Noninvasive identification of metabolic dysfunction-associated steatohepatitis (INFORM MASH): a retrospective cohort and disease modeling study.

Background: Using common clinical parameters, we aimed to noninvasively identify and predict metabolic dysfunction-associated steatohepatitis (MASH)/MASH with clinically significant fibrosis.

Research design and methods: Patients aged ≥18 with electronic health record (EHR) documented liver function tests and liver biopsies between 2016 and 2021 were retrospectively identified from the Geisinger Health System Research Liver Registry. MASH cases were confirmed using the nonalcoholic fatty liver disease (NAFLD) activity score. Training and validation datasets were used to create an algorithm/predictive model assessing EHR-derived predictors of MASH/MASH with clinically significant fibrosis (fibrosis stage F2-F4). Predictive accuracy was evaluated using the area under the curve.

Results: The analysis included 2698 patients. We created a composite likelihood score using variables significant for MASH and/or MASH with clinically significant fibrosis: liver enzymes (alanine aminotransferase [ALT], aspartate aminotransferase [AST]), prior year AST, metabolic disease, pulse (heart rate), and body mass index. The score had higher sensitivity and specificity for predicting MASH than Fibrosis-4 (FIB-4) Index, AST to platelet ratio index (APRI), and NAFLD fibrosis score (NFS); sensitivity and specificity were comparable to FIB-4 and APRI for predicting MASH with clinically significant fibrosis but superior to NFS.

Conclusion: The composite likelihood score could potentially be a tool for early MASH screening.