714P Impact of viral aetiology in the phase III HIMALAYA study of tremelimumab (T) plus durvalumab (D) in unresectable hepatocellular carcinoma (uHCC)

Abstract

In HIMALAYA (NCT03298451), a single, high priming dose of T plus D (STRIDE) significantly improved overall survival (OS) vs sorafenib (S), and D was noninferior to S in uHCC (Abou-Alfa et al. J Clin Oncol 2022;40(suppl 4). Abs 379). Viral aetiology is associated with hepatic impairment in HCC development, and may influence immunotherapy activity. Thus, we analysed the impact of viral aetiology on clinical outcomes. This exploratory analysis assessed STRIDE, D and S in patients (pts) with HBV (presence of HBsAg and/or anti-HBcAb with detectable HBV DNA), HCV or nonviral/other (NV) aetiology. OS hazard ratios (HRs) were calculated using a Cox proportional hazards model. As subsets were not sized for formal comparisons, no multiplicity adjustments were made. A post hoc multivariate analysis was used to identify chance imbalances in key prognostic factors that may bias estimated treatment effects. Baseline demographic and disease characteristics were similar across treatment arms in the HBV and NV subsets. However, in the HCV group, multivariate analysis identified imbalances in two prognostic variables: extrahepatic spread (EHS; more frequent for STRIDE than S) and ALBI (score ≥2 more frequent for STRIDE and D than S). OS and progression-free survival were improved with STRIDE vs S in the HBV and NV groups, but not in the HCV group (Table). Using a stratified Cox proportional hazards model to account for imbalances in EHS and ALBI in the HCV subset, OS HRs favoured STRIDE vs S. OS HRs continued to favour STRIDE when adjusting for EHS and ALBI in the other groups. Results for D vs S showed similar trends to those for STRIDE vs S (Table).Table: 714POSHR vs S; 95% CIOSCox-stratifieda HR vs S; 95% CIObjective response, %Disease control rate, %Median time to response, monthsMedian duration of response, monthsHBVSTRIDE n=1220.64; 0.48–0.860.64; 0.47–0.8621.359.01.925.7D n=1190.78; 0.58–1.040.78; 0.58–1.0414.349.61.99.5S n=1195.048.72.817.0HCVSTRIDE n=1101.06; 0.76–1.490.89; 0.63–1.2535.565.53.613.5D n=1071.05; 0.75–1.480.93; 0.66–1.3122.457.92.012.9S n=1049.670.27.315.7NVSTRIDE n=1610.74; 0.57–0.950.77; 0.59–1.00b18.057.12.113.2D n=1630.82; 0.64–1.050.80; 0.62–1.0319.056.43.713.8S n=1666.063.33.76.0aAdjusted for EHS and ALBI. bn=160; one pt with missing ALBI score excluded. Open table in a new tab aAdjusted for EHS and ALBI. bn=160; one pt with missing ALBI score excluded. In HIMALAYA, OS favoured STRIDE vs S (HR <1) for all aetiologies when subsets were adjusted for prognostic factor imbalances in the HCV cohort; similar trends were observed with D vs S across subsets. These results confirm the benefits of STRIDE and D in pts with uHCC, irrespective of underlying aetiology.