Article PDF first page preview

Article PDF first page preview

Introduction: Given suboptimal performance of ultrasound-based surveillance for early hepatocellular carcinoma (HCC) detection in patients with cirrhosis, there is interest in alternative surveillance strategies, including blood-based biomarkers. We aimed to evaluate cost-effectiveness of biomarker-based surveillance in patients with cirrhosis. Methods: We constructed a decision-analytic model to compare ultrasound/alpha fetoprotein (AFP) and biomarker-based surveillance strategies in 1,000,000 simulated patients with compensated cirrhosis. Model inputs for adherence, benefits, and harms of each strategy were based on literature review, and costs were derived from the Medicare fee schedule. Primary outcomes were quality-adjusted life-years (QALY) and incremental cost-effectiveness ratio (ICER) of the surveillance strategies, with cost-effectiveness assessed at a threshold of $150,000 per QALY. We performed sensitivity analyses for HCC incidence, test performance characteristics, surveillance adherence, and biomarker costs. Results: In the base case, both ultrasound/AFP and biomarker-based surveillance were cost effective versus no surveillance, with ICERs of $105,620, and $101,295, per QALY, respectively. Biomarker-based surveillance was also cost-effective versus ultrasound/AFP, with an ICER of $14,800 per QALY. Biomarker sensitivity exceeding 80%, cost below $210, or adherence exceeding 58% were necessary for biomarker-based screening to be cost effective versus ultrasound/AFP. In two-way sensitivity analyses, biomarker costs were directly related with test sensitivity and adherence, whereas sensitivity and adherence were inversely related. In a probabilistic sensitivity analysis, biomarker-based screening was the most cost-effective strategy in most (65%) simulations. Conclusion: Biomarker-based screening appears cost-effective for HCC screening, but results are sensitive to test sensitivity, adherence, and costs.

This content is only available via PDF.