Received: 19 August 2022 | Revised: 24 October 2022 | Accepted: 7 November 2022 DOI: 10.1111/cas.15650 ORIGINAL ARTICLE Pembrolizumab in Asian patients with microsatellite-­instability-­ high/mismatch-­repair-­deficient colorectal cancer Takayuki Yoshino1 | Thierry Andre2 | Tae Won Kim3 | Wei Peng Yong4 | Kai-­Keen Shiu5 | Benny Vittrup Jensen6 | Lars Henrik Jensen7 | Cornelis J. A. Punt8,9 | Denis Smith10 | Rocio Garcia-­Carbonero11 | Julia Alcaide-­Garcia12 | Peter Gibbs13 | Christelle de la Fouchardiere14 | Fernando Rivera15 | Elena Elez16 | Dung T. Le17 | Noriaki Adachi18 | David Fogelman18 | Patricia Marinello18 | Luis A. Diaz Jr19 1 National Cancer Center Hospital East, Kashiwa, Chiba, Japan 2 Department of Medical Oncology, Sorbonne University, Saint-­Antoine Hospital, AP-­HP, INSERM 938, SIRIC CURAMUS, Paris, France 3 Asan Medical Center, University of Ulsan, Seoul, South Korea 4 National University Hospital, National University Cancer Institute, Singapore, Singapore 5 University College Hospital, NHS Foundation Trust, London, UK 6 Herlev and Gentofte Hospital, Herlev, Denmark 7 University Hospital of Southern Denmark, Vejle, Denmark 8 Amsterdam University Medical Center, University of Amsterdam, Amsterdam, Netherlands 9 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, Netherlands 10 Hôpital Haut-­Lévêque, Centre Hospitalier Universitaire de Bordeaux, Pessac, France 11 Hospital Universitario 12 de Octubre, Imas12, Universidad Complutense de Madrid (UCM), Madrid, Spain 12 Unidad de Gestión Clínica Intercentros de Oncología Médica, Hospitales Universitarios Regional y Virgen de la Victoria, IBIMA, Malaga, Spain 13 Western Hospital, Footscray, Victoria, Australia 14 Centre Léon Bérard, Lyon, France 15 Fernando Rivera, University Hospital Marqués de Valdecilla, IDIVAL, Santander, Spain 16 Vall d'Hebron Barcelona Hospital Campus, Vall D'Hebron Institute of Oncology, Universitat Autònoma de Barcelona, Barcelona, Spain 17 The Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins Medicine, Baltimore, Maryland, USA 18 Merck & Co., Inc., Rahway, New Jersey, USA 19 Division of Solid Tumor Oncology, Memorial Sloan Kettering Cancer Center, New York, New York, USA Correspondence Takayuki Yoshino, National Cancer Abstract Center Hospital East, 6-­5-­1 Kashiwanoha, The phase 3 KEYNOTE-­177 study evaluated pembrolizumab versus chemother- Kashiwa, Chiba 277-­8577, Japan. Email: tyoshino@east.ncc.go.jp apy with or without bevacizumab or cetuximab in patients with newly diagnosed, microsatellite-­instability-­high (MSI-­H)/mismatch-­repair-­deficient (dMMR) metastatic Funding information Merck Sharp and Dohme colorectal cancer (mCRC). Primary endpoints were progression-­free survival (PFS) per RECIST v1.1 by blinded independent central review (BICR) and overall survival (OS). Secondary endpoints were overall response rate (ORR) per RECIST v1.1 by BICR and safety. Here, we report results from the post hoc analysis of patients who were en- rolled in Asia from the final analysis (FA) of KEYNOTE-­177. A total of 48 patients from This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2022 Merck Sharp & Dohme LLC and The Authors. Cancer Science published by John Wiley & Sons Australia, Ltd on behalf of Japanese Cancer Association. 1026 |  wileyonlinelibrary.com/journal/cas Cancer Science. 2023;114:1026–1036. YOSHINO et al. | 1027 Japan, Korea, Singapore, and Taiwan (pembrolizumab, n = 22; chemotherapy, n = 26) were included. At FA, median time from randomization to data cutoff (February 19, 2021) was 45.3 (range 38.1–­57.8) months with pembrolizumab and 43.9 (range 36.6–­ 55.1) months with chemotherapy. Median PFS was not reached (NR; 95% confidence interval [CI] 1.9 months–­NR) with pembrolizumab versus 10.4 (95% CI 6.3–­22.0) months with chemotherapy (hazard ratio [HR] 0.56, 95% CI 0.26–­1.20). Median OS was NR (range 13.8 months–­NR) versus 30.0 (14.7–­NR) months (HR 0.65, 95% CI 0.27–­1.55) and ORR was 50% (95% CI 28–­72) versus 46% (95% CI 27–­67). Grade 3/4 treatment-­related adverse events (TRAEs) were reported by two patients (9%) in the pembrolizumab arm and 20 (80%) in the chemotherapy arm. Immune-­mediated ad- verse events or infusion reactions were reported by six patients (27%) and 10 patients (40%), respectively. No deaths due to TRAEs occurred. These data support first-­line pembrolizumab as a standard of care for patients from Asia with MSI-­H/dMMR mCRC. Clini​calTr​ials.gov identifier: NCT02563002. KEYWORDS Asia, colorectal cancer, microsatellite instability, mismatch-­repair deficiency, pembrolizumab 1 | I NTRO D U C TI O N demonstrated substantial antitumor activity in patients with pre- viously treated MSI-­H/dMMR mCRC.8–­11 As a consequence, treat- Colorectal cancer (CRC) was the third most commonly diagnosed ment guidelines for Asia were updated to recommend routine MMR cancer worldwide and the second leading cause of cancer-­related testing because of the strong predictive value of MMR status for the mortality in 2020.1 Although outcomes have improved with early use of ICIs in mCRC.6 Recent data from the phase 3 KEYNOTE-­177 detection and advances in treatment, the 5-­year overall survival study showed that pembrolizumab has robust and durable anti- (OS) rate for patients with metastatic CRC (mCRC) remains less than tumor activity as first-­line therapy in patients with MSI-­H/dMMR 2,3 15%. The incidence of CRC has risen in most countries over the mCRC.13,14 Pembrolizumab provided significantly longer median 4 last three decades, with a substantial increase observed in Asia. progression-­free survival (PFS) compared with chemotherapy (16.5 Particularly high rates of CRC have