PURPOSE: Urinary tract cancer can be pure urothelial carcinoma, pure nonurothelial carcinoma or variant urothelial carcinoma (defined here as mixed urothelial carcinoma). Little is known regarding outcomes for patients with variant urothelial carcinoma receiving immune checkpoint inhibitors. We hypothesized that variant urothelial carcinoma does not compromise immune checkpoint inhibitor efficacy in patients with advanced urothelial carcinoma. MATERIALS AND METHODS: We performed a retrospective cohort study across 18 institutions. Demographic, clinicopathological, treatment and outcomes data were collected for patients with advanced urothelial carcinoma who received immune checkpoint inhibitors. Patients were divided into pure vs variant urothelial carcinoma subgroups, with variant urothelial carcinoma further divided by type of variant (ie squamous, neuroendocrine etc). We compared overall response rate using univariate and multivariate logistic regression and progression-free survival and overall survival using Kaplan-Meier and univariate and multivariate Cox proportional hazards. RESULTS: Overall 519 patients were identified, with 395, 406 and 403 included in overall response rate, overall survival and progression-free survival analyses, respectively. Overall response rate to immune checkpoint inhibitors between patients with pure vs variant urothelial carcinoma was comparable (28% vs 29%, p=0.90) without significant differences for individual subtypes vs pure urothelial carcinoma. Median overall survival for patients with pure urothelial carcinoma was 11.0 months vs 10.1 months for variant urothelial carcinoma (p=0.60), but only 4.6 months for patients with neuroendocrine features (9 patients, HR 2.75, 95% CI 1.40-5.40 vs pure urothelial carcinoma, p=0.003). Median progression-free survival was 4.1 months for pure vs 5.2 months for variant urothelial carcinoma (p=0.43) and 3.7 months for neuroendocrine features (HR 1.87, 95% CI 0.92-3.79 vs pure urothelial carcinoma, p=0.09). CONCLUSIONS: Overall response rate to immune checkpoint inhibitors was comparable across histological types. However, overall survival was worse for patients with tumors containing neuroendocrine features. Variant urothelial carcinoma should not exclude patients from receiving immune checkpoint inhibitors.