Pembrolizumab plus chemotherapy in Japanese patients with triple‐negative breast cancer: Results from KEYNOTE‐355

Abstract Pembrolizumab plus chemotherapy improved progression‐free survival (PFS) and overall survival (OS) compared with placebo plus chemotherapy in patients with previously untreated locally recurrent inoperable or metastatic triple‐negative breast cancer with tumor programmed cell death ligand 1 (PD‐L1) combined positive score (CPS) ≥10 in the global, phase 3, randomized controlled trial KEYNOTE‐355. We report results for patients enrolled in Japan. Patients were randomized 2:1 to pembrolizumab 200 mg or placebo Q3W for 35 cycles plus chemotherapy (nab‐paclitaxel, paclitaxel, or gemcitabine–carboplatin). Primary endpoints were PFS per RECIST version 1.1 by blinded independent central review and OS in patients with PD‐L1 CPS ≥10, PD‐L1 CPS ≥1, and the intention‐to‐treat (ITT) population. No alpha was assigned to this exploratory analysis. Eighty‐seven patients were randomized in Japan (pembrolizumab plus chemotherapy, n = 61; placebo plus chemotherapy, n = 26), 66 (76%) had PD‐L1 CPS ≥1, and 28 (32%) had PD‐L1 CPS ≥10. Median time from randomization to data cutoff (June 15, 2021) was 44.7 (range, 37.2–52.9) months in the ITT population. Hazard ratios (HRs; 95% CI) for OS were 0.36 (0.14–0.89), 0.52 (0.30–0.91), and 0.46 (0.28–0.77) in the PD‐L1 CPS ≥10, PD‐L1 CPS ≥1, and ITT populations, respectively. HRs (95% CI) for PFS were 0.52 (0.20–1.34), 0.61 (0.35–1.06), and 0.64 (0.39–1.05). Grade 3 or 4 treatment‐related adverse events occurred in 85% of patients in each group (no grade 5 events). Consistent with the global population, pembrolizumab plus chemotherapy tended to show improvements in OS and PFS with manageable toxicity versus placebo plus chemotherapy in Japanese patients and supports this combination in this setting.


| INTRODUCTION
Breast cancer was the leading cause of cancer among women in 2020, both globally and in Japan. 1,2 Breast cancer is associated with high mortality and was the fifth leading cause of cancer-related death among women in Japan in 2019. 2 However, Asian patients are sometimes underrepresented in global, phase 3 trials. Evidence has suggested that differences in patient characteristics and tumor pathophysiologic characteristics (such as luminal A to luminal B ratio, frequency of TP53 mutations, and tumor-infiltrating lymphocyte gene signatures) between Asian and nonAsian patients may affect the observed response to treatment in clinical trials. 3 Additionally, safety findings from certain trials of targeted therapies in patients with breast cancer have reported higher frequencies of adverse events (such as neutropenia and thrombocytopenia) in Asian patients than in non-Asian patients. 3 Given these considerations, it is important to assess response to treatments specifically in Japanese patients.
Evidence has suggested that Japanese patients may have better survival outcomes following treatment than American patients with breast cancer. 4 Such differences in treatment outcomes may be due to biological factors, treatment patterns, lifestyle, and differences in the gut microbiome (potentially due to differences in diet). [4][5][6] Triple-negative breast cancer (TNBC), defined as breast cancer that is estrogen receptor-, progesterone receptor-, and human epidermal growth factor receptor 2-negative, accounts for 10%-15% of all breast cancers in Japan and is associated with a poor prognosis. 4,7,8 Cytotoxic chemotherapy has historically been the mainstay standard of care for metastatic TNBC. [9][10][11] Pembrolizumab, a humanized anti-programmed cell death protein 1 (PD-1) antibody, has demonstrated efficacy in patients with TNBC. [12][13][14][15] In patients with early-stage TNBC, pembrolizumab in combination with standard-of-care chemotherapy demonstrated statistically significant and clinically meaningful improvement in pathological complete response and event-free survival (primary endpoints) in the phase 3 KEYNOTE-522 study. 14,15 Among patients with programmed cell death ligand 1 (PD-L1)-positive metastatic TNBC, pembrolizumab plus chemotherapy improved progression-free survival (PFS) and overall survival (OS) outcomes compared with placebo plus chemotherapy. 12,13 In the global phase 3 KEYNOTE-355 (NCT02819518) study of pembrolizumab plus chemotherapy versus placebo plus chemotherapy in patients with previously untreated locally recurrent inoperable or metastatic TNBC, pembrolizumab plus chemotherapy significantly improved the study's primary endpoints of PFS and OS versus placebo plus chemotherapy in patients with metastatic TNBC whose tumors expressed PD-L1 (combined positive score [CPS] ≥10). 12,13 Among the 847 patients enrolled in KEYNOTE-355 worldwide, 87 patients were from Japan. Given the potential for differences in response to treatment among Japanese patients, it is important to assess outcomes in this setting. Here, we report results for the subset of patients in KEYNOTE-355 enrolled in Japan.

