Pembrolizumab plus chemotherapy as first-line therapy for advanced HER2-negative gastric or gastroesophageal junction adenocarcinoma: 4.5-year follow-up from the randomized phase III KEYNOTE-859 study Authors: SY Rha, LS Wyrwicz, P Yañez, Y Bai, M-H Ryu, J Lee, F Rivera, G Vasconcelos Alves, M Garrido, K-K Shiu, M González Fernández, J Li, MA Lowery, T Çil, F Melo Cruz, D-Y Oh, A Wang, P Leconte, S Qin Published: 2026 ABSTRACT AbstractBackgroundKEYNOTE-859 showed a favorable benefit–risk profile for pembrolizumab plus chemotherapy compared with placebo plus chemotherapy, regardless of programmed death-ligand 1 (PD-L1) status, in participants with untreated locally advanced or metastatic human epidermal growth factor receptor 2 (HER2)-negative gastric or gastroesophageal junction (GEJ) adenocarcinoma. Outcomes after a median study follow-up of 4.5 years are reported.Patients and methodsOverall, 1579 participants were randomly assigned 1 : 1 to pembrolizumab 200 mg or placebo plus chemotherapy every 3 weeks for ≤35 cycles. The primary endpoint was overall survival (OS). Secondary endpoints included progression-free survival, objective response rate, and duration of response, all per RECIST v1.1 by blinded independent central review, and safety.ResultsThe median study follow-up was 54.8 months (Q1-Q3, 46.8-62.1) at data cut-off (27 September 2024). In the intention-to-treat population, the median OS was 12.9 months for pembrolizumab plus chemotherapy compared with 11.5 months for placebo plus chemotherapy [hazard ratio (HR) 0.78, 95% confidence interval (CI) 0.70-0.86]. Median OS was longer in the PD-L1 combined positive score (CPS) ≥1 (13.0 compared with 11.4 months; HR 0.74, 95% CI 0.66-0.84) and CPS ≥10 (15.8 compared with 11.8 months; HR 0.64, 95% CI 0.53-0.77) populations. Grade 3 or 4 treatment-related adverse events (AEs) occurred in 458 participants (58.3%) receiving pembrolizumab plus chemotherapy and 388 (49.3%) receiving placebo plus chemotherapy; grade 5 treatment-related AEs occurred in 8 participants (1.0%) and 16 participants (2.0%), respectively.ConclusionsExtended follow-up confirms that first-line pembrolizumab plus chemotherapy improves efficacy and maintains a manageable safety profile compared with placebo plus chemotherapy, regardless of PD-L1 status, supporting this combination as a first-line treatment option for locally advanced or metastatic HER2-negative gastric or GEJ adenocarcinoma.Key words: gastric cancer, gastroesophageal adenocarcinoma, HER2-negative, pembrolizumab Introduction In the first-line setting, management of human epidermal growth factor receptor 2 (HER2)-negative, locally advanced or metastatic gastric or gastroesophageal junction (GEJ) adenocarcinoma involves fluoropyrimidine (FP; fluorouracil or capecitabine) and platinum-based (oxaliplatin or cisplatin) chemotherapy, regardless of programmed death-ligand 1 (PD-L1) expression. In the phase III KEYNOTE-859 study (NCT03675737), pembrolizumab combined with chemotherapy was assessed against placebo plus chemotherapy as a first-line therapeutic approach for participants with HER2-negative gastric or GEJ adenocarcinoma.1 The interim analysis revealed clinically meaningful improvements in overall survival [OS; hazard ratio (HR) 0.78, 95% confidence interval (CI) 0.70-0.87, P < 0.0001], progression-free survival (PFS; HR 0.76, 95% CI 0.67-0.85, P < 0.0001), and objective response rate (ORR; 51.3% compared with 42.0%; difference 9.3%, 95% CI 4.4% to 14.1%, P < 0.0001) compared with pembrolizumab plus chemotherapy.1 The efficacy outcomes for participants with a PD-L1 combined positive score (CPS) ≥1 or CPS ≥10 were consistent with those of the overall population.1 Importantly, the safety profile remained consistent, with similar rates of grade 3-5 treatment-related adverse events (AEs) observed across both treatment groups (59.4% compared with 51.1%).1 For patients with a PD-L1 CPS ≥1, clinical guidelines recommend the addition of agents targeting programmed cell death protein 1 (PD-1), such as nivolumab, pembrolizumab, or tislelizumab.2, 3, 4 In this article, we report efficacy and safety outcomes after a median study follow-up of 4.5 years. Methods Study design and participants The study design of KEYNOTE-859 has been previously described.1 Briefly, eligible participants were aged ≥18 years with previously untreated locally advanced unresectable or