Background: Recent data suggest that modulating the microbiome may affect response to ICI. Using electronic health record (EHR) data, we examined whether use of ABX influences OS in pts with aNSCLC and aMel treated with ICI. Methods: We performed two retrospective cohort studies (aNSCLC, aMel) of pts who received ICI +/- chemotherapy as first-line treatment in the nationwide Flatiron Health EHR-derived database. The impact of ABX exposure on OS was assessed using Cox regression, adjusting for age, sex, body mass index (BMI), performance status (ECOG PS), Elixhauser Comorbidity Index, glucocorticoid (GC) use, and smoking. In the primary analysis, ABX exposure was defined as use from -6 to +4 weeks around start of ICI. In secondary analyses, we separately evaluated ABX use within 6 weeks pre- and 4 weeks post-initiation of ICI. Results: Baseline characteristics are displayed (Table). In the primary analysis, there was no association between ABX use and OS in aNSCLC (HR 1.16, 95% CI 0.54-2.47, p = 0.7) or in aMel (HR 0.70, 95% CI 0.17-2.81, p = 0.6). The only secondary analysis to reveal a negative impact of ABX on OS was the time window of 0 to +4 weeks after ICI start in aNSCLC (HR 3.41, 95% CI 1.38-8.39, p = 0.008). Conclusions: In this large, real-world dataset, ABX exposure around the start of front-line ICI did not influence OS in pts with aNSCLC or aMel. However, ABX exposure was associated with inferior OS in the subgroup of aNSCLC pts who received ABX within 4 weeks following start of ICI. Further characterization of completeness of ABX exposure data in the oncology EHR will assist with interpretation of these results.