been reported in high-­income vs. 8.2 months, hazard ratio [HR] 0.60, 95% confidence interval [CI] countries in East and Southeast Asia, suggested to be largely attrib- 0.45–­0.80, P = 0.0002) at the second interim analysis and a nonsta- utable to the adoption of a Westernized lifestyle, the prevalence of tistically significant but clinically meaningful improvement in OS at smoking, and the aging population.5 In 2017, the age-­standardized final analysis (FA; median not reached [NR] vs. 36.7 months, HR 0.74, incidence rate of CRC (per 100,000 person-­years) was 48.0 in 95% CI 0.53–­1.03, P = 0.036).13,14 Fewer treatment-­related adverse Taiwan, 45.0 in Japan, 34.9 in Singapore, and 32.5 in South Korea events (TRAEs) were reported with pembrolizumab compared with 4 compared with a global rate of 23.2. Therefore, there is substantial chemotherapy, with clinically meaningful improvement in health-­ interest in the effectiveness of emerging treatments for CRC in pa- related quality of life.14,15 However, data are limited on the efficacy tients from Asia. of ICIs in Asian patients with MSI-­H/dMMR mCRC. Here, we report Treatment options for patients with unresectable CRC world- results from the FA of the KEYNOTE-­177 study for patients who wide have until recently been limited to chemotherapy and targeted were enrolled in Asia. agents.6,7 Evidence from a number of clinical trials has shown that CRC tumors characterized by DNA mismatch-­repair deficiency (dMMR) and a high degree of microsatellite instability (MSI-­H) 2 | M ATE R I A L S A N D M E TH O DS are susceptible to treatment with immune checkpoint inhibitors (ICIs).8–­11 Tumors that are MSI-­H/dMMR have an increased tumor 2.1 | Study design and participants mutational burden that leads to the production of neoantigens, re- sulting in MSI-­H/dMMR tumors being more immunogenic than those KEYNOTE-­177 (NCT02563002) was an international, multicenter, 12 with proficient DNA mismatch repair. As a result, MSI-­H/dMMR open-­label, phase 3 study. Details of the study were previously tumors are associated with high numbers of tumor-­infiltrating lym- published.13 In brief, eligible patients were aged at least 18 years 12 phocytes, whose activity can be enhanced by ICI. The programmed and had MSI-­H/dMMR stage IV CRC. Patients had measurable dis- death 1 (PD-­1) inhibitors pembrolizumab and nivolumab have both ease per RECIST v1.1, an ECOG performance status of 0 or 1, and 1028 | YOSHINO et al. adequate organ function. Patients must not have received prior safety. Duration of response (DOR) per RECIST v1.1 by BICR was systemic therapy for stage IV CRC but may have received prior ad- included as an exploratory endpoint. juvant chemotherapy for CRC completed at least 6 months before randomization. Patients were randomly assigned (1:1) to receive pembrolizumab 2.4 | Statistical analyses 200 mg i.v. every 3 weeks or the investigator's choice of chemother- apy with or without bevacizumab or cetuximab every 2 weeks. No This unplanned post hoc subgroup analysis included only patients stratification factors were used for randomization. The choice of enrolled in KEYNOTE-­177 from the Asian region, defined as Japan, chemotherapeutic regimen was determined prior to randomization. Korea, Singapore, and Taiwan. Efficacy was assessed in the intention-­ Chemotherapy options included modified FOLFOX6 (mFOLFOX6; to-­treat (ITT) population, comprising all patients enrolled in Asia who oxaliplatin 85 mg/m2 i.v. over 2 h, day 1; leucovorin 400 mg/m2 i.v. were randomly assigned to treatment. Safety was assessed in the over 2 h, day 1; 5-­fluoropyrimidine 400 mg/m2 i.v. bolus, day 1, then as-­treated population, comprising all patients enrolled in Asia who 1200 mg/m2/day for 2 days for a total of 2400 mg/m2 delivered by were randomly assigned to treatment and received at least one dose continuous infusion over 46–­48 h), mFOLFOX6 plus bevacizumab of study treatment. (5 mg/kg i.v., day 1), mFOLFOX6 plus cetuximab (400 mg/m2 i.v. PFS, OS, and DOR were estimated using the Kaplan–­Meier over 2 h, then 250 mg/m2 i.v. over 1 h, weekly), FOLFIRI (irinotecan method. For the PFS analysis, data for patients who were alive with 180 mg/m2 i.v. over 30–­90 min, day 1; leucovorin 400 mg/m2 i.v. in- no PD were censored at the time of the last imaging assessment, 2 fusion to match irinotecan, day 1; 5-­fluoropyrimidine 400 mg/m i.v. and data for patients who underwent surgery with curative intent bolus, day 1, then 1200 mg/m2/day for 2 days for a total of 2400 mg/ were censored at the surgical date. For the OS analysis, missing data m2 delivered by continuous infusion over 46–­48 h), FOLFIRI plus bev- for patients were censored at the date of last known contact. A Cox acizumab (5 mg/kg i.v., day 1), and FOLFIRI plus cetuximab (400 mg/ proportional-­hazards model with Efron's method of tie handling was 2 2 m i.v. over 2 h, then 250 mg/m i.v. over 1 h, weekly). Treatment with used to estimate HRs and associated 95% CIs. pembrolizumab was continued for ≤2 years (~35 cycles) or until pro- Two interim analyses and a FA were prespecified in the protocol gressive disease (PD), unacceptable toxicity, or physician or patient for the global population.14 The FA was planned to be performed decision to withdraw. Eligible patients randomly assigned to chemo- after 190 OS events had occurred across both arms or 12 months therapy who experienced PD per RECIST v1.1 as assessed by the after the second interim analysis, whichever occurred first. investigator and confirmed by blinded independent central review Additional details are provided in Appendix S1 (Additional Statistical (BICR) could cross over to receive pembrolizumab for ≤35 cycles at Analyses Methods, Data Availability Statement; Table S1). SAS ver- the investigator's discretion. sion 9.4 was used for all statistical analyses. 