Lenvatinib plus pembrolizumab versus lenvatinib plus placebo for advanced hepatocellular carcinoma (LEAP-002): a randomised, double-blind, phase 3 trial

Abstract

Background Systemic therapies have improved the management of hepatocellular carcinoma, but there is still a need to further enhance overall survival in first-line advanced stages. This study aimed to evaluate the addition of pembrolizumab to lenvatinib versus lenvatinib plus placebo in the first-line setting for unresectable hepatocellular carcinoma.Methods In this global, randomised, double-blind, phase 3 study (LEAP-002), patients aged 18 years or older with unresectable hepatocellular carcinoma, Child Pugh class A liver disease, an Eastern Cooperative Oncology Group performance status of 0 or 1, and no previous systemic treatment were enrolled at 172 global sites. Patients were randomly assigned (1:1) with a central interactive voice-response system (block size of 4) to receive lenvatinib (bodyweight <60 kg, 8 mg/day; bodyweight >= 60 kg, 12 mg/day) plus pembrolizumab (200 mg every 3 weeks) or lenvatinib plus placebo. Randomisation was stratified by geographical region, macrovascular portal vein invasion or extrahepatic spread or both, alpha-fetoprotein concentration, and Eastern Cooperative Oncology Group performance status. Dual primary endpoints were overall survival (superiority threshold at final overall survival analysis, one-sided p=0019; final analysis to occur after 532 events) and progression-free survival (superiority threshold one-sided p=0002; final analysis to occur after 571 events) in the intention-to-treat population. Results from the final analysis are reported. This study is registered with ClinicalTrials.gov, NCT03713593, and is active but not recruiting.Findings Between Jan 17, 2019, and April 28, 2020, of 1309 patients assessed, 794 were randomly assigned to lenvatinib plus pembrolizumab (n=395) or lenvatinib plus placebo (n=399). Median age was 660 years (IQR 570-720), 644 (81%) of 794 were male, 150 (19%) were female, 345 (43%) were Asian, 345 (43%) were White, 22 (3%) were multiple races, 21 (3%) were American Indian or Alaska Native, 21 (3%) were Native Hawaiian or other Pacific Islander, 13 (2%) were Black or African American, and 46 (6%) did not have available race data. Median follow up as of data cutoff for the final analysis (June 21, 2022) was 321 months (IQR 294-353). Median overall survival was 212 months (95% CI 190-236; 252 [64%] of 395 died) with lenvatinib plus pembrolizumab versus 190 months (172-217; 282 [71%] of 399 died) with lenvatinib plus placebo (hazard ratio [HR] 084; 95% CI 071-100; stratified log-rank p=0023). As of data cutoff for the progression-free survival final analysis (April 5, 2021), median progression-free survival was 82 months (95% CI 64-84; 270 events occurred [42 deaths; 228 progressions]) with lenvatinib plus pembrolizumab versus 80 months (63-82; 301 events occurred [36 deaths; 265 progressions]) with lenvatinib plus placebo (HR 087; 95% CI 073-102; stratified log-rank p=0047). The most common treatment-related grade 3-4 adverse events were hypertension (69 [17%] of 395 patients in the lenvatinib plus pembrolizumab group vs 68 [17%] of 395 patients) in the lenvatinib plus placebo group), increased aspartate aminotransferase (27 [7%] vs 17 [4%]), and diarrhoea (25 [6%] vs 15 [4%]).Treatment-related deaths occurred in four (1%) patients in the lenvatinib plus pembrolizumab group (due to gastrointestinal haemorrhage and hepatorenal syndrome [n=1 each] and hepatic encephalopathy [n=2]) and in three (1%) patients in the lenvatinib plus placebo group (due to gastrointestinal haemorrhage, hepatorenal syndrome, and cerebrovascular accident [n=1 each]).Interpretation In earlier studies, the addition of pembrolizumab to lenvatinib as first-line therapy for advanced hepatocellular carcinoma has shown promising clinical activity; however, lenvatinib plus pembrolizumab did not meet prespecified significance for improved overall survival and progression-free survival versus lenvatinib plus placebo. Our findings do not support a change in clinical practice.