| Patient eligibility
Eligibility criteria for KEYNOTE-355 have been previously published. 12,13 Briefly, eligible patients were ≥18 years of age, with centrally confirmed TNBC (as defined by No alpha was assigned to this exploratory analysis. Eighty-seven patients were randomized in Japan (pembrolizumab plus chemotherapy, n = 61; placebo plus chemotherapy, n = 26), 66 (76%) had PD-L1 CPS ≥1, and 28 (32%) had PD-L1 CPS ≥10.  16,17 that was locally recurrent inoperable or metastatic, previously untreated with chemotherapy, and could not be treated with curative intent. Patients  had completed treatment for stage I-III breast cancer (if indicated), with ≥6 months having elapsed between completion of treatment with curative intent (i.e., date of primary breast tumor surgery or date of last adjuvant chemotherapy administration including capecitabine, whichever occurred last) and first documented local or distant disease recurrence; measurable disease per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 as assessed by the investigator; provided a newly obtained tumor sample from a locally recurrent inoperable or metastatic site for central determination of TNBC status and PD-L1 expression; an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1; and adequate organ function. Exclusion criteria included an active autoimmune disease that had required systemic treatment in the past 2 years, diagnosis of immunodeficiency or receipt of systemic steroid therapy or any other immunosuppressive therapy within 7 days before randomization, active central nervous system metastases and/or carcinomatous meningitis, and receipt of prior therapy with an anti-PD(L)-1 agent, anti-PD-L2 agent, or an agent directed to another coinhibitory T-cell receptor.
The trial protocol was approved by an institutional review board or independent ethics committee at each site, and the trial was conducted in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki. Written informed consent was obtained from all patients.

| Study design
As previously described, 12,13 KEYNOTE-355 (Clini calTr ials.gov, NCT02819518) was a randomized, placebocontrolled, double-blind, global, phase 3 study of pembrolizumab plus chemotherapy versus placebo plus chemotherapy in patients with locally recurrent inoperable or metastatic TNBC that was previously untreated with chemotherapy and could not be treated with curative intent.
Eligible patients were randomized 2:1 in a doubleblind manner to receive pembrolizumab plus chemotherapy or placebo plus chemotherapy. Pembrolizumab 200 mg IV or saline placebo was administered every 3 weeks for up to 35 cycles (approximately 2 years). Chemotherapy comprised the investigator's choice of nab-paclitaxel 100 mg/m 2 IV on days 1, 8, and 15 every 28 days; paclitaxel 90 mg/m 2 IV on days 1, 8, and 15 every 28 days; or gemcitabine 1000 mg/m 2 and carboplatin AUC 2 on days 1 and 8 every 21 days. Randomization was stratified by investigator's choice of chemotherapy (taxane vs. gemcitabine-carboplatin), tumor PD-L1 status (CPS <1 vs. CPS ≥1), and prior treatment with the same class of chemotherapy in the neoadjuvant or adjuvant setting (yes vs. no). Study treatment continued until completion of 35 cycles (approximately 2 years) of pembrolizumab or placebo (treatment with chemotherapy could have been continued at the investigator's discretion), disease progression, unacceptable toxicity, intercurrent illness that prevented further treatment administration, investigator decision, or patient withdrawal of consent. Crossover between treatment arms was not permitted.