metastatic HER2-negative gastric or GEJ adenocarcinoma. Additional inclusion criteria included measurable disease per RECIST v1.1 by investigator, known PD-L1 expression status, and an Eastern Cooperative Oncology Group performance status of 0 or 1. The study was conducted in accordance with principles of Good Clinical Practice and was approved by the appropriate institutional review boards and regulatory agencies. All participants provided informed consent. Randomization and treatment Participants were randomly assigned 1 : 1 to receive pembrolizumab 200 mg or placebo, administered intravenously (i.v.) every 3 weeks (Q3W) for ≤35 cycles, in combination with investigator’s choice of chemotherapy. Chemotherapy regimens included either cisplatin (80 mg/m2 i.v. on day 1 Q3W) and 5-fluorouracil (800 mg/m2/day i.v. continuous on days 1-5 Q3W; FP regimen), or capecitabine (1000 mg/m2 by mouth twice daily on days 1-14 Q3W) and oxaliplatin (130 mg/m2 i.v. on day 1 Q3W; CAPOX regimen). The choice of chemotherapy backbone was determined by the investigator before patients were randomly assigned to treatment. Randomization was stratified by geographic region (Australia/Europe/Israel/North America, Asia, or rest of the world), PD-L1 tumor expression status (CPS <1 or ≥1), and chemotherapy regimen (FP or CAPOX). In accordance with local guidelines, cisplatin and oxaliplatin could be limited to a maximum of six cycles. Treatment continued until disease progression, investigator decision, completion of the maximum number of cycles, unacceptable toxicity, withdrawal of consent, or other reasons. Assessments A computed tomography scan (preferred) or contrast-enhanced magnetic resonance imaging of the chest, abdomen, and pelvis was conducted within 4 weeks before before patients were randomly assigned to treatment, 6 weeks after treatment assignment, and subsequently every 6 weeks until confirmed disease progression [as per RECIST v1.1 by blinded independent central review (BICR)], initiation of a new anticancer therapy, or withdrawal of consent. After treatment discontinuation, survival was monitored every 12 weeks by telephone or scheduled visits until death, consent withdrawal, or study completion. AEs were graded per the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0. Treatment-related AEs were determined by the investigator to be related to study treatment. Immune-mediated AEs and infusion reactions were reported based on a list of terms specified by the sponsor intended to capture known risks of pembrolizumab and were considered regardless of attribution to study treatment by the investigator. Outcomes The primary endpoint was OS (time from randomization to death from any cause). The secondary endpoints included PFS (time from randomization to first documented onset of progressive disease or death from any cause, whichever occurred first), ORR (proportion of participants with a complete response or partial response), and duration of response (DOR; time from first complete response or partial response until progressive disease or death from any cause, whichever occurred first), all per RECIST v1.1 by BICR, and safety. Statistical analysis Efficacy endpoints were evaluated in the intention-to-treat (ITT) population, which included all randomly assigned participants and participants with PD-L1 CPS ≥1 and CPS ≥10. Safety was assessed in all randomly assigned participants who received ≥1 dose of study treatment. OS, PFS, and DOR were estimated using the nonparametric Kaplan–Meier method, with between-group differences in OS and PFS assessed using a stratified log-rank test. The HRs and associated 95% CIs for OS and PFS were assessed using a Cox proportional hazards model with the Efron method of handling ties. The difference in ORR (and corresponding 95% CI) between the two treatment groups was determined using the Miettinen and Nurminen method. The data cut-off date was 27 September 2024. Results Participants From 8 November 2018 to 11 June 2021, 1579 participants were enrolled and randomly assigned (n = 790 in the pembrolizumab plus chemotherapy group and n = 789 in the placebo plus chemotherapy group; Supplementary Figure S1, available at https://doi.org/10.1016/j.esmoop.2026.108536). The median time from randomization to the data cut-off date was 54.8 months (Q1-Q3, 46.8-62.1). Baseline demographics and clinical characteristics were generally