IMbrave150: Exploratory analysis to examine the association between treatment response and overall survival (OS) in patients (pts) with unresectable hepatocellular carcinoma (HCC) treated with atezolizumab (atezo) + bevacizumab (bev) versus sorafenib (sor).

Abstract

4071 Background: Based on IMbrave150 (NCT03434379) results, atezo + bev has been approved in > 60 countries for pts with unresectable HCC who have not received prior systemic therapy (Finn RS, NEJM 2020). OS and objective response rate (ORR) improvements with atezo + bev vs sor were maintained with an additional 12 mo of follow up since primary analysis. Updated median OS was 19.2 mo with atezo + bev vs 13.4 mo with sor (stratified HR, 0.66; 95% CI: 0.52, 0.85). Updated ORR was 30% with atezo + bev vs 11% with sor by independently-assessed (IRF) RECIST 1.1 (Finn RS, ASCO GI 2021). Here, we report an exploratory analysis examining the association of response by RECIST 1.1 with OS and independent predictors of survival. Methods: Pts in this Ph III study were systemic treatment–naive with unresectable HCC, ≥1 measurable untreated lesion (RECIST 1.1), Child-Pugh class A liver function and ECOG PS 0/1. Pts were randomized 2:1 to atezo 1200 mg IV q3w + bev 15 mg/kg IV q3w or sor 400 mg bid until unacceptable toxicity or loss of clinical benefit per investigator. ORR was determined by IRF RECIST 1.1. Kaplan-Meier analyses of OS by response status were conducted without landmark and with 4- and 6-mo landmarks. Multivariate analysis was conducted using Cox modeling with time-dependent covariate (responder [yes/no]) with backwards elimination. These analyses only included pts treated with atezo + bev. Results: IMbrave150 enrolled 501 pts, including 336 treated with atezo + bev. Median follow-up was 15.6 mo. OS was longer in pts with confirmed response per RECIST 1.1 (responders, CR + PR) vs non-responders by Kaplan-Meier analyses without landmark and with 4- and 6-mo landmarks (Table). By multivariate analysis, in addition to responder (yes/no), 5 of the 10 initially included predictors of OS remained in the final Cox model ( P< 0.10): ECOG PS (0/1), geographic region (Asia excluding Japan/rest of the world), etiology (hepatitis B/hepatitis C/non-viral), macrovascular invasion and/or extrahepatic spread (yes/no), and baseline alpha-fetoprotein ( < 400 ng/mL/≥400 ng/mL). Conclusions: Atezo + bev is the new standard of care for pts with previously untreated, unresectable HCC. Here we showed that in IMbrave150 pts treated with atezo + bev, response by RECIST 1.1 was associated with OS, suggesting that confirmed response is an independent predictor of OS in these pts. Clinical trial information: NCT03434379. [Table: see text]