The prevalence of chronic liver disease (CLD) is increasing in the Veterans Health Administration. Access to specialty care is important for patients with CLD, and hence accurate risk stratification algorithms are needed for consult management.

Indexes such as Child-Turcotte-Pugh (CTP) score from electronic medical records have been utilized but only performed well for patients with decompensated liver disease. Imaging biomarkers, which reflect body composition and frailty, have been shown to be strongly associated with mortality.

In this talk, we will introduce how we analyze CT images and extract biomarkers automatically through modern deep-learning algorithms. We then utilize these imaging biomarkers together with clinical data to predict mortality.


Peng Zhang, Ph.D. is an Associate Research Scientist in the Department of Surgery and Director of Advanced Analytics at the Morphomic Analysis Group. Dr. Zhang received his undergraduate degree in Mathematics from Peking University in Beijing, China, and his doctoral degree in Biostatistics from Harvard University. His primary research interest is to design personalized treatment decision support and strategies through analytic morphomics, which are anatomically indexed information extracted from Computed Tomography (CT) images. Dr. Zhang's research interests include an array of clinical outcomes, such as injuries in motor vehicle crashes, long/short-term surgical outcomes, cancer occurence/recurrence (lung, adrenal, liver, esophageal, colon), and metabolic diseases (type-2 diabetes and non-alcoholic fatty liver disease).

Sameer Saini, MD, MS is a Professor in the Division of Gastroenterology at University of Michigan and Director of the VA Ann Arbor HSR&D Center for Clinical Management Research (CCMR). He also serves as Co-Director of the NIH T32 GI Epidemiology Training Program and Director of the Gastroenterology Early Career Faculty Mentoring Program at the University of Michigan. His research focuses on the appropriate use of resource-limited medical procedures, such as colonoscopy and upper endoscopy, the development of “next-generation” performance measures, and the development and evaluation of complex interventions to reduce the use of low-value care. He uses a variety of methodological approaches in his work, including simulation modeling, mixed methods, and implementation science. He has a long list of current and former post-doctoral mentees.


The seminar series “Providing Better Healthcare through Systems Engineering” is presented by the U-M Center for Healthcare Engineering and Patient Safety (CHEPS): Our mission is to improve the safety and quality of healthcare delivery through a multi-disciplinary, systems-engineering approach.

For the Zoom link and password, and to be added to the weekly e-mail for the series, RSVP or contact genehkim@umich.edu