consistent between the treatment groups as well as between the ITT population and participants with PD-L1 CPS ≥1 or CPS ≥10 (Table 1). A total of 384 participants (48.6%) in the pembrolizumab plus chemotherapy group and 394 (49.9%) in the placebo plus chemotherapy group received any subsequent systemic anticancer therapy (Supplementary Table S1, available at https://doi.org/10.1016/j.esmoop.2026.108536). Table 1Baseline demographics and clinical characteristicsBaseline characteristicPembrolizumab + chemotherapyPlacebo + chemotherapyITT n = 790PD-L1 CPS ≥1 n = 618PD-L1 CPS ≥10 n = 280ITT n = 789PD-L1 CPS ≥1 n = 617PD-L1 CPS ≥10 n = 273Age, years, median (range)61 (23-86)62 (24-86)63 (26-84)62 (21-85)63 (25-85)63 (25-82)Age, ≥65 years304 (38.5)241 (39.0)118 (42.1)310 (39.3)253 (41.0)113 (41.4)Male sex527 (66.7)422 (68.3)193 (68.9)544 (68.9)448 (72.6)206 (75.5)ECOG PS 0281 (35.6)223 (36.1)99 (35.4)301 (38.1)228 (37.0)104 (38.1)Primary tumor locationa GEJ148 (18.7)123 (19.9)65 (23.2)185 (23.4)164 (26.6)74 (27.1) Stomach641 (81.1)494 (79.9)215 (76.8)603 (76.4)453 (73.4)199 (72.9)Disease statusb Locally advanced28 (3.5)26 (4.2)14 (5.0)30 (3.8)24 (3.9)11 (4.0) Metastatic761 (96.3)591 (95.6)266 (95.0)759 (96.2)593 (96.1)262 (96.0)MSI statusc MSI-high39 (4.9)35 (5.7)20 (7.1)35 (4.4)31 (5.0)16 (5.9) Non-MSI-high641 (81.1)503 (81.4)228 (81.4)639 (81.0)500 (81.0)225 (82.4) Missing110 (13.9)80 (12.9)32 (11.4)114 (14.4)85 (13.8)31 (11.4)Liver metastasesd314 (39.7)258 (41.7)119 (42.5)311 (39.4)253 (41.0)110 (40.3)Prior gastrectomy or esophagectomye172 (21.8)109 (17.6)48 (17.1)162 (20.5)105 (17.0)41 (15.0)Histological subtypef Diffuse318 (40.3)236 (38.2)103 (36.8)301 (38.1)220 (35.7)90 (33.0) Intestinal284 (35.9)239 (38.7)111 (39.6)273 (34.6)215 (34.8)99 (36.3) Indeterminate186 (23.5)141 (22.8)65 (23.2)215 (27.2)182 (29.5)84 (30.8)Choice of chemotherapy at randomization CAPOX682 (86.3)528 (85.4)243 (86.8)681 (86.3)528 (85.6)236 (86.4) FP108 (13.7)90 (14.6)37 (13.2)108 (13.7)89 (14.4)37 (13.6)Data are n (%) unless otherwise specified.CAPOX, capecitabine–oxaliplatin; CPS, combined positive score; ECOG PS, Eastern Cooperative Oncology Group performance status; FP, fluoropyrimidine; GEJ, gastroesophageal junction; ITT, intention-to-treat; MSI, microsatellite instability; PD-L1, programmed death-ligand 1.aMissing for one participant in the pembrolizumab plus chemotherapy group and designated ‘other’ for one participant in the placebo plus chemotherapy group.bMissing for one participant in the pembrolizumab plus chemotherapy group.cUnknown for one participant in the placebo plus chemotherapy group.dMissing for one participant in the pembrolizumab plus chemotherapy group.eMissing for four participants in the pembrolizumab plus chemotherapy group and five in the placebo plus chemotherapy group.fMissing and unknown for one participant each in the pembrolizumab plus chemotherapy group. Baseline demographics and clinical characteristics Data are n (%) unless otherwise specified. CAPOX, capecitabine–oxaliplatin; CPS, combined positive score; ECOG PS, Eastern Cooperative Oncology Group performance status; FP, fluoropyrimidine; GEJ, gastroesophageal junction; ITT, intention-to-treat; MSI, microsatellite instability; PD-L1, programmed death-ligand 1. Missing for one participant in the pembrolizumab plus chemotherapy group and designated ‘other’ for one participant in the placebo plus chemotherapy group. Missing for one participant in the pembrolizumab plus chemotherapy group. Unknown for one participant in the placebo plus chemotherapy group. Missing for one participant in the pembrolizumab plus chemotherapy group. Missing for four participants in the pembrolizumab plus chemotherapy group and five in the placebo plus chemotherapy group. Missing and unknown for one participant each in the pembrolizumab plus chemotherapy group. At the data cut-off date, 1572 participants had received at least one dose of study treatment (n = 785 in the pembrolizumab plus chemotherapy group and n = 787 in the placebo plus chemotherapy group; Supplementary Figure S1, available at https://doi.org/10.1016/j.esmoop.2026.108536). Of the 785 participants in the pembrolizumab plus chemotherapy group, 83 (10.6%) had completed treatment, none remained on treatment, and 703 (89.6%) had discontinued treatment, most commonly due to progressive disease (n = 436; 62.0%). Of the 787 participants in the