MATCHING-ADJUSTED INDIRECT COMPARISONS OF NIVOLUMAB + IPILIMUMAB VS DURVALUMAB + TREMELIMUMAB FOR FIRST-LINE UNRESECTABLE HEPATOCELLULAR CARCINOMA: ASIAN SUBGROUP RESULTS
Author(s)
Kiran Dave, MSc1, Aidan N. Franklin, BS2, Jessica Thackway, MSc3, Morgan Smith, PharmD, BCOP4, Qiong Lou, PhD5, Tristan Curteis, MSc6.
1Bristol Myers Squibb, Uxbridge, United Kingdom, 2Costello Medical, Boston, MA, USA, 3Costello Medical, London, United Kingdom, 4Bristol Myers Squibb, Lawrence Township, NJ, USA, 5Bristol Myers Squibb, Shanghai, China, 6Costello Medical, Manchester, United Kingdom.
1Bristol Myers Squibb, Uxbridge, United Kingdom, 2Costello Medical, Boston, MA, USA, 3Costello Medical, London, United Kingdom, 4Bristol Myers Squibb, Lawrence Township, NJ, USA, 5Bristol Myers Squibb, Shanghai, China, 6Costello Medical, Manchester, United Kingdom.
OBJECTIVES: In the CheckMate 9DW trial, the combination immunotherapy (IO) nivolumab + ipilimumab (N+I) demonstrated significantly improved overall survival (OS) vs sorafenib/lenvatinib in patients with unresectable hepatocellular carcinoma (uHCC). In the absence of head-to-head trials for N+I vs durvalumab + tremelimumab (D+T), we compared their efficacy via matching-adjusted indirect comparisons (MAICs). Given higher prevalence of hepatitis B virus (HBV)-related uHCC in Asian vs Western populations, we evaluated comparative efficacy in patients enrolled in Asia.
METHODS: MAIC analyses compared Asian subgroups from CheckMate 9DW and HIMALAYA trials. Individual patient data from the CheckMate 9DW N+I arm was weighted to match aggregate baseline characteristics of the HIMALAYA D+T arm within Asian patient subgroups (excluding Japan). Given control arm differences, pairwise unanchored analyses were conducted, adjusting for treatment effect modifiers and prognostic factors including age, sex, and HBV/HCV. Cox proportional hazards models generated hazard ratios (HRs) for OS for N+I vs D+T. For OS, piecewise analyses generated HRs for follow-up periods of 0-6 months and post 6-months. Logistic regression models generated odds ratios (ORs) for objective response rate (ORR).
RESULTS: In Asian subgroups, MAIC analyses of OS numerically favoured N+I vs D+T in the overall period (HR: 0.74 [0.52, 1.04]), but this did not reach statistical significance. In piecewise OS analyses, N+I was numerically favoured vs D+T in 0-6 month (HR: 0.59 [0.32, 1.12]) and ≥6 month intervals (HR: 0.80 [0.53, 1.22]). In ORR analyses, N+I was numerically favoured vs D+T (OR: 1.47 [0.86, 2.50]). MAIC analyses of the trial Asian subgroups were directionally consistent with the intention-to-treat (ITT) population.
CONCLUSIONS: With the caveats associated with limitations of unanchored subgroup results, these unanchored MAIC analyses in Asian subgroups were directionally consistent with the ITT population, in which potential OS and ORR benefits of N+I vs D+T for first-line uHCC treatment were previously demonstrated.
METHODS: MAIC analyses compared Asian subgroups from CheckMate 9DW and HIMALAYA trials. Individual patient data from the CheckMate 9DW N+I arm was weighted to match aggregate baseline characteristics of the HIMALAYA D+T arm within Asian patient subgroups (excluding Japan). Given control arm differences, pairwise unanchored analyses were conducted, adjusting for treatment effect modifiers and prognostic factors including age, sex, and HBV/HCV. Cox proportional hazards models generated hazard ratios (HRs) for OS for N+I vs D+T. For OS, piecewise analyses generated HRs for follow-up periods of 0-6 months and post 6-months. Logistic regression models generated odds ratios (ORs) for objective response rate (ORR).
RESULTS: In Asian subgroups, MAIC analyses of OS numerically favoured N+I vs D+T in the overall period (HR: 0.74 [0.52, 1.04]), but this did not reach statistical significance. In piecewise OS analyses, N+I was numerically favoured vs D+T in 0-6 month (HR: 0.59 [0.32, 1.12]) and ≥6 month intervals (HR: 0.80 [0.53, 1.22]). In ORR analyses, N+I was numerically favoured vs D+T (OR: 1.47 [0.86, 2.50]). MAIC analyses of the trial Asian subgroups were directionally consistent with the intention-to-treat (ITT) population.
CONCLUSIONS: With the caveats associated with limitations of unanchored subgroup results, these unanchored MAIC analyses in Asian subgroups were directionally consistent with the ITT population, in which potential OS and ORR benefits of N+I vs D+T for first-line uHCC treatment were previously demonstrated.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
CO169
Topic
Clinical Outcomes, Study Approaches
Topic Subcategory
Comparative Effectiveness or Efficacy
Disease
Gastrointestinal Disorders, Oncology