Developing Personalized Treatment for Diabetes Through Novel Imaging Methods

University Of Michigan At Ann Arbor, Ann Arbor MI

Investigators

Linked publications & trials

Abstract

Project Summary/Abstract Type 2 diabetes mellitus (T2DM) is a heterogeneous disease, resulting from genetic susceptibility, environment, lifestyles, and other factors. Heterogeneity in T2DM poses significant difficulties in designing personalized prognostic models and treatment algorithms. In this proposal, the candidate seeks to leverage vast amounts of additional quantitative data in computed tomography (CT) images with a new analysis method, Analytic Mor- phomics. Existing imaging studies had small sample sizes or only analyzed few variables due to low- throughput methods. In contrast, analytic morphomics we have developed is a novel high-throughput process to obtain granular intricacies about the morphological characteristics through CT images. It can greatly ad- vance the current imaging research in T2DM. Analytic morphomics can distinguish subtle nuances that lead to heterogeneity in clinical presentation and pathophysiologic processes of T2DM, and thus can signi?cantly im- prove the current risk prediction models over other `omics. The candidate has strong training in statistics and computations, and will leverage this proposal to broaden his knowledge in the foundational medicine, especial- ly in obesity and diabetes, to facilitate his development into an independent clinical scientist and designing per- sonalized prognostic tools for diabetes and its related complications on a broad scale. The research plan has two Aims. Aim 1: To quantitatively determine the impact of diabetes on musculoskeletal quality and adiposity distribution through analytic morphomics. He will conduct a cross-sectional investigation and determine inde- pendent differences in musculoskeletal quality and adiposity distribution by comparing 1251 patients with T2DM vs 1709 non-diabetic controls. Aim 2: To determine the independent associations between analytic morphomic factors and clinical outcomes in diabetes, such as biochemical markers and disease progression. He will perform a historical cohort study to determine the association between analytic morphomics, processed from patients' past CT scans, and their current medical outcomes, retrieved from patient electronic medical records. These studies will have an important impact on the design of individualized clinical and public health interventions for diabetes. Quantifying the exact quantitative changes in key body elements such as bone/muscle quality and adipose tissues can provide a better understanding of T2DM pathophysiology, and can also improve the screening for and management of diabetes. More accurate personalized prognostic mod- els can dramatically a?ect the diagnosis, classi?cation, and management of diabetes. Developing personalized treatment plans would have a large impact on clinical practice, which has the power to improve the treatment guidelines and deliver the most suitable treatments to patients. Analytic morphomics enables us to extract highly predictive and high-dimensional data from medical images and facilitate the development of better per- sonalized medicine. It is scalable and relevant to a broad range of clinical conditions, and will help to reduce healthcare burden for T2DM, one of the most costly diseases.