Comparative efficacy of cabozantinib and ramucirumab after sorafenib for patients with advanced hepatocellular carcinoma and AFP ≥ 400 ng/ml

Cabozantinib (CAB) and ramucirumab (RAM) are approved for the treatment of adults with hepatocellular carcinoma (HCC) following sorafenib (SOR); RAM is restricted for use in patients with a serum α-fetoprotein (AFP) ≥400 ng/ml. The matching-adjusted indirect comparison (MAIC) evaluated the efficacy and safety of both agents after SOR in HCC patients with AFP≥400 ng/ml.

The CELESTIAL (CAB) and REACH-2 (RAM) trials were used for the MAIC. Population-level data were available for REACH-2, individual patient data (IPD) for CELESTIAL. To align with REACH-2, CELESTIAL IPD were limited to patients who received first-line SOR only with AFP≥400 ng/ml and were weighted to balance the distribution of 11 effect-modifying baseline characteristics. Overall survival (OS) and progression-free survival (PFS) were compared for the matching-adjusted CELESTIAL and the REACH-2 populations using weighted Kaplan-Meier (KM) curves and parametric modeling (OS, Weibull; PFS, log-logistic). Rates of treatment discontinuation due to adverse events (AEs) were also compared.

After matching and weighting, baseline characteristics were balanced between trial populations (REACH-2, n = 292; CELESTIAL, effective sample size=105). Weighted KM estimates for median (95 % CI) OS and PFS were longer for CAB than for RAM, significantly for PFS: OS 10.6 (9.5-17.3) vs 8.7 (7.3-10.8) months (p=0.104); PFS 5.5 (4.6-7.4) vs 2.8 (2.7-4.1) months (p=0.016). Results from the parametric modeling mirrored the weighted KM analysis: CAB vs RAM; OS 12.0 (9.6-14.5) vs 9.6 (8.4-10.8) months; PFS 5.2 (4.1-6.4) vs 3.2 (2.8-3.6) months. There were no significant differences in OS or PFS estimates for the placebo arms suggesting that the matching and adjusting removed effect-modifying differences between the CELESTIAL and REACH-2 populations (hazard ratio [95 % CI]: OS 1.05 [0.65-1.68], p=0.854; PFS 0.94 [0.66 -1.33], p=0.713). There was no significant difference in rates of treatment discontinuation due to AEs in the matching-adjusted groups (p=0.271).

CAB significantly prolonged PFS, by an additional 2.7 months, compared with RAM after SOR in HCC patients with AFP≥400 ng/ml; AE-related treatment discontinuation rates were similar. MAIC analyses may help to guide decision-making in the absence of direct trial evidence.

Publication History

Article published online:
07 September 2021