Rohan Dhamdhere
BME PhD Proposal Presentation

Date: 2026-02-04
Time: 10:00 AM to 11:00 AM
Location / Meeting Link: In-person location: HSRB2, 6th Floor Room S-645 Zoom Meeting Link: https://emory.zoom.us/j/96271688220

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


Title: Enabling Opportunistic Preventive Care: Computational Characterization of Subclinical Cardiovascular Pathophysiology from Routine Clinical Assessments

Abstract:
Cardiovascular morbidity and mortality in high-risk populations are driven by discrete events that are typically preceded by months-to-years of subclinical remodeling in the myocardium and vasculature. Prevention is most effective during this early window, yet routine workflows often detect disease only after dysfunction becomes apparent on coarse clinical metrics, while advanced imaging and molecular testing remain difficult to deploy at scale. This PhD proposal seeks to enable earlier prevention by computationally detecting and quantifying subclinical change from routinely acquired clinical data. A central objective is to develop three independent, modality-specific pipelines that generate three distinct signatures from retinal fundus photographs (microvascular architecture), echocardiography (cardiac structure and function), and longitudinal ECGs (electrophysiologic remodeling). Each pipeline will include consistent preprocessing, data quality checks, and signature construction designed to perform well across the variation seen in routine clinical acquisitions. These signatures will characterize complementary aspects of early cardiac change, while also supporting principled multimodal integration when it aligns with clinical workflow and a combined estimate would be clinically useful. Evaluation is planned in two high-risk cohorts with clinically actionable endpoints: (i) chronic kidney disease, where retinal and echocardiographic phenotypes will be assessed for prognostic utility for major adverse cardiovascular events and incremental value beyond routine variables and established risk calculators; and (ii) heart transplantation, where delta-ECG signatures referenced to an individual baseline will be assessed for early detection of biopsy-confirmed rejection and near-term rejection risk across surveillance horizons. The proposed research will deliver validated imaging- and signal-based risk signatures that identify subclinical disease in routine assessments, supporting timely preventive intervention without increasing testing burden.