adjuvant ADT (goserelin monthly or bilateral orchiectomy) versus RP + salvage ADT Median OS: 13.9 y versus 11.3 y; HR, 1.84 (1.01–3.35); P = 0.04 Median DSS: NR versus 12.3 y; HR, 4.09 (1.76–9.49); P = 0.0004 Median PFS: 13.9 y versus 2.4 y; HR, 3.42 (1.96–5.98); P < 0.0001 Southwest Oncology Group (SWOG) S992117, 20 (2000–2007) 2 Prospective, randomized 983 High‐risk features at RP (GS ≥ 8; preop PSA ≥ 15 ng/mL; stage T3b, T4, or N1; or GS = 7 + preop PSA ≥ 10 ng/mL or a positive margin) Adjuvant ADT (goserelin + bicalutamide) alone or in combination with mitoxantrone chemotherapy for 2 y 10‐y DFS: 72% versus 72%; HR, 1.01 (0.80–1.27); P = 0.94 10‐y OS: 87% versus 86%; HR, 1.06 (0.79–1.43); P = 0.70 Deaths due to other cancer: 18% versus 36%; P = 0.011 Deaths due to leukemia: 0.2% versus 1.0% SWOG 91092 (1993–1996) 2 Prospective, phase 2 62 Locally advanced (T3–4, N0M0) PC Neoadjuvant ADT (goserelin [1 mo] + flutamide [4 mo]) followed by RP Median PFS: 7.5 y 10‐y PFS: 40% (27–53%) Median OS: NR 10‐y OS: 68% (56–80%) (2004–2012)8 3 Retrospective 156 High‐risk (T1c–3) PC Neoadjuvant therapy (LHRH agonist + estramustine for 6 mo) followed by RP versus neoadjuvant ADT for ≥6 mo followed by RT (3D conformal, 70–76 Gy in 2 Gy fractions) 3‐y OS: 98.3% versus 92.1% (P = 0.156) 3‐y BFS: 86.4% versus 89.4% (P = 0.878)

Median OS: NR 10‐y OS: 68% (56–80%) (2004–2012)8 3 Retrospective 156 High‐risk (T1c–3) PC Neoadjuvant therapy (LHRH agonist + estramustine for 6 mo) followed by RP versus neoadjuvant ADT for ≥6 mo followed by RT (3D conformal, 70–76 Gy in 2 Gy fractions) 3‐y OS: 98.3% versus 92.1% (P = 0.156) 3‐y BFS: 86.4% versus 89.4% (P = 0.878) (2000–2014)9 3 Retrospective 518 High‐risk PC Neoadjuvant therapy (LHRH agonist and for 6 mo + l‐PLND and RP versus e‐PLND and RP only) 5‐y BFS: 84.9% (80.4–59.4%) versus 54.7% (53.9–62.5%) (2002–2013)10 3 Retrospective 111 High‐risk PC Neoadjuvant hormonal therapy followed by RP Six pts with pT0: no recurrence after median follow‐up of 59 mo 105 pts with non‐pT0: 57.1% developed BCR within a median of 14 mo (2000–2014)11 3 Retrospective 116 Initially inoperable PC Neoadjuvant ADT for ≥3 mo or until PSA nadir reached followed by RP Median OS: 10 y, comparable with that of patients with initially operable high‐risk PC (2000–2006)18 3 Retrospective 128 Locally advanced (pT3N0M0) PC Immediate adjuvant ADT for ≥5 y 10‐y hormone‐refractory BFS: 88.3% 10‐y DSS: 96.3% 10‐y OS: 85.7% (1990–1999)19 3 Matched cohort 8290 Pathological lymph node‐negative PC RP + adjuvant ADT versus RP alone 10‐y systemic PFS: 95% versus 90%; P < 0.001 10‐y DSS: 98% versus 95%; P = 0.009 10‐y OS: 84% versus 83%; P = 0.427 (1989–2005)21 3 Retrospective 372 High risk (PSA > 20 ng/mL, ≥T2c, or GS ≥ 8) PC RP + adjuvant ADT (LHRH agonist, LHRH agonist/orchiectomy + oral antiandrogen, or orchiectomy alone) if seminal vesicle invasion or lymph node metastases were present versus RP alone 5‐y BFS: 76.6% 10‐y BFS: 56.2%

10‐y OS: 84% versus 83%; P = 0.427 (1989–2005)21 3 Retrospective 372 High risk (PSA > 20 ng/mL, ≥T2c, or GS ≥ 8) PC RP + adjuvant ADT (LHRH agonist, LHRH agonist/orchiectomy + oral antiandrogen, or orchiectomy alone) if seminal vesicle invasion or lymph node metastases were present versus RP alone 5‐y BFS: 76.6% 10‐y BFS: 56.2% BFS with versus without ADT: P = 0.0019 5‐y OS: 84.3% 10‐y OS: 72.1% OS with versus without ADT: P = 0.0821 (2004–2012)26 3 Retrospective 132 High‐risk PC (pelvic lymph node invasion, lymphovascular invasion, high tumor grade, or high preop PSA) Adjuvant RT + adjuvant ADT (LHRH agonist or bicalutamide 150 mg/d) versus adjuvant RT alone following RP; duration of ADT left to the discretion of the physician Among 56 patients treated with RT + ADT: 5‐y BFS: 90.5% 5‐y MFS: 95.9% 5‐y DSS: 100% 5‐y OS: 90.6% Median duration of ADT: 24 mo (6–36) ADT, androgen deprivation therapy; BCR, biochemical recurrence; BFS, biochemical progression‐free survival; DFS, disease‐free survival; DSS, disease‐specific survival; GS, Gleason score; HR, Hazards ratio; LHRH, luteinizing hormone‐releasing hormone; MFS, metastasis‐free survival; NR, not reported; OS, overall survival; OR, odds ratio; preop, preopeartive; PC, prostrate cancer; PFS, progression‐free survival; PLND, pelvic lymph node dissection; PSA, prostate‐specific antigen; RP, radical prostatectomy; RT, radiotherapy. a Level of evidence determined by study design: 1, meta‐analysis or systematic review; 2, randomized controlled trial; and 3, cohort study.

ADT, androgen deprivation therapy; BCR, biochemical recurrence; BFS, biochemical progression‐free survival; DFS, disease‐free survival; DSS, disease‐specific survival; GS, Gleason score; HR, Hazards ratio; LHRH, luteinizing hormone‐releasing hormone; MFS, metastasis‐free survival; NR, not reported; OS, overall survival; OR, odds ratio; preop, preopeartive; PC, prostrate cancer; PFS, progression‐free survival; PLND, pelvic lymph node dissection; PSA, prostate‐specific antigen; RP, radical prostatectomy; RT, radiotherapy. a Level of evidence determined by study design: 1, meta‐analysis or systematic review; 2, randomized controlled trial; and 3, cohort study. John Wiley & Sons, Ltd.The phase 2 Southwest Oncology Group (SWOG) 9109 trial (N = 62) investigated neoadjuvant ADT plus radical prostatectomy for patients with locally advanced prostate cancer and demonstrated long survival.2 Median PFS was 7.5 years, and the 10‐year PFS rate was 40% (95% confidence interval [CI], 27–53%). Median OS was not reached, but the overall 10‐year OS rate was 68% (56–80%).

62) investigated neoadjuvant ADT plus radical prostatectomy for patients with locally advanced prostate cancer and demonstrated long survival.2 Median PFS was 7.5 years, and the 10‐year PFS rate was 40% (95% confidence interval [CI], 27–53%). Median OS was not reached, but the overall 10‐year OS rate was 68% (56–80%). A retrospective study (N = 156) compared neoadjuvant therapy for 6 months followed by radical prostatectomy versus neoadjuvant ADT for ≥6 months followed by radiotherapy for patients with high‐risk prostate cancer.8 Biochemical PFS and OS rates were similar for both treatment groups. The 3‐year OS rate was 98.3% for neoadjuvant therapy plus radical prostatectomy versus 92.1% for neoadjuvant therapy plus radiotherapy (P = 0.156), and the 3‐year biochemical PFS rate was 86.4% versus 89.4% (P = 0.878). A larger retrospective analysis (N = 518) assessed whether neoadjuvant therapy with extended (e‐) pelvic lymph node dissection (PLND) conferred a benefit for high‐risk prostate cancer patients compared with neoadjuvant ADT with luteinizing hormone‐releasing hormone agonist (LHRHa) plus estramustine.9 Five‐year biochemical recurrence‐free survival rates were 84.9% and 54.7% for ADT and e‐PLND, respectively (P < 0.0001).

ic lymph node dissection (PLND) conferred a benefit for high‐risk prostate cancer patients compared with neoadjuvant ADT with luteinizing hormone‐releasing hormone agonist (LHRHa) plus estramustine.9 Five‐year biochemical recurrence‐free survival rates were 84.9% and 54.7% for ADT and e‐PLND, respectively (P < 0.0001). Neoadjuvant ADT can also eradicate high‐risk prostate cancer. A retrospective analysis of men with high‐risk prostate cancer (N = 111) treated with neoadjuvant ADT followed by radical prostatectomy found that 57.1% of men with non‐pT0 disease (residual tumor) developed biochemical relapse within a median of 14 months.10 However, among the six patients who had pT0 disease, none of them experienced recurrence after a median follow‐up of 59 months. Another retrospective analysis of men with initially inoperable prostate cancer (N = 116) treated with neoadjuvant ADT for ≥3 months, or until PSA nadir was reached (whichever was the sooner), found that median OS was 10 years, which is comparable to that of patients with initially operable high‐risk prostate cancer.11

Another retrospective analysis of men with initially inoperable prostate cancer (N = 116) treated with neoadjuvant ADT for ≥3 months, or until PSA nadir was reached (whichever was the sooner), found that median OS was 10 years, which is comparable to that of patients with initially operable high‐risk prostate cancer.11 A longer duration (either 6 or 8 months) of neoadjuvant hormonal therapy, compared with short‐term (usually 3 months) treatment, prior to radical prostatectomy, has demonstrated increased clinical benefit. This benefit usually presents as lower positive margin rates after prostatectomy and decreased PSA recurrence risk after 2–5 years.7, 12, 13, 14, 15 Between 3 and 8 months of neoadjuvant ADT therapy, prostate tumors may still undergo pathological and biochemical regression, which might result from prolonged duration of apoptosis of prostate tumor cells.12 However, due to the absence of long‐term survival results, the optimal duration of neoadjuvant therapy before radical prostatectomy is still to be elucidated.

py, prostate tumors may still undergo pathological and biochemical regression, which might result from prolonged duration of apoptosis of prostate tumor cells.12 However, due to the absence of long‐term survival results, the optimal duration of neoadjuvant therapy before radical prostatectomy is still to be elucidated. Summary of findings for neoadjuvant ADT plus radical prostatectomy Neoadjuvant ADT followed by radical prostatectomy is feasible in patients with localized or locally advanced prostate cancer with intermediate‐ or high‐risk features, or men with initially inoperable prostate cancer. Neoadjuvant ADT significantly reduced recurrence and positive surgical margin rates. A moderate level of evidence currently suggests that 6‐8 months of neoadjuvant ADT before radical prostatectomy provide clinical benefit but no studies have yet demonstrated an OS benefit. Neoadjuvant ADT plus radical prostatectomy warrants further exploration, particularly to determine the optimal duration of treatment. As of July 2018, no clinical trials have been listed on Clinicaltrials.gov comparing short‐term and long‐term neoadjuvant ADT prior to radical prostatectomy. However, there are currently seven studies that include the use of neoadjuvant ADT followed by radical prostatectomy (Clinicaltrials.gov identifiers: NCT01696877, NCT01409200, NCT01542021, NCT03358563, NCT00589472, NCT03228810, and NCT00430183), each using neoadjuvant ADT for different durations. A meta‐analysis of these studies in the future may provide more evidence for the optimal duration of neoadjuvant ADT.

radical prostatectomy (Clinicaltrials.gov identifiers: NCT01696877, NCT01409200, NCT01542021, NCT03358563, NCT00589472, NCT03228810, and NCT00430183), each using neoadjuvant ADT for different durations. A meta‐analysis of these studies in the future may provide more evidence for the optimal duration of neoadjuvant ADT. 3.1.2 Adjuvant ADT following radical prostatectomy A number of studies have demonstrated that adjuvant ADT following radical prostatectomy results in excellent PFS, OS, and disease‐specific survival in patients with high‐risk localized or locally advanced prostate cancer (Table 1).15, 16, 17, 18 Several of these studies investigated the optimal timing and duration of ADT following radical prostatectomy. A matched cohort study (N = 8290) compared outcomes of patients with lymph node‐negative prostate cancer who were treated with radical prostatectomy with or without adjuvant ADT. Adjuvant ADT improved 10‐year rates for systemic PFS (95% vs 90%; P < 0.001) and disease‐specific survival (98% vs 95%; P = 0.009) compared with ADT following PSA increase; however, 10‐year OS was similar (84% vs 83%; P = 0.427).19

rostate cancer who were treated with radical prostatectomy with or without adjuvant ADT. Adjuvant ADT improved 10‐year rates for systemic PFS (95% vs 90%; P < 0.001) and disease‐specific survival (98% vs 95%; P = 0.009) compared with ADT following PSA increase; however, 10‐year OS was similar (84% vs 83%; P = 0.427).19 The Eastern Cooperative Oncology Group study EST 3886 study (n = 98) found that immediate, continuous adjuvant ADT following prostatectomy and lymphadenectomy significantly improved outcomes in men with node‐positive prostate cancer compared with ADT at clinical recurrence.16 Median PFS was 13.9 years for men who received immediate ADT versus 2.4 years for those who received salvage ADT (hazard ratio [HR], 3.42; P < 0.0001). Median disease‐specific survival was 12.3 years with salvage ADT and not reached with immediate ADT (HR, 4.09; P = 0.0004). OS was also significantly improved, with a median of 13.9 years versus 11.3 years with immediate ADT versus salvage ADT (HR, 1.84; P = 0.04). The duration of adjuvant ADT varied across studies; however, key studies have demonstrated excellent survival with long‐term adjuvant ADT.