placebo plus chemotherapy group, 36 (4.6%) had completed treatment, 1 (0.1%) remained on treatment, and 750 (95.3%) had discontinued treatment, most commonly due to progressive disease (n = 492; 65.6%). Efficacy In the ITT population, 1417 of 1579 participants died [688 of 790 (87.1%) in the pembrolizumab plus chemotherapy group and 729 of 789 (92.4%) in the placebo plus chemotherapy group]. The median OS was 12.9 months (95% CI 11.9-14.0) in the pembrolizumab plus chemotherapy group compared with 11.5 months (95% CI 10.6-12.1) in the placebo plus chemotherapy group (HR 0.78, 95% CI 0.70-0.86; Figure 1A). Estimated 24-month OS rates were 28% compared with 19%, 36-month OS rates were 19% compared with 10%, and 48-month OS rates were 14% compared with 8%, respectively. Disease progression or death was experienced by a total of 1199 of 1579 participants [584 of 790 (73.9%) in the pembrolizumab plus chemotherapy group and 615 of 789 (77.9%) in the placebo plus chemotherapy group]. The median PFS was 6.9 months (95% CI 6.3-7.2) in the pembrolizumab plus chemotherapy group compared with 5.6 months (95% CI 5.5-5.7) in the placebo plus chemotherapy group (HR 0.76, 95% CI 0.68-0.85; Figure 2A). Estimated 24-month PFS rates were 17% and 9%, respectively. The ORR was 51.1% (95% CI 47.6% to 54.7%) in the pembrolizumab plus chemotherapy group compared with 42.0% (95% CI 38.5% to 45.5%) in the placebo plus chemotherapy group, and the median DOR was 8.0 months (range 1.2+ to 66.3+) compared with 5.7 months (range 1.3+ to 58.1+), respectively (Figure 3A and Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2026.108536). Kaplan–Meier estimates of OS in the (A) ITT, (B) PD-L1 CPS ≥1, and (C) PD-L1 CPS ≥10 populations. CPS, combined positive score; HR, hazard ratio; ITT, intention-to-treat; OS, overall survival; PD-L1, programmed death-ligand 1. Kaplan–Meier estimates of PFS per RECIST v1.1 by BICR in the (A) ITT, (B) PD-L1 CPS ≥1, and (C) PD-L1 CPS ≥10 populations. BICR, blinded independent central review; CPS, combined positive score; HR, hazard ratio; ITT, intention-to-treat; PD-L1, programmed death-ligand 1; PFS, progression-free survival. Kaplan–Meier estimates of DOR per RECIST v1.1 by BICR in the (A) ITT, (B) PD-L1 CPS ≥1, and (C) PD-L1 CPS ≥10 populations. ‘+’ indicates there was no PD at the time of last disease assessment. BICR, blinded independent central review; CPS, combined positive score; DOR, duration of response; ITT, intention-to-treat; PD, progressive disease; PD-L1, programmed death-ligand 1. In the PD-L1 CPS ≥1 population, 1098 of 1235 participants died [529 of 618 (85.6%) in the pembrolizumab plus chemotherapy group and 569 of 617 (92.2%) in the placebo plus chemotherapy group]. The median OS was 13.0 months (95% CI 11.6-14.2) in the pembrolizumab plus chemotherapy group compared with 11.4 months (95% CI 10.5-12.0) in the placebo plus chemotherapy group (HR 0.74, 95% CI 0.66-0.84; Figure 1B). Estimated 24-month OS rates were 29% compared with 18%, 36-month OS rates were 20% compared with 10%, and 48-month OS rates were 16% compared with 8%, respectively. Disease progression or death was experienced by a total of 939 of 1235 participants [452 of 618 (73.1%) in the pembrolizumab plus chemotherapy group and 487 of 617 (78.9%) in the placebo plus chemotherapy group]. The median PFS was 6.9 months (95% CI 6.0-7.2) in the pembrolizumab plus chemotherapy group compared with 5.6 months (95% CI 5.4-5.7) in the placebo plus chemotherapy group (HR 0.72, 95% CI 0.64-0.82; Figure 2B). Estimated 24-month PFS rates were 19% and 8%, respectively. The ORR was 51.9% (95% CI 47.9% to 55.9%) in the pembrolizumab plus chemotherapy group compared with 42.6% (95% CI 38.7% to 46.6%) in the placebo plus chemotherapy group, and the median DOR was 8.3 months (range 1.2+ to 66.3+) compared with 5.6 months (range >1.3+ to 58.1+), respectively (Figure 3B and Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2026.108536). In the PD-L1 CPS ≥10 population, 465 of 553 participants died [218 of 280 (77.9%) in the pembrolizumab plus chemotherapy group and 247 of 273 (90.5%) in the placebo plus chemotherapy group]. The median OS was 15.8 months (95% CI 14.0-19.3) in the pembrolizumab plus chemotherapy group compared with 11.8 months (95% CI 10.3-12.7) in the placebo plus chemotherapy group (HR 0.64, 95% CI 0.53-0.77; Figure 1C). Estimated 24-month OS rates were 38% compared