234 Background: Nivolumab (Nivo) is a standard therapy after ramucirumab (RAM)-based second-line chemotherapy in patients (pts) with advanced gastric cancer (AGC). However, the relationship between angiogenesis-related factors and the efficacies of immune-checkpoint inhibitors after anti-angiogenic chemotherapy is unclear. Methods: We retrospectively evaluated pts with AGCs who received Nivo after RAM-based second-line chemotherapy at a single institution from Nov 2017 to Aug 2019. Clinical benefit of Nivo was defined as CR, PR, or >4 months of SD and non-CR/non-PD. In addition, preserved serum samples at time points of pre-RAM, pre-Nivo, and post-Nivo (within 12 weeks from the start of Nivo) were analyzed by using the designated panel, containing vascular endothelial growth factor-A/D (VEGF-A/D), placental growth factor (PlGF), soluble vascular endothelial growth factor receptor-1/2/3 (sVEGFR-1/2/3), Interleukin-6/8 (IL-6/8), soluble intercellular adhesion molecule-1 (sICAM-1), soluble vascular cell adhesion molecule-1 (sVCAM-1), tissue inhibitor of metalloproteinase-1 (TIMP-1), angiopoietin-2, hepatocyte growth factor (HGF), interferon gamma (INF-γ), osteopontin (OPN), soluble neuropilin-1, and thrombospondin-2 (TSP-2). We investigated the association between angiogenesis-related factors and clinical outcomes with Nivo. Results: We evaluated 167 samples from 58 pts, including 10 pts (17.2%) exhibiting clinical benefit. Characteristics of pts were as follows: median age, 68; male, 74%; PS, 0/1-2, 17/83%; HER2 positive, 16%; prior gastrectomy, 40%; histology, intestinal/diffuse, 16/84%; disease status, metastatic/recurrent, 69/31%; lung metasstasis, 14%; liver metasstasis, 26%; peritoneum metastasis, 67%; lymph node metastasis, 31%. The median progression-free survival (PFS) and overall survival (OS) of Nivo were 1.7 and 6.2 months, respectively, with a median follow-up of 5.0 months. There were no correlations between pre-Nivo levels of angiogenesis-related factors and clinical benefit; however, the median post-Nivo/pre-Nivo VEGF-A ratio was significantly lower in pts with clinical benefit than in those without clinical benefit (0.44 versus 0.94, p = 0.008). By multivariate analysis using characteristics of pts and data from pre-Nivo samples, VEGF-A above the median level of 1400 pg/mL and sVCAM-1 below the median level of 1640000 pg/mL were independent prognostic factors for poor PFS (hazard ratio, 1.90, p = 0.033) and OS (hazard ratio, 2.02, p = 0.037), respectively. Conclusions: Our study suggests that rapid decline of serum VEGF-A level after Nivo and higher VEGF-A before Nivo were respectively associated with the clinical benefit of Nivo and poor survival in pts with AGC treated with Nivo after RAM-based second-line chemotherapy.