h immediate ADT (HR, 4.09; P = 0.0004). OS was also significantly improved, with a median of 13.9 years versus 11.3 years with immediate ADT versus salvage ADT (HR, 1.84; P = 0.04). The duration of adjuvant ADT varied across studies; however, key studies have demonstrated excellent survival with long‐term adjuvant ADT. Although ADT has been shown to significantly improve OS following radical prostatectomy, patients with high‐risk prostate cancer still experience worse disease progression and shorter OS than patients with lower‐risk disease. Chemotherapy may improve outcomes in a number of solid tumors, and the addition of chemotherapy to adjuvant ADT following radical prostatectomy was investigated in the SWOG S9921 study. Patients with high‐risk features at radical prostatectomy (N = 961) received 2 years of CAB alone or in combination with mitoxantrone chemotherapy and prednisone.17, 20 Survival results were greater than expected and similar for both treatment regimens. Ten‐year disease‐free survival was 72% in both treatment groups (HR, 1.01; P = 0.94), and 10‐year OS was 87% in the CAB group and 86% in the CAB plus mitoxantrone group (HR, 1.06; P = 0.70). However, mitoxantrone and prednisone when added to CAB significantly increased the risk of leukemia and other cancers, making CAB alone preferable.

ree survival was 72% in both treatment groups (HR, 1.01; P = 0.94), and 10‐year OS was 87% in the CAB group and 86% in the CAB plus mitoxantrone group (HR, 1.06; P = 0.70). However, mitoxantrone and prednisone when added to CAB significantly increased the risk of leukemia and other cancers, making CAB alone preferable. A retrospective study of men with pT3N0M0 prostate cancer who received adjuvant ADT following radical prostatectomy (N = 128) found that immediate long‐term ADT for ≥5 years is feasible in patients with locally advanced prostate cancer.18 The 10‐year disease‐specific survival and hormone‐refractory biochemical PFS rates were 96.3% and 88.3%, respectively. Another study retrospectively analyzed patients with high‐risk prostate cancer (N = 372) who received radical prostatectomy; patients with seminal vesicle invasion or lymph node metastases also received adjuvant ADT.21 Five‐ and 10‐year biochemical PFS rates were 76.6% and 56.2%, respectively. Despite having more advanced prostate cancer, patients who received adjuvant ADT had significantly longer biochemical PFS (P = 0.0019). Five‐ and 10‐year OS rates were 84.3% and 72.1%, respectively. OS was similar between the two patient groups (P = 0.0821). These findings demonstrate that stage‐dependent adjuvant therapy for patients receiving radical prostatectomy is a viable therapeutic option.

had significantly longer biochemical PFS (P = 0.0019). Five‐ and 10‐year OS rates were 84.3% and 72.1%, respectively. OS was similar between the two patient groups (P = 0.0821). These findings demonstrate that stage‐dependent adjuvant therapy for patients receiving radical prostatectomy is a viable therapeutic option. Summary of findings for adjuvant ADT following radical prostatectomy Findings from these studies show that long‐term adjuvant ADT immediately following radical prostatectomy can benefit men with high‐risk prostate cancer. Several studies have demonstrated progression‐free and disease‐specific survival benefits, and one study showed that immediate, continuous adjuvant ADT significantly prolonged OS compared with delaying ADT until progression. Adjuvant ADT following radical prostatectomy is supported by strong evidence (meta‐analysis of over 10 000 patients) for men with high‐risk localized or locally advanced prostate cancer, particularly those with positive lymph nodes. Further study is needed to determine the optimal duration of treatment; however, selecting a control arm and conducting randomized clinical trials with very long follow‐up times is challenging. Clinical studies are also needed to determine whether adjuvant ADT following radical prostatectomy benefits patients with intermediate‐risk prostate cancer.

ermine the optimal duration of treatment; however, selecting a control arm and conducting randomized clinical trials with very long follow‐up times is challenging. Clinical studies are also needed to determine whether adjuvant ADT following radical prostatectomy benefits patients with intermediate‐risk prostate cancer. 3.1.3 Adjuvant radiotherapy plus ADT following radical prostatectomy Radiotherapy as adjuvant therapy following radical prostatectomy was also reported to improve outcomes in a number of studies.22, 23, 24 Compared with radical prostatectomy alone, adjuvant radiotherapy has been shown to significantly improve biochemical PFS (10‐year PFS: 56% vs 35%; P < 0.0001),22, 23, 24 median metastasis‐free survival (14.7 years vs 12.9 years; HR, 0.71; P = 0.016),25 and median OS (15.2 years vs 13.3 years; HR, 0.72; P = 0.023).25 The combination of adjuvant ADT plus adjuvant radiotherapy following radical prostatectomy was analyzed in a retrospective study (N = 132), which found that men with high‐risk prostate cancer who received adjuvant radiotherapy plus ADT following prostatectomy had excellent outcomes (Table 1).26 Five‐year biochemical relapse‐free, metastasis‐free, disease‐specific, and OS rates were 90.5, 95.9, 100, and 90.6%, respectively. The median duration of ADT was 24 months (6–36). Further investigation into this combination regimen is therefore warranted.

ADT following prostatectomy had excellent outcomes (Table 1).26 Five‐year biochemical relapse‐free, metastasis‐free, disease‐specific, and OS rates were 90.5, 95.9, 100, and 90.6%, respectively. The median duration of ADT was 24 months (6–36). Further investigation into this combination regimen is therefore warranted. Summary of findings for adjuvant radiotherapy plus ADT following radical prostatectomy A strong level of evidence supports combining adjuvant ADT with adjuvant radiotherapy following radical prostatectomy, which meta‐analysis has shown to improve outcomes compared with radical prostatectomy alone, particularly in men with high‐risk prostate cancer. Five‐year biochemical relapse‐free, metastasis‐free, disease‐specific, and OS rates were >90% with long‐term ADT in combination with adjuvant radiotherapy. 3.1.4 Biochemical recurrence following radical prostatectomy The standard of care for patients with biochemical recurrence is ADT. However, the optimal timing of ADT (early or late) and the optimal adjuvant regimen remain controversial. Three phase 3 trials have investigated these issues (Table 1). The timing of antigen deprivation (TOAD) study (n = 293) investigated immediate treatment with ADT versus delayed ADT for men with PSA relapse following radical prostatectomy or radiotherapy, and men with incurable prostate cancer.27 The investigated timing of delayed ADT was ≥2 years after biochemical recurrence, unless earlier treatment was clinically indicated. Immediate ADT significantly improved OS (5‐year OS was 91.2% vs 86.4%; P = 0.047) compared with delayed ADT.

dical prostatectomy or radiotherapy, and men with incurable prostate cancer.27 The investigated timing of delayed ADT was ≥2 years after biochemical recurrence, unless earlier treatment was clinically indicated. Immediate ADT significantly improved OS (5‐year OS was 91.2% vs 86.4%; P = 0.047) compared with delayed ADT. The French genito‐urinary group and the French association of urology (GETUG‐AFU) 16 study (n = 743) compared salvage radiotherapy plus short‐term ADT versus salvage radiotherapy alone for men with rising PSA following radical prostatectomy.28 This study showed that 6 months of ADT plus salvage radiotherapy significantly improved PFS (5‐year PFS was 80% vs 62%; HR, 0.50; P < 0.0001). However, OS was similar between the two treatment groups; 5‐year OS was 96% (93‐98%) vs 95% (92‐97%; HR, 0.7; P = 0.18). The RTOG 9601 study (n = 760) investigated long‐term ADT plus salvage radiotherapy versus salvage radiotherapy alone in patients with rising PSA following radical prostatectomy.29 The addition of 24 months of ADT improved 12‐year OS (76.3% vs 71.3%; HR, 0.77; P = 0.04) and reduced the 12‐year rates of metastatic prostate cancer (14.5% vs 23.0%; HR, 0.63; P = 0.005) and prostate cancer‐specific mortality (5.8% vs 13.4%; HR, 0.49; P < 0.001). In addition, subgroup analyses indicated that men with a GS ≥7, PSA = 0.7‐4.0 ng/mL, or positive surgical margins were most likely to benefit.

4) and reduced the 12‐year rates of metastatic prostate cancer (14.5% vs 23.0%; HR, 0.63; P = 0.005) and prostate cancer‐specific mortality (5.8% vs 13.4%; HR, 0.49; P < 0.001). In addition, subgroup analyses indicated that men with a GS ≥7, PSA = 0.7‐4.0 ng/mL, or positive surgical margins were most likely to benefit. Summary of findings for biochemical recurrence following radical prostatectomy ADT with or without radiotherapy at the time of biochemical relapse significantly improved outcomes for patients who previously received radical prostatectomy as a primary therapy, which is strongly supported by three phase 3, randomized, prospective trials. Among the salvage treatment options that have been studied, radiotherapy plus long‐term ADT provided an OS benefit. Thus, radiotherapy plus long‐term ADT is recommended at the time of biochemical recurrence, especially for men with GS ≥7, PSA = 0.7‐4.0 ng/mL, or positive surgical margins. 3.2 ADT for patients who received radiotherapy as a primary treatment Radiotherapy, either external beam or brachytherapy, can be administered alone, following radical prostatectomy, or in combination with ADT to patients with prostate cancer. In patients with low‐risk localized prostate cancer, radiotherapy alone has been shown to provide durable control, with 73% each disease‐free survival rates at 15, 20, and 25 years.30

brachytherapy, can be administered alone, following radical prostatectomy, or in combination with ADT to patients with prostate cancer. In patients with low‐risk localized prostate cancer, radiotherapy alone has been shown to provide durable control, with 73% each disease‐free survival rates at 15, 20, and 25 years.30 Many studies have shown that radiotherapy plus adjuvant ADT provides a benefit for patients with intermediate‐risk, high‐risk, or locally advanced disease (Table 2). Results from a meta‐analysis indicated that adjuvant ADT following radiotherapy significantly improves 5‐year OS (odds ratio [OR], 1.46; P = 0.0009), disease‐specific survival (OR, 2.10; P = 0.00001), and disease‐free survival (OR, 2.53; P < 0.00001).7 Table 2 Summary of results from important clinical trials that investigated radiotherapy as primary therapy with or without neoadjuvant or adjuvant ADT Study (years) Level of evidencea Study type N Patient characteristics Treatment and duration Outcomes (95% CI) 1966–20067 1 Meta‐analysis 11 149 Localized or locally advanced PC with or without lymph node involvement (T1–4 N1, M0) Adjuvant ADT following RT versus RT alone 5‐y OS: OR, 1.46 (1.17–1.83); P = 0.0009) 10‐y OS: OR, 1.44 (1.13–1.84); P = 0.003 5‐y DSS: OR, 2.10 (1.53–2.88); P = 0.00001 5‐y DFS: OR, 2.53 (2.05–3.12); P < 0.00001 DFCI 9509631 (1995–2001) 2 Prospective, randomized 206 Localized (T1b‐T2b) but unfavorable‐risk PC RT plus 6 mo ADT (LHRHa + flutamide) versus RT alone 8‐y OS: 74% (64–82%) versus 61% (49–71%); P = 0.01

10‐y OS: OR, 1.44 (1.13–1.84); P = 0.003 5‐y DSS: OR, 2.10 (1.53–2.88); P = 0.00001 5‐y DFS: OR, 2.53 (2.05–3.12); P < 0.00001 DFCI 9509631 (1995–2001) 2 Prospective, randomized 206 Localized (T1b‐T2b) but unfavorable‐risk PC RT plus 6 mo ADT (LHRHa + flutamide) versus RT alone 8‐y OS: 74% (64–82%) versus 61% (49–71%); P = 0.01 Among men with no or minimal comorbidity, 8‐y OS: 90% (79–95%) versus 64% (49–75%); P < 0.001 Among men with moderate or severe comorbidity, 8‐y OS: 25% (9–44%) versus 54% (32–72%); P = 0.08 EORTC 2299132 (2001–2008) 2 Prospective, randomized 819 Localized (T1b‐T2aN0M0) or locally advanced (T2b‐T4) PC RT + concomitant/adjuvant ADT for 6 mo (goserelin) versus RT alone 5‐y biochemical DFS: 82.6% (78.4–86.1%) versus 69.8% (64.9–74.2%); HR, 0.52 (0.41–0.66); P < 0.001 5‐y clinical DFS: 88.7% (82.1–85.2%) versus 80.8% (76.5–84.3%); HR, 0.63 (0.48–0.84); P = 0.001 RTOG 94–0833 (1994–2001) 2 Prospective, randomized, phase 3 1979 T1b‐T2b PC with PSA ≤20 ng/mL 4 mo CAB beginning 2 mo before RT (46.8 Gy to pelvis and 19.8 Gy to prostate) versus RT alone 10‐y OS: 62% versus 57%; HR, 1.17; P = 0.03 10‐y DSM: 8% versus 4%; HR, 1.87; P = 0.001 RTOG 991034 (2000–2004) 2 Prospective, randomized 1579 Intermediate‐risk PC Neoadjuvant CAB (8 wk vs 28 wk) + 8 wk CAB during RT 10‐y DSS: 95% (93.3–97.0) versus 96% (94.6–98.0); HR, 0.81; P = 0.45 10‐y OS: 66% (62.0–69.9) versus 67% (63.0–70.8); HR, 0.95; P = 0.62) 10‐y locoregional progression: 6% (4.3–8.0) versus 4% (2.5–5.7); HR, 0.65; P = 0.07 10‐y distant metastasis: 6% (4.0–7.7) versus 6% (4.0–7.6; HR, 1.07; P = 0.80)