2.2 | Efficacy and safety assessments 2.5 | Ethics approval MMR or MSI status was determined locally. MMR status was deter- Written informed consent was obtained from all patients prior to mined by immunohistochemical analysis, with tumors classified as enrollment. The study protocol and all amendments were reviewed dMMR by the absence of at least one of four MMR enzymes (MLH1, and approved by the appropriate institutional review board or eth- MSH2, MSH6, and PMS2). Polymerase chain reaction-­based analysis ics committee at each study center, and the study was conducted of tumor microsatellite loci was also undertaken locally, with MSI-­H in accordance with the ethical principles of the Declaration of status defined as the detection of at least two allele shifts among the Helsinki. three to five allele shifts analyzed. Imaging by computed tomogra- phy or magnetic resonance imaging was performed every 9 weeks. Tumor response was assessed per RECIST v1.1 by BICR. Adverse 3 | R E S U LT S events (AEs) were evaluated throughout the study and for 30 days after treatment discontinuation (90 days for serious AEs). AEs Between February 11, 2016, and February 19, 2018, of the 307 were graded according to the National Cancer Institute Common patients randomly assigned to treatment in KEYNOTE-­177, 48 Terminology Criteria for Adverse Events, version 4.0. were from Asia (Japan [n = 22], Korea [n = 14], Singapore [n = 5], or Taiwan [n = 7]), of whom 22 received pembrolizumab and 26 re- ceived chemotherapy (Figure 1). At data cutoff (February 19, 2021), 2.3 | Outcomes of the 22 patients who were assigned to receive pembrolizumab, all received ≤1 dose of study treatment, 10 (45%) completed study The primary endpoints were PFS per RECIST v1.1 by BICR and OS. treatment, and 12 (55%) discontinued study treatment. Reasons Secondary endpoints included overall response rate (ORR; complete for discontinuation in the pembrolizumab arm were radiologic response [CR] or partial response [PR]) per RECIST v1.1 by BICR and PD (n = 7), AEs (n = 2), clinical progression (without radiologic YOSHINO et al. | 1029 F I G U R E 1 Patient disposition in the 307 patients randomly assigned Asia subgroup Global Population Global Population 153 assigned to pembrolizumab 154 assigned to chemotherapy Asia Subgroup Asia Subgroup 22 assigned to pembrolizumab 26 assigned to chemotherapy 22 received pembrolizumab 25 received chemotherapy Asia Subgroup Asia Subgroup 10 (45%) completed treatment 0 (0%) completed treatment 12 (55%) discontinued 23 (92%) discontinued 7 (32%) progressive disease 17 (68%) progressive disease 2 (9%) adverse event 2 (8%) adverse event 1 (5%) clinical progression 0 (0%) clinical progression 1 (5%) complete response 1 (4%) complete response 1 (5%) physician decision 1 (4%) physician decision 0 (0%) withdrawal by patient 2 (8%) withdrawal by patient 0 (0%) still receiving treatment 2 (8%) still receiving treatment evidence) (n = 1), CR (n = 1), and physician decision (n = 1). Of the chemotherapy group received subsequent therapies other than the 26 patients who were assigned to receive chemotherapy, 25 an anti-­PD-­1/L1 therapy. received ≤1 dose of study treatment, two (8%) were continuing to receive study treatment, and 23 (92%) had discontinued study treatment. Reasons for study discontinuation in the chemotherapy 3.2 | Progression-­free survival arm were PD (n = 17), AEs (n = 2), withdrawal by the patient or physician (n = 3), and CR (n = 1). At data cutoff, the median PFS was NR (95% CI 1.9 months–­NR) in The median time from randomization to data cutoff was 45.3 the pembrolizumab arm and 10.4 (6.3–­22.0) months in the chemo- (range 38.1–­57.8) months in the pembrolizumab arm and 43.9 (range therapy arm (HR 0.56, 95% CI 0.26–­1.20) (Figure 2). The 12-­month 36.6–­55.1) months in the chemotherapy arm. PFS rate was 62% (95% CI 38–­79) in the pembrolizumab arm com- pared with 46% (95% CI 25–­6 4) in the chemotherapy arm. The 24-­ month PFS rates were 52% (95% CI 29–­71) and 28% (95% CI 11–­46), 3.1 | Demographics and baseline characteristics respectively. Demographics and baseline characteristics were generally balanced between treatment arms (Table 1). The median (range) age was 65.5 3.3 | Overall survival (24–­83) years in the pembrolizumab arm and 64.0 (31–­90) years in the chemotherapy arm. Compared with chemotherapy, a higher pro- The median OS was NR (95% CI 13.8 months-­NR) in the pembroli- portion of patients receiving pembrolizumab had an ECOG perfor- zumab arm and 30.0 (14.7–­NR) months in the chemotherapy arm (HR mance status of 0 (13 [59%] vs. 10 [38%]), had right-­sided tumors (14 0.65, 95% CI 0.27–­1.55) (Figure 3). The 12-­month OS rate was 77% [64%] vs. 13 [50%]), and had received prior (neo)adjuvant therapy (8 (95% CI 54–­90) in the pembrolizumab arm compared with 73% (95% [36%] vs. 5 [19%]) (Table 1). CI 52–­86) in the chemotherapy arm. The 24-­month OS rates were At data cutoff, two patients (9%) in the pembrolizumab arm re- 68% (95% CI 45–­83) and 50% (95% CI 30–­67), respectively. ceived subsequent anti-­PD-­1/L1 therapies. A total of 10 patients (38%) assigned to chemotherapy had crossed over to pembroli- zumab after PD was confirmed, and an additional four patients 3.4 | Radiographic response (15%) received subsequent anti-­PD-­1/L1 therapies outside the study, for an effective crossover rate of 54% among patients initially The ORR was 50% (95% CI 28–­72) in the pembrolizumab arm and randomly assigned