AI-Based Screening for At-Risk MASLD and Advanced Fibrosis in the US Population: A Cost-Effectiveness Analysis

Metabolic dysfunction-associated steatotic liver disease (MASLD) is highly prevalent and often underdiagnosed until advanced fibrosis, leading to increased morbidity and costs. While early identification can enable timely intervention, the cost-effectiveness of screening strategies for at-risk adults remains uncertain.

To evaluate the cost-effectiveness of an artificial intelligence (AI)-based risk stratification tool combined with transient elastography (AI + TE) compared with three alternative screening strategies and no screening for detecting advanced fibrosis in at-risk adults in the United States.

A decision-analytic Markov model simulated MASLD progression over 5-year and 10-year horizons from the US payer perspective, using a 3% discount rate. Four strategies were compared (AI + TE, FIB-4 + TE, TE-only, and no screening), followed by treatment with semaglutide or resmetirom for confirmed advanced fibrosis. Main outcomes were quality-adjusted life years (QALYs), total costs, and Incremental Cost-Effectiveness Ratios (ICERs).

Over 10 years, AI + TE versus no screening provided an incremental gain of 0.107 QALYs at an ICER of $38,916/QALY (semaglutide) and $76,141/QALY (resmetirom). FIB-4 + TE and TE-only were less efficient, with ICERs above $72,500 for semaglutide. At the $100,000/QALY threshold, the probability that AI + TE was cost-effective exceeded 90% for semaglutide. Semaglutide consistently dominated or was more cost-effective than resmetirom across all strategies.

AI-based screening followed by transient elastography is a cost-effective strategy for identifying advanced fibrosis in at-risk adults in the US, particularly when paired with semaglutide therapy. These findings support payer and policy consideration of AI-enabled screening programs.