RTOG 991034 (2000–2004) 2 Prospective, randomized 1579 Intermediate‐risk PC Neoadjuvant CAB (8 wk vs 28 wk) + 8 wk CAB during RT 10‐y DSS: 95% (93.3–97.0) versus 96% (94.6–98.0); HR, 0.81; P = 0.45 10‐y OS: 66% (62.0–69.9) versus 67% (63.0–70.8); HR, 0.95; P = 0.62) 10‐y locoregional progression: 6% (4.3–8.0) versus 4% (2.5–5.7); HR, 0.65; P = 0.07 10‐y distant metastasis: 6% (4.0–7.7) versus 6% (4.0–7.6; HR, 1.07; P = 0.80) 10‐y PSA recurrence: 27% (23.1–29.8) versus 27% (23.4–30.3); HR, 0.97; P = 0.77) TROG 96.0135 (1996–2000) 2 Prospective, randomized 818 T2b, T2c, T3, or T4, N0, M0 PC Neoadjuvant ADT (goserelin + flutamide, 6 mo vs 3 mo) + RT (66 Gy in 33 fractions over 6.5‐7 wk) versus RT alone 10‐y PSA progression: 52.8% (46.5–58.7) versus 60.4% (54.2–66.1) versus 73.8% (68.1–78.7); P (6 mo vs RT) < 0.0001; P (3 mo vs RT) = 0.0009 10‐y DSM: 11.4% (7.9–15.6) versus 18.9% (14.4–23.9) versus 22.0% (17.2–27.2); P = 0.0002; P = 0.394 10‐y overall mortality: 29.2% (24.1–35.1) versus 36.7% (31.1–42.9) versus 42.5% (36.7–48.7); P = 0.0005; P = 0.198 DART01/05 GICOR36 (2005–2010) 2 Prospective, randomized, phase 3 355 Clinical stage T1c‐T3b N0M0 PC with intermediate‐ or high‐risk factors Short‐term ADT: Neoadjuvant and concomitant ADT for 4 mo +radiotherapy (3D conformal) versus long‐term ADT: the same treatment + adjuvant ADT for 24 mo 5‐y BFS: 90% (87–92) versus 81% (78–85); HR, 1.88 (1.12–3.15); P = 0.01) 5‐y OS: 95% (93–97) versus 86% (83–89); HR, 2.48 (1.31–4.68); P = 0.009 5‐y MFS: 94% (92–96) versus 83% (80–86); HR, 2.31 (1.23–3.85); P = 0.01

DART01/05 GICOR36 (2005–2010) 2 Prospective, randomized, phase 3 355 Clinical stage T1c‐T3b N0M0 PC with intermediate‐ or high‐risk factors Short‐term ADT: Neoadjuvant and concomitant ADT for 4 mo +radiotherapy (3D conformal) versus long‐term ADT: the same treatment + adjuvant ADT for 24 mo 5‐y BFS: 90% (87–92) versus 81% (78–85); HR, 1.88 (1.12–3.15); P = 0.01) 5‐y OS: 95% (93–97) versus 86% (83–89); HR, 2.48 (1.31–4.68); P = 0.009 5‐y MFS: 94% (92–96) versus 83% (80–86); HR, 2.31 (1.23–3.85); P = 0.01 EORTC 2286337 (1987–1995) 2 Prospective, randomized, phase 3 415 High‐risk T1–4 PC Long‐term ADT: 36 mo goserelin plus external RT (5 days/wk for 7 wk, total dose 50 Gy to whole pelvis plus additional 20 Gy to prostate and seminal vesicles) versus RT alone 10‐y clinical DFS: 47.7% (39.0–56.0) versus 22.7% (16.3–29.7); HR, 0.42 (0.33–0.55); P < 0.0001 10‐y OS: 58.1% (49.2–66.0) versus 39.8% (31.9–47.5); HR, 0.60 (0.45–0.80); P = 0.0004 RTOG 861038 (1987–1991) 2 Prospective, randomized, phase 3 456 Locally advanced (T2‐4) PC with or without lymph node involvement EBRT + neoadjuvant CAB (goserelin + flutamide) for 2 mo before and concurrent with EBRT versus EBRT alone 10‐y OS: 42.6% (35.9–49.3) versus 33.8% (27.5–40.1); P = 0.12 10‐y DFS: 11.2% (7.0–15.6) versus 3.4% (1.0–5.8); P < 0.0001 10‐y DSM: 23.3% (17.6–29.1) versus 35.6% (29.2–42.0); P = 0.01 10‐y distant metastases: 34.9% (28.5–41.3) versus 46.9% (40.3–53.5); P = 0.006 10‐y biochemical failure: 65.1% (58.6–71.6) versus 80.0% (74.7–85.4); P < 0.0001

RTOG 861038 (1987–1991) 2 Prospective, randomized, phase 3 456 Locally advanced (T2‐4) PC with or without lymph node involvement EBRT + neoadjuvant CAB (goserelin + flutamide) for 2 mo before and concurrent with EBRT versus EBRT alone 10‐y OS: 42.6% (35.9–49.3) versus 33.8% (27.5–40.1); P = 0.12 10‐y DFS: 11.2% (7.0–15.6) versus 3.4% (1.0–5.8); P < 0.0001 10‐y DSM: 23.3% (17.6–29.1) versus 35.6% (29.2–42.0); P = 0.01 10‐y distant metastases: 34.9% (28.5–41.3) versus 46.9% (40.3–53.5); P = 0.006 10‐y biochemical failure: 65.1% (58.6–71.6) versus 80.0% (74.7–85.4); P < 0.0001 EORTC 2296139 (1997–2001) 2 Prospective, randomized 1113 Locally advanced (T1c–T2a–b, pN1–2, M0 or T2c–4, cN0–2, M0) PC EBRT (3D conformal, 50 Gy for first target volume, an additional 20 Gy for the second target volume, 5 d/wk for 7 wk) + ADT (LHRH analog) for 6 mo or 3 y 5‐y overall mortality: 19.0% versus 15.2%; HR, 1.42 Trials 23, 24, and 2540 (NR) 2 Three prospective, randomized trials 8113 Localized (T1–2 N0/Nx M0) or locally advanced (T3–4 and any N, or any T and N+; M0) PC Standard care plus either bicalutamide 150 mg daily or placebo PFS: HR, 0.85 (0.79–0.91); P = 0.001; patients with locally advanced disease derived PFS benefit OS: HR, 1.01 (0.94–1.09); P = 0.77; patients with locally advanced disease who received RT derived OS benefit: HR, 0.70 (0.51–0.97); P = 0.031

EORTC 2296139 (1997–2001) 2 Prospective, randomized 1113 Locally advanced (T1c–T2a–b, pN1–2, M0 or T2c–4, cN0–2, M0) PC EBRT (3D conformal, 50 Gy for first target volume, an additional 20 Gy for the second target volume, 5 d/wk for 7 wk) + ADT (LHRH analog) for 6 mo or 3 y 5‐y overall mortality: 19.0% versus 15.2%; HR, 1.42 Trials 23, 24, and 2540 (NR) 2 Three prospective, randomized trials 8113 Localized (T1–2 N0/Nx M0) or locally advanced (T3–4 and any N, or any T and N+; M0) PC Standard care plus either bicalutamide 150 mg daily or placebo PFS: HR, 0.85 (0.79–0.91); P = 0.001; patients with locally advanced disease derived PFS benefit OS: HR, 1.01 (0.94–1.09); P = 0.77; patients with locally advanced disease who received RT derived OS benefit: HR, 0.70 (0.51–0.97); P = 0.031 RTOG 85–3141 (1987–1992) 2 Prospective, randomized, phase 3 945 Locally advanced (T3 or regional lymphatic involvement) PC RT (1.8‐2.0 Gy) daily for 4‐5 times/wk for a total of 44–46 Gy plus additional 20‐25 Gy to prostate) plus adjuvant ADT (goserelin) until progression or RT alone followed by salvage ADT 10‐y OS: 49% versus 39%; P = 0.002 10‐y local failure: 23% versus 38%; P < 0.0001 10‐y distant metastases: 24% versus 39%; P < 0.001 10‐y DSM: 16% versus 22%; P = 0.0052

RTOG 85–3141 (1987–1992) 2 Prospective, randomized, phase 3 945 Locally advanced (T3 or regional lymphatic involvement) PC RT (1.8‐2.0 Gy) daily for 4‐5 times/wk for a total of 44–46 Gy plus additional 20‐25 Gy to prostate) plus adjuvant ADT (goserelin) until progression or RT alone followed by salvage ADT 10‐y OS: 49% versus 39%; P = 0.002 10‐y local failure: 23% versus 38%; P < 0.0001 10‐y distant metastases: 24% versus 39%; P < 0.001 10‐y DSM: 16% versus 22%; P = 0.0052 RTOG 92‐0242 (1992–1995) 2 Prospective, randomized 1554 Locally advanced (T2c‐T4, N0‐X) PC with no extra pelvic lymph node involvement and PSA < 150 ng/mL ADT (goserelin + flutamide) for 4 mo before and during RT (45 Gy to pelvic nodes and 65‐70 Gy to the prostate) with or without an additional 2 y goserelin after RT 10‐y DFS: 22.5% versus 13.2%; P < 0.0001 10‐y DSS: 88.7% versus 83.9%; P = 0.0042 10‐y local progression: 12.3% versus 22.2%; P < 0.0001 10‐y distant metastasis: 14.8% versus 22.8%; P < 0.0001 10‐y BFS: 51.9% versus 68.1%; P ≤ 0.0001 10‐y OS: 53.9% versus 51.6%; P = 0.36 10‐y OS among patients with GS 8‐10: 45.1% versus 31.9%; P = 0.0061 SWOG‐JPR743 (1999–2005) 2 Prospective, randomized 1386 Rising PSA > 3 ng more than 12 mo after primary or salvage RT Intermittent salvage ADT (LHRHa + nonsteroidal antiandrogen in 8‐mo cycles versus continuous salvage ADT (LHRHa + nonsteroidal antiandrogen or orchiectomy) Median OS: 8.8 y versus 9.1 y; HR, 1.02 (0.86–1.21); P value for noninferiority = 0.009 Median DSS: HR, 1.18; P = 0.24 Time to castration‐resistant disease: HR, 0.80; P = 0.02

SWOG‐JPR743 (1999–2005) 2 Prospective, randomized 1386 Rising PSA > 3 ng more than 12 mo after primary or salvage RT Intermittent salvage ADT (LHRHa + nonsteroidal antiandrogen in 8‐mo cycles versus continuous salvage ADT (LHRHa + nonsteroidal antiandrogen or orchiectomy) Median OS: 8.8 y versus 9.1 y; HR, 1.02 (0.86–1.21); P value for noninferiority = 0.009 Median DSS: HR, 1.18; P = 0.24 Time to castration‐resistant disease: HR, 0.80; P = 0.02 ADT, androgen deprivation therapy; BFS, biochemical progression‐free survival; CAB, combined androgen blockade; DFS, disease‐free survival; DSM, disease‐specific mortality; DSS, disease‐specific survival; EBRT, external beam radiation therapy; GS, Gleason score; LHRH, luteinizing hormone‐releasing hormone; NR, not reported; OS, overall survival; PFS, progression‐free survival; PSA, prostate‐specific antigen; RP, radical prostatectomy; RT, radiotherapy. a Level of evidence determined by study design: 1, meta‐analysis or systematic review and 2, randomized controlled trial. John Wiley & Sons, Ltd.3.2.1 Radiotherapy plus ADT for patients with intermediate‐risk disease The DFCI 95096 study (N = 206) showed that 6 months of CAB plus radiotherapy significantly improved OS compared with radiotherapy alone in men with intermediate‐risk localized prostate cancer (8‐year OS was 74% [64–82%] vs 61% [49–71%]; P = 0.01).31 Men with no or minimal comorbidities benefited from ADT; however, ADT may have a negative impact on survival for men with moderate or severe comorbidities.

antly improved OS compared with radiotherapy alone in men with intermediate‐risk localized prostate cancer (8‐year OS was 74% [64–82%] vs 61% [49–71%]; P = 0.01).31 Men with no or minimal comorbidities benefited from ADT; however, ADT may have a negative impact on survival for men with moderate or severe comorbidities. In the EORTC 22991 study (N = 819), radiotherapy plus 6 months of concomitant and adjuvant ADT improved biochemical and disease‐free survival compared with radiotherapy alone.32 Five‐year biochemical disease‐free survival was 82.6% (78.4–86.1%) vs 69.8% (64.9–74.2%); HR, 0.52 (0.41–0.66); P < 0.001. Five‐year clinical disease‐free survival was 88.7% (82.1–85.2%) vs 80.8% (76.5–84.3%); HR, 0.63; P = 0.001. In the phase 3 RTOG 94‐08 study (n = 1979), short‐term CAB for 4 months before and during radiotherapy was associated with a decreased disease‐specific mortality and increased OS.33 Patients with intermediate‐risk prostate cancer benefited from neoadjuvant and concurrent ADT; however, there was no benefit for those with low‐ or high‐risk disease. A fourth study, the RTOG 9910 study (n = 1,579), investigated treatment with 8 weeks versus 28 weeks of neoadjuvant CAB therapy plus 8 weeks of CAB during radiotherapy.34 Outcomes were similar for the two treatment groups, suggesting that CABs for 8 weeks before and 8 weeks during radiotherapy are preferred for patients with intermediate‐risk prostate cancer.