to chemotherapy (Table 2). A total of eight pa- 46% (95% CI 27–­67) in the chemotherapy arm (Table 3). In the pem- tients (36%) in the pembrolizumab group and six patients (23%) in brolizumab arm, four patients (18%) had CR, seven patients (32%) 1030 | YOSHINO et al. TA B L E 1 Baseline characteristics of the Asia subgroup TA B L E 2 Subsequent anticancer therapy in the Asia subgroup Pembrolizumab vs. chemotherapy Pembrolizumab, Chemotherapy, n = 22 n = 26 Pembrolizumab, Chemotherapy, n = 22 n = 26 Any anti-­PD-­1/PD-­L1 2 (9) 14 (54) therapy, n (%) Age, median (range), years 65.5 (24–­83) 64.0 (31–­90) On-­protocol therapy 1 (5) 10 (38) ≥65 years 11 (50) 11 (42) –­pembrolizumaba <65 years 11 (50) 15 (58) Off-­protocol therapies 1 (5) 4 (15) Sex, n (%) Any non-­anti–­PD-­1/PD-­L1 8 (36) 6 (23) Male 9 (41) 14 (54) therapy, n (%) Female 13 (59) 12 (46) Chemotherapy 7 (32) 4 (15) Race, n (%) Folic acid derivative 4 (18) 2 (8) Asian 22 (100) 25 (96) VEGF inhibitor 4 (18) 3 (12) White 0 (0) 1 (4) EGFR inhibitor 3 (14) 1 (4) ECOG performance status, n (%) TIM3 inhibitor 1 (5) 1 (4) 0 13 (59) 10 (38) Abbreviations: EGFR, epithelial growth factor receptor; PD-­1, 1 9 (41) 16 (62) programmed death 1; PD-­L1, programmed death ligand 1; TIM3, T-­cell Stage, n (%) immunoglobulin and mucin-­domain containing molecule 3; VEGF, Recurrent metachronousa 12 (55) 11 (42) vascular endothelial growth factor. a Including a second course of treatment for patients randomly assigned Newly diagnosed 10 (45) 15 (58) to the pembrolizumab arm. Metastases location, n (%) Hepatic and/or pulmonary 9 (41) 11 (42) Other 13 (59) 15 (58) had PR, and two patients (9%) had stable disease. In the chemo- Site of primary tumor, n (%) therapy arm, three patients (12%) had CR, nine patients (35%) had Right 14 (64) 13 (50) PR, and nine patients (35%) had stable disease. More patients in the Left 7 (32) 10 (38) pembrolizumab arm had PD compared with the chemotherapy arm Other/missing b 1 (5) 3 (12) (eight [36%] vs. four [15%]). Prior systemic therapy, n (%) Adjuvant only 6 (27) 2 (8) Neoadjuvant with or without 2 (9) 3 (12) 3.5 | Duration of response adjuvant None 14 (64) 21 (81) The median time to response (CR or PR) was 2.1 (range 1.9–­33.2) Mutation status, n (%) months in the pembrolizumab arm and 2.1 (range 1.9–­24.9) months BRAF/KRAS/NRAS all wild 4 (18) 5 (19) in the chemotherapy arm (Table 3). The median DOR was NR (range type 4.4+ to 45.7+; + indicates there was no PD by the time of the last KRAS/NRAS mutant and 5 (23) 8 (31) disease assessment) in the pembrolizumab arm and 18.6 (range 3.4+ BRAFV600E not mutant to 47.9+) months in the chemotherapy arm (Figure 4 and Table 3). BRAFV600E mutant and KRAS/ 5 (23) 5 (19) Seven patients (88%) in the pembrolizumab arm and four (44%) in NRAS not mutant the chemotherapy arm had an estimated DOR of at least 24 months. Could not be evaluated for 8 (36) 8 (31) BRAF, KRAS, NRASc MSI-­highd 22 (100) 26 (100) 3.6 | Safety Abbreviations: BRAF, v-­R AF murine sarcoma viral oncogene homolog; KRAS, Kirsten rat sarcoma viral oncogene homolog; MSI, microsatellite The median duration of treatment exposure was 20.7 (range 0.0–­ instability; NRAS, neuroblastoma RAS viral oncogene homolog. a 25.3) months in the pembrolizumab arm and 8.4 (range 1.6–­49.2) Recurrence was defined as a secondary colorectal cancer occurring at least 6 months after the index cancer. months in the chemotherapy arm. AEs occurred in 22 patients b The tumor site was classified as ‘Other’ if primary tumors were located (100%) in the pembrolizumab arm and 25 patients (100%) in the on both the left and right sides. chemotherapy arm. Grade 3–­5 AEs were reported in 10 patients c Patients could not be evaluated for BRAF, KRAS, or NRAS, if no (46%) receiving pembrolizumab and 22 (88%) receiving chemo- BRAF V600E, KRAS, or NRAS mutation was present, and if ≥1 mutation therapy. Grade 3–­5 AEs occurring in at least 10% of patients were status was undetermined or missing or the type of BRAF mutation was increased gamma-­glutamyl transferase levels in three patients (14%) not BRAF V600E. d MSI-­high status was determined locally by means of a polymerase in the pembrolizumab arm and decreased neutrophil count in 10 pa- chain reaction or immunohistochemical testing. tients (40%), decreased white blood cell count in four patients (16%), YOSHINO et al. | 1031 F I G U R E 2 Kaplan–­Meier estimates of Median PFS, Events, n (%) months (95% CI)a HR (95% CI)b progression-­free survival (PFS) in the Asia 100 Pembrolizumab 10 (45%) NR (1.9-NR) 0.56 (0.26-1.20) subgroup (n = 48). aFrom the product-­ 12-mo rate Chemotherapy 20 (77%) 10.4 (6.3-22.0) 90 limit (Kaplan–­Meier) method for censored 62% 46% 24-mo rate Progression-free survival, % data. bBased on the Cox regression model 80 52% with Efron's method of tie handling with 70 28% treatment as a covariate. CI, confidence 60 interval; HR, hazard ratio; mo, month; NR, not reached 50 40 30 20 10 0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 Time, months No. at Risk Pembrolizumab 22 13 12 12 11 10 10 10 10 10 6 4 1 1 0 0 Chemotherapy 26 19 14 10 7 7 6 6 5 5 4 3 2 0 0 0 F I G U R E 3 Kaplan–­Meier estimates of Median OS, Events, n (%) months (95% CI)a HR (95% CI)b overall survival (OS) in the Asia subgroup 100 Pembrolizumab 8 (36%) NR (13.8-NR) 0.65 (0.27-1.55) (n = 48). aFrom the product-­limit (Kaplan–­ 12-mo rate Chemotherapy 14 (54%) 30.0 (14.7-NR) 90 Meier) method for censored data. bBased 77% 73% 24-mo rate on the Cox regression model with Efron's 80 68% method of tie handling with treatment as 50% Overall survival, % 70 a covariate. CI, confidence interval; HR, 60 hazard ratio; mo, month; NR, not