AI-Based Hepatic Fibre Phenotyping in MASLD: Hepatology | April 2026

Introduction

Metabolically–dysfunction–associated steatotic liver disease (MASLD) is a complex and heterogeneous condition where fibrosis plays a central role in disease progression and hepatocellular carcinoma (HCC) risk. Traditional histological fibrosis staging, although widely used, often fails to capture the nuanced microstructural and molecular heterogeneity of liver tissue. With advances in artificial intelligence and multi-omics, there is an opportunity to redefine fibrosis beyond conventional staging and link structural changes with biological pathways.

Problem Statement

Current fibrosis assessment lacks precision in predicting disease progression, molecular alterations, and HCC risk in MASLD. There is a critical need for advanced phenotyping tools that can integrate tissue architecture with molecular signatures to improve risk stratification and therapeutic targeting.

Summary

This study introduces an AI-based platform (FibroNest) that identifies detailed hepatic fiber morphologies and classifies them into distinct phenotypic components (FibroPCs). These AI-derived patterns were superior to conventional fibrosis staging in capturing key molecular pathways, including IL-6 signaling and therapeutic responsiveness (e.g., resmetirom). Notably, a specific phenotype (FibroPC4) was strongly associated with future HCC risk and revealed a unique microenvironment driven by interactions between hepatic stellate cells and periportal endothelial cells. This work establishes AI-driven fibrosis phenotyping as a powerful tool for precision hepatology, enabling better prediction of disease progression and targeted therapy in MASLD.