| Assessments
Tumor imaging was performed at baseline, at weeks 8, 16, and 24 from randomization, then every 9 weeks for the first year and every 12 weeks thereafter. Responses were assessed by RECIST version 1.1 by blinded independent central review. Survival was assessed every 12 weeks until withdrawal of consent or end of study. Adverse events (AEs) were monitored from the time of randomization through 30 days following cessation of treatment (90 days for serious AEs) and were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. Baseline tumor PD-L1 expression was assessed using PD-L1 IHC 22C3 pharmDx (Agilent Technologies) at a central laboratory and reported as CPS, defined as the number of PD-L1-positive tumor cells, lymphocytes, and macrophages divided by the total number of tumor cells multiplied by 100.

| Endpoints
The dual primary endpoints were PFS per RECIST version 1.1 as assessed by blinded independent central review and OS in patients in the intention-to-treat population, patients with PD-L1 CPS ≥10 tumors, and patients with PD-L1 CPS ≥1 tumors. The primary endpoints were amended to include PFS and OS in patients with PD-L1 CPS ≥10 tumors following the completion of enrollment and the first interim analysis based on data from other studies that reported increased clinical benefit with PD-L1 enrichment. 12,[18][19][20][21] Key secondary endpoints were safety in all treated patients and objective response rate (ORR) assessed by blinded independent central review in the intention-to-treat population, patients with PD-L1 CPS ≥10 tumors, and patients with PD-L1 CPS ≥1 tumors. Other secondary endpoints included duration of response and disease control rate per RECIST version 1.1 by blinded independent central review, each in the intention-to-treat population, in patients with PD-L1 CPS ≥10 tumors, and in patients with PD-L1 CPS ≥1 tumors. Analysis of outcomes among patients enrolled in Japan was exploratory.

| Statistical analysis
Statistical considerations for the global KEYNOTE-355 study have been previously described. 12,13 This study was not designed to test hypotheses in the Japanese subset and therefore lacks power for inferential purposes in this population. No alpha was assigned to this analysis. Efficacy was assessed in patients with PD-L1 CPS ≥10 tumors, patients with PD-L1 CPS ≥1 tumors, and the intention-to-treat population (i.e., all randomized patients). Safety was assessed in all treated patients (i.e., all randomized patients who received ≥1 dose of study treatment), and patients were included in the treatment group corresponding to the study treatment received. PFS and OS were estimated by the Kaplan-Meier method, and treatment differences were assessed by the stratified log-rank test. HRs and corresponding 95% CIs were assessed using a Cox proportional hazards model with the Efron method of tie handling to assess the magnitude of treatment differences. Differences in ORR were assessed using the stratified Miettinen and Nurminen method.
In Japanese patients, the median time from randomization to the database cutoff date of June 15, 2021 (i.e., the same database cutoff date as the protocol-specified final analysis for the global KEYNOTE-355 population reported previously 13 ) was 44.7 months (range, 37.2-52.9 months) in the intention-to-treat population. At the time of data cutoff, 54 patients (89%) in the pembrolizumab plus chemotherapy group had discontinued all study treatments, 5 patients (8%) had completed 35 cycles (approximately 2 years) of pembrolizumab, and 2 patients (3%) were continuing study treatment. The reasons for discontinuation were progressive disease (n = 40), AEs (n = 5), withdrawal of consent (n = 5), complete response (n = 2), and physician decision (n = 2). In the placebo plus chemotherapy group, 25 patients (96%) had discontinued all study treatments due to progressive disease, 1 patient (4%) had completed treatment, and no patients were continuing study treatment ( Figure 1). T A B L E 1 Demographics and baseline disease characteristics in the intention-to-treat population. Median duration of treatment (for any component of study medication) in the pembrolizumab plus chemotherapy group and placebo plus chemotherapy group was 7.4 months (range, 0.3-44.3 months) and 7.1 months (range, 1.2 to 25.8 months), respectively. A summary of on-study treatment exposure is provided in Table S1 and  Table S2.

