Apatinib Plus PD-1 Blockade versus Chemotherapy After Prior Immunotherapy in Advanced Gastric or Gastroesophageal Junction Adenocarcinoma: A Retrospective Cohort Study
Abstract
Objective To compare clinical outcomes and safety of apatinib plus programmed death-1 (PD-1) blockade versus single-agent chemotherapy in patients with advanced gastric or gastroesophageal junction (G/GEJ) adenocarcinoma previously exposed to immune checkpoint inhibitors (ICIs). Methods This single-center retrospective cohort included 117 adults with HER2-negative advanced G/GEJ adenocarcinoma treated after discontinuation of an ICI-containing regimen. 59 patients received apatinib plus PD-1 blockade (AP group) and 58 received single-agent chemotherapy (CT group). The primary endpoint was overall survival (OS), secondary endpoints included progression-free survival (PFS), objective response rate (ORR), disease control rate (DCR), duration of response (DoR), and treatment-related adverse events (TRAEs). Survival was analyzed using Kaplan–Meier methods and Cox regression, stabilized inverse probability of treatment weighting (IPTW) and prespecified sensitivity analyses were used to assess robustness to measured confounding. Results ORR was numerically higher in the AP group than in the CT group (20.3% vs 8.6%, P=0.072). The DCR was significantly higher in AP group (71.2% vs 36.2%, P<0.001). Among the responders, median DoR was 5.2 months in the AP group and 2.6 months in the CT group (P<0.001). Median PFS was 4.5 months in AP group and 2.8 months in CT group (HR=0.58, 95% CI: 0.39–0.88, P=0.008). Median OS was 11.3 and 6.4 months, respectively (HR=0.55, 95% CI: 0.36–0.83, P=0.004). Multivariable Cox and stabilized IPTW analyses yielded directionally consistent survival associations (IPTW: OS HR=0.53, 95% CI: 0.36–0.79; PFS HR=0.57, 95% CI: 0.39–0.85). Any-grade TRAEs occurred in 86.4% and 91.4% of patients in AP and CT group, respectively. Grade ≥3 TRAEs occurred in 39.0% and 55.2%, respectively. Conclusion Apatinib plus PD-1 blockade was associated with longer survival and greater disease control than single-agent chemotherapy. Given the single-center retrospective design, limited sample size, treatment heterogeneity, incomplete biomarker characterization, and potential residual confounding, these findings are hypothesis-generating, do not establish causality, and require prospective validation.