Precision medicine and AI: future treatment and diagnostic strategies for MASLD

Abstract

Artificial intelligence (AI) has transformed the diagnosis and management of metabolically dysfunction-associated fatty liver disease (MASLD). Machine learning models integrate clinical data, biochemical biomarkers, and imaging to non-invasively detect and stratify the disease, with algorithms achieving AUCs as high as 0.91 for advanced fibrosis³.

Precision medicine incorporates multi-omics data—genomics, lipidomics, transcriptomics, proteomics—to identify biomarkers of early mechanistic risk and design personalized interventions^1,2. AI also enables mapping the systemic interconnection between MASLD, insulin resistance, and type 2 diabetes, improving cardiovascular and renal risk stratification⁴. The use of explainable methods such as SHpley Additive exPlanations (SHAP) increases clinical confidence in these models¹.

Key challenges include high data dimensionality, cohort heterogeneity, and the need for external validation in diverse populations to ensure generalizability^2,3. The transition to causal inference models and the integration of dynamic monitoring into clinical decision support algorithms represent the most promising prospects^1,2.