PD-L1 CPS ≥1
Intention-to-treat population  No assessment includes subjects without postbaseline assessment on the data cutoff date.
c From product-limit (Kaplan-Meier) method for censored data.

| Safety
Treatment-related AEs occurred in all 61 patients (100%) and 26 patients (100%) in the pembrolizumab plus chemotherapy and placebo plus chemotherapy groups, respectively. The most common treatment-related AEs in both treatment groups were decreased white blood cell count (pembrolizumab plus chemotherapy, 75%; placebo plus chemotherapy, 85%), decreased neutrophil count (72% and 81%), anemia (66% and 62%), and nausea (52% and 62%). Grade 3 or 4 treatment-related AEs occurred in 52 patients (85%) in the pembrolizumab plus chemotherapy group and 22 (85%) in the placebo plus chemotherapy group. The most frequently occurring grade 3 or 4 treatment-related AEs in the pembrolizumab plus chemotherapy group versus placebo plus chemotherapy group were decreased neutrophil count (59% vs. 62%), decreased white blood cell count (49% vs. 58%), and anemia (26% vs. 23%). Nineteen patients (31%) in the pembrolizumab plus chemotherapy group and four patients (15%) in the placebo plus chemotherapy group discontinued any treatment due to a treatment-related AE. No patients in either group died due to a treatment-related AE. Immune-mediated AEs and infusion reactions of any grade occurred in 20 patients (33%) in the pembrolizumab plus chemotherapy group and four patients (15%) in the placebo plus chemotherapy group. The most frequently occurring immune-mediated AEs (incidence ≥5%) in the pembrolizumab plus chemotherapy group were hypothyroidism (n = 7, 11%) and adrenal insufficiency (n = 5, 8%). In the placebo plus chemotherapy group, immunemediated AEs were hyperthyroidism and vasculitis (n = 1 each, 3.8%). Ten patients (16%) in the pembrolizumab plus chemotherapy group and two patients (8%) in the placebo plus chemotherapy group had infusion reactions. Grade 3 or 4 immune-mediated AEs and infusion reactions occurred in four patients (7%) in the pembrolizumab plus chemotherapy group (grade 3 adrenal insufficiency, n = 2; grade 4 infusion reaction, n = 1; grade 3 severe skin reaction, n = 1); there were no grade 3 or 4 events in the placebo plus chemotherapy group. Five patients (8%) in the pembrolizumab plus chemotherapy group and two patients (8%) in the placebo plus chemotherapy group discontinued any treatment due to an immune-mediated AE. No patients in either group died due to an immunemediated AE or infusion reaction (Table 3).

