Imaging Biomarkers for Diabetes Medications
Type 2 diabetes mellitus is a common disease affecting about 35 million adults and the seventh leading cause of death in the U.S. With its slow onset of symptoms over several years, its diagnosis is often delayed until complications have already arisen. This motivated a previous study in our lab, where we developed imaging biomarkers to detect type 2 diabetes mellitus on computed tomography (CT) scans, enabling opportunistic screening and potentially, early diagnosis of diabetes. These biomarkers include CT based measurements focused on the liver and the pancreas, including intrapancreatic fat percentage and the average liver CT attenuation. In this follow-up study, we intend to determine the long-term effects of different diabetes medications on the imaging biomarkers of diabetes. Under the supervision of a staff scientist, the student will use deep learning-based AI tools to segment organs of interest such as the liver and pancreas, and compute the relevant biomarkers on CT imaging from a large retrospective cohort of patients diagnosed with and being treated for diabetes. We will then conduct a longitudinal analysis on the interval change of biomarkers across multiple scans over time from individual patients. At the end of the internship, the intern will have a taste of the full computational research workflow, from data curation and using state-of-the-art AI tools, to hypothesis generation and statistical analysis. Hopefully, we will also have new insight into the trajectory of diabetes management from an imaging perspective.