Background: Despite response to systemic therapy, few metastatic urothelial carcinoma (mUC) patients (pts) achieve durable responses. Mutations in the COMPASS-related proteins KMT2D, KMT2C, and KDM6A are found in 66% of UC patients suggesting histone regulation may be an acquired mechanism of tumor viability. We previously found regression of UC in mice when treated with EZH2 inhibitor, tazemetostat. In vivo administration of tazemtostat enhanced the immune response with direct regulation of cytokines and antigen recognition machinery (MHCI and MHCII). We hypothesized that restoration of epigenetic imbalance, combined with immunotherapy, may improve outcomes in mUC pts. Methods: ETCTN 10183 (NCT03854474) is a pilot study evaluating the efficacy of tazemetostat 800 mg BID + pembrolizumab 200 mg every 3 weeks for cisplatin-refractory (Cohort A) or cisplatin/chemo-ineligible advanced UC (Cohort B) pts, with N=12 pts in each cohort. The primary objective is to identify the recommended phase two dose (RP2D) of tazemetostat in combination with pembrolizumab. Secondary objectives include RECIST-based response rate, safety, and progression-free survival. The maximum planned treatment duration is 2 years. Translational objectives include defining total and COMPASS-related genes tumor mutational burden, tumor subtyping, TCR clonality, T cell infiltration, and PDL-1 expression. Results: In the safety lead-in phase, 6 mUC pts were treated. There were no dose-limiting toxicities (DLTs). The RP2D for tazemetostat was established at 800 mg BID. Here we report results of fully accrued cohort A (N=12); 67% males, 33% females, 83% white, 8.3% Black, and 25% Hispanic/Latino ethnicity. The primary sites of UC were: bladder (n=9, 75%), renal pelvis (n=2, 17%), and ureter (n=1, 8.3%). The median (interquartile range, IQR) time since primary diagnosis was 1.1 years (0.7, 1.9); 11/12 patients had non-nodal metastatic disease; 10/12 had 1 or more visceral/bone metastasis. Median number of treatment cycles was 5 (IQR: 3-11; range 1-35). Median duration on protocol treatment was 12 weeks (IQR: 7-30; range 3-107 weeks). Three pts had treatment related grade 3/4 AEs: grade 4 sepsis (n=1), grade 3 lymphopenia (n=2); anemia, increased alkaline-phosphatase, and HSV oral infection (n=1 each). Best response was partial response (PR) in 3 pts (25%) and 3 (25%) had stable disease. Median PFS was 3.1 months (95%CI: 2.3-NA), and median overall survival was 8.0 months (95% CI: 4.7 – NA). Conclusions: The RP2D dose for tazemetostat is 800 mg + 200 mg Pembrolizumab q 3 weeks. The combination was feasible, well tolerated, and resulted in durable responses in poor-risk chemo-refractory UC patients. Clinical trial information: 03854474.