reached 50 40 30 20 10 0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 Time, months No. at Risk Pembrolizumab 22 20 18 17 16 15 15 15 15 14 11 7 2 2 1 0 Chemotherapy 26 25 22 19 17 16 13 13 13 12 10 7 2 0 0 0 and decreased appetite in three patients (12%) in the chemotherapy and hypothyroidism in two patients (8%), pneumonitis in two (8%), arm. TRAEs occurred in 14 patients (64%) in the pembrolizumab arm and colitis in one (4%) in the chemotherapy arm. Infusion reactions compared with 25 patients (100%) in the chemotherapy arm. Grade occurred in one patient (5%) in the pembrolizumab arm and five pa- 3/4 TRAEs were reported for two patients (9%) and 20 patients tients (20%) in the chemotherapy arm (all grade 1/2). (80%) in the pembrolizumab and chemotherapy arms, respectively; there were no grade 5 TRAEs (Tables 4 and S1). Two patients (9%) in the pembrolizumab arm and one patient (4%) in the chemotherapy 4 | DISCUSSION arm discontinued treatment due to TRAEs (increased alanine ami- notransferase levels and psoriasis in the pembrolizumab arm, and In this post hoc subgroup analysis of patients from Asia with MSI-­H/ pneumonitis in the chemotherapy arm). dMMR mCRC enrolled in KEYNOTE-­177, first-­line treatment with Immune-­mediated AEs or infusion reactions occurred in six pa- pembrolizumab provided numerically longer PFS, OS, and DOR tients (27%) in the pembrolizumab arm and 10 patients (40%) in the compared with chemotherapy. The ORR was similar between the chemotherapy arm (Table 5). These included hypothyroidism in three two arms. These results are generally consistent with those ob- patients (14%), hyperthyroidism in two (9%), adrenal insufficiency in served in the global population evaluated in KEYNOTE-­177.13,14 one (5%), and pneumonitis in one (5%) in the pembrolizumab arm, In the global population, first-­line treatment with pembrolizumab 1032 | YOSHINO et al. TA B L E 3 Summary of antitumor Pembrolizumab, Chemotherapy, responses in the Asia subgroup n = 22 n = 26 Overall responsea No. of patients 11 12 % (95% CI) 50.0 (28.2–­71.8) 46.2 (26.6–­66.6) Best overall response, n (%) Complete response 4 (18) 3 (12) Partial response 7 (32) 9 (35) Stable disease 2 (9) 9 (35) Progressive disease 8 (36) 4 (15) Not evaluable/no assessmentb 1 (5) 1 (4) Time to response, median (range), months 2.1 (1.9–­33.2) 2.1 (1.9–­24.9) Duration of responsec median (range), NR (4.4+ to 45.7+) 18.6 (3.4+ to 47.9+) months Response duration of ≥24 months, n (%)c 7 (88) 4 (44) Note: + Indicates there was no progressive disease by the time of the last disease assessment. Abbreviations: CI, confidence interval; NR, not reached. a Overall response was defined as a confirmed complete response or partial response. b Patients for whom no postbaseline imaging was available. c The Kaplan–­Meier method for censored data was used to calculate median of duration. 100 F I G U R E 4 Kaplan–­Meier estimates for duration of response (DOR) in patients Patients remaining in response, % 90 from the Asia subgroup with a confirmed 80 response (n = 23). + indicates there was 70 no progressive disease by the time of last disease assessment. NR, not reached 60 Median DOR, months (range) 50 NR (4.4-45.7+) 18.6 (3.4-47.9+) 40 30 20 10 0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 Time, months No. at Risk Pembrolizumab 11 11 9 7 7 7 7 7 6 6 3 1 0 0 0 0 Chemotherapy 12 11 9 7 5 4 4 4 3 2 2 1 0 0 0 0 provided a significantly longer PFS, higher ORR, and prolonged re- Among patients in the Asia subgroup, the Kaplan–­Meier curves sponse duration compared with chemotherapy. OS was also longer of PFS with pembrolizumab and chemotherapy cross shortly after with pembrolizumab versus chemotherapy in the global popula- treatment initiation and continue to diverge, with higher PFS rates tion; however, this difference was not statistically significant. This observed with pembrolizumab compared with chemotherapy (52% result is partially due to the number of deaths observed at FA. vs. 28%) at 24 months. PFS improvement was clinically meaningful The protocol-­specified FA of OS was planned after 190 deaths had for pembrolizumab compared with chemotherapy (HR 0.56, 95% CI occurred across both arms or 12 months after the second interim 0.26–­1.20), which was consistent with the HR benefit seen in the analysis. As the latter cutoff was reached first, the analysis was global population (HR 0.59, 95% CI 0.45–­0.79).14 13 conducted when only 140 deaths were observed. It might also Similarly, in the Asia subgroup, the Kaplan–­Meier curves for OS be related to the high effective crossover rate (60%) of patients in with pembrolizumab and chemotherapy cross each other and then the chemotherapy arm who received off-­s tudy anti-­PD-­1/L1 ther- continue to diverge. OS in the Asia subgroup was numerically longer apy leading to improved survival and a nonsignificant difference in for pembrolizumab compared with chemotherapy (HR 0.65, 95% CI OS between the two arms.13 0.27–­1.55), which was also consistent with findings from the global YOSHINO et al. | 1033 TA B L E 4 Treatment-­related adverse Pembrolizumab, n = 22 Chemotherapy, n = 25 events of any grade that occurred in at least 10% of patients in either treatment Treatment-­related adverse Any gradea Grade 3–­5a Any gradea Grade 3–­5a arm in the Asia subgroup event n (%) n (%) n (%) n (%) Any 14 (64) 2 (9) 25 (100) 20 (80) Malaise 4 (18) 0 4 (16) 0 Diarrhea 3 (14) 0 11 (44) 2 (8) Hypothyroidism 3 (14) 0 0 0 Pyrexia 3 (14) 0 3 (12) 0 ALT level increased 2 (9) 1 (5) 5 (20) 0 AST level increased 2 (9) 1 (5) 5 (20) 1 (4) Decreased appetite 2 (9) 0 13 (52) 2 (8) Fatigue 2 (9) 0 13 (52) 2 (8) Alopecia 1 (5) 0 9 (36) 0 Anemia 1 (5) 0 3 (12) 1 (4) Rash 1 (5) 0 3 (12) 0 Nausea 0 0 15 (60) 0 Neutrophil count decreased 0 0 15 (60) 10 (40) Peripheral sensory 0 0 14 (56) 1 (4) neuropathy Stomatitis 0 0 11 (44) 0 WBC count decreased 0 0 11 (44) 4 (16) Vomiting 0 0 9 (36) 0 PPE syndrome 0 0 6 (24) 0 Epistaxis 0 0 5 (20) 0 Hypersensitivity 0 0 5 (20) 0 Platelet count decreased 0 0 4 (16) 0 Proteinuria 0 0 4 (16) 1 (4) Skin hyperpigmentation 0 0 4 (16) 0 Hypoesthesia 0 0 4 (16) 0 Dermatitis acneiform 0 0 3 (12) 0 Hypertension 0 0 3 (12) 2 (8) Neutropenia 0 0 3 (12) 2 (8) Paronychia 0 0 3 (12) 1 (4) Urticaria 0 0 3 (12) 0 Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; PPE, palmar-­ plantar erythrodysesthesia; WBC, white blood cell. a Grades are based on National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0. population (HR 0.74, 95% CI 0.53–­1.03, P = 0.036).14 Similar effective in the global population (45% vs. 33%). CR rates in the Asian sub- crossover rates were also observed between the Asia subgroup and population for pembrolizumab-­ and chemotherapy-­t reated pa- the global population (54% and 60%), which likely contributed to the tients (18% vs. 12%, respectively) were somewhat greater than in lack of statistical significance in OS. At FA, 10 of 26 patients (38%) the global study (13% vs. 4%, respectively). However, comparisons in the Asia subgroup who were randomly assigned to chemotherapy between the Asia subgroup and the global population should be met crossover criteria and were treated with pembrolizumab. A total viewed with caution given the small number of patients in the Asia of four additional patients (15%) received anti-­PD-­1/L1 therapies subgroup. As we observed in the global population (29% vs. 12%), outside the study, for an effective crossover rate of 54%. there was a trend toward a greater number of patients in the pem- In the Asian subgroup, the ORR was similar between the pem- brolizumab versus chemotherapy arm who experienced PD as best brolizumab and chemotherapy treatment arms (50% vs. 46%, re- response (36% vs. 15%).14 In the Asia subgroup, pembrolizumab spectively), whereas the ORR was improved with pembrolizumab treatment led to longer-­lasting responses than chemotherapy, 1034 | YOSHINO et al. TA B L E 5 Immune-­mediated adverse events and infusion pembrolizumab in the Asia subgroup had an ECOG performance reactions of any grade that occurred in at least one patient in any treatment arm in the Asia subgroupa status of 0 compared with the global population (59% vs. 49%). Similarly, a higher proportion of patients receiving chemotherapy in Pembrolizumab, Chemotherapy, the Asia subgroup had left-­sided tumors, which are generally asso- Immune-­mediated n = 22 n = 25 adverse event n (%) n (%) ciated with a better prognosis, compared with the global population (38% vs. 27%). However, given the small number of patients in the Any 6 (27) 10 (40) Asian subgroup, these data should be interpreted with caution. Adrenal insufficiency 1 (5) 0 The results from the global population of KEYNOTE-­177 have Colitis 0 1 (4) led to a paradigm shift in the treatment of patients with MSI-­H/ Hyperthyroidism 2 (9) 0 dMMR mCRC, and pembrolizumab is now approved by the Japan Hypothyroidism 3 (14) 2 (8) Pharmaceuticals and Medical Devices Agency as monotherapy for Infusion reactions 1 (5) 5 (20) the treatment of patients with unresectable, advanced, or recurrent Pneumonitis 1 (5) 2 (8) MSI-­H CRC, by the US Food and Drug Administration for the treat- a Immune-­mediated adverse events and infusion reactions were derived ment of unresectable or metastatic MSI-­H/dMMR CRC, and by the from a list of terms specified by the sponsor, regardless of attribution to European Medicines Agency for the first-­line treatment of meta- any trial treatment by investigators. All events were reported. static MSI-­H/dMMR CRC.19–­21 The current analysis is the only report of the efficacy and safety of a first-­line checkpoint inhibitor in Asian which was consistent with the results of the global population and patients with MSI-­H/dMMR mCRC. The results show that first-­line durable responses documented with pembrolizumab across sev- pembrolizumab provides longer PFS, more durable responses, and eral tumor types.14,16 improved safety compared with chemotherapy, supporting the use The safety profile of pembrolizumab was favorable compared of pembrolizumab monotherapy for MSI-­H/dMMR mCRC in Asian with chemotherapy in the Asia subgroup, as was observed in the patients. global population. The difference in the incidence of grade 3–­5 TRAEs was particularly notable (9% with pembrolizumab vs. 80% AU T H O R C O N T R I B U T I O N S with chemotherapy). The most common immune-­mediated AEs Conception, design, or planning of the study: T.Y., B.V.J., C.D.L.F., in the pembrolizumab arm were hypothyroidism (14%) and hyper- D.T.L., P.M., and L.A.D. Acquisition of the data: T.Y., T.A., W.P.Y., thyroidism (9%), which are well-­documented AEs associated with K.K.S., B.V.J., L.H.J., D.S., R.G.C., J.A.G., C.D.L.F., F.R., E.E., and D.T.L. checkpoint inhibitors.17 Infusion reactions were less common with Analysis of the data: T.Y., T.A., B.V.J., D.S., C.D.L.F., D.T.L., N.A., D.F., pembrolizumab (5%) compared with chemotherapy (20%). and L.A.D. Interpretation of the results: T.Y., T.A., T.W.K., K.K.S., To date, limited data are available regarding the efficacy B.V.J., C.J.A.P., R.G.C., J.A.G., P.G., C.D.L.F., F.R., E.E., D.T.L., N.A., and safety of ICIs among patients of Asian descent with MSI-­H/ D.F., and L.A.D. Drafting of the manuscript: T.Y., T.A., B.V.J., J.A.G., dMMR mCRC. A subgroup analysis of