Background: KEYNOTE-240 (NCT02702401) examined the anti?PD-1 antibody pembro and showed improvement in OS and PFS vs pbo in pts with aHCC previously treated with sorafenib. The study did not meet prespecified statistical significance criteria for OS and PFS. Median OS (final analysis) was 13.9 mo for pembro vs 10.6 mo for pbo (HR 0.781; 95% CI 0.611-0.998). At the first interim analysis when testing for PFS and ORR was prespecified, median PFS was 3.0 mo for pembro vs 2.8 mo for pbo (HR 0.775; 95% CI 0.609-0.987) and ORR was 16.9% (CR, n = 3) for pembro and 2.2% (CR, n = 0) for pbo. AEs were consistent with the known safety profile of pembro. Longer term data from KEYNOTE-240 after ̃1.5 y of additional follow-up are reported. Methods: Adults with confirmed aHCC for whom sorafenib therapy failed (progression or intolerance) were randomly assigned 2:1 to receive pembro 200 mg IV Q3W + best supportive care (BSC) or pbo + BSC for ≤35 cycles or until confirmed progression/unacceptable toxicity, pt withdrawal of consent, or investigator decision to withdraw pt. Dual primary end points were OS and PFS, assessed by blinded independent central review (BICR) per RECIST v1.1. Secondary end points included ORR, DOR, DCR, TTP (all assessed by BICR per RECIST v1.1), and safety. Results: Of 413 pts, 278 were randomized to receive pembro; 135, to pbo.As of July 13, 2020, median time (range) from randomization to data cutoff was 39.6 mo (31.7-48.8) for pembro and 39.8 mo (31.7-47.8) for pbo. Median OS (95% CI) was 13.9 mo (11.6-16.0) for pembro and 10.6 mo (8.3-13.5) for pbo (HR 0.771; 95% CI 0.617-0.964). Estimated OS rates at 24 and 36 mo for pembro and pbo were 28.8% and 20.4% and 17.7% and 11.7%, respectively. Median PFS (95% CI) was 3.3 mo (2.8-4.1) for pembro and 2.8 mo (1.6-3.0) for pbo (HR 0.703; 95% CI 0.559-0.885). Estimated PFS rate at 24 mo was 11.8% for pembro and 4.8% for pbo. ORR (95% CI) was 18.3% (14.0-23.4) for pembro and 4.4% (1.6-9.4) for pbo. Median time to response (95% CI) was 2.7 mo (1.2-16.9) for pembro and 2.9 mo (1.1-6.9) for pbo. Median DOR (range) was 13.9 mo (1.5+ to 41.9+) for pembro and 15.2 mo (2.8-21.9) for pbo; 53.7% of responders in the pembro arm and 50.0% of responders in the pbo arm had DOR ≥12 mo. DCR was 61.9% for pembro and 53.3% for pbo. Best overall responses were 10 CR, 41 PR, 121 SD, and 85 PD for pembro and 0 CR, 6 PR, 66 SD, and 54 PD for pbo. Median TTP (95% CI) was 4.0 mo (2.8-5.3) for pembro and 2.8 mo (1.6-3.0) for pbo. No new or unexpected AEs occurred. The frequency of sponsor-assessed immune-mediated hepatitis events did not increase with additional follow-up. There continued to be no HBV or HCV viral flare events. Conclusions: In previously treated pts with aHCC, improvement in OS and PFS was maintained over time with pembro vs pbo, and the safety profile remained consistent. These data support the benefit:risk profile of pembro. Clinical trial information: NCT02702401