Background
Phase III open-label RCT, N=668 (randomized 1:1), patients ≥18 years with unresectable HCC, no prior systemic therapy, Child-Pugh score 5–6, ECOG PS 0–1, ≥1 measurable lesion per RECIST 1.1. Enrolled at 163 sites across 25 countries Jan 2020–Nov 2021. Stratified by aetiology, macrovascular invasion/EHS, and AFP. Investigator chose lenvatinib or sorafenib as the comparator. Primary endpoint OS at preplanned interim analysis after 394 events.
Interventions and follow up
Arm A: Nivolumab 1mg/kg + ipilimumab 3mg/kg IV q3wk ×4 doses → nivolumab 480mg IV q4wk maintenance (n=335)
Arm B: Investigator's choice lenvatinib 8–12mg oral daily OR sorafenib 400mg oral twice daily (n=333)
Primary endpoint: Overall survival
Median follow up: 35.2mo
Arm B: Investigator's choice lenvatinib 8–12mg oral daily OR sorafenib 400mg oral twice daily (n=333)
Primary endpoint: Overall survival
Median follow up: 35.2mo
Results
OS: 23.7 vs 20.6mo, HR 0.79 (95% CI 0.65–0.96), P=.018
24-mo OS: 49% vs 39%; 36-mo OS: 38% vs 24%
ORR: 36% vs 13% (BICR); median DoR 30.4 vs 12.9mo
Note: Early KM curve crossing — HR 1.65 (1.12–2.43) in first 6 months favoring comparator, then sustained reversal HR 0.61 (0.48–0.77) thereafter
24-mo OS: 49% vs 39%; 36-mo OS: 38% vs 24%
ORR: 36% vs 13% (BICR); median DoR 30.4 vs 12.9mo
Note: Early KM curve crossing — HR 1.65 (1.12–2.43) in first 6 months favoring comparator, then sustained reversal HR 0.61 (0.48–0.77) thereafter
Adverse events
Overall/fatal: Grade 3–4 TRAEs 41% (nivo+ipi) vs 42% (len/sor); treatment-related deaths 12 (4%) vs 3 (1%).
By regimen: Most common grade 3–4 with nivo+ipi were immune-mediated hepatitis, diarrhoea, rash; with len/sor, hypertension and hand-foot syndrome.
By regimen: Most common grade 3–4 with nivo+ipi were immune-mediated hepatitis, diarrhoea, rash; with len/sor, hypertension and hand-foot syndrome.
Conclusions
Nivolumab plus ipilimumab significantly improved OS versus lenvatinib or sorafenib as first-line therapy for unresectable HCC, with a median OS benefit of 3.1 months and markedly higher ORR (36% vs 13%). The landmark survival improvement at 36 months (38% vs 24%) and durable response rates suggest a subset achieving long-term benefit, a pattern characteristic of immune checkpoint inhibitor combinations.
Key Limitations
Early KM crossing indicates higher short-term mortality with nivo+ipi (HR 1.65 in first 6 months), likely reflecting immune-mediated toxicity leading to early deaths (12 treatment-related vs 3). Patients with poor baseline functional reserve may be harmed. The open-label design and investigator choice of comparator (lenvatinib vs sorafenib) introduces heterogeneity in the control arm. The 3.1-month median OS benefit is modest in absolute terms, though the tail of the survival curve shows greater separation. Cross-trial comparison with atezo+bev and STRIDE is limited.
Clinical Context
FDA approved nivolumab plus ipilimumab for first-line unresectable HCC in 2025 (CheckMate 9DW), adding to atezo+bev (IMbrave150) and STRIDE/durvalumab+tremelimumab (HIMALAYA) as approved first-line IO combinations. The 4× induction regimen followed by nivo maintenance parallels the CheckMate 142 design in CRC. The early mortality signal requires careful patient selection — patients with Child-Pugh A5, good ECOG PS, and willingness to accept immune-mediated risk are most appropriate candidates.
References