Cost-Effectiveness of AI-Enabled Electrocardiography for Screening MASLD-Associated Advanced Chronic Liver Disease in the US Population.

BACKGROUND: MASLD is the most prevalent chronic liver disease worldwide, with pooled global prevalence estimates of approximately 30%. Current noninvasive pathways have improved risk stratification, but they depend on stepwise testing and follow-through in routine care. Against this background, artificial intelligence-enabled electrocardiography (AI-ECG) has emerged as a potentially scalable, opportunistic tool that could use an already familiar test to identify patients who may benefit from liver-specific assessment. METHODS: A decision-analytic Markov model was developed using TreeAge Pro Healthcare 2026 to evaluate screening strategies for MASLD-associated ACLD/cirrhosis in the U.S. POPULATION: Four strategies were compared: AI-ECG + TE, FIB-4 + TE, sequential AI-ECG + FIB-4 + TE, and no systematic screening. The model used annual cycles over 5-year and 10-year horizons. Analyses were conducted from the payer perspective. Costs and outcomes were accrued annually and discounted at 3% per year. RESULTS: Across all modeled scenarios, screening increased total costs and QALYs compared with no systematic screening over both 5-year and 10-year horizons. In the MASLD base case at 5 years, AI-ECG + TE had an ICER of $11,655 per QALY, while FIB-4 + TE and sequential AI-ECG + FIB-4 + TE had ICERs of $11,835 and $12,251 per QALY, respectively. At 10 years, AI-ECG + TE had an ICER of $10,615 per QALY, FIB-4 + TE had an ICER of $10,480 per QALY, and sequential AI-ECG + FIB-4 + TE had an ICER of $10,833 per QALY. Similar patterns were observed in the MASLD-T2D and MASLD-obesity scenarios. CONCLUSION: AI-ECG + TE demonstrated cost-effectiveness broadly comparable to FIB-4 + TE across modeled scenarios, without clear economic dominance. AI-ECG is best framed as a potential complementary first-step pathway for MASLD-associated ACLD/cirrhosis screening, rather than as a replacement for established noninvasive fibrosis algorithms.