1,579), investigated treatment with 8 weeks versus 28 weeks of neoadjuvant CAB therapy plus 8 weeks of CAB during radiotherapy.34 Outcomes were similar for the two treatment groups, suggesting that CABs for 8 weeks before and 8 weeks during radiotherapy are preferred for patients with intermediate‐risk prostate cancer. Summary of findings for men with intermediate‐risk disease Compelling evidence supports the use of short‐term ADT in combination with an LHRHa and an antiandrogen to provide benefit to men with intermediate‐risk prostate cancer, particularly, those with unfavorable intermediate‐risk disease with no or minimal comorbidities. Short‐term ADT (neoadjuvant/concomitant ADT for 4 months, concomitant/adjuvant ADT for 4‐6 months, or adjuvant ADT for 6 months) is recommended for intermediate‐risk patients who received radiotherapy as a primary therapy. 3.2.2 Radiotherapy plus ADT for high‐risk localized and locally advanced prostate cancer For high‐risk disease, several studies have investigated various lengths of ADT therapy in combination with radiotherapy. The results demonstrate better outcomes with long‐term use of ADT, which has become the standard of care. The TROG 96.01 study (N = 818) compared three different treatment groups: radiotherapy preceded by 6 months of neoadjuvant ADT, radiotherapy plus 3 months of neoadjuvant ADT, and radiotherapy alone in patients with high‐risk, localized prostate cancer.35 The study showed that 6 months of ADT significantly improved outcomes compared with radiotherapy alone, and also improved outcomes compared with 3 months of ADT.

by 6 months of neoadjuvant ADT, radiotherapy plus 3 months of neoadjuvant ADT, and radiotherapy alone in patients with high‐risk, localized prostate cancer.35 The study showed that 6 months of ADT significantly improved outcomes compared with radiotherapy alone, and also improved outcomes compared with 3 months of ADT. In the phase 3 DART01/05 GICOR study (N = 355), men with intermediate‐ or high‐risk prostate cancer were treated with 4 months of neoadjuvant plus concurrent ADT combined with radiotherapy (short‐term ADT), or with the same treatment followed by 24 months of adjuvant ADT (long‐term ADT).36 Long‐term ADT plus radiotherapy improved biochemical control and OS compared with short‐term ADT. The 5‐year OS rate was 95% (93–97%) in the long‐term ADT group versus 86% (83–89%) in the short‐term ADT group (HR, 2.48; P = 0.009). The phase 3 EORTC 22863 clinical trial in patients with high‐risk prostate cancer (N = 415) showed that adjuvant ADT with an LHRHa during, and for 3 years following, radiotherapy significantly improved 10‐year clinical disease‐free survival and OS compared with radiotherapy alone.37 The 10‐year OS rate was 58.1% (49.2–66.0%) in the group that received adjuvant ADT versus 39.8% (31.9–47.5%) in the group that received radiotherapy alone (HR, 0.60; P = 0.0004). In addition, 10‐year prostate cancer mortality was decreased in high‐risk patients receiving adjuvant ADT.

ompared with radiotherapy alone.37 The 10‐year OS rate was 58.1% (49.2–66.0%) in the group that received adjuvant ADT versus 39.8% (31.9–47.5%) in the group that received radiotherapy alone (HR, 0.60; P = 0.0004). In addition, 10‐year prostate cancer mortality was decreased in high‐risk patients receiving adjuvant ADT. For locally advanced disease, many studies have demonstrated that the addition of long‐term adjuvant therapy following radiotherapy improved outcomes. The phase 3 RTOG 8610 study (N = 456) investigated 2 months of neoadjuvant CAB plus 2 months of ADT during external beam radiotherapy versus radiotherapy alone for men with locally advanced prostate cancer with or without lymph node involvement.38 The addition of ADT improved 10‐year disease‐free survival, disease‐specific mortality, distant metastases, and biochemical failure. There was also a trend toward improved OS. Furthermore, results from the EORTC 22961 study in patients with locally advanced prostate cancer (N = 970) demonstrated that radiotherapy plus 3 years of adjuvant ADT decreased mortality compared with radiotherapy plus 6 months of adjuvant ADT.39 The 5‐year overall mortality rate was 19.0% (15.5–23.0%) for short‐term ADT versus 15.2% (12.1–18.9%) for long‐term ADT (HR, 1.42). The 5‐year prostate cancer‐specific mortality rate was 4.7% (2.7–6.7%) versus 3.2% (1.6–4.8%; HR, 1.71; P = 0.002).

ed mortality compared with radiotherapy plus 6 months of adjuvant ADT.39 The 5‐year overall mortality rate was 19.0% (15.5–23.0%) for short‐term ADT versus 15.2% (12.1–18.9%) for long‐term ADT (HR, 1.42). The 5‐year prostate cancer‐specific mortality rate was 4.7% (2.7–6.7%) versus 3.2% (1.6–4.8%; HR, 1.71; P = 0.002). Trials 23, 24, and 25 of the bicalutamide Early Prostate Cancer program investigated treatment with radiotherapy, radical prostatectomy, or watchful waiting followed by 150 mg bicalutamide or placebo in men with localized or locally advanced prostate cancer (N = 8113). A combined analyses showed that, among patients treated with radiotherapy, the addition of bicalutamide improved PFS (HR, 0.62; P = 0.001) and OS (HR, 0.70; P = 0.031) in men with locally advanced disease.40 Results from the RTOG 85‐31 study in locally advanced prostate cancer (N = 945) patients showed that OS improved with radiotherapy plus adjuvant goserelin until progression compared with radiotherapy and goserelin at relapse.41 The 10‐year OS rate was 49% in the adjuvant ADT group versus 39% in the ADT at relapse group (P = 0.002); the benefit was significant in men with GS 7 or GS 8‐10 but not those with GS 2‐6.

that OS improved with radiotherapy plus adjuvant goserelin until progression compared with radiotherapy and goserelin at relapse.41 The 10‐year OS rate was 49% in the adjuvant ADT group versus 39% in the ADT at relapse group (P = 0.002); the benefit was significant in men with GS 7 or GS 8‐10 but not those with GS 2‐6. In the RTOG 92‐02 study of men with locally advanced disease (N = 1554), 10‐year disease‐free survival, disease‐specific survival, local progression, distant metastasis, and biochemical failure rates were all improved with 4 months of goserelin and flutamide neoadjuvant therapy, radiotherapy, and 24 months of goserelin adjuvant therapy compared with neoadjuvant therapy plus radiotherapy without long‐term adjuvant therapy.42 Summary of findings for men with high‐risk localized and locally advanced disease The results from these studies strongly indicate that the higher the patient's risk of recurrence, the longer the duration of ADT should be used as adjuvant therapy in combination with radiotherapy. Long‐term use of ADT (2–3 years) concomitant or adjuvant with LHRHa, with or without an antiandrogen, is correlated with improved PFS and OS in patients with high‐risk or locally advanced nonmetastatic prostate cancer. Alternatively, a shorter course of ADT (4–6 months) may be sufficient for patients with intermediate‐risk prostate cancer.

(2–3 years) concomitant or adjuvant with LHRHa, with or without an antiandrogen, is correlated with improved PFS and OS in patients with high‐risk or locally advanced nonmetastatic prostate cancer. Alternatively, a shorter course of ADT (4–6 months) may be sufficient for patients with intermediate‐risk prostate cancer. 3.2.3 Biochemical recurrence following radiotherapy The SWOG‐JPR7 study investigated intermittent versus continuous ADT in men with elevated PSA following primary or salvage radiotherapy (N = 1386; Table 2).43 Patients were treated with an LHRHa plus a nonsteroidal antiandrogen either continuously or in 8‐month cycles. Nontreatment periods were determined by PSA levels. Median OS was similar: 8.8 years in the intermittent group and 9.1 years in the continuous group (HR, 1.02). However, certain quality‐of‐life issues improved with intermittent ADT. Thus, intermittent ADT was noninferior to continuous therapy at the time of biochemical relapse following radiotherapy but improved quality of life, particularly, during nontreatment phases. Further research is warranted to determine the optimal PSA level to undergo intermittent ADT. The difficulties in determining the optimal PSA level for reinitiating ADT may be resolved by conducting studies with more stringent patient stratification in terms of prognostic parameters. However, this presents a difficulty in limiting the number of patients who fit the specified criteria, resulting in a small sample size.

ficulties in determining the optimal PSA level for reinitiating ADT may be resolved by conducting studies with more stringent patient stratification in terms of prognostic parameters. However, this presents a difficulty in limiting the number of patients who fit the specified criteria, resulting in a small sample size. 3.3 Unfit or unwilling to receive primary treatment ADT can be used in men with localized or locally advanced prostate cancer who refuse, or are not candidates for, primary treatment, including those with limited life expectancy, advanced tumor stage, or other serious comorbidities. In the EORTC 30891 study, men with newly diagnosed localized or locally advanced prostate cancer who were not suitable for primary treatment were treated with buserelin, an LHRH analog, immediately or at symptomatic disease progression (n = 985).44 Immediate treatment significantly improved OS (HR, 1.25; noninferiority P > 0.1). Trials 23, 24, and 25, as described above, found that patients with locally advanced disease derived a greater improvement in PFS from 150 mg bicalutamide compared to watchful waiting patients (HR, 0.67; P < 0.001).39

ression (n = 985).44 Immediate treatment significantly improved OS (HR, 1.25; noninferiority P > 0.1). Trials 23, 24, and 25, as described above, found that patients with locally advanced disease derived a greater improvement in PFS from 150 mg bicalutamide compared to watchful waiting patients (HR, 0.67; P < 0.001).39 3.4 Real‐life implications of ADT There is a lack of ongoing studies investigating the real‐life implications of ADT in prostate cancer patients with regard to overall benefit of the therapy. Of the 4033 ongoing prostate cancer studies listed in Clinicaltrials.gov, only one (Clinicaltrials.gov identifier NCT02895230) is an observational study investigating the real‐life implications of ADT, enrolling all men receiving treatment for prostate cancer over an 18‐month period. While the primary objective of this study is to investigate the association between ADT and vascular stroke, it would also be of interest to determine whether different modalities of ADT have an impact on survival or not. Previous studies have established a survival benefit associated with the use of ADT in patients with high‐risk or locally advanced disease, with up to 26% and 24% more patients still alive at 5 and 10 years, respectively.45 There are two ongoing prostate cancer studies taking place in China (Clinicaltrials.gov identifier NCT03177551 and NCT03507597), but these do not focus on real‐life implications of ADT. However, the Asian Prostate Cancer Study, a database established through an alliance of 12 countries and regions in Asia, is a promising source of real‐world data on the implications of prostate cancer, including ADT.46

ov identifier NCT03177551 and NCT03507597), but these do not focus on real‐life implications of ADT. However, the Asian Prostate Cancer Study, a database established through an alliance of 12 countries and regions in Asia, is a promising source of real‐world data on the implications of prostate cancer, including ADT.46 3.5 Adverse events associated with ADT While neoadjuvant and/or adjuvant ADT offers significant improvements to survival, it can cause significant morbidity and negatively affect quality of life. Most adverse events are not dose limiting and can be managed through pharmacological or other interventions. ADT is associated with a decrease in bone mineral density, resulting in an increased risk of bone fractures.47 Denosumab and toremifene have both been shown to increase bone mineral density and decrease risk of fractures in randomized controlled studies with over 1000 participants.48, 49 Prospective and population‐based studies have associated the use of ADT with the development of metabolic syndrome, a group of cardiovascular risk factors related to insulin resistance. An observational study by Keating et al analyzed a population‐based cohort of 73 196 men and found an increased risk of insulin resistance and development of diabetes following ADT (HR, 1.44; P < 0.001).50 Exercise and prophylactic use of metformin have been trialed and are promising strategies to prevent the development of metabolic syndrome in men undergoing ADT.51, 52 However, this has only been trialed in small populations and further research with larger cohorts is required to validate these preliminary findings. Increased risk of cardiovascular disease (CVD) and CVD‐related death has been associated with ADT.50 CVD events can differ between different modalities of ADT. A study involving 3578 Chinese patients found that orchiectomy results in more ischemic events than gonadotropin‐releasing hormone agonists.53

