Identification of plasma lipids as non-invasive signatures for metabolic dysfunction-associated steatotic liver disease

Abstract

Metabolic dysfunction-associated steatotic liver disease (MASLD) has emerged as the leading cause of chronic liver diseases worldwide, which may progress to cirrhosis and hepatocellular carcinoma (HCC). To address this growing public health concern, non-invasive diagnostic approaches are needed to identify individuals with MASLD for early detection. We used a liquid chromatography-mass spectrometry-based method to quantify plasma lipid species from 99 samples collected from individuals with MASLD (No-MASH, n = 35; MASH, n = 43) and healthy individuals ( n = 21) from All India Institute of Medical Sciences (AIIMS) and National Liver Disease Biobank (NLDB) at Institute of Liver and Biliary Sciences (ILBS), New Delhi, India, between 2021 and 2024. We performed statistical analysis to identify lipid signatures, which were confirmed by machine learning (ML) model (Random Forest). Here, we identified 10 lipid signatures for No-MASH and 8 lipid signatures for MASH to distinguish respective groups from healthy individuals, and additionally, 2 lipid signatures that distinguish individuals with No-MASH from MASH on the discovery set ( n = 69) with subsequent confirmation on the internal validation set ( n = 30) using statistical analysis. Using these signature lipids, ML model achieved combined AUC of 1.0 (95% CI: 1.00-1.00) for healthy versus No-MASH and healthy versus MASH comparisons, whereas an AUC of 0.97 (95% CI: 0.89-1.00) for No-MASH versus MASH comparison. This cross-sectional study highlights remodelling of the MASLD lipidome and demonstrates the potential of plasma-based lipids as non-invasive signatures for disease stratification upon future validation in independent cohorts.