| DISCUSSION
In this subset analysis of patients in the phase 3 KEYNOTE-355 study enrolled in Japan, pembrolizumab plus chemotherapy improved outcomes versus placebo plus chemotherapy in patients with previously untreated locally recurrent inoperable or metastatic TNBC. Baseline characteristics were generally consistent among Japanese patients and those enrolled in the global population. 12,13 A slightly higher proportion of patients enrolled in Japan had an ECOG PS of 0 (pembrolizumab plus chemotherapy, 85%; placebo plus chemotherapy, 81%) compared with the global population (59% and 62%), and slightly higher proportions had received gemcitabine-carboplatin as chemotherapy on-study (pembrolizumab plus chemotherapy, 80%; placebo plus chemotherapy, 69%) compared with the global study (55% and 55%). A broad body of evidence supports the efficacy of gemcitabine-platinum combinations in patients with metastatic TNBC, 12,22-24 which may explain, at least in part, why Japan investigators were interested in the gemcitabine-carboplatin combination and used it more at a higher rate than in the study overall, despite it not being a standard-of-care therapy in Japan. 9 Treatment with pembrolizumab plus chemotherapy improved both OS and PFS, and the 18-month OS rates and 12-month PFS rates were higher in the pembrolizumab plus chemotherapy group than in the placebo plus chemotherapy group. The ORR and disease control rates were generally similar between the treatment groups; however, the duration of response was more than doubled among responders in the pembrolizumab plus chemotherapy group compared with the placebo plus chemotherapy group. Overall, toxicity was manageable. Findings from this analysis are generally consistent with those from the global KEYNOTE-355 population and support the use of pembrolizumab plus chemotherapy as a new standard-ofcare treatment regimen for patients with previously untreated locally recurrent inoperable or metastatic TNBC in Japan. 12,13 The magnitude of treatment benefit with pembrolizumab plus chemotherapy versus placebo plus chemotherapy appeared greater in Japanese patients than in the global study across each of the PD-L1 CPS populations (i.e., PD-L1 CPS ≥10, PD-L1 CPS ≥1, and the intention-totreat population). 12,13 Notably, the 18-month OS rates were higher in patients in the pembrolizumab plus chemotherapy group versus the placebo plus chemotherapy group in the Japanese subset; the 18-month OS rates were approximately doubled in each of the PD-L1 CPS populations. However, given the wide 95% CIs for OS and PFS in the Japan subset due to smaller patient numbers than in the global study, particularly among Japanese patients with PD-L1 CPS ≥10 (n = 19 and n = 9, respectively), we are unable to draw definitive conclusions from this data set. The reasons for the differences in OS and PFS outcomes for patients enrolled in Japan compared with the overall study population are uncertain; possible contributing factors may include the small number of patients and differences  in patient baseline characteristics. The greater proportion of patients with ECOG PS of 0 in the Japanese subgroup compared with the global population (84% vs. 60%) may have contributed to the longer median OS times among the Japanese population versus the global population. The ORRs and disease control rates were generally similar between the treatment groups in the Japan subset and the global study in each of the PD-L1 CPS populations. 12,13 The duration of response was slightly longer in the Japan subset than in the global study, particularly in patients with PD-L1 CPS ≥10 tumors; however, it is important to note the very small number of responders with PD-L1 CPS ≥10 tumors in the Japan subset (pembrolizumab plus chemotherapy, n = 8; placebo plus chemotherapy, n = 3).
The toxicity profile was generally consistent with the profile reported for the global study. 12,13 In the Japan subset, a slightly higher incidence of grade 3 or 4 treatmentrelated AEs was reported in both treatment groups (Japan, 85% in both treatment groups; global study, 68% in pembrolizumab plus chemotherapy group vs. 67% in placebo plus chemotherapy group), and more patients in the pembrolizumab plus chemotherapy group discontinued due to treatment-related AEs (31%) compared with the global study (18%). The higher incidence of grade 3 or 4 treatment-related AEs and discontinuations due to AEs in the Japan subset may be attributed to differences in clinical characteristics; in particular, a higher proportion of patients in the Japan subset had received gemcitabinecarboplatin as chemotherapy on-study. Additionally, in the Japan subgroup, 8% of patients in the pembrolizumab plus chemotherapy group experienced adrenal insufficiency (grade 3 or 4, 3%) compared with none in the placebo plus chemotherapy. Although immune-mediated AEs and infusion reactions occurred more frequently in the pembrolizumab plus chemotherapy group than in the placebo plus chemotherapy group few of these were grade 3 or 4 and, notably, a similar proportion of patients in either treatment arm discontinued any treatment due to an immune-mediated AE or infusion reaction.
In a subgroup analysis of Japanese patients (n = 65) enrolled in the phase 3 IMpassion130 study that evaluated first-line atezolizumab (an anti-PD-L1 inhibitor) in combination with nab-paclitaxel in patients with locally advanced or metastatic TNBC, the median PFS was 7.4 months (95% CI, 5.  25 As with results from KEYNOTE-355, these outcomes provide evidence for activity of anti-PD-(L)1 agents in Japanese patients with advanced TNBC.
A limitation of this study is that the Japanese subset comprises a portion of patients enrolled globally in KEYNOTE-355 (87 of 847 patients) and no alpha was allocated to this analysis. Secondly, given the small number of patients in this analysis, we were not able to perform any meaningful subset analyses, including assessments of clinical benefit with different routinely used chemotherapy regimens, assessment of the influence of differences in clinical characteristics on outcomes, or differences in treatment patterns. In the current study, the allocation of patients to nabpaclitaxel, paclitaxel, or gemcitabine-carboplatin was by investigator's choice. At the time of writing, there is no unequivocal evidence for whether either of these chemotherapy regimens provides more clinical benefit than the other in this setting. 11 The investigators' choice of chemotherapy regimen may have been driven, at least in part, by the regulatory and clinical guideline status of these drugs in Japan; in particular, it is important to note that platinum chemotherapeutics are not currently included as standard of care for metastatic TNBC in Japan. 9 Nonetheless, it is apparent that approval of the combination of pembrolizumab plus chemotherapy (based on results from KEYNOTE-355) has added to the treatment options for patients with PD-L1-positive TNBC in Japan. 26 In Japanese patients enrolled in KEYNOTE-355, pembrolizumab plus chemotherapy tended to show improvements in OS and PFS with manageable toxicity versus placebo plus chemotherapy, consistent with the global population. These findings provide support for the use of pembrolizumab plus chemotherapy in Japanese patients with previously untreated locally recurrent inoperable or metastatic TNBC.

