Performance of non-invasive biomarkers in identifying patients with at-risk metabolic dysfunction-associated steatohepatitis

Motivation: MRE-assessed liver stiffness, MRI-derived fat fraction, AST, and FIB-4 are promising predictors for identifying at-risk MASH, which is defined as MASH with fibrosis stage 2 or higher. Some investigators have also proposed an additional criterion, NAS ≥ 4 with at least 1 point in each category. Goal(s): Assess prediction models for diagnostic accuracy in identifying at-risk MASH under both definitions. Approach: AUC and its 95% confidence intervals were used to assess the overall diagnostic accuracy. Results: The additional criterion influenced the performance of prediction models, but liver stiffness remains the most reliable single predictor among non-invasive biomarkers. Impact: At-risk MASH is defined as a MASH diagnosis with fibrosis stage ≥2. Adding a criterion of NAS≥4 with ≥1 point in each category impacts diagnostic accuracy but liver stiffness remains the most reliable single predictor.