Japanese patients in the C.D.L.F., and D.T.L. Critically reviewing or revising the manuscript KEYNOTE-­164 study showed that pembrolizumab provided durable for important intellectual content: T.Y., T.A., T.W.K., W.P.Y., K.K.S., antitumor activity and an acceptable safety profile in patients with B.V.J., L.H.J., C.J.A.P., D.S., R.G.C., J.A.G., P.G., C.D.L.F., F.R., E.E., previously treated MSI-­H/dMMR advanced CRC.18 The ORR among D.T.L., N.A., D.F., P.M., and L.A.D. Japanese patients was 29% (2/7) for patients who had received ≤2 prior lines of standard therapy and 67% (4/6) for patients who had re- AC K N OW L E D G M E N T S ceived ≤1 prior line of therapy. Median DOR was NR in either group. This work was funded by Merck Sharp & Dohme LLC, a subsidiary of It was also reported that TRAEs tended to occur more frequently Merck & Co., Inc., Rahway, NJ, USA. We thank the patients and their in the Japanese subgroup compared with the global population. No families and caregivers for participating in this trial, and all investi- data are available for nivolumab because only one patient of Asian gators and site personnel. Medical writing and editorial assistance descent was included in the CheckMate 142 study that investigated were provided by Jemimah Walker, PhD, Mehak Aggarwal, PharmD, nivolumab in patients with previously treated MSI-­H/dMMR CRC.11 and Doyel Mitra, PhD, CMPP, of ApotheCom (Yardley, PA, USA). This The current subgroup analysis is limited by the post hoc nature assistance was funded by Merck Sharp & Dohme LLC, a subsidiary of of the analysis and the lack of statistical power to show differences Merck & Co., Inc., Rahway, NJ, USA. in outcome between the treatment arms among patients enrolled from Asia. The number of patients in each arm was also small, which F U N D I N G I N FO R M AT I O N limits definitive conclusions. The high crossover rate seen in the The study was designed under the responsibility of Merck Sharp chemotherapy arm may have contributed to the antitumor activity & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, seen in the chemotherapy arm in the survival analyses, as in the in conjunction with the steering committee. The study was funded case of the global population. Differences in baseline characteris- by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., tics between the Asia subgroup and the global population may also Rahway, NJ, USA. Pembrolizumab was provided by Merck Sharp & have influenced results. A higher proportion of patients receiving Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. All YOSHINO et al. | 1035 Se'er, Delfi, Kinnate, and Neophore; has a family member who holds authors had full access to all of the data in the study and had final equity in Amgen; holds a patent; and is a member of the board of di- responsibility for the decision to submit for publication. rectors of Jounce Therapeutics and Epitope. C O N FL I C T O F I N T E R E S T B.V.J., C.J.A.P., D.S., J.A.-­G., L.H.J., P.G., T.W.K., and W.P.Y have no E T H I C A L A P P O R OVA L conflicts of interest. T.Y. has received lecture fees and/or honoraria The KEYNOTE-­177 trial is registered on Clini​calTr​ials.gov from Taiho Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., (NCT02563002). The study protocol and all amendments were re- Eli Lilly Japan K.K., Merck Biopharma Co., Ltd., Bayer Yakuhin, Ltd., viewed and approved by the appropriate institutional review board Ono Pharmaceutical Co., Ltd., and MSD K.K.; and has also received re- or ethics committee at each study center, and the study was con- search funds from Ono Pharmaceutical Co., Ltd., Sanofi K.K., Daiichi ducted in accordance with the ethical principles of the Declaration Sankyo Co., Ltd., PAREXEL International Inc., Pfizer Japan Inc., Taiho of Helsinki. Written informed consent was obtained from all patients Pharmaceutical Co., Ltd., MSD K.K., Amgen K.K., Genomedia Inc., prior to enrollment. Sysmex Corporation, Chugai Pharmaceutical Co., Ltd., and Nippon Boehringer Ingelheim Co., Ltd. T.A. reports attending advisory board ORCID meetings and receiving consulting fees from AstraZeneca, Astellas, Takayuki Yoshino https://orcid.org/0000-0002-0489-4756 Bristol Myers Squibb, Gritstone Oncology, GamaMabs Pharma Sa, GlaxoSmithKline, Merck & Co. Inc., Nordic Oncology, Pierre Fabre, REFERENCES Seagen, Servier and Transgène; honoraria from AstraZeneca, Bristol 1. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: glo- Myers Squibb, GlaxoSmithKline, Merck & Co. Inc., Pierre Fabre, bocan estimates of incidence and mortality worldwide for 36 can- Roche/Ventana, Sanofi, Seagen and Servier; and travel expenses of cers in 185 countries. CA Cancer J Clin. 2021;71:209-­249. 2. Rawla P, Sunkara T, Barsouk A. Epidemiology of colorectal cancer: US$500 or more from Bristol Myers Squibb and Merck & Co and incidence, mortality, survival, and risk factors. Prz Gastroenterol. Servier. K.K.S. has received lecture fees and/or honoraria from Daiichi 2019;14:89-­103. Sankyo, Merck KGaA, MSD, Mirati Therapeutics, Roche, and Servier; 3. Wang J, Li S, Liu Y, Zhang C, Li H, Lai B. Metastatic patterns and sur- and has received research funds from MSD, Roche, AstraZeneca, and vival outcomes in patients with stage IV colon cancer: a population-­ based analysis. Cancer Med. 2020;9:361-­373. Adaptimmune Therapeutics. R.G.C. has received annual remunera- 4. The global, regional, and national burden of colorectal cancer and tion of US$500 or more from AAA, Ipsen, BMS, MSD, Novartis, Ph its attributable risk factors in 195 countries and territories, 1990-­ MAR, Merck, and Roche; and a family member has received annual 2017: a systematic analysis for the global burden of disease study remuneration of US$500 or more from Amgen, AstraZeneca, Bayer, 2017. Lancet Gastroenterol Hepatol. 