with 22%, 36-month OS rates were 30% compared with 13%, and 48-month OS rates were 26% compared with 10%, respectively. Disease progression or death was experienced by a total of 409 of 553 participants [196 of 280 (70.0%) in the pembrolizumab plus chemotherapy group and 213 of 273 (78.0%) in the placebo plus chemotherapy group]. The median PFS was 7.8 months (95% CI 6.8-8.5) in the pembrolizumab plus chemotherapy group compared with 5.6 months (95% CI 5.4-6.7) in the placebo plus chemotherapy group (HR 0.62, 95% CI 0.51-0.76; Figure 2C). Estimated 24-month PFS rates were 25% and 8%, respectively. The ORR was 60.4% (95% CI 54.4% to 66.1%) in the pembrolizumab plus chemotherapy group compared with 43.2% (95% CI 37.3% to 49.3%) in the placebo plus chemotherapy group, and the median DOR was 10.0 months (range 1.2+ to 66.3+) compared with 5.7 months (range 1.4+ to 55.0+), respectively (Figure 3C and Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2026.108536). Safety All-cause AEs of any grade were reported in 776 participants (98.9%) in the pembrolizumab plus chemotherapy group compared with 771 (98.0%) in the placebo plus chemotherapy group (Supplementary Table S3, available at https://doi.org/10.1016/j.esmoop.2026.108536). Grade 3-5 all-cause AEs occurred in 595 participants (75.8%) compared with 550 (69.9%), respectively. Discontinuation of study treatment due to an all-cause AE was reported in 264 participants (33.6%) in the pembrolizumab plus chemotherapy group compared with 206 (26.2%) in the placebo plus chemotherapy group. Deaths due to all-cause AEs occurred in 65 participants (8.3%) compared with 58 (7.4%), respectively. Treatment-related AEs of any grade were reported in 751 participants (95.7%) in the pembrolizumab plus chemotherapy group compared with 736 (93.5%) in the placebo plus chemotherapy group (Table 2). Grade 3-5 treatment-related AEs occurred in 466 participants (59.4%) compared with 404 (51.3%), respectively. Discontinuation of study treatment due to treatment-related AEs was reported in 212 participants (27.0%) in the pembrolizumab plus chemotherapy group compared with 160 (20.3%) in the placebo plus chemotherapy group. Deaths due to treatment-related AEs occurred in eight participants (1.0%) compared with 16 (2.0%), respectively. Table 2Treatment-related AEs summaryTreatment-related AEsPembrolizumab + chemotherapy n = 785Placebo + chemotherapyn = 787Any treatment-related AE751 (95.7)736 (93.5) Grade 3-5466 (59.4)404 (51.3) Led to discontinuation212 (27.0)160 (20.3) Led to death8 (1.0)a16 (2.0)bIncidence ≥15%Any gradeGrade 3-5Any gradeGrade 3-5 Nausea325 (41.4)26 (3.3)326 (41.4)29 (3.7) Diarrhea252 (32.1)46 (5.9)214 (27.2)37 (4.7) Anemia245 (31.2)63 (8.0)212 (26.9)53 (6.7) Vomiting215 (27.4)35 (4.5)175 (22.2)32 (4.1) Platelet count decreased197 (25.1)55 (7.0)177 (22.5)37 (4.7) Neutrophil count decreased195 (24.8)72 (9.2)171 (21.7)58 (7.4) Palmar-plantar erythrodysesthesia syndrome189 (24.1)25 (3.2)167 (21.2)14 (1.8) Decreased appetite170 (21.7)15 (1.9)169 (21.5)14 (1.8) Fatigue158 (20.1)28 (3.6)164 (20.8)32 (4.1) Neuropathy peripheral149 (19.0)10 (1.3)163 (20.7)25 (3.2) Neutropenia142 (18.1)55 (7.0)135 (17.2)60 (7.6) Aspartate aminotransferase increased141 (18.0)11 (1.4)104 (13.2)8 (1.0) Peripheral sensory neuropathy137 (17.5)22 (2.8)131 (16.6)8 (1.0)Data are n (%).AE, adverse event.aOne each death, diarrhea, peripheral embolism, pneumonitis, pulmonary hemorrhage, sepsis, septic shock, and thrombotic thrombocytopenic purpura.bThree septic shock; two acute myocardial infarction; one each cerebrovascular accident, cerebral hemorrhage, diarrhea, gastric perforation, abnormal hepatic function, neurotoxicity, pneumonitis, pulmonary embolism, sepsis, sudden death, and urosepsis. Treatment-related AEs summary Data are n (%). AE, adverse event. One each death, diarrhea, peripheral embolism, pneumonitis, pulmonary hemorrhage, sepsis, septic shock, and thrombotic thrombocytopenic purpura. Three septic shock; two acute myocardial infarction; one each cerebrovascular accident, cerebral hemorrhage, diarrhea, gastric perforation, abnormal hepatic function, neurotoxicity, pneumonitis, pulmonary embolism, sepsis, sudden death, and urosepsis. Immune-mediated AEs and infusion reactions were reported in 251 participants (32.0%) in the pembrolizumab plus chemotherapy group compared with 108 (13.7%) in the placebo plus chemotherapy