iminary findings. Increased risk of cardiovascular disease (CVD) and CVD‐related death has been associated with ADT.50 CVD events can differ between different modalities of ADT. A study involving 3578 Chinese patients found that orchiectomy results in more ischemic events than gonadotropin‐releasing hormone agonists.53 Sexual dysfunction affects over 90% of men receiving ADT, due to the reduction in testosterone. This results in loss of libido, as well as erectile dysfunction. For men with rising PSA but no metastases, intermittent ADT may be a feasible option. In a randomized, noninferiority study involving 1386 patients, there was no difference in OS between men receiving continuous versus intermittent ADT (HR, 1.03; P = 0.009), and the group receiving intermittent ADT experienced significant improvements in libido, fatigue, and overall quality of life.43 Other effects of ADT affecting quality of life include gynecomastia, fatigue, and hot flashes.54 Gynecomastia and breast pain are common adverse events, affecting up to 85% of men receiving high‐dose ADT, and strategies to mitigate these effects are currently being investigated. Although ADT is associated with a number of adverse side effects, a cohort study of 13 368 patients from the Taiwan National Health Insurance Research Database identified a reduced risk of dementia in men with prostate cancer who underwent chemical castration (HR, 0.79; P < 0.001).55

s are currently being investigated. Although ADT is associated with a number of adverse side effects, a cohort study of 13 368 patients from the Taiwan National Health Insurance Research Database identified a reduced risk of dementia in men with prostate cancer who underwent chemical castration (HR, 0.79; P < 0.001).55 3.6 New predictive biomarkers for ADT and personalized therapy Prognostic and predictive biomarkers have the potential for optimizing therapy through personalization of treatment regimens. The use of blood or urine‐based prognostic markers, present a minimally invasive method for determining treatment response, allowing more responsive adjustments to therapy when necessary. Inactivating phosphate and tensin homolog (PTEN) mutations are commonly detected in prostate cancers and are associated with a poorer prognosis. PTEN expression may also be able to predict response to ADT; this is currently being investigated in patients with intermediate‐ and high‐risk prostate cancer (Clinicaltrials.gov identifier NCT01542021). What's more, subtyping based on luminal and basal lineage could also potentially serve as a predictive biomarker.56 Results from an analysis of 1567 prostate cancer samples from high‐risk patients treated with prostatectomy showed that luminal B prostate cancers were significantly associated with response to ADT: 10‐year metastasis occurred in 33% for those treated with ADT versus 55% for those untreated (P = 0.006), suggesting that luminal/basal subtyping may be useful in the selection of patient treatment.56

s treated with prostatectomy showed that luminal B prostate cancers were significantly associated with response to ADT: 10‐year metastasis occurred in 33% for those treated with ADT versus 55% for those untreated (P = 0.006), suggesting that luminal/basal subtyping may be useful in the selection of patient treatment.56 Circulating tumor cells (CTCs) are increasingly investigated prognostic biomarkers in many cancers, including prostate cancer. High levels of CTCs are correlated with unfavorable prognosis and worse clinical outcomes.57, 58 Protein expression in CTCs can also be used as a predictive biomarker of treatment response, where epidermal growth factor receptor (EGFR) expression in CTCs is associated with poor response to ADT (time to progression of 5 months [EGFR+] compared to 11 months [EGFR−]; P < 0.05).59 By using a blood test instead of radiographic imaging, monitoring can be performed more frequently, allowing prompt treatment before metastatic tumors become clinically detectable. Currently, ongoing clinical trials are investigating the use of urine metabolomic profiling, as well as the expression of tumor markers in CTCs as biomarkers for predicting response to therapy. These studies are being explored in small populations and will require further validation before they can be applied in a clinical setting.

going clinical trials are investigating the use of urine metabolomic profiling, as well as the expression of tumor markers in CTCs as biomarkers for predicting response to therapy. These studies are being explored in small populations and will require further validation before they can be applied in a clinical setting. 4 CONCLUSIONS The addition of ADT to primary therapy significantly improves outcomes for certain men with intermediate‐ or high‐risk prostate cancer and is recommended in these settings in our proposed roadmap for treatment (Figure 1). Among men who undergo radical prostatectomy as primary therapy, neoadjuvant ADT is feasible in those with localized or locally advanced prostate cancer and warrants further exploration. Long‐term adjuvant ADT is recommended immediately following radical prostatectomy for men with high‐risk localized or locally advanced prostate cancer, particularly those with positive lymph nodes. Radiotherapy plus long‐term ADT provided an OS benefit and is recommended at the time of biochemical recurrence, particularly for men with GS ≥7, PSA 0.7–4.0 ng/mL, or positive surgical margins. ADT with an LHRHa and an antiandrogen is recommended for 4–6 months following radiotherapy as a primary therapy, for patients with intermediate‐risk disease, while ADT with an LHRHa with or without an antiandrogen is recommended for 2–3 years in high‐risk localized or locally advanced disease. Intermittent ADT was noninferior to continuous therapy following biochemical recurrence but it improved quality of life, particularly during nontreatment phases. Further research is warranted to determine the optimal PSA level at which to begin intermittent ADT; however, the inherent challenges associated with undertaking such studies makes identifying optimal PSA levels an elusive target. Several adverse events are frequently associated with ADT. While they are not dose limiting, some of these events can cause serious morbidity such as death from CVD or bone fractures due to loss of bone mineral density. As such, careful monitoring of patients is required during use of ADT. A predictive biomarker may be helpful to identify patients who would benefit from ADT. It is still at the stage of exploration and research. Further clinical studies are needed to confirm and validate the clinical value of these predictive biomarkers, so as to guide clinical practice in future.

uired during use of ADT. A predictive biomarker may be helpful to identify patients who would benefit from ADT. It is still at the stage of exploration and research. Further clinical studies are needed to confirm and validate the clinical value of these predictive biomarkers, so as to guide clinical practice in future. Figure 1 ADT treatment roadmap [Color figure can be viewed at http://wileyonlinelibrary.com] CONFLICTS OF INTEREST The authors have no conflicts of interest to disclose. ACKNOWLEDGMENTS All authors contributed to the study design and were responsible for the final interpretation of the data and development of the manuscript. All authors read and approved the final manuscript. Editorial assistance in manuscript preparation and data collation for this review was provided by Julie Shilane, PhD, of Articulate Science LLC and Joyce Lee, PhD, of Nucleus Global, Shanghai, China, funded by AstraZeneca.

26.4) 31 (23.8) 1 (10.0) 2 (10.5) 3 (10.3) >3 6 (14.0) 9 (10.3) 15 (11.5) 2 (20.0) 3 (15.8) 5 (17.2) Metastatic disease, n (%) Liver metastasis only 14 (32.6) 26 (29.9) 40 (30.8) 5 (50.0) 7 (36.8) 12 (41.1) Other metastasis 28 (65.1) 61 (70.1) 89 (68.5) 5 (50.0) 12 (63.2) 17 (58.6) No metastasis 1 (2.3) 0 1 (0.8) 0 0 0 BRAF status, n (%) Wild‐type 39 (90.7) 85 (97.7) 124 (95.4) 9 (90.0) 13 (68.4) 22 (75.9) Mutated 4 (9.3) 2 (2.3) 6 (4.6) 1 (10.0) 6 (31.6) 7 (24.1) Prior therapy, n (%) Chemotherapy 20 (46.5) 18 (20.7) 38 (29.2) 2 (20.0) 3 (15.8) 5 (17.2) Radiotherapy 8 (18.6) 10 (11.5) 18 (13.8) 1 (10.0) 0 1 (3.4) Surgery 32 (74.4) 62 (71.3) 94 (72.3) 4 (40.0) 18 (94.7) 22 (75.9) Vaccines 0 0 0 1 (10.0) 0 1 (3.4) Other 2 (4.7) 2 (2.3) 4 (3.1) 1 (10.0) 1 (5.3) 2 (6.9) Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; FOLFOX, infusional fluorouracil, oxaliplatin, and leucovorin; FOLFIRI, infusional fluorouracil, leucovorin, and irinotecan; mCRC, metastatic colorectal cancer.

s 0 0 0 1 (10.0) 0 1 (3.4) Other 2 (4.7) 2 (2.3) 4 (3.1) 1 (10.0) 1 (5.3) 2 (6.9) Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; FOLFOX, infusional fluorouracil, oxaliplatin, and leucovorin; FOLFIRI, infusional fluorouracil, leucovorin, and irinotecan; mCRC, metastatic colorectal cancer. John Wiley & Sons, Ltd.Median PFS for the total population of patients with RAS wt left‐sided mCRC in the APEC study (n = 130) was 14.0 months (95% [confidence interval] CI, 11.4–14.9), and median OS was 30.6 months (95% CI, 25.7–34.3). Median PFS for the total population of patients with RAS wt right‐sided mCRC in the APEC study (n = 29) was 8.9 months (95% CI, 5.5–15.4), and median OS was 24.6 months (95% CI, 13.8–31.2; Figure 1). Median PFS in patients with left‐sided tumors was 12.8 months (95% CI, 9.7–14.9) with cetuximab plus FOLFIRI and 14.2 months (95% CI, 11.2–16.2) with cetuximab plus FOLFOX. Median PFS in patients with right‐sided mCRC was 15.4 months (95% CI, 3.6–20.3) and 8.3 months (95% CI, 3.7–13.3) with cetuximab plus FOLFIRI or FOLFOX, respectively (Table 2, Figure 2). Median OS was 31.7 months (95% CI, 18.5–40.5) and 30.6 months (95% CI, 24.5–36.8) for patients with left‐sided tumors who received cetuximab plus FOLFIRI or FOLFOX, respectively. Median OS was 32.1 months (95% CI, 6.8–not evaluable) and 21.8 months (95% CI, 7.8–29.7) with cetuximab plus FOLFIRI or FOLFOX, respectively, in patients with right‐sided mCRC (Table 2, Figure 2).

1 BACKGROUND Patients who present with metastatic colorectal cancer (mCRC) receive systemic therapy consisting of a targeted monoclonal antibody (mAb) and a chemotherapy backbone, and therapy selection is generally guided by one or more biomarkers (most commonly, mutations in the rat sarcoma [RAS] family genes). It is recommended that patients with mCRC that is RAS wild‐type (wt) receive an anti–epidermal growth factor receptor (EGFR) targeted therapy.1

ntibody (mAb) and a chemotherapy backbone, and therapy selection is generally guided by one or more biomarkers (most commonly, mutations in the rat sarcoma [RAS] family genes). It is recommended that patients with mCRC that is RAS wild‐type (wt) receive an anti–epidermal growth factor receptor (EGFR) targeted therapy.1 Cetuximab, an immunoglobulin G1 (IgG1)‐isotype anti‐EGFR mAb, has been shown in several randomized, phase III trials to combine successfully, with similar efficacy and safety, with either of the doublet chemotherapy regimens available for the treatment of mCRC: FOLFIRI (infusional fluorouracil [5‐FU]/leucovorin/irinotecan) or FOLFOX (infusional 5‐FU/oxaliplatin/folinic acid [leucovorin]).2, 3 Currently, median survival in trials for patients with RAS wt mCRC receiving systemic therapy is ≥30 months,2, 3 although certain patient subpopulations have worse prognoses.4, 5 The location of the primary tumor within the colorectal tract (right vs left) has significant prognostic value for patient survival. Indeed, patients with right‐sided RAS wt tumors have a worse prognosis—that is, much poorer survival outcomes than patients with left‐sided tumors, regardless of the treatment received. In addition, right‐sidedness appears to have predictive value, as patients with right‐sided RAS wt mCRC have been shown to derive less benefit from treatment with cetuximab than patients with left‐sided RAS wt mCRC. By comparison, patients with left‐sided RAS wt tumors appear to fare better if treated with anti‐EGFR therapy plus chemotherapy versus bevacizumab plus chemotherapy.3, 6

patients with right‐sided RAS wt mCRC have been shown to derive less benefit from treatment with cetuximab than patients with left‐sided RAS wt mCRC. By comparison, patients with left‐sided RAS wt tumors appear to fare better if treated with anti‐EGFR therapy plus chemotherapy versus bevacizumab plus chemotherapy.3, 6 The open‐label, nonrandomized, multicenter, phase II APEC trial demonstrated that first‐line cetuximab administered once every 2 weeks in combination with investigator's choice of either FOLFOX or FOLFIRI yielded good response and survival outcomes in an Asian patient population with RAS wt mCRC, with no new or unexpected safety findings.7 Furthermore, the APEC trial found no evidence for differences in efficacy between FOLFIRI and FOLFOX chemotherapy when combined with cetuximab in patients with RAS wt mCRC.7 In this subgroup analysis, we present the efficacy outcomes of the APEC trial by chemotherapy backbone for each tumor location. 2 METHODS Detailed design and methodology for the nonrandomized, phase II APEC study (NCT00778830) were previously described.7 The trial was conducted in accordance with the Declaration of Helsinki. The protocol was approved by the ethics committees of all participating centers. All patients gave written informed consent before trial entry.

ign and methodology for the nonrandomized, phase II APEC study (NCT00778830) were previously described.7 The trial was conducted in accordance with the Declaration of Helsinki. The protocol was approved by the ethics committees of all participating centers. All patients gave written informed consent before trial entry. Briefly, 289 patients with previously untreated KRAS exon 2 (codon 12/13) wt mCRC and an Eastern Cooperative Oncology Group performance status of 0 or 1 were enrolled and assigned to a treatment by participating investigators (cetuximab plus FOLFIRI, n = 101; cetuximab plus FOLFOX, n = 188). Of these, 167 patients had RAS wt disease according to extended RAS analysis (cetuximab plus FOLFIRI, n = 57; cetuximab plus FOLFOX, n = 110). In this population, 159 patients were evaluable for outcome analysis by tumor location. Transverse colon tumors were included in the analysis and classified as right‐sided.

