Efficacy of Second-Line Lenvatinib After Atezolizumab–Bevacizumab and Durvalumab–Tremelimumab in Unresectable Hepatocellular Carcinoma: Association with Prior Immunotherapy Response

Background/Objectives: The efficacy of lenvatinib after different first-line immune checkpoint inhibitor (ICI)-based regimens for hepatocellular carcinoma (HCC) remains unclear. Materials and Methods: In this retrospective single-center study, we analyzed 56 patients with unresectable HCC who received lenvatinib as second-line therapy after atezolizumab plus bevacizumab (Atz/Bev; n = 41) or durvalumab plus tremelimumab (Dur/Tre; n = 15). Antitumor response was evaluated using RECIST v1.1 and modified RECIST (mRECIST). Results: Patients in the Dur/Tre group demonstrated significantly lower disease control rates (46.7% vs. 82.9%, p = 0.014) and shorter progression-free survival (PFS) (median, 56 vs. 210 days; p = 0.015) during prior ICI therapy, indicating more refractory disease. Despite this, lenvatinib demonstrated clinically meaningful antitumor activity after both Atz/Bev and Dur/Tre. Objective response rates were 12.2% vs. 26.7% by RECIST v1.1 and 41.5% vs. 60.0% by mRECIST in the Atz/Bev and Dur/Tre groups, respectively. Similar findings were observed in the Child–Pugh class A subgroup, in which mRECIST response rates were numerically higher in the Dur/Tre group. PFS and overall survival (OS) after lenvatinib initiation were comparable between groups. Objective responses were observed even in patients with progressive disease during prior ICI therapy. Multivariate analysis identified ECOG performance status and mALBI grade, but not prior ICI regimen or antitumor response, as independent prognostic factors for OS. Conclusions: Lenvatinib demonstrated clinically meaningful efficacy after both Atz/Bev and Dur/Tre, and no significant association was observed between prior ICI response and subsequent lenvatinib efficacy. These findings suggest that resistance to immunotherapy does not necessarily confer resistance to lenvatinib in patients with unresectable HCC. The observed differences in response patterns according to prior treatment exposure warrant further investigation and may have implications for treatment sequencing in the current ICI era. However, these findings should be considered hypothesis-generating and require validation in larger multicenter studies.