Abstract Non-alcoholic fatty liver disease (NAFLD) prevalence is estimated at >25% in the U.S., making it the most common cause of chronic liver disease. Non-alcoholic steatohepatitis (NASH), the progressive form of NAFLD, affects 1-3% of the U.S. population and is expected to double by 2030. Currently, there is no effective pharmacotherapy for NAFLD, but there are numerous promising drug candidates being evaluated in clinical trials. However, the primary endpoint for these trials, histologic fibrosis, has shortcomings including sampling error and use of a subjective five category scale to quantify a continuous variable. Our team has pioneered the use of 3D open-top light-sheet (OTLS) microscopy, which enables rapid, high-throughput imaging of large clinical samples. In combination with cutting-edge machine learning techniques, we hypothesize that 3D OTLS microscopy can provide more accurate and consistent assessment of fibrosis in liver biopsies from NASH patients. We will test this hypothesis by developing a multiplex staining protocol and machine learning analysis pipeline, which will be piloted on 20 archived FFPE liver biopsies. Results from our assay will be correlated with currently used primary and secondary outcome measures to motivate further studies with sing archived samples from clinical trials with responsivity and outcomes data.