Machine Learning Fibrosis Models Based on Liver Histology Images Accurately Characterize the Heterogeneity of Cirrhosis Due to Nonalcoholic Steatohepatitis

♦ Nonalcoholic steatohepatitis (NASH) cirrhosis is
characterized by heterogeneity in histology, clinical
presentation, and prognosis1

♦ In patients with NASH, the presence of cirrhosis is
associated with an increased risk of liver-related
and all-cause mortality, but identification of those
at risk for liver-related clinical events may be
challenging2-4

♦ Although machine learning (ML) approaches have
been used to evaluate liver histology in NASH,5,6
the utility of these approaches to characterize
fibrosis and risk stratify patients with cirrhosis
requires evaluation