Diabetes Metab Res Rev. 2026 Feb;42(2):e70129. doi: 10.1002/dmrr.70129.
ABSTRACT
Metabolic Dysfunction-associated Steatotic Liver Disease (MASLD) is a prevalent liver disease worldwide, with its prevalence rising alongside the increase in metabolic syndrome (MetS), obesity and ageing. Machine learning (ML), as a powerful analysis tool to handle and analyse massive data/information, has been employed to enhance and refine the diagnosis, risk assessment, non-invasive screening, and treatment options against MASLD. This review thoroughly explores the application of ML in identifying MASLD-related genes and lipidomic biomarkers, non-invasive screening technologies such as ultrasound and imaging, and predicting the risk of disease progression to metabolic dysfunction-associated steatohepatitis (MASH) or more advanced stages, such as cirrhosis. Additionally, ML models have shown potential and definitive performance in accurately predicting and effectively managing the risk of comorbidities in relation to MASLD. By integrating clinical data, biochemical markers, imaging techniques, and an individual's biochemical metrics, ML offers a personalised medical approach that improves therapeutic strategies and holds promise for significant contributions to public health in the future.
PMID:41631499 | DOI:10.1002/dmrr.70129