AI digital pathology as a key tool providing in-depth understanding of the progression and regression of MASH and fibrosis in male mouse models

Abstract Optimizing murine models for studying Metabolic Dysfunction-Associated Steatohepatitis (MASH) and fibrosis is crucial for understanding disease mechanisms and evaluating therapies. In this study, we characterized diet-induced male murine models of MASH and liver fibrosis using an AI-digital pathology (DP) pipeline for zone-specific analysis and spatial (co)-localization of key MASH features within the liver microarchitecture, providing insights beyond standard histology. Model characterization included an integrative omics-based approach, standard blood-chemistry, and traditional histology. AI-DP revealed previously unknown temporal events leading to MASH, emphasizing the interplay between inflammation and dysmetabolism in disease progression. We also noted distinct morphometric characteristics of granulomas and their correlation with fibrosis. In efficacy studies of clinically-validated treatments, Semaglutide (GLP-1RA), Resmetirom (THRβ-agonist), and MK-4074 (Acetyl-CoA-carboxylase inhibitor), AI-DP demonstrated differential effects on macrosteatosis, microsteatosis, and colocalized fibrosis. Overall, integrating AI enables identification of fit-for-purpose disease models for therapeutic testing, and facilitates robust preclinical study designs for advancing effective therapeutic strategies.