Rohan Narendra Dhamdhere
BME PhD Defense Presentation

Date: 2026-11-25
Time: 1:00 pm 
Location / Meeting Link: HSRB 2 Floor 6 Room N600. Zoom link - https://emory.zoom.us/j/93587995605

Committee Members:
Anant Madabhushi, PhD (advisor); Sadeer Al-Kindi, MD; Hanjoong Jo, PhD; Lakshmi P. Dasi, PhD; Ran Xiao, PhD.


Title: Computational Characterization of Subclinical Cardiovascular Pathophysiology to Inform Preventive Care

Abstract:
Cardiovascular disease often develops through a prolonged subclinical phase, during which changes in cardiac structure and function accumulate before the onset of symptoms or clinical events. Patients are frequently assessed during this phase through imaging and electrocardiography acquired as part of routine care. A result within the normal range on these tests is generally considered clinically unremarkable, yet patients with such results may nonetheless harbor residual cardiovascular risk unaccounted for by conventional assessment. The underlying images and signals, however, encode information beyond the summary measurements reported in clinical practice. This dissertation investigates computational phenotypes, defined here as quantitative descriptors of that information. The central hypothesis of this work is that these computational phenotypes describing cardiovascular structure and function, and their change over time, provide prognostic information incremental to conventional clinical risk assessment, including among patients with normal test results. Aim 1 establishes the computational methods required for the derivation of these phenotypes. Aim 2 evaluates the prognostic value of retinal and echocardiographic phenotypes for major adverse cardiovascular events in patients with chronic kidney disease, and of splenic phenotypes derived from calcium-scoring computed tomography for incident heart failure. After heart transplantation, serial electrocardiograms, biopsy findings and clinical assessments constitute a longitudinal record that permits extension of the same phenotyping framework to temporal change. Rejection of the transplanted heart may arise in the intervals between scheduled biopsies. Aim 3 therefore evaluates phenotypes of change across serial visits for the noninvasive stratification of allograft rejection risk, with the objective of reducing reliance on invasive endomyocardial biopsy.