Identification of potential metabolic biomarkers and immune cell infiltration for metabolic associated steatohepatitis by bioinformatics analysis and machine learning

Background: Metabolic associated steatohepatitis (MASH) represents a severe subtype of metabolic associated fatty liver disease (MASLD), with an increased risk of progression to cirrhosis and hepatocellular carcinoma. The nomenclature shift from nonalcoholic steatohepatitis (NASH)/nonalcoholic fatty liver disease (NAFLD) to MASH/MASLD, underscores the pivotal role of metabolic factors in disease progression. Diagnosis of MASH currently hinges on liver biopsy, a procedure whose invasive nature limits its clinical utility. This study aims to identify and validate metabolism-related genes (MRGs) markers for the non-invasive diagnosis of MASH. Methods: This study extracted multiple datasets from the GEO database to identify metabolism-related differentially expressed genes (MRDEGs). Protein-Protein Interaction (PPI) network and machine learning algorithms, including Least Absolute Shrinkage and Selection Operator (LASSO) regression, Support Vector Machine-Recursive Feature Elimination (SVM-RFE), and Random Forest (RF), were applied to screen for signature MRDEGs. The diagnostic performance of these MRDEGs was evaluated using the Receiver Operating Characteristic (ROC) curve and further validated using independent external datasets. Additionally, enrichment analysis was performed to uncover key driver pathways in MASH. The infiltration levels of various immune cell types were assessed using single sample Gene Set Enrichment Analysis (ssGSEA). Finally, Spearman correlation analysis confirmed the association between signature genes and immune cells. Results: We successfully identified seven signature MRDEGs, including CYP7A1, GCK, AKR1B10, HPRT1, GPD1, FADS2, and ENO3, through PPI network analysis and machine learning algorithms. The gene model displayed exceptional diagnostic performance in the training and validation cohorts, as evidenced by the area under ROC curve (AUC) exceeding 0.9. Further enrichment analysis revealed that signature MEDEGs were primarily involved in multiple biological pathways related to glucose and lipid metabolism. Immune infiltration analysis indicated a significant increase in the infiltration levels of activated CD8 T cells, gamma-delta T cells, natural killer cells, and CD56bright NK cells in patients with MASH. Conclusion: This study successfully identified seven signature MRDEGs as significant diagnostic biomarkers for MASH. The findings not only offer novel strategies for non-invasive diagnosis of MASH but also highlight the substantial role of immune cell infiltration in the progression of MASH.