group (Supplementary Table S4, available at https://doi.org/10.1016/j.esmoop.2026.108536). Grade 3-5 immune-mediated AEs and infusion reactions occurred in 79 participants (10.1%) compared with 16 (2.0%), respectively. Discontinuation of study treatment due to immune-mediated AEs and infusion reactions was reported in 43 participants (5.5%) in the pembrolizumab plus chemotherapy group compared with 14 (1.8%) in the placebo plus chemotherapy group. Deaths due to immune-mediated AEs and infusion reactions occurred in one participant (0.1%) each in both groups. Discussion Gastric cancer remains a leading cause of morbidity and mortality. Despite therapeutic advances, there is a continued need for therapies that offer sustained responses and prolong survival. KEYNOTE-859 showed that pembrolizumab plus chemotherapy significantly improved efficacy outcomes with a manageable safety profile for participants with locally advanced or metastatic HER2-negative gastric or GEJ adenocarcinoma. After approximately 4.5 years of additional follow-up, pembrolizumab plus chemotherapy maintained a favorable benefit–risk profile compared with placebo plus chemotherapy. OS, PFS, and ORR continued to be longer or higher with pembrolizumab plus chemotherapy after a median follow-up of 4.5 years compared with placebo plus chemotherapy in all study populations. The study also reported a prolonged DOR and continued to show a manageable safety profile. The magnitude of treatment benefit varied according to PD-L1 expression, with the greatest OS improvement observed among participants with PD-L1 CPS ≥10 tumors. This finding is consistent with previous studies of PD-1 inhibitors in gastroesophageal cancers, in which higher PD-L1 expression has generally been associated with greater clinical benefit from immunotherapy.5, 6, 7 Notably, the OS Kaplan–Meier curves showed sustained separation through 48 months, with a modest plateau emerging in the PD-L1 CPS ≥10 population. Pembrolizumab plus chemotherapy also showed favorable OS outcomes across all prespecified PD-L1 populations, including the overall study population (HR 0.78, 95% CI 0.70-0.86). The sustained separation of the OS Kaplan–Meier curves in the overall population further supports a clinically meaningful survival benefit irrespective of PD-L1 expression level. Collectively, these findings suggest that PD-L1 expression may enrich for benefit while not excluding the potential for benefit in broader patient populations. Furthermore, the study was not powered for formal subgroup comparisons, and PD-L1 expression should be interpreted as a potential enrichment factor rather than a strict selection criterion. The KEYNOTE-811 study focused on HER2-positive advanced gastric or GEJ adenocarcinoma, evaluating pembrolizumab in combination with trastuzumab and chemotherapy. The study showed a significant improvement in ORR and durable responses, with subsequent analyses confirming a statistically significant OS benefit, particularly in participants with PD-L1 CPS ≥1 tumors.8, 9, 10 Together, the positive results from KEYNOTE-859 and KEYNOTE-811 demonstrate the consistency of added clinical benefit with pembrolizumab combined with standard-of-care treatment, regardless of HER2 status, in the first-line treatment of advanced or metastatic gastric or GEJ adenocarcinoma. Other PD-1 inhibitors have also been studied in similar populations. The CheckMate 649 study demonstrated that nivolumab plus chemotherapy significantly improved OS and PFS, especially in patients with a PD-L1 CPS ≥5.5, 6, 7 Its large global cohort and consistent benefit across subgroups established nivolumab as a standard first-line option. The ATTRACTION-4 study, conducted in East Asia, also evaluated nivolumab plus chemotherapy.11, 12, 13 Although it showed improvements in PFS and ORR, it did not demonstrate a statistically significant OS benefit. The ORIENT-16 study, conducted in China, evaluated sintilimab plus chemotherapy and demonstrated a significant OS benefit in both the overall population and in patients with a PD-L1 CPS ≥5.14,15 Similarly, the global RATIONALE-305 study assessed tislelizumab plus chemotherapy and reported meaningful improvements in OS and PFS, regardless of PD-L1 expression.16 Together, KEYNOTE-859, CheckMate 649, ORIENT-16, and RATIONALE-305 support the use of PD-1 inhibitors plus chemotherapy in the first-line setting, demonstrating