Background
Phase III double-blind RCT, N=480 (randomized 1:1), patients with unresectable non-metastatic HCC amenable to TACE, ECOG PS 0–1, Child-Pugh A. Enrolled at 137 sites in 33 countries/regions May 2020–Jan 2023. Stratified by study site, AFP level, ECOG PS, albumin-bilirubin grade, and tumour burden. Dual primary endpoints PFS (threshold one-sided P=.025) and OS (threshold one-sided P=.0012).
Interventions and follow up
Arm A: TACE + lenvatinib (12mg/day if ≥60kg, 8mg/day if <60kg) + pembrolizumab 400mg IV q6wk for up to 2 years (n=237)
Arm B: TACE + oral placebo + IV placebo (dual placebo) (n=243)
Primary endpoint: PFS and OS (co-primary)
Median follow up: 25.6mo
Arm B: TACE + oral placebo + IV placebo (dual placebo) (n=243)
Primary endpoint: PFS and OS (co-primary)
Median follow up: 25.6mo
Results
PFS: 14.6 vs 10.0mo, HR 0.66 (95% CI 0.51–0.84), P=.0002 — threshold met
OS (interim): 24-mo OS rate 75% vs 69%, HR 0.80 (0.57–1.11), P=.087 — threshold P=.0012 not met; OS immature
OS (interim): 24-mo OS rate 75% vs 69%, HR 0.80 (0.57–1.11), P=.087 — threshold P=.0012 not met; OS immature
Adverse events
Overall/cardiovascular: Grade ≥3 treatment-related AEs 71% vs 32% (Arm A vs B); most common grade 3–4 hypertension 24% vs 7% and platelet count decreased 11% vs 6%.
Fatal: Treatment-related deaths 4 (2%) vs 1 (<1%), including hepatic failure, GI haemorrhage, myositis, and immune-mediated hepatitis.
Fatal: Treatment-related deaths 4 (2%) vs 1 (<1%), including hepatic failure, GI haemorrhage, myositis, and immune-mediated hepatitis.
Conclusions
TACE plus lenvatinib and pembrolizumab significantly improved PFS versus TACE plus placebo in unresectable non-metastatic HCC. The OS endpoint was not met at the interim analysis and requires longer follow-up. LEAP-012 demonstrates that adding systemic lenvatinib+pembrolizumab to TACE meaningfully delays progression, though at the cost of substantially higher toxicity (71% vs 32% grade ≥3 TRAEs).
Key Limitations
OS remains immature and did not meet its pre-specified threshold; the PFS improvement has not translated to demonstrated survival benefit. The substantially higher grade ≥3 AE rate (71% vs 32%) raises tolerance concerns in a population with underlying liver disease — higher than in typical systemic HCC trials. The OS threshold was very stringent (one-sided P=.0012) reflecting multiple-testing correction, potentially underpowering detection of a moderate survival benefit, raising uncertainty about the long-term benefit-risk profile.
Clinical Context
LEAP-012 and EMERALD-1 represent parallel attempts to enhance TACE with systemic immunotherapy. Both showed PFS benefit; neither has demonstrated OS benefit. LEAP-012 received regulatory approval in China (lenvatinib+pembro+TACE for unresectable non-metastatic HCC, 2025) by NMPA. Whether PFS constitutes sufficient evidence for approval in Western jurisdictions (FDA/EMA) without OS benefit remains under active evaluation. TACE+durvalumab+bevacizumab (EMERALD-1) is the competing approach with a different safety profile.