Machine Learning Models Identify Novel Histologic Features Predictive of Clinical Disease Progression in Patients With Advanced Fibrosis Due to Nonalcoholic Steatohepatitis

♦ Fibrosis is the primary determinant of disease progression
in patients with nonalcoholic steatohepatitis (NASH), but the
prognostic value of other histologic features is unclear

♦ Human pathologist staging of fibrosis and NAFLD Activity
Score (NAS) are limited by sampling variability, and intra- and
inter-reader variability

♦ Machine learning (ML) approaches to interpretation of liver
histology may enable more reliable and quantitative assessment
of both traditional and novel histologic features, with potential
prognostic relevance in NASH