AUTHOR CONTRIBUTIONS
Masaya Hattori: Investigation (equal); validation (equal); writing -review and editing (equal). Norikazu Masuda: Investigation (equal); validation (equal); writing -review and editing (equal). Toshimi Takano: Investigation (equal); validation (equal); writing -review and editing (equal). Koichiro Tsugawa: Investigation (equal); validation (equal); writing -review and editing (equal). Kenichi Inoue: Investigation (equal); writing -review and editing (equal). Koji Matsumoto: Investigation (equal); resources (equal); validation (equal); writing -review and editing (equal). Takashi Ishikawa: Investigation (equal); resources (equal); writing -review and editing (equal). Mitsuya Itoh: Investigation (equal); resources (equal); writing -review and editing (equal). Hiroyuki Yasojima: Investigation (equal); writingoriginal draft (equal); writing -review and editing (equal). Yuko Tanabe: Investigation (equal); validation (equal); writing -review and editing (equal). Keiko Yamamoto: Investigation (equal); validation (equal); writing -review and editing (equal). Masato Suzuki: Formal analysis (equal); writing -review and editing (equal). Wilbur Pan: Formal analysis (equal); investigation (equal); validation (equal); writing -review and editing (equal). Javier Cortes: Conceptualization (equal); writing -review and editing (equal). Hiroji Iwata: Investigation (equal); validation (equal); writing -review and editing (equal). the requestors. In line with data privacy legislation, submitters of approved requests must enter into a standard datasharing agreement with MSD before data access is granted. Data will be made available for request after product approval in the US and EU or after product development is discontinued. There are circumstances that may prevent MSD from sharing requested data, including country or region-specific regulations. If the request is declined, it will be communicated to the investigator. Access to genetic or exploratory biomarker data requires a detailed, hypothesisdriven statistical analysis plan that is collaboratively developed by the requestor and MSD subject matter experts; after approval of the statistical analysis plan and execution of a data-sharing agreement, MSD will either perform the proposed analyses and share the results with the requestor or will construct biomarker covariates and add them to a file with clinical data that is uploaded to an analysis portal so that the requestor can perform the proposed analyses.

ETHICS STATEMENT
The trial protocol was approved by an institutional review board or independent ethics committee at each site, and the trial was conducted in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki.

PATIENT CONSENT STATEMENT
Written informed consent was obtained from all patients.

PERMISSION TO REPRODUCE MATERIAL FROM OTHER SOURCES
Not applicable.