In the RESORCE trial, regorafenib was shown to provide overall survival (OS) benefit for patients with hepatocellular carcinoma (HCC) that has progressed on sorafenib treatment. Subsequently, it was approved by the Therapeutic Goods Administration for the treatment of patients with HCC who were previously treated with sorafenib; however, regorafenib is still not recommended by the Pharmaceutical Benefits Advisory Committee in Australia. We aimed to assess the cost effectiveness of regorafenib as a second-line therapy for patients with HCC who progressed on sorafenib from an Australian healthcare perspective.

We developed a Markov model to compare the cost effectiveness of regorafenib with best supportive care (BSC) as a second-line therapy for HCC after treatment with sorafenib. The health outcomes of life-years and quality-adjusted life-years (QALYs) were derived from the RESORCE trial. Survival benefits sourced from the RESORCE trial were fitted with the parametric model to estimate survival beyond the follow-up period. Drug costs and costs associated with adverse events (AEs) were sourced from published literature and the Independent Health and Aged Care Pricing Authority cost report. Model validity was verified using probabilistic sensitivity analyses.

The incremental monthly cost of treatment with regorafenib was AUD19,273 (USD13,374), with an incremental life-year gain of 0.38, compared with BSC. The incremental QALYs gained with regorafenib were 0.24, resulting in a base-case incremental cost-effectiveness ratio (ICER) of AUD80,511 (USD55,872) per QALY. In the probabilistic sensitivity analyses across scenarios, the ICER remained above the conventional threshold of AUD50,000 (USD34,698) per QALY, with a zero probability of being cost effective at this willingness-to-pay threshold.

At the current price, second-line treatment with regorafenib in patients with HCC that has progressed on sorafenib was not cost effective at the conventional willingness-to-pay threshold from an Australian health-system perspective.