consistent survival benefits and manageable safety profiles. However, direct comparisons between these studies should be approached with caution due to differences in study design, participant demographics, disease characteristics, sample sizes, follow-up durations, chemotherapy regimens, and the assays and techniques used to assess PD-L1 expression. A limitation of the KEYNOTE-859 study is the small sample size in certain subgroups, such as participants with GEJ adenocarcinoma. Furthermore, the availability of postprogression therapy varies globally and may influence long-term OS outcomes. Nevertheless, with a median study follow-up of approximately 4.5 years (54.8 months), pembrolizumab plus chemotherapy continued to show improved OS, PFS, and ORR compared with placebo plus chemotherapy, regardless of PD-L1 status. These findings further support pembrolizumab plus chemotherapy as a first-line treatment option for locally advanced or metastatic HER2-negative gastric or GEJ adenocarcinoma. Acknowledgements The authors thank the participants and their families for their involvement in the study, all investigators, and site personnel. The authors would also like to acknowledge the contributions of the following employees of Merck Sharp & Dohme (MSD) LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA: Viktoria Youreva and Grace Kuo. The authors also thank Pooja Bhagia, a former employee of MSD LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, for her contributions. Medical writing and/or editorial assistance was provided by Bresler Swanepoel, PhD, and Cindi A. Hoover, PhD, of ApotheCom (Yardley, PA, USA). This assistance was funded by MSD LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Data availability Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA (MSD) is committed to providing qualified scientific researchers access to anonymized data and clinical study reports from the company’s clinical trials for the purpose of conducting legitimate scientific research. MSD is also obligated to protect the rights and privacy of trial participants and, as such, has a procedure in place for evaluating and fulfilling requests for sharing company clinical trial data with qualified external scientific researchers. The MSD data-sharing website (available at: https://externaldatasharing-msd.com/) outlines the process and requirements for submitting a data request. Applications will be promptly assessed for completeness and policy compliance. Feasible requests will be reviewed by a committee of MSD subject matter experts to assess the scientific validity of the request and the qualifications of the requestors. In line with data privacy legislation, submitters of approved requests must enter into a standard data-sharing agreement with MSD before data access is granted. Data will be made available for request after product approval in the United States and European Union 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, hypothesis-driven 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 carry out 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 SAS portal so that the requestor can carry out the proposed analyses. Funding This work was supported by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. The funder participated in study design, data collection, data analysis, data interpretation, writing of the report, and the decision to submit the report for publication. All authors had full access to the study data. Disclosure SYR reports advisory/consultancy roles for Amgen, Arcus, Astellas, AstraZeneca, BeOne, Boehringer Ingelheim, Bristol Myers Squibb (BMS), Daiichi Sankyo, Gilead, Indivumed, Jazz Pharmaceuticals, LG Biochemical, Merck Sharp & Dohme (MSD), Ono Pharmaceutical, and Qureator; speakers bureau for Amgen, Astellas, AstraZeneca, BeOne, BMS, Daiichi Sankyo, MSD, and Ono Pharmaceutical; and research funding from Amgen, Astellas, BeOne, Boehringer Ingelheim, BMS, Daiichi Sankyo, Eli Lilly, Gilead, Ipsen, Jazz Pharmaceuticals, MSD, Roche, Sanofi, and Taiho Pharmaceutical. LSW reports advisory/consultancy roles for Agenus, AstraZeneca, GlaxoSmithKline (GSK), and Servier; speakers bureau for BMS, MSD, and Servier; honoraria from BeiGene, BMS, MSD, and Servier; and travel/accommodation/expenses from Amgen and Servier. PY reports advisory/consultancy roles for AstraZeneca, BMS/Medarex, and MSD Oncology; speakers