88). Of these, 167 patients had RAS wt disease according to extended RAS analysis (cetuximab plus FOLFIRI, n = 57; cetuximab plus FOLFOX, n = 110). In this population, 159 patients were evaluable for outcome analysis by tumor location. Transverse colon tumors were included in the analysis and classified as right‐sided. Patients received cetuximab (500 mg/m2) on day 1 of every 14‐day treatment cycle over 120 min for the first infusion, 90 min at the second infusion and 60 min at subsequent infusions. Based on investigator's choice, patients received either FOLFOX (oxaliplatin 100 mg/m2, leucovorin 200 mg/m2 L‐form or 400 mg/m2 racemic, then 5‐FU as a 400‐mg/m2 IV bolus and a 2400‐mg/m2 continuous infusion over 46 hours) or FOLFIRI (irinotecan 180 mg/m2, leucovorin 200 mg/m2 L‐form or 400 mg/m2 racemic, then 5‐FU as a 400‐mg/m2 IV bolus and a 2400‐mg/m2 continuous infusion over 46 hours). Treatment was continued until disease progression, unacceptable toxicity or withdrawal of patient consent. All patients included in this analysis received ≥1 dose of treatment. The primary endpoint of the APEC trial was best overall response rate (ORR) as determined by Response Evaluation Criteria In Solid Tumors (RECIST) 1.0. Secondary endpoints included progression‐free survival (PFS), overall survival (OS) and safety. Statistical analysis was performed as previously described.7

Patients received cetuximab (500 mg/m2) on day 1 of every 14‐day treatment cycle over 120 min for the first infusion, 90 min at the second infusion and 60 min at subsequent infusions. Based on investigator's choice, patients received either FOLFOX (oxaliplatin 100 mg/m2, leucovorin 200 mg/m2 L‐form or 400 mg/m2 racemic, then 5‐FU as a 400‐mg/m2 IV bolus and a 2400‐mg/m2 continuous infusion over 46 hours) or FOLFIRI (irinotecan 180 mg/m2, leucovorin 200 mg/m2 L‐form or 400 mg/m2 racemic, then 5‐FU as a 400‐mg/m2 IV bolus and a 2400‐mg/m2 continuous infusion over 46 hours). Treatment was continued until disease progression, unacceptable toxicity or withdrawal of patient consent. All patients included in this analysis received ≥1 dose of treatment. The primary endpoint of the APEC trial was best overall response rate (ORR) as determined by Response Evaluation Criteria In Solid Tumors (RECIST) 1.0. Secondary endpoints included progression‐free survival (PFS), overall survival (OS) and safety. Statistical analysis was performed as previously described.7 3 RESULTS A total of 130 patients with left‐sided RAS wt mCRC were identified (43 received every‐2‐weeks cetuximab plus FOLFIRI and 87 received every‐2‐weeks cetuximab plus FOLFOX). Twenty‐nine patients with right‐sided RAS wt mCRC were identified; 10 patients received cetuximab plus FOLFIRI and 19 received cetuximab plus FOLFOX. Baseline characteristics such as number of metastatic sites, presence of liver‐limited disease, and median age were reasonably balanced between tumor‐location subgroups who received cetuximab plus FOLFIRI or cetuximab plus FOLFOX. 4.6% and 24.1% of patients with left‐ and right‐sided mCRC, respectively, had BRAF mutations. Detailed baseline characteristics by tumor location and treatment arm are shown in Table 1.

and median age were reasonably balanced between tumor‐location subgroups who received cetuximab plus FOLFIRI or cetuximab plus FOLFOX. 4.6% and 24.1% of patients with left‐ and right‐sided mCRC, respectively, had BRAF mutations. Detailed baseline characteristics by tumor location and treatment arm are shown in Table 1. Table 1 Baseline characteristics by tumor sidedness in the phase II APEC study Patients with left‐sided mCRC Patients with right‐sided mCRC Characteristic Cetuximab + FOLFIRI (n = 43) Cetuximab + FOLFOX (n = 87) Total (n = 130) Cetuximab + FOLFIRI (n = 10) Cetuximab + FOLFOX (n = 19) Total (n = 29)

and median age were reasonably balanced between tumor‐location subgroups who received cetuximab plus FOLFIRI or cetuximab plus FOLFOX. 4.6% and 24.1% of patients with left‐ and right‐sided mCRC, respectively, had BRAF mutations. Detailed baseline characteristics by tumor location and treatment arm are shown in Table 1. Table 1 Baseline characteristics by tumor sidedness in the phase II APEC study Patients with left‐sided mCRC Patients with right‐sided mCRC Characteristic Cetuximab + FOLFIRI (n = 43) Cetuximab + FOLFOX (n = 87) Total (n = 130) Cetuximab + FOLFIRI (n = 10) Cetuximab + FOLFOX (n = 19) Total (n = 29) Sex, n (%) Male 27 (62.8) 59 (57.8) 86 (66.2) 7 (70.0) 11 (57.9) 18 (62.1) Female 16 (37.2) 28 (32.2) 44 (33.8) 3 (30.0) 8 (42.1) 11 (37.9) Age, years Median 56.0 59.0 57.5 59.5 53.0 56.0 Min–max 31–87 28–81 28–87 41–70 31–78 31–78 Race, n (%) White 6 (14.0) 16 (18.4) 22 (16.9) 2 (20.0) 2 (10.5) 4 (13.8) Asian (Chinese) 31 (72.1) 42 (48.3) 73 (56.2) 7 (70.0) 11 (57.9) 18 (62.1) Asian (non‐Chinese) 6 (14.0) 29 (33.3) 35 (26.9) 1 (10.0) 6 (31.6) 7 (24.1) ECOG PS, n (%) 0 28 (65.1) 61 (70.1) 89 (68.5) 8 (80.0) 10 (52.6) 18 (62.1) 1 15 (34.9) 26 (29.9) 41 (31.5) 2 (20.0) 9 (47.4) 11 (37.9) Number of metastatic sites, n (%) 0 1 (2.3) 0 1 (0.8) 0 0 0 1 10 (23.3) 23 (26.4) 33 (25.4) 2 (20.0) 3 (15.8) 5 (17.2) 2 18 (41.9) 32 (36.8) 50 (38.5) 5 (50.0) 11 (57.9) 16 (55.2) 3 8 (18.6) 23 (26.4) 31 (23.8) 1 (10.0) 2 (10.5) 3 (10.3) >3 6 (14.0) 9 (10.3) 15 (11.5) 2 (20.0) 3 (15.8) 5 (17.2) Metastatic disease, n (%) Liver metastasis only 14 (32.6) 26 (29.9) 40 (30.8) 5 (50.0) 7 (36.8) 12 (41.1) Other metastasis 28 (65.1) 61 (70.1) 89 (68.5) 5 (50.0) 12 (63.2) 17 (58.6) No metastasis 1 (2.3) 0 1 (0.8) 0 0 0

months (95% CI, 24.5–36.8) for patients with left‐sided tumors who received cetuximab plus FOLFIRI or FOLFOX, respectively. Median OS was 32.1 months (95% CI, 6.8–not evaluable) and 21.8 months (95% CI, 7.8–29.7) with cetuximab plus FOLFIRI or FOLFOX, respectively, in patients with right‐sided mCRC (Table 2, Figure 2). Figure 1 Progression‐free survival (A) and overall survival (B) according to sidedness group treated with either chemotherapy backbone in the phase II APEC study [Color figure can be viewed at http://wileyonlinelibrary.com] Table 2 Efficacy outcomes for cetuximab + FOLFOX and cetuximab + FOLFIRI by tumor sidedness in patients with RAS wt mCRC in the phase II APEC study Patients with left‐sided mCRC Patients with right‐sided mCRC Efficacy outcome Cetuximab + FOLFIRI (n = 43) Cetuximab + FOLFOX (n = 87) Total (n = 130) Cetuximab + FOLFIRI (n = 10) Cetuximab + FOLFOX (n = 19) Total (n = 29) Progression‐free survival, months Median 12.8 14.2 14.0 15.4 8.3 8.9 95% CI 9.7–14.9 11.2–16.2 11.4–14.9 3.6–20.3 3.7–13.3 5.5–15.4 Overall survival, months Median 31.7 30.6 30.6 32.1 21.8 24.6 95% CI 18.5–40.5 24.5–36.8 25.7–34.3 6.8–NE 7.8–29.7 13.8–31.2 Response rate, % Overall response rate 74.4 65.5 68.5 50.0 52.6 51.7 95% CI 58.8–86.5 54.6–75.4 59.7–76.3 18.7–81.3 28.9–75.6 32.5–70.6 Resection rate, n (%) 1 (2.3) 13 (14.9) 14 (10.8) 1 (10.0) 2 (10.5) 3 (10.3) Abbreviations: FOLFOX, infusional fluorouracil, oxaliplatin, and leucovorin; FOLFIRI, infusional fluorouracil, leucovorin, and irinotecan; mCRC, metastatic colorectal cancer; NE, not evaluable; wt, wild type.

76.3 18.7–81.3 28.9–75.6 32.5–70.6 Resection rate, n (%) 1 (2.3) 13 (14.9) 14 (10.8) 1 (10.0) 2 (10.5) 3 (10.3) Abbreviations: FOLFOX, infusional fluorouracil, oxaliplatin, and leucovorin; FOLFIRI, infusional fluorouracil, leucovorin, and irinotecan; mCRC, metastatic colorectal cancer; NE, not evaluable; wt, wild type. John Wiley & Sons, Ltd.Figure 2 Progression‐free survival (A) and overall survival (B) according to treatment and sidedness group in the phase II APEC study [Color figure can be viewed at http://wileyonlinelibrary.com] The ORR, which included complete and partial responses, observed with cetuximab plus FOLFIRI or FOLFOX in patients with left‐sided mCRC was 74.4% (95% CI, 58.8–86.5) and 65.5% (95% CI, 54.6–75.4), respectively. The ORR with cetuximab plus FOLFIRI or FOLFOX in patients with right‐sided mCRC was 50.0% (95% CI, 18.7–81.3) and 52.6% (95% CI, 28.9–75.6), respectively. Few patients in the APEC trial underwent surgery for metastatic disease, but resection rates were similar between the left‐sided and right‐sided subgroups (10.8% and 10.3%, respectively). The APEC study revealed no new safety findings for first‐line, every‐2‐weeks cetuximab plus doublet chemotherapy, mirroring the results from previous (including phase III) studies;2, 7, 8, 9 no safety analysis was performed for this tumor‐location subgroup analysis because, per data from the CRYSTAL (unpublished) and TAILOR trials,10 there is no evidence that the safety profile differs meaningfully between right‐ and left‐sided mCRC.

roring the results from previous (including phase III) studies;2, 7, 8, 9 no safety analysis was performed for this tumor‐location subgroup analysis because, per data from the CRYSTAL (unpublished) and TAILOR trials,10 there is no evidence that the safety profile differs meaningfully between right‐ and left‐sided mCRC. 4 DISCUSSION Here we present the tumor‐location subgroup analysis of the phase II APEC study population of patients with RAS wt mCRC. This is the first study to provide tumor‐location analyses for subgroups treated with the combination of every‐2‐weeks cetuximab plus FOLFIRI and cetuximab plus FOLFOX in parallel, and this thus serves as a hypothesis‐generating landmark study. Because the APEC study was not randomized, the distribution of patients with left‐sided versus right‐sided tumors as well as that of patients receiving FOLFOX versus FOLFIRI may be unbalanced; therefore, outcomes by tumor side and by chemotherapy backbone are not directly comparable. Other limitations of the study were the small sample size and the lack of a bevacizumab comparator arm.

nts with left‐sided versus right‐sided tumors as well as that of patients receiving FOLFOX versus FOLFIRI may be unbalanced; therefore, outcomes by tumor side and by chemotherapy backbone are not directly comparable. Other limitations of the study were the small sample size and the lack of a bevacizumab comparator arm. Chemotherapy backbone did not appear to affect the outcomes in patients with left‐sided RAS wt tumors because median PFS and OS were similar in both groups. Furthermore, both treatment combinations yielded a median OS of >30 months for these patients, in agreement with observations made in pivotal phase III studies in which cetuximab was administered on a weekly schedule.2, 3 Curiously, the data for both PFS and OS observed in patients with right‐sided RAS wt tumors suggested better outcomes with the FOLFIRI backbone than FOLFOX. Indeed, while the number of patients with right‐sided RAS wt mCRC in the APEC study was very low, every‐2‐weeks cetuximab plus FOLFIRI was associated with a median OS of >30 months compared with 21.8 months with FOLFOX in this subgroup. Additionally, first‐line cetuximab in combination with either chemotherapy backbone yielded ORRs of ≥50.0% in patients with right‐sided tumors and a similar rate of resection of metastases to that observed in patients with left‐sided tumors, which is consistent with the ORRs reported for patients with right‐sided RAS wt mCRC in the CRYSTAL (42% with cetuximab plus FOLFIRI), FIRE‐3 (53% with cetuximab plus FOLFIRI) and TAILOR (44% with cetuximab plus FOLFOX) studies.3, 10 Therefore, this study provides further evidence that this may potentially be a preferred combination therapy for right‐sided mCRC when cytoreduction is a key treatment goal. Thus, first‐line cetuximab plus FOLFOX or FOLFIRI can be considered for right‐sided RAS wt mCRC when the patient requires rapid tumor shrinkage to improve symptoms or facilitate resection. Finally, in agreement with previous observations, both cetuximab plus FOLFIRI and cetuximab plus FOLFOX yielded high response rates in patients with left‐sided RAS wt mCRC.3