2019;4:913-­933. BMS, Glaxo, Janssen, Merck, Novartis, Pfizer, Roche, Sanofi, Takeda, 5. Onyoh EF, Hsu WF, Chang LC, Lee YC, Wu MS, Chiu HM. The rise of colorectal cancer in Asia: epidemiology, screening, and manage- and US$10,000 or more from MSD, Lilly, and Genomics. C.D.L.F. has ment. Curr Gastroenterol Rep. 2019;21:36. received lecture fees, honoraria, or other fees from MSD and BMS 6. Yoshino T, Arnold D, Taniguchi H, et al. Pan-­A sian adapted ESMO and has received research funds from MSD. F.R. has received lecture consensus guidelines for the management of patients with met- fees and/or honoraria from MSD, BMS, Roche, Lilly, Servier, Sanofi, astatic colorectal cancer: a JSMO-­ESMO initiative endorsed by CSCO, KACO, MOS, SSO and TOS. Ann Oncol. 2018;29:44-­70. Astra-­Zeneca; research funds from MSD; and annual remuneration 7. Van Cutsem E, Cervantes A, Nordlinger B, Arnold D. Metastatic of US$500 or more from MSD. E.E. has received personal financial colorectal cancer: ESMO clinical practice guidelines for diagnosis, interest for consulting/advisory role and/or honoraria, travel grants treatment and follow-­up. Ann Oncol. 2014;25(Suppl 3):iii1-­iii9. and research grants from Amgen, Bayer, Hoffman-­La Roche, Merck 8. Le DT, Uram JN, Wang H, et al. PD-­1 blockade in tumors with mismatch-­repair deficiency. N Engl J Med. 2015;372:2509-­2520. Serono, Sanofi, Pierre Fabre, MSD, Organon, Novartis, and Servier. 9. Le DT, Durham JN, Smith KN, et al. Mismatch repair deficiency She has received institutional financial interest in the form of finan- predicts response of solid tumors to PD-­1 blockade. Science. cial support for clinical trials or contracted research from Amgen Inc, 2017;357:409-­413. Array Biopharma Inc, AstraZeneca Pharmaceuticals LP, BeiGene, 10. Le DT, Kim TW, Van Cutsem E, et al. Phase II open-­label study of pembrolizumab in treatment-­refractory, microsatellite instability-­ Boehringer Ingelheim, Bristol Myers Squibb, Celgene, Debiopharm high/mismatch repair-­deficient metastatic colorectal cancer: International SA, F. Hoffmann-­La Roche Ltd., Genentech Inc, KEYNOTE-­164. J Clin Oncol. 2020;38:11-­19. HalioDX SAS, Hutchison MediPharma International, Janssen-­Cilag 11. Overman MJ, McDermott R, Leach JL, et al. Nivolumab in patients SA, MedImmune, Menarini, Merck Health KGAA, Merck Sharp & with metastatic DNA mismatch repair-­deficient or microsatellite instability-­high colorectal cancer (CheckMate 142): an open-­label, Dohme, Merus NV, Mirati, Novartis Farmacéutica SA, Pfizer, Pharma multicentre, phase 2 study. Lancet Oncol. 2017;18:1182-­1191. Mar, Sanofi, Aventis Recherche & Développement, Servier, and Taiho 12. Sahin IH, Akce M, Alese O, et al. Immune checkpoint inhibitors for Pharma USA Inc. D.T.L. has received research funding to run clinical the treatment of MSI-­H/MMR-­D colorectal cancer and a perspec- trials. D.F., N.A., and P.M. are employees of Merck Sharp & Dohme tive on resistance mechanisms. Br J Cancer. 2019;121:809-­818. LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. D.F. and P.M. 13. André T, Shiu KK, Kim TW, et al. Pembrolizumab in microsatellite-­ instability-­high advanced colorectal cancer. N Engl J Med. are stockholders of Merck & Co., Inc., Rahway, NJ, USA. L.A.D. has 2020;383:2207-­2218. served as a consultant for PetDx, Innovatus CP, Se′er, Delfi, Kinnate, 14. Diaz LA, Shiu K-­K , Kim T-­W, et al. Pembrolizumab versus che- and Neophore; holds equity in Epitope, Jounce Therapeutics, PetDx, motherapy for microsatellite instability-­high or mismatch 1036 | YOSHINO et al. repair-­deficient metastatic colorectal cancer (KEYNOTE-­177): final merck​s -­k ey tr​u da-­p embr​o lizu​m ab-­r ecei​v es-­t wo-­n ew-­a ppro​ analysis of a randomised, open-­label, phase 3 study. Lancet Oncol. vals-­in-­japan/ 2022;23:659-­670. 20. Keytruda [package insert]. Merck Sharp & Dohme LLC, Rahway, NJ, 15. Andre T, Amonkar M, Norquist JM, et al. Health-­related quality USA; 2022. of life in patients with microsatellite instability-­high or mismatch 21. KEYTRUDA 25 mg/mL concentrate for solution for infusion. (SPC). repair deficient metastatic colorectal cancer treated with first-­line 2022. MSD B.V.: The Netherlands, 2022; 116. pembrolizumab versus chemotherapy (KEYNOTE-­177): an open-­ label, randomised, phase 3 trial. Lancet Oncol. 2021;22:665-­677. 16. Pons-­Tostivint E, Latouche A, Vaflard P, et al. Comparative analy- S U P P O R T I N G I N FO R M AT I O N sis of durable responses on immune checkpoint inhibitors versus Additional supporting information can be found online in the other systemic therapies: a pooled analysis of phase III trials. JCO Supporting Information section at the end of this article. Precision Oncologia. 2019;3:1-­10. 17. Barroso-­Sousa R, Barry WT, Garrido-­C astro AC, et al. Incidence of endocrine dysfunction following the use of different immune checkpoint inhibitor regimens: a systematic review and meta-­ analysis. JAMA Oncol. 2018;4:173-­182. How to cite this article: Yoshino T, Andre T, Kim TW, et al. 18. Taniguchi H, Hara H, Yoshino T, et al. Pembrolizumab (pembro) Pembrolizumab in Asian patients with microsatellite-­ in mismatch repair-­deficient (dMMR) advanced colorectal cancer instability-­high/mismatch-­repair-­deficient colorectal cancer. (CRC): KEYNOTE-­164 Japan subgroup. Ann Oncol. 2019;30(suppl 6):vi85. Cancer Sci. 2023;114:1026-1036. doi:10.1111/cas.15650 19. Merck's KEYTRUDA® (Pembrolizumab) Receives Two New Approvals in Japan. Merck & Co., Inc; 2021 https://www.merck.com/news/