Automated segmentation and quantification of histological liver features for MASH/MASLD scoring

Abstract The increasing global prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD) including metabolic dysfunction-associated steatohepatitis (MASH) creates an urgent need for objective histopathological assessment. Conventional histological approaches are time-consuming and subjective. Therefore, we propose a fully automated, deep-learning-based pipeline for the segmentation and characterisation of histological liver features of MASH/MASLD. Segmentation was applied to H&E sections from 45 mice and 44 humans with MASH/MASLD. Our qHisto (quantitative histology) method, is nnU-Net-based and quantifies key histological components of the MASH-score, including steatosis, fibrosis, inflammation and ballooning. Additionally, we determined novel features that are inaccessible through manual histology, like the distribution of fat droplet sizes, nuclear aspect ratio and heatmaps. qHisto parameters showed strong positive correlations with conventional histology scores (fat: R = 0.91, inflammation: R = 0.70, ballooning: R = 0.47) and with hepatic fat fraction by magnetic resonance imaging ( R = 0.87). Our novel scores showed that nuclei deformation is driven by large fat droplets rather than the overall amount of fat. A key advantage of our method is spatially resolved, precise histological quantification to provide detailed assessment of disease severity. By automating time-consuming and repetitive readouts, qHisto improves standardisation and reproducibility of MASH/MASLD feature quantification and provides scalable, slide-wide analysis that supports histopathologists and enhances clinical assessment.