Background: Initial results from KEYNOTE-224 (NCT02702414), an open label, phase 2 trial showed that pembro, an anti-PD-1 antibody, was active and safe in pts with advanced HCC previously treated with sorafenib. Here we report updated clinical results and biomarker studies. Methods: Eligible pts were age ≥18 y with histologically confirmed HCC, radiographic progression on/intolerance to sorafenib and disease not amenable to curative treatment, Child Pugh A, ECOG PS 0-1 and BCLC stage C or B. Pts received pembro 200 mg IV Q3W for 2 y or until disease progression, unacceptable toxicity, withdrawal of consent or investigator decision. Response was assessed every 9 wk. Primary endpoint was ORR (RECIST v1.1, central review). Secondary endpoints were DOR, DCR, PFS, OS and safety. Exploratory endpoint was relationship of PD-L1 IHC combined positive and tumor proportion scores (CPS and TPS; n = 52), and T cell inflamed gene expression profile (GEP; n = 42) with response. Data cutoff date was Nov 24, 2017. Results: Of 104 treated pts, 18 remained on therapy. Median age was 68 y (range 43-87), 21.2% were HBV⁺, 26% were HCV⁺, 94.2% were Child Pugh A, 79.8% had PD on sorafenib and 64.4% had extrahepatic disease. ORR was 16.3% (95% CI, 9.8 to 24.9, n = 17) and similar across differing etiologies; 66% of responders had duration ≥12 mo (Kaplan Meier) and median response duration was not reached (3.1 - 12.5+ mo). Best response was CR in 1 pt (1.0%), PR in 16 (15.4%), SD in 47 (45.2%) and PD in 34 (32.7%); DCR was 61.5%. Median PFS (95% CI) was 4.9 mo (3.4 to 7.0) and OS was 12.9 mo (9.7 to NA). PFS and OS 12 mo rates were 25.4% and 53.6%, respectively. Safety was similar to that observed for pembrolizumab in other indications. Immune-mediated hepatitis occurred in 3 (2.9%) pts; no cases of HBV/HCV flare were seen. Higher PD-L1 CPS in tumor and immune cells was associated with higher ORR and longer PFS, while PD-L1 TPS in tumor cells alone and GEP showed less robust associations with outcomes. Conclusions: These results confirm that pembro may be a promising treatment option for pts with advanced HCC. PD-L1 CPS was associated with clinical response to pembro while results for PD-L1 TPS and GEP were less robust; further study is needed to better define these relationships. Clinical trial information: NCT02702414