AI-Based Discovery of Exosomal miRNA Markers for Metabolic Dysfunction-Associated Steatotic Liver Disease

Identifying Novel Biomarkers for MASH Diagnosis with AI-Driven Analysis of EV miRNAs

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a progressive condition marked by inflammation, potentially advancing to cirrhosis and liver failure in its steatohepatitis (MASH) stage.

Monitoring the progression from MASLD to MASH is crucial, but effective non-invasive alternatives to liver biopsy are still lacking.

This study sought to identify differentially expressed extracellular vesicle (EV) miRNAs in MASH patients and validate their potential as novel biomarkers for hepatic fibrosis diagnosis using an artificial intelligence (AI)-based approach.

Serum EV Isolation and miRNA Expression Analysis

Sample Collection:

Serum-derived extracellular vesicles (EVs) were isolated from:

40 biopsy-confirmed MASH patients.

20 healthy controls.

miRNA Analysis:

EV miRNAs were analyzed using:

3D-Gene human miRNA oligo chips.

Next-generation sequencing (NGS).

Data Analysis:

Concordance between fibrosis stages (New Inuyama or Brunt) and miRNA expression profiles was assessed using:

Multiple LASSO regression models.

Ordinal logistic regression models.

EV miRNAs as Biomarkers for Fibrosis and Brunt Staging in MASH

Fourteen extracellular vesicle (EV) miRNAs were identified from MASH patient analyses, demonstrating high accuracy in determining fibrosis and Brunt stages. Six differentially expressed EV miRNAs were associated with Brunt scores.

Ordinal union analysis revealed strong concordance between miRNA-based determination and fibrosis/Brunt stages (kappa statistics: New Inuyama = 0.87; Brunt = 0.99).

Pathway analysis indicated that these miRNAs primarily regulate inflammatory pathways involving myofibroblast activation and cytokine production, including IFN-γ, hepatocyte growth factor (HGF), TGF-β, IL-1β, IL-6, and IL-8.

Key Insight

This study demonstrates that specific EV miRNAs, identified through robust analytical models, can accurately differentiate stages of liver disorder progression from MASLD to MASH. These findings underscore the potential of EV miRNA expression levels as novel, non-invasive biomarkers for predicting liver pathology and advanced fibrosis.

By - Eeshan Aggarwal

Reference: Hepatology. Volume 80, Issue S1. Abstract Supplement for The Liver Meeting by the American Association for the Study of Liver Diseases (AASLD), November 15-19, 2024, San Diego, CA.