PROJECT SUMMARY
Ventricular arrhythmias remain the leading cause of death in patients with cardiomyopathy and account for up
to 300,000 deaths per year in the United States. However, the current classifications of these rhythms is based
largely on whether the cardiomyopathy is due to obstructed coronary arteries and poorly stratifies patient
response to therapy, arrhythmic risk, and pathophysiology.
The goal of this project is to develop an actionable classification scheme for ventricular arrhythmias in patients
with cardiomyopathy that is based on the interplay between both structural and electrical abnormalities
measured from each patient’s heart. Such a classification, based in measurements of pathophysiology, would
inform the clinical approach to risk assessment, interventional therapies, and medications.
The proposal outlines three Specific Aims: 1) To identify electrical fingerprints of endocardial, mid-myocardial
and epicardial scar using machine learning of endocardial high-density contact electrograms trained to the
ground truth of regional delayed gadolinium fibrosis on magnetic resonance imaging, from our large patient
registry. 2) To develop and validate a mapping strategy that could be used at clinical electrophysiology to
measure ventricular refractory period, a measure of electrical remodeling that indicates ability to sustain VA, by
machine learning of high density electrical signals from the heart of a porcine model labeled by repolarization
indices from the gold standard, simultaneously recorded, monophasic action potentials. And 3) To derive novel
phenotypes of arrhythmogenic cardiomyopathy in patients with VA based on regional distributions of fibrosis
and electrical remodeling, and associate these with acute response to ablation and recurrence in a well-
characterized patient registry.
To successfully complete the proposed project, training objectives include 1) advanced MRI processing and
segmentation, 2) machine learning models for multimodal data analysis, 3) translational interventional
procedures, and 4) translational clinical electrophysiology. The proposed NHLBI K23 award will provide
protected time for the candidate to obtain this advanced training, to disseminate new knowledge via written and
spoken communication, and to build the foundation for an independent research program focused on
ventricular arrhythmia diagnosis, prevention, and therapy in a supportive environment of established
mentorship, collaborators, and interdