9503 Background: The anti–PD-1 antibody pembro (pembro; MK-3475) prevents PD-1 from binding to its ligands, PD-L1 and PD-L2, and is approved for treating advanced melanoma at a dose of 2 mg/kg every 3 wk (Q3W). Pembro demonstrated superior PFS over chemotherapy for ipilimumab (ipi)-refractory melanoma (KEYNOTE-002) and superior OS and PFS over ipi for advanced melanoma (KEYNOTE-006). We present 3-year OS data for all patients (pts) with melanoma enrolled in the phase 1b KEYNOTE-001 study (NCT01295827). Methods: Pts were enrolled in ipi-naive and ipi-treated cohorts and received pembro 2 or 10 mg/kg Q3W or 10 mg/kg Q2W until intolerable toxicity, progression, or investigator decision. Clinically stable pts with radiographic progression could remain on pembro until progression was confirmed. Response was assessed by RECIST v1.1 every 12 wk. After pembro discontinuation, pts were contacted to assess survival every 3 mo. OS was estimated using the Kaplan-Meier method. Results: Of the 655 pts enrolled, 24% had BRAFV600mutation, 78% had stage M1c disease, 38% had elevated LDH, 75% had ≥1 prior therapy, and 52% had prior ipi. As of the Sep 18, 2015, data cutoff date, median follow-up duration was 32 mo (range 24-46) and 358 (55%) pts had died. The 36-mo OS rate was 40% and median OS was 23.8 mo (95% CI, 20.2-29.0), with similar results for each dose (Table). 36-mo OS rates were 41% in both ipi-treated and ipi-naive pts and 45% in treatment-naive pts (Table). Examination of the OS curve suggests a long-term OS benefit for a fraction of pts treated with pembro. Additional data, including PFS, ORR, duration of response, and safety, will be available for presentation. Conclusions: Pembro provides long-term survival benefit in pts with ipi-naive and ipi-treated advanced melanoma, with 40% of pts alive at 3 years. These data support the use of pembrolizumab in pts with advanced melanoma regardless of prior treatment. Clinical trial information: NCT01295827.