Background: PD1/PD-L1 inhibitors have been evaluated in trials enrolling patients (pts) with pure urothelial or mixed urothelial histology containing non-urothelial components. However, any differential impact of pure vs. mixed urothelial histology on ICI benefit is unclear. We conducted a retrospective study to evaluate the impact of pure vs. mixed urothelial carcinoma histology on outcomes with ICIs in pts with metastatic urothelial carcinoma (mUC). Methods: We obtained data from 120 pts with mUC from a single institution (DFCI) who received ICI therapy. Demographic, clinical variables and outcomes (overall response rate [ORR], overall survival [OS]) were collected. Histology was reviewed at DFCI for all pts and recorded as pure urothelial if only urothelial carcinoma was seen or mixed urothelial if components of any other histology were observed in addition to urothelial. A Cox regression analysis was done to study the association of prognostic variables and histology with objective response. Results: Data was obtained from 120 pts, of whom 110 (91.7%) received a single agent PD1/PD-L1 inhibitor (pembrolizumab=58, atezolizumab=52, nivolumab=4, nivolumab + ipilimumab=3, nivolumab + vaccine=2, durvalumab+tremelimumab=1). The median age was 66, 70.8% were male and 72.5% had received prior chemotherapy. 79 (65.8%) tumors originated from the bladder, 39 (32.5%) from the upper tract, 2 (1.67%) had unknown site of origin. 91 (76.6%) had pure urothelial and 28 (23.3%) had mixed urothelial histology. On univariable analysis, pure vs. mixed urothelial histology was not associated with response (HR 1.52 [95% CI 0.59-3.98, p=0.39]). On multivariable analysis, upper tract vs. bladder primary (HR 3.06 [95% CI 1.10-8.49], p=0.032) and higher blood neutrophil to lymphocyte ratio (HR 0.35 [95% CI 0.17-0.72], p=0.004) were associated with lower response rate. Conclusions: In this hypothesis-generating study, pure vs. mixed urothelial carcinoma histology did not appear to significantly impact response to ICI therapy for mUC. The impact of proportion of non-urothelial histology, pure non-urothelial histology and site of primary on response warrants further study.