FOLFIRI can be considered for right‐sided RAS wt mCRC when the patient requires rapid tumor shrinkage to improve symptoms or facilitate resection. Finally, in agreement with previous observations, both cetuximab plus FOLFIRI and cetuximab plus FOLFOX yielded high response rates in patients with left‐sided RAS wt mCRC.3 Although the higher rate of BRAF mutations in cetuximab plus FOLFOX–treated patients with right‐sided disease may partially explain dissimilarities in PFS and OS between the FOLFOX and FOLFIRI cohorts, another possible explanation for the observed trends may be that the synergy between cetuximab and irinotecan observed in preclinical models11, 12 is sufficient to overcome some of the molecular mechanisms that limit the benefits of antitumor therapy in right‐sided mCRC. This was reinforced clinically in the BOND study, in which patients who were resistant to irinotecan had improved ORR and PFS when treated with cetuximab and irinotecan compared to cetuximab alone.13 Further investigation of these findings is required.

s that limit the benefits of antitumor therapy in right‐sided mCRC. This was reinforced clinically in the BOND study, in which patients who were resistant to irinotecan had improved ORR and PFS when treated with cetuximab and irinotecan compared to cetuximab alone.13 Further investigation of these findings is required. Consistent with pivotal phase III studies in first‐line weekly cetuximab plus chemotherapy, a prognostic effect of tumor sidedness was observed in patients with RAS wt mCRC receiving first‐line every‐2‐weeks cetuximab plus doublet chemotherapy in the APEC study. We suggest that the results of the tumor‐location subgroup analysis of the APEC trial confirm published observations on the efficacy of first‐line cetuximab in combination with either FOLFIRI or FOLFOX in patients with left‐sided tumors. Indeed, there appears to be no difference in survival between chemotherapy backbones on the left side. Additionally, the ORR in patients with right‐sided RAS wt mCRC was ≥50.0% in both treatment subgroups, providing evidence for the use of first‐line cetuximab plus FOLFIRI or FOLFOX in this patient population when tumor shrinkage/cytoreduction is a key treatment goal. Finally, the results of this subgroup analysis raise the possibility of a clinically relevant synergy between cetuximab and irinotecan, which may be associated with prolonged PFS and OS in patients with right‐sided RAS wt tumors; further investigation of this observation is warranted. In summary, these results serve to generate interesting hypotheses regarding the choice of chemotherapy backbone with first‐line cetuximab for the treatment of patients with right‐sided mCRC.

ed with prolonged PFS and OS in patients with right‐sided RAS wt tumors; further investigation of this observation is warranted. In summary, these results serve to generate interesting hypotheses regarding the choice of chemotherapy backbone with first‐line cetuximab for the treatment of patients with right‐sided mCRC. CONFLICT OF INTEREST R.E. is an employee of Merck Healthcare KGaA, and holds shares of Merck Healthcare KGaA. W.C. is an employee of Merck Serono, China. P.G. has received research funding support from Merck Healthcare KGaA. All other authors declare no conflict of interest. ACKNOWLEDGMENTS Medical writing assistance was provided by Clinical Thinking, Inc. and funded by Merck Healthcare KGaA, in accordance with Good Publication Practice (GPP3) guidelines (http://www.ismpp.org/gpp3). Clinical trial registration: NCT00778830. AVAILABILITY OF DATA AND MATERIAL All authors had access to the primary data. AUTHORS’ CONTRIBUTIONS All authors contributed equally to the conception of the intellectual content, interpretation of the data and writing of the manuscript. All authors also reviewed any revisions that were made and provided their final approval of the manuscript.

1 INTRODUCTION Following breast‐conserving surgery for breast cancer, a positive surgical margin is widely recognized to increase the risk of local recurrence by at least twofold compared with a negative margin.1, 2, 3 For patients with positive margins after breast‐conserving surgery, the National Comprehensive Cancer Network guidelines and other guidelines recommend re‐excision to ensure that the margins are indeed negative.4, 5, 6 One report has indicated that re‐excision resulted in the detection of residual tumors in 24% to 84% of patients according to the extent of positive margins after breast‐conserving surgery.7 The above‐mentioned guidelines recommend that re‐excision should also be performed to secure negative margins in patients with positive margins after mastectomy.6 Re‐excision after mastectomy, however, is invasive and frequently difficult to perform because it requires the excision of muscle and ribs in addition to the mammary tissue already removed. Re‐excision of the skin may occasionally be performed if a positive margin is detected after nipple‐sparing mastectomy (NSM). Furthermore, re‐excision increases the risk of complications such as postoperative infection and may delay the initiation of subsequent adjuvant chemotherapy.8, 9 Thus, for patients with positive or “close” margins after breast‐conserving surgery, boost irradiation to the primary tumor bed is often added to whole‐breast radiation therapy (WBRT).

creases the risk of complications such as postoperative infection and may delay the initiation of subsequent adjuvant chemotherapy.8, 9 Thus, for patients with positive or “close” margins after breast‐conserving surgery, boost irradiation to the primary tumor bed is often added to whole‐breast radiation therapy (WBRT). Some reports indicate that additional irradiation can decrease the local recurrence rate in patients with positive/close margins to a rate as low as that in patients with negative margins, whereas other reports indicate that the recurrence rate, even after additional irradiation, remains higher than that in patients with negative margins.10, 11, 12 Moreover, in patients with positive/close margins after mastectomy, including those with axillary lymph node metastasis, postmastectomy RT (PMRT) of the chest wall or regional lymph nodes has widely been recognized to reduce the local recurrence rate and has become standard treatment.13, 14 Because the number of patients with positive/close margins without axillary lymph node metastasis after mastectomy is small, however, the significance of PMRT in this patient population is unknown. No recommendations have been offered, even in the American Society of Clinical Oncology guidelines.15 In the electron intraoperative therapy (ELIOT) study, NSM was combined with intraoperative RT with electrons as a strategy for partial chest wall RT (PCWRT), although patients with positive margins were excluded from the study population.16 We performed PCWRT in patients with positive/close margins and no pathological evidence of axillary lymph node metastasis and compared the outcomes with those of patients who did not undergo re‐excision or PCWRT.

l chest wall RT (PCWRT), although patients with positive margins were excluded from the study population.16 We performed PCWRT in patients with positive/close margins and no pathological evidence of axillary lymph node metastasis and compared the outcomes with those of patients who did not undergo re‐excision or PCWRT. 2 PATIENTS AND METHODS Among patients who underwent modified radical mastectomy plus either axillary dissection or sentinel node biopsy for breast cancer between January 2009 and December 2017, 40 patients with positive/close margins who were pathologically shown to have no axillary lymph node metastasis were included in this retrospective study. Postoperatively, 22 patients underwent RT, and 18 did not. None of the 40 patients underwent re‐excision. This study was approved by the Institutional Review Board of Nihon University School of Medicine. Informed consent was obtained from all patients. 2.1 Surgical margin status As is the case after breast‐conserving surgery, a positive margin was defined as an invasive carcinoma or a ductal carcinoma in situ (DCIS) microscopically identified ink on tumor tissue (defined as abnormal tissue just touching the ink).6 A close margin was defined as having a tumor ≤1 mm from the inked resection edge. This definition relied on a shorter distance than that previously accepted (≤10 mm or ≤2 mm).17, 18 2.2 Nuclear grading Severity was classified according to a three‐point scale (grades 1‐3) based on a composite picture of nuclear atypia and mitotic counts on hematoxylin and eosin‐stained carcinomatous tissue specimens.

2.1 Surgical margin status As is the case after breast‐conserving surgery, a positive margin was defined as an invasive carcinoma or a ductal carcinoma in situ (DCIS) microscopically identified ink on tumor tissue (defined as abnormal tissue just touching the ink).6 A close margin was defined as having a tumor ≤1 mm from the inked resection edge. This definition relied on a shorter distance than that previously accepted (≤10 mm or ≤2 mm).17, 18 2.2 Nuclear grading Severity was classified according to a three‐point scale (grades 1‐3) based on a composite picture of nuclear atypia and mitotic counts on hematoxylin and eosin‐stained carcinomatous tissue specimens. 2.3 Immunohistological subtypes Estrogen receptors (ERs) and human epidermal growth factor receptor 2 (HER2) were evaluated by immunohistochemistry (IHC) analysis. HER2 was judged positive when the IHC score was 3+ or there was a positive fluorescence in situ hybridization test result. The Ki‐67 antibody MIB‐1 clone (Dako, Glostrup, Denmark) was used to detect Ki‐67 expression. The Ki‐67 proliferation index was defined as the percentage of cells with positive nuclear Ki‐67 immunostaining in a section of confirmed carcinomatous tissue.

ere was a positive fluorescence in situ hybridization test result. The Ki‐67 antibody MIB‐1 clone (Dako, Glostrup, Denmark) was used to detect Ki‐67 expression. The Ki‐67 proliferation index was defined as the percentage of cells with positive nuclear Ki‐67 immunostaining in a section of confirmed carcinomatous tissue. 2.4 PCWRT parameters The primary tumor bed was contoured by reference to preoperative computed tomography or magnetic resonance imaging, or to a pathological specimen. The clinical target volume (CTV) was defined as the primary tumor bed plus a margin of at least 2 cm with clipping patient and muscle surface and fat. For example, three‐dimensional (3D) conformal RT (CRT) consisting of two tangential beams was planned for the CTV. The field size was set as the partial chest wall measuring ≥9.5 cm in the cephalocaudal dimension (Figure 1). Using the thickness of the chest wall as the basis, irradiation of 4‐ to 10‐MV photons or 5‐ to 8‐MeV electrons with or without a bolus to the skin was planned, and 95% of the prescribed isodose volume was used to irradiate the CTV to the maximum extent. Figure 1 Representative digitally reconstructed radiograph for partial chest wall radiation therapy. The primary tumor bed (red) and clinical target volume (light red) were contoured. The cephalocaudal length was 12 cm, and did not include the entire chest wall [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 1 Representative digitally reconstructed radiograph for partial chest wall radiation therapy. The primary tumor bed (red) and clinical target volume (light red) were contoured. The cephalocaudal length was 12 cm, and did not include the entire chest wall [Color figure can be viewed at http://wileyonlinelibrary.com] 2.5 Statistical methods SPSS version 21.0 (IBM, Armonk, NY, USA) was used for statistical analysis. Clinicopathological characteristics were compared between groups with and without PCWRT using Pearson's χ 2 test.

Figure 1 Representative digitally reconstructed radiograph for partial chest wall radiation therapy. The primary tumor bed (red) and clinical target volume (light red) were contoured. The cephalocaudal length was 12 cm, and did not include the entire chest wall [Color figure can be viewed at http://wileyonlinelibrary.com] 2.5 Statistical methods SPSS version 21.0 (IBM, Armonk, NY, USA) was used for statistical analysis. Clinicopathological characteristics were compared between groups with and without PCWRT using Pearson's χ 2 test. 3 RESULTS Table 1 summarizes the patient characteristics with detailed surgical margin status. The median age of patients at mastectomy was 58.5 years in the PCWRT group and 69 years in the no‐RT group. Premenopausal patients accounted for 40.9% of patients in the PCWRT group and 29.4% in the no‐RT group, with the difference not reaching significance. The prevalence of positive margins was 77.3% in the PCWRT group and 27.8% in the no‐RT group, with the difference being statistically significant (P = 0.002, Pearson's χ 2 test). In the PCWRT group, significantly more tumors at the margins were DCIS. The most common area of positivity on the positive/close margins was the anterior margin in the PCWRT group and the posterior margin in the no‐RT group. In the no‐RT group, there were significantly more tumors at the posterior margins. Four patients in the PCWRT group and one in the no‐RT group had multiple positive/close areas at the margins. There were no significant differences between the two groups regarding other clinicopathological characteristics. The median total dose of PCWRT was 60 Gy (range 50‐66 Gy). We tended to select a dose ≥60 Gy in the case of invasive carcinoma at the positive margin (in 4 out of 5 patients). Acute RT‐related toxicity was observed in only one patient who had undergone surgical removal of a seroma and manifested as an abscess caused by a bacterial infection after postoperative RT. The abscess resolved following drainage and antibiotic therapy. Other RT‐related toxicities included grade 1 dermatitis and grade 1‐2 pneumonitis (National Cancer Institute Common Terminology Criteria for Adverse Events 4.03).19 In both groups, the majority of patients (91.0% and 88.9%, respectively) underwent hormonal therapy, with only a few given chemotherapy. The median follow‐up after mastectomy was 38 months (range 6‐116 months) in the PCWRT group and 48 months (range 23‐60 months) in the no‐RT group. In both groups, the local recurrence‐free survival (RFS) and disease‐free survival (DFS) values were 100%. None of the patients developed a local recurrence.

