Background: Platinum-based systemic chemotherapy (CTx) is recommended for poorly differentiated advanced gastrointestinal neuroendocrine carcinoma (GI-NEC). As second-line CTx various regimens are adopted according to the guidelines of GI adenocarcinoma or small cell lung cancer in daily practice. Recently, some clinical and preclinical reports suggest that vascular endothelial growth factor (VEGF) inhibitors are effective against NEC. In this study, we evaluated the efficacy of ramucirumab (RAM, anti-VEGF receptor 2 antibody)-containing CTx. Methods: We retrospectively evaluated gastric (G-) or colorectal (C-) NEC patients previously treated with first-line platinum-based CTx followed by second-line CTx between March 2015 and June 2020 (G-NEC) or between May 2016 and June 2020 (C-NEC). The main inclusion criteria were ECOG performance status (PS) of 0–2, pathologically diagnosed as G or C- NEC or mixed adenoneuroendocrine carcinoma. We compared the efficacy of RAM-containing CTx as second- or later-line treatment (Group A) with that of CTx without RAM as second-line treatment (Group B). A Cox proportional hazard model and inverse probability weighting of propensity scores were used to adjust the background factors between the two groups. Results: From 25 facilities, 139 patients were included in this study. Group A and B contained 50 (G/C, 43/7 patients) and 89 (G/C, 58/31 patients) patients, respectively. Patient characteristics were as follows (Group A/Group B): median age (range), 70 (28–83)/68 (39–86); ECOG PS 1–2, 66/69%, more than two prior regimens, 46/0%; response to previous platinum-based CTx, 46/38%. The regimen in combination with RAM in Group A was paclitaxel/nab-paclitaxel/FOLFIRI (62/26/12%). Second-line CTx in Group B were commonly amrubicin (58%) and irinotecan (11%). The efficacy (Group A/Group B) after a median follow-up time of 7.5/5.5 months were as follows: median overall survival (mOS), 8.6/5.8 months (HR, 0.68; 95% CI, 0.45–1.01; p = 0.058); median progression-free survival, 4.3/1.9 months (HR, 0.57; 95% CI, 0.39–0.84; p = 0.004); objective response rate, 50/8% (p< 0.0001); disease control rate, 70/24% (p< 0.0001); mOS in G-NEC, 8.6/5.3 months (HR, 0.62; 95% CI, 0.39–0.99; p = 0.047); and mOS in C-NEC, 5.4/6.8 months (HR, 0.66; 95% CI, 0.26–1.66; p = 0.378). The most common grade 3 or higher treatment-related adverse event that occurred in more than 5% of patients in Group A was neutropenia (32%). Conclusions: This study suggests that RAM-containing CTx as second- or later-line treatment is effective for GI-NEC patients, especially for G-NEC. Central pathological review and biomarker analyses using pretreated tissue specimens are ongoing. We will present updated data in the congress. Clinical trial information: UMIN000043200.