Clinical outcomes of first-line immune checkpoint inhibitors with chemotherapy in advanced EBV-associated gastric cancer.

4049 Background: EBV-associated gastric cancer (EBVaGC) is a distinct subgroup of GC with high PD-L1/PD-L2 expression and immune cell infiltration. While several rationales support the potential efficacy of immune checkpoint inhibitors (ICIs) in EBVaGC, clinical evidence from large-scale trials is limited, and their effectiveness remains inconclusive. This retrospective study evaluated the clinical outcomes of palliative first-line chemotherapy with or without the addition of ICIs in patients with EBVaGC. Methods: We identified 217 patients diagnosed with EBVaGC via in situ hybridization and evaluated their baseline characteristics, including HER2 expression, MSI status, and PD-L1 expression. Among these, 83 patients with metastatic disease received palliative first-line treatment. Of the 77 HER2-negative patients, 23 were treated with ICIs (nivolumab or pembrolizumab) combined with cytotoxic chemotherapy (ICI+chemo), while 54 received chemotherapy alone. Survival outcomes and response rates were compared between ICI+chemo and chemotherapy-alone groups. Results: Among 217 EBVaGC patients, 8.5% were HER2-positive, 2.6% were MSI-High, and PD-L1 CPS ≥10 and ≥5 was observed in 59.1% and 84.1%, respectively. In HER2-negative EBVaGC patients, ICI+chemo showed significantly prolonged progression free survival (PFS) compared to chemotherapy alone (median: 9.28 vs. 6.07 months; HR = 0.48; p = 0.031). There was no significant difference in overall survival (OS) (median: 20.72 vs. 18.89 months; HR = 0.78; p = 0.508). The objective response rate (ORR) and disease control rate (DCR) were significantly higher in the ICI+chemo group than in the chemotherapy-alone group (ORR, 78.3% vs. 51.9%; OR = 3.29; p = 0.042; DCR, 100% vs. 79.6%; p = 0.028). In the ICI+chemo group, PD-L1 expression (cutoff: CPS 5 or 10) was not significantly associated with survival outcomes or response rates. For CPS ≥5, the median PFS was 7.84 months compared to 8.99 months in the CPS < 5 group (p = 0.944), the median OS was 20.75 vs. 16.51 months (p = 0.131), and the ORR was 76.5% vs. 75.0%. For CPS ≥10, the median PFS was 13.9 months compared to 8.69 months in CPS < 10 group (p = 0.580), the median OS was 24.95 months vs. 20.33 months (p = 0.061), and the ORR was 81.8% vs. 70.0%. Conclusions: To our knowledge, this study represents one of the largest cohorts including patients with EBVaGC who received first-line ICI+chemo. The addition of ICIs to chemotherapy showed clinical benefit, including survival and response rates, compared to chemotherapy alone in EBVaGC. Interestingly, the clinical benefit of ICI+chemo was observed in PD-L1 low group as well as PD-L1 high group. Our results highlight the clinical potential of ICI addition to chemotherapy in EBVaGC, and PD-L1 expression alone is insufficient as a predictive biomarker, warranting further exploration of alternative predictors for immunotherapy efficacy in this subgroup.