n chemotherapy. The median follow‐up after mastectomy was 38 months (range 6‐116 months) in the PCWRT group and 48 months (range 23‐60 months) in the no‐RT group. In both groups, the local recurrence‐free survival (RFS) and disease‐free survival (DFS) values were 100%. None of the patients developed a local recurrence. Table 1 Patient characteristics with positive/close margins after mastectomy with pathological N0 status PCWRT Yes No P value Characteristics N = 22 (%) N = 18 (%) Age at mastectomy (years) Median (range) 58.5 (38‐75) 69 (33‐84) 0.062* Premenopausal 9 (40.9) 5 (27.8) 0.386 Histology DCIS 5 (22.7) 2 (11.1) 0.336 Invasive ductal 10 (45.5) 12 (66.7) Invasive lobular 2 (9.0) 3 (16.7) Invasive micropapillary 1 (4.6) 0 Mucinous 3 (13.6) 1 (5.5) Apocrine 1 (4.6) 0 Pathological size (UICC) Tis 5 (22.7) 2 (11.1) 0.336 T1 9 (40.9) 10 (55.6) T2 6 (27.2) 6 (33.3) T3 1 (4.6) 0 T4 1 (4.6) 0 Surgical margin Positive 17 (77.3) 5 (27.8) 0.002 Close (>0 ≤1 mm) 5 (22.7) 13 (72.2) Tumor at positive/close margin Noninvasive 15 (68.2) 6 (33.3) 0.028 Invasive 7 (31.8) 12 (66.7) Tumor at positive margin only Noninvasive 12 (54.5) 3 (16.7) 0.655 Invasive 5 (22.7) 2 (11.1) Location of positive/close margin Anterior 7 (31.8) 4 (22.2) 0.499 Superior 0 0 Medial 4 (18.2) 0 0.057 Lateral 6 (27.2) 1 (5.5) 0.072 Inferior 3 (13.6) 1 (5.5) 0.397 Posterior 7 (31.8) 13 (72.2) 0.011

Invasive 7 (31.8) 12 (66.7) Tumor at positive margin only Noninvasive 12 (54.5) 3 (16.7) 0.655 Invasive 5 (22.7) 2 (11.1) Location of positive/close margin Anterior 7 (31.8) 4 (22.2) 0.499 Superior 0 0 Medial 4 (18.2) 0 0.057 Lateral 6 (27.2) 1 (5.5) 0.072 Inferior 3 (13.6) 1 (5.5) 0.397 Posterior 7 (31.8) 13 (72.2) 0.011 Nuclear grade 1/2 21 (95.4) 17 (94.5) 0.884 3 1 (4.6) 1 (5.5) Immunohistological subtypes ER+/HER2− 18 (82.0) 11 (61.2) 0.145 ER+/HER2+ 2 (9.0) 2 (11.1) ER−/HER2+ 0 2 (11.1) ER−/HER2− 2 (9.0) 1 (5.5) ER+/HER2 borderline 0 2 (11.1) Ki‐67 labeling index <10% 5 (22.7) 6 (33.3) 0.152 ≥10% 13 (59.1) 12 (66.7) Unknown 4 (18.2) 0 Hormone therapy Yes 20 (91.0) 16 (88.9) 0.832 No 2 (9.0) 2 (11.1) Chemotherapy Yes 2 (9.0) 2 (11.1) 0.832 No 20 (91.0) 16 (88.9) Trastuzumab treatment Yes 2 (9.0) 5 (27.8) 0.122 No 20 (91.0) 13 (72.2) Abbreviations: DCIS, ductal carcinoma in situ; ER, estrogen receptor; HER2, human epidermal growth factor 2; PCWRT, partial chest wall radiation therapy; RT, radiation therapy; UICC, Union for International Cancer Control. Clinicopathological P value was calculated using Pearson's χ 2 test for categorical covariates. *P value was calculated using Mann‐Whitney U test. Bold type indicates statistical significance.

Nuclear grade 1/2 21 (95.4) 17 (94.5) 0.884 3 1 (4.6) 1 (5.5) Immunohistological subtypes ER+/HER2− 18 (82.0) 11 (61.2) 0.145 ER+/HER2+ 2 (9.0) 2 (11.1) ER−/HER2+ 0 2 (11.1) ER−/HER2− 2 (9.0) 1 (5.5) ER+/HER2 borderline 0 2 (11.1) Ki‐67 labeling index <10% 5 (22.7) 6 (33.3) 0.152 ≥10% 13 (59.1) 12 (66.7) Unknown 4 (18.2) 0 Hormone therapy Yes 20 (91.0) 16 (88.9) 0.832 No 2 (9.0) 2 (11.1) Chemotherapy Yes 2 (9.0) 2 (11.1) 0.832 No 20 (91.0) 16 (88.9) Trastuzumab treatment Yes 2 (9.0) 5 (27.8) 0.122 No 20 (91.0) 13 (72.2) Abbreviations: DCIS, ductal carcinoma in situ; ER, estrogen receptor; HER2, human epidermal growth factor 2; PCWRT, partial chest wall radiation therapy; RT, radiation therapy; UICC, Union for International Cancer Control. Clinicopathological P value was calculated using Pearson's χ 2 test for categorical covariates. *P value was calculated using Mann‐Whitney U test. Bold type indicates statistical significance. John Wiley & Sons, Ltd.4 DISCUSSION Some patients with breast cancer who undergo a mastectomy are node‐negative, and understandably, there may be a reluctance to recommend RT of the entire chest wall. Previous studies have shown that patient characteristics can vary in terms of margin status, presence of axillary lymph node metastasis, and use of PMRT (Table 2). The association between margin status and local recurrence after mastectomy without PMRT was investigated in two studies conducted during the 1980s. Both studies addressed the association between the posterior margin (the distance from the tumor margin to the pectoral fascia) and local recurrence.20, 21 One of the studies included patients with axillary lymph node metastasis,20 and the other focused on patients without axillary lymph node metastasis.21 Neither study showed any correlation between the distance from the tumor to the posterior margin and local recurrence. In a study that included patients with axillary lymph node metastasis who did not undergo PMRT, the presence of positive/close margins (<5 mm) after postoperative chemotherapy was associated with an increased risk of locoregional recurrence.22 In another study in patients at pathological stage T1‐2 without axillary lymph node metastasis and who did not undergo PMRT, the local recurrence rate was extremely low (1.68%). The study protocol, however, excluded patients with positive margins from the analysis.23 In a study involving patients at pathological stage T3 without axillary lymph node metastasis who did not undergo PMRT, the locoregional recurrence rate was 7.1%.

undergo PMRT, the local recurrence rate was extremely low (1.68%). The study protocol, however, excluded patients with positive margins from the analysis.23 In a study involving patients at pathological stage T3 without axillary lymph node metastasis who did not undergo PMRT, the locoregional recurrence rate was 7.1%. Cancer recurred on the chest wall in 85.7% of the patients with recurrence, but the surgical margins were not examined.24 In a study conducted in patients aged ≤50 years with positive/close margins (≤6 mm) who did not undergo PMRT, including those with axillary lymph node metastasis, the local recurrence rate was high (23%), indicating that younger patients with positive/close margins might benefit more from PMRT than older patients.17 In studies conducted in high‐risk patients, such as those with tumors ≥5 cm who underwent PMRT and those with invasion of the skin or fascia, the locoregional recurrence rate decreased from 17%‐23% to 3%‐6% after PMRT was performed along with chemotherapy or hormone therapy, even in patients without axillary lymph node metastasis. However, the surgical margins were not examined in these studies.25, 26 In a study involving patients at pathological stage T3 without axillary lymph node metastasis after PMRT and who were selected from the U.S. National Cancer Data Base, PMRT was shown to improve outcomes, regardless of whether the margins were positive or negative.27 In that study, the RT field was also analyzed, and regional lymph node area irradiation in combination with chest wall RT was not associated with the outcomes. In a study conducted in patients with positive/close margins (≤2 mm), including those with axillary lymph node metastasis after PMRT, the presence of positive/close margins did not correlate with locoregional recurrence when a 15 or 20 Gy boost dose was delivered to the tumor bed in addition to PMRT.18 Regarding the site of local recurrence, some reports have shown that the cancer recurred in the surgical field in 33% to 67% of patients after breast‐conserving surgery.28, 29 For partial breast irradiation, which has recently been implemented based on these reports, the CTV is defined as an area covering the tumor bed cavity and at least 1 cm of surrounding mammary gland.30 In patients who have undergone mastectomy, however, the CTV for RT is difficult to establish because the surrounding mammary gland has been removed.

, which has recently been implemented based on these reports, the CTV is defined as an area covering the tumor bed cavity and at least 1 cm of surrounding mammary gland.30 In patients who have undergone mastectomy, however, the CTV for RT is difficult to establish because the surrounding mammary gland has been removed. Local chest wall recurrence after mastectomy has reportedly been detected within 7 cm of the inferior margin of the operation's transverse scar in ≥97% of patients.31 In our patients, who did not have risk factors (eg, axillary lymph node metastasis) other than positive/close margins, the RT field was not expanded to cover the entire chest wall in a uniform manner. Instead, we selected PCWRT with a field having at least 2‐cm margins to the primary tumor bed and measuring ≥9.5 cm in the cephalocaudal dimension. In all previous studies of PMRT, the RT field was all of the chest wall with or without the regional lymph node area. To the best of our knowledge, the current study is the first to report the selection of only part of the chest wall for PMRT. Regarding the sites of positivity on the margin, one study showed that the proportion of patients with residual tumor did not correlate with the presence of four positive mammary gland margin directions (medial, lateral, superior, and inferior) or the presence of two nonmammary gland margin directions (anterior and posterior) in patients undergoing breast‐conserving surgery.32 Another study showed that positive nonmammary gland margins are associated with a significantly lower local recurrence rate than positive peripheral mammary gland margins, suggesting that subcutaneous and retromammary fat should be retained during breast‐conserving surgery.33 In the present study, the absence of local recurrence might have been attributable to the fact that all patients had undergone mastectomy and had positive/close nonbreast parenchymal margins.

mammary gland margins, suggesting that subcutaneous and retromammary fat should be retained during breast‐conserving surgery.33 In the present study, the absence of local recurrence might have been attributable to the fact that all patients had undergone mastectomy and had positive/close nonbreast parenchymal margins. RT doses, which have been investigated in patients with positive/close margins (≤2 mm), showed no correlation between the margin status and locoregional recurrence when a 15‐ or 20‐Gy boost dose was delivered to the tumor bed in addition to PMRT at 50 Gy.18 Another study showed that when patients with positive margins after breast‐conserving surgery received a 10‐ or 26‐Gy boost dose in addition to WBRT at 50 Gy, there was no difference in the local recurrence rate in the two groups.34 After evaluating the results of these studies, we irradiated only part of the chest wall with a total dose of 50‐66 Gy in the present study and achieved good local control. However, local control was also satisfactory in patients without irradiation, and therefore the significance of PCWRT remains unclear. In the present study, the baseline characteristics varied between the two groups. For example, the PCWRT group included significantly more patients with positive margins. When the breast parenchyma was removed by mastectomy, locoregional control might have been more strongly affected by the presence or absence of lymph node metastasis than by the surgical margin status. Limitations of the present study include the relatively short follow‐up period. However, during our review of previous studies on patients’ undergoing mastectomy for breast cancer, one study investigated any recurrence that developed within 2 years after the procedure,26 and another showed that locoregional recurrence was detected in 88% of patients within 3 years.27 Thus, the median follow‐up period in the present study may be considered acceptable. In the present study, both the PCWRT group and the no‐RT group showed equally good local control without re‐excision. We therefore suggest that PCWRT be considered for patients with node‐negative breast cancer after mastectomy if they have positive or close margins, although further studies are required to verify these results.

esent study, both the PCWRT group and the no‐RT group showed equally good local control without re‐excision. We therefore suggest that PCWRT be considered for patients with node‐negative breast cancer after mastectomy if they have positive or close margins, although further studies are required to verify these results. Table 2 Summary of previous studies with/without PMRT after mastectomy for positive/close margins Author Year Years of patient accrual No. of positive/close margin patients Definition of positive/close margin Axillary lymph node metastasis PMRT Local recurrence rate Mentzer et al.20 1986 1974‐1982 100 ≤30 mm Included Botha 8.0% Ahlborn et al.21 1988 1975‐1980 88 ≤4 mm Excluded No 5.7% Katz et al.22 2001 1975‐1994 29 <5 mm Included No 45% Freedman et al.17 1998 1984‐1993 34 ≤6 mm Included No 14.7% Feigenberg et al.18 2003 1978‐1998 49 ≤2 mm Included Yes 6.1% Abbreviation: PMRT, postmastectomy radiation therapy. a This study included patients with or without PMRT. John Wiley & Sons, Ltd.5 DECLARATIONS 5.1 Availability of data and material The data used during this study are available from the corresponding author on reasonable request. COMPETING INTERESTS Drs. Ishibashi, Nishimaki, Maebayashi, Adachi, Sakurai, Masuda, Hata, and Okada declare that they have no competing interests. ACKNOWLEDGMENTS We thank Nancy Schatken, BS, MT(ASCP), and Clare Cox, PhD, from Edanz Group (http://www.edanzediting.com/ac) for editing a draft of this manuscript.