ABSTRACT: BACKGROUND:Efficacy of immune checkpoint inhibitors (ICIs) in metastatic gastric/gastroesophageal junction (GEJ) adenocarcinoma is inconsistent. Whether the efficacy of ICIs is comparable across different subgroups remains unknown. METHODS:We identified randomized controlled trials (RCTs) that compared standard treatment for metastatic gastric/GEJ adenocarcinoma to ICIs. Hazard ratios (HRs) and 95% confidence intervals (CI) for overall survival (OS) were extracted and pooled in a meta-analysis. Prespecified subgroups were included as follows: age at randomization (65 years), gender (female vs male), ethnicity (Asians vs non-Asians), performance-status (0 vs 1), tumor location (gastric vs GEJ), and histological subtype (diffuse vs others). OS in patients with programmed death ligand (PD-L1) positive and with microsatellite instability-high (MSI-H) were also extracted and pooled in a meta-analysis. RESULTS:Five RCTs comprising 2,264 patients were analyzed. Compared to standard therapy, ICIs did not improve OS (HR = 0.86, 95% CI 0.71-1.03, P = .10) and the effect of ICIs on OS was similar in all subgroups. Nonsignificantly greater effect sizes were seen in older patients (HR = 0.85 vs 0.88, P = .81), male (HR = 0.82 vs 0.99, P = .16), Asians (HR = 0.86 vs 0.96, P = .55), performance-status 0 (HR = 0.84 vs 0.88, P = .81), GEJ tumors (HR = 0.78 vs 0.90, P = .37), and nondiffuse subtype (HR = 0.71 vs 0.79, P = .62). ICIs were associated with significantly improved OS in patients with MSI-H (HR = 0.33, P = .001), but not in PD-L1 positive disease (HR = 0.86, P = .06). CONCLUSIONS:Compared to standard treatment, ICIs in metastatic gastric/GEJ adenocarcinoma did not improve OS. None of the evaluated subgroups has shown increased magnitude of effect to ICIs, aside of the small group with MSI-H tumors.