bureau for AstraZeneca and BMS Chile; and honoraria from AstraZeneca, BMS Chile, MSD, and Sanofi/Regeneron. YB reports no conflict of interest. M-HR reports advisory/consultancy roles and honoraria from Astellas, AstraZeneca, BMS, DAEHWA Pharmaceutical, Daiichi Sankyo, Lilly, MSD, Novartis, Ono Pharmaceutical, and Taiho Pharmaceutical and research funding from AstraZeneca, BMS, and Ono Pharmaceutical. JLee reports advisory/consultancy roles for AstraZeneca, Guardant Health AMEA, Mirati Therapeutics, Oncologie, and Trutino Biosciences and research funding from AstraZeneca and MSD. FR reports advisory/consultancy roles and speakers bureau for Amgen, Astellas, BeiGene, BMS, MSD, and Roche; research funding from AstraZeneca and MSD; and travel/accommodation/expenses from MSD and Roche. GVA reports research funding from Amgen, AstraZeneca, BioGene, BMS, Exelixis, Immunocore, Incyte Corporation, Janssen, Merck Serono, MSD, Parexel, Pfizer, Roche, and Samsung. MG reports advisory/consultancy roles and expert testimony for AstraZeneca, BMS, and MSD; honoraria from AstraZeneca, BMS, MSD, and Pfizer; and travel/accommodation/expenses from AstraZeneca, MSD, and Pfizer. K-KS reports honoraria from Bayer, BMS, Merck Serono, MSD Oncology, and Servier; consulting/advisory roles for BMS, GSK, Mirati Therapeutics, MSD Oncology, Nouscom, Roche, Seagen, and Takeda; and research funding from AstraZeneca (Inst), BioNTech SE (Inst), MSD Oncology (Inst), Nouscom (Inst), and Roche (Inst). MGF reports no conflict of interest. JLi reports advisory/consultancy roles for Boehringer Ingelheim. MAL reports advisory/consultancy roles for Agios, Astellas, AstraZeneca, and Servier and research funding from Amgen, Astellas, Daiichi Sankyo, Exelixis, MSD, and Zymeworks. TÇ reports advisory/consultancy role for AstraZeneca and speakers bureau for Deva, Eczacibasi, MSD, Novartis, and Roche. FMC reports advisory/consultancy roles for Astellas, Bayer, and Johnson & Johnson and travel/accommodation/expenses from Astellas, Bayer, and BMS. D-YO reports advisory/consultancy roles for AbbVie, Alligator Biosciences AB, Arcus Biosciences, Aslan, Astellas, AstraZeneca, Basilea, Bayer, BeiGene, BMS/Celgene, Eutilex, Genentech/Roche, Hana Pharm, Halozyme, Hanmi, Idience, IQVIA, J-Pharma, LG Chem, Merck Serono, Mirati Therapeutics, Moderna, MSD, Novartis, Revolution Medicine, Tallac Therapeutics, Taiho Pharmaceutical, Turning Point, Yuhan, and Zymeworks and research funding from Array, AstraZeneca, BeiGene, Eli Lilly, Handok, Hanmi, MSD, Novartis, and Servier. AW and PL report employment, travel/accommodation/expenses from MSD LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and shareholder/stockholder/stock options from Merck & Co., Inc., Rahway, NJ, USA. SQ reports no conflict of interest. Supplementary data Supplementary Tables S1-S4 and Figure S1 References Associated Data Supplementary Materials Supplementary Tables S1-S4 and Figure S1 Data Availability Statement Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA (MSD) is committed to providing qualified scientific researchers access to anonymized data and clinical study reports from the company’s clinical trials for the purpose of conducting legitimate scientific research. MSD is also obligated to protect the rights and privacy of trial participants and, as such, has a procedure in place for evaluating and fulfilling requests for sharing company clinical trial data with qualified external scientific researchers. The MSD data-sharing website (available at: https://externaldatasharing-msd.com/) outlines the process and requirements for submitting a data request. Applications will be promptly assessed for completeness and policy compliance. Feasible requests will be reviewed by a committee of MSD subject matter experts to assess the scientific validity of the request and the qualifications of the requestors. In line with data privacy legislation, submitters of approved requests must enter into a standard data-sharing agreement with MSD before data access is granted. Data will be made available for request after product approval in the United States and European Union 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, hypothesis-driven 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 carry out 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 SAS portal so that the requestor can carry out the proposed analyses.