290 Background: Atezo (anti–PD-L1) has demonstrated safety and efficacy in a broad range of cancers and is approved in the United States for mUC previously treated with platinum-based chemotherapy. Here we report long-term results in mUC from Phase Ia study NCT01375842 (PCD4989g). Methods: Previously treated mUC patients received atezo 15 mg/kg or 1200 mg IV q3w. Enrollment in this Phase Ia expansion cohort initially required PD-L1–selected status and later opened to patients regardless of PD-L1 expression on tumor-infiltrating immune cells. The primary endpoint was safety/tolerability. Secondary endpoints included investigator-assessed RECIST v1.1 ORR (confirmed), DOR and OS. Results: 95 patients were safety evaluable (Table). Median age was 66 years, 76% were male and 80% had primary bladder tumors. 61% had ECOG PS 1. 52% received ≥ 3 prior systemic therapies for mUC (70% platinum). Median treatment duration was 3 months (range: 0-32 months); 24% were treated for ≥ 1 year. Treatment-related AEs occurred in 66% (all Grade) and 8% (Grade 3-4) of patients. No treatment-related deaths were reported. In 94 objective response–evaluable patients (follow-up ≥ 12 weeks), the ORR was 27% (95% CI: 18, 37%), and the CR rate was 10%; the SD rate was 19%. mDOR was 22.1 months (95% CI: 12.1, NE months) in all patients; 56% of responses (7/9 CRs and 7/16 PRs) were ongoing at the December 15, 2015 data cutoff. With a 24-month median follow-up duration (range: 1+ to 32 months), the 1-year OS rate was 47% (95% CI: 36, 58%), and the 2-year rate was 29% (19, 40%); mOS is in the Table. Updated clinical data with further follow-up and analyses by PD-L1 status will be presented. Conclusions: Long-term treatment with atezo was well tolerated, without new safety signals in heavily pre-treated mUC patients. The durability of responses, including CRs, along with extended OS, confirm atezo as a new standard for previously treated mUC patients. Clinical trial information: NCT01375842. [Table: see text]