72MO - Clinical outcomes and response (R) with atezolizumab plus bevacizumab (AB) or lenvatinib (L) in hepatocellular carcinoma (HCC)

Background

The aim of this multicenter study is to compare R obtained by AB or L and identify any factors that may predict the emergence of this R to one of two treatments in a real-world setting.

Methods

Study population derived from retrospective analysis of prospectively collected patients (P) treated with AB or L as first-line treatment for BCLC C or BCLC B HCC deemed not eligible for locoregional therapies. Overall cohort included Western and Eastern patient populations from 46 centres in 4 countries (Italy, Germany, Japan, and Republic of Korea). ORR with AB vs L was the primary endpoint. The secondary endpoint compared OS with L vs AB in terms of best R.

Results

1312 patients were treated with L, and 823 were treated with AB. ORR was 38.6% for patients receiving L, and 27.3% for patients receiving AB [p < 0.01; odds ratio (OR) 0.60)]. In responder patients (CR + PR), OS was 22.3 months (m) for patients receiving L, and 22.5 m for patients treated with AB (p 0.20; HR 0.81, reference L). In non-responder patients (SD + PD), OS was 10.8 m for patients receiving L, and 11.5 m for patients receiving AB (p 0.36; HR 0.85, reference L). For patients who achieved CR, OS was not reached in both arms, but the result from univariate Cox regression model showed 62% reduction of death risk for p treated with AB (p 0.05). In all multivariate analyses, the treatment arm was not found to be an independent factor conditioning OS. Comparing ORR achieved in the two arms, L was shown to be statistically significant in all subgroup patients except for Child Pugh B, presence of portal vein thrombosis (PT), αFP ≥ 400 ng/mL, presence of extrahepatic disease (EHD), ALBI 2, and no previous locoregional procedures. Patients who achieved CR were compared to the rest of the population in the same way. The only statistically significant difference was in favor of L in patients with NLR > 3 (p 0.03; OR 0.39). Finally, we evaluated separately which patients had more R in AB arm and then in L arm. Only in L arm, 3 factors favoring R were highlighted: absence of EHD (p < 0.01; OR 1.88), BCLC B (p < 0.01; OR 0.54), absence of PT (p 0.05; OR 1.34), and NLR ≤ 3 (p < 0.01; OR 0.70).

Conclusions

L achieves higher ORR in all patient subgroups. Patients who achieve CR with AB can achieve OS so far never recorded in HCC patients. This study did not highlight any factors that could identify patient subgroups capable of obtaining CR.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.