Plasma proteomic profiles predictive of metabolic dysfunction-associated steatotic liver disease and severe liver outcomes

The circulating proteome could improve the prediction of metabolic diseases and offer insights into the biological mechanisms. The role of plasma proteomics in the development and progression of chronic liver disease is less clear. The analysis included 43,844 participants from the UK Biobank. Multivariable Cox regression models were applied to assess the associations of 2923 unique plasma proteins with incident metabolic dysfunction-associated steatotic liver disease (MASLD), over a median follow-up of 12.7 years. The significant proteins were then fed into a trained light gradient boosting machine classifier to identify the key proteins predictive of MASLD. Receiver operating characteristic analyses were conducted to evaluate the performance of these important proteins in predicting incident MASLD and severe liver outcomes (i.e., fibrosis and cirrhosis of liver, hepatocellular carcinoma, and liver-related mortality). There were 215 proteins associated with incident MASLD after Bonferroni corrections, among which ACY1, PON3, APOF, and IL1RN ranked the highest in protein importance. A panel of these top proteins showed AUCs of 0.758 for predicting MASLD and 0.746 for severe liver outcomes, and significantly improved prediction of both endpoints beyond traditional risk factors (both P values < 0.01). Notably, ACY1 alone, compared with the fatty liver index, showed comparable predictive ability for MASLD (AUC: 0.712 vs. 0.719) and severe liver outcomes (AUC: 0.767 vs. 0.744). Plasma protein biomarkers, particularly ACY1, significantly enhanced the prediction of MASLD and severe liver outcomes, which may facilitate early identification of high-risk individuals and support personalized risk stratification and future management strategies.