The immune checkpoint inhibitors (ICI) tremelimumab (anti–CTLA-4) and durvalumab (anti–PD-L1) in combination exhibited tolerability and a promising objective response rate (ORR) in the initial cohort of this study (NCT02519348). Subsequent evaluation of increasing doses of tremelimumab in solid tumors suggested priming with a higher dose may induce a stronger immune response and enhance anti-tumor activity. Thus, the randomized expansion cohorts included 4 arms evaluating tremelimumab and durvalumab as monotherapies and 2 tremelimumab+durvalumab regimens, including a novel tremelimumab+durvalumab regimen featuring a single, priming dose of tremelimumab. Here, we present a detailed safety analysis along with the clinical activity for all regimens. ICI-naïve patients with advanced hepatocellular carcinoma (aHCC) who progressed on, were intolerant to, or refused sorafenib were randomized to one of two tremelimumab+durvalumab combinations: T300+D (T 300 mg + D 1500 mg [1 dose] followed by durvalumab every 4 weeks [Q4W]) or T75+D (T 75 mg Q4W + D 1500 mg Q4W [4 doses] followed by durvalumab Q4W); or single-agent durvalumab (D, 1500 mg Q4W) or tremelimumab (T, 750 mg Q4W). Safety was the primary endpoint. ORR by blinded, independent central review using RECIST v1.1, circulating lymphocytes, and overall survival (OS) were assessed. At data cutoff (02-Sept-2019), 332 patients were enrolled (T300+D, n=75; D, n=104; T, n=69; T75+D, n=84). Median follow-ups were 11.7 months for T300+D, 14.6 (T75+D), 8.9 (D), and 15.8 (T). Grade 3/4 treatment-related adverse events occurred in 28.9% of patients (T300+D, 35.1%; T75+D, 25.6%; D, 19.8%; T, 42.0%;). Grade ≥3 hepatobiliary disorders were infrequent; abnormal hepatic function occurred in the T300+D (n=1), T75+D (n=2), and D (n=4) arms, and hepatitis was reported in T75+D, D and T arms (n=1 each). Grade ≥3 gastrointestinal disorders were also uncommon (T300+D, n=6; T75+D, n=5; D, n=4; T, n=9) and grade ≥3 infections/infestations were rare (T300+D, n=2; T75+D, n=0; D, n=1; T, n=1). Treatment-related discontinuation rates were 10.8% for T300+D, 6.1% for T75+D 7.9% for D, and 11.6% for T. No treatment-related deaths were reported for T300+D or T; there was 1 death for T75+D (hepatic failure) and 3 for D (abnormal hepatic function, hepatic failure, pneumonitis). ORR (95% CI) was 22.7% (13.8-33.8) for T300+D, 9.5% (4.2-17.9) for T75+D, 9.6% (4.7-17.0) for D, and 7.2% (2.4-16.1) for T. Median OS (95% CI) was 18.7 (10.8-NR) months with T300+D, 11.3 (8.4-14.6) months with T75+D, 11.7 (8.5-16.9) months with D, and 17.1 (10.9-NR) months with T. Moreover, a unique proliferative T cell profile was identified for patients treated with T300+D, suggesting additive biologic activity for this combination, and showed patients with an objective response exhibited higher cytotoxic (CD8) counts. The encouraging safety profile combined with the observed clinical activity suggests that T300+D provides the best benefit-risk profile versus T75+D or monotherapies. Additionally, the distinctive pharmacodynamic activity observed with the T300+D regimen further supports its use in aHCC. T300+D and durvalumab are being evaluated in the ongoing phase 3 HIMALAYA study (NCT03298451) in first-line HCC versus sorafenib.