Multiple DP/AI Quantification approaches are complimentary in identifying histology features: cross sectional study in a Japanese cohort

MASH (metabolic dysfunction–associated steatohepatitis) drug development currently requires pathologist-read, biopsy-based evidence of improvement in fibrosis without worsening of steatohepatitis and/or resolution of steatohepatitis with no worsening of liver fibrosis.

Liver biopsy is the reference standard for screening and efficacy in MASH drug development, with CRN staging of fibrosis and NAS features (steatosis, inflammation, and ballooning) as key components.

However, pathologist-read scoring systems have many limitations, such as inter- and intra-reader variability with ordinal scoring systems. This has led to the establishment of many digital pathology (DP) approaches to evaluate fibrosis and disease activity in MASH.

Here, we evaluated a direct comparison of multiple digital pathology approaches (PathAI, HistoIndex, and HALO) to measure liver fibrosis, steatosis, ballooning, and inflammation in a Japanese cohort.