Racial/ethnic disparities in the effectiveness of atezolizumab plus bevacizumab (A+B) vs. tyrosine kinase inhibitors (TKIs) among veterans with unresectable hepatocellular carcinoma (uHCC).

Abstract

448 Background: The etiology of liver disease and HCC often differs by race and ethnicity, which may lead to potential differential effectiveness of immunotherapy (IO) by race/ethnicity. However, clinical trials for HCC have historically had limited racial diversity. Emerging data has characterized the real-world effectiveness of A+B, the standard of care for first-line (1L) uHCC. This study assessed whether there is racial/ethnic difference in the effectiveness of A+B compared to TKIs, including sorafenib (S) or lenvatinib (L). Methods: A population-based retrospective analysis was conducted using data from the Veterans Health Administration (VHA) Data Warehouse. Non-Hispanic White, African American (AA), and Hispanic patients diagnosed with uHCC between 2017 and 2022 who initiated 1L A+B (on or after 2020), S, or L were selected. Kaplan-Meier analyses and multivariable Cox regression analysis with race/ethnicity as a treatment modifier (adjusting for age, region, extrahepatic metastases, etiology, presence of cirrhosis, comorbidity index, prior local therapies, Child-Pugh [CP] class, and Albumin-Bilirubin [ABLI] grades) were conducted to investigate the racial/ethnic differences in the effectiveness of A+B vs S or L. Results: Of 1,738 patients, 1,140 were White (20% A+B, 56% S, 24% L), 437 were AA (25% A+B, 50% S, 25% L), and 161 were Hispanic (22% A+B, 57% S, 21% L). Compared to Whites, a significantly higher proportion of AAs had underlying hepatitis C (87% vs 56%, p<0.05). There was a trend towards a lower proportion of CP B and C and less ascites in AAs compared to Whites. Table shows median OS (mOS) by race and regimen, along with hazard ratios (HRs) from the multivariate analysis. The survival benefit associated with A+B compared with S or L was numerically greater in AA patients versus White patients, although the difference did not achieve statistical significance. Conclusions: In this real-world cohort of 1L uHCC patients, A+B demonstrated a significant survival advantage over TKIs across racial/ethnic groups. Understanding the factors contributing to the effectiveness of A+B across racial/ethnic groups could provide valuable insights for guiding future trials and initiatives aimed at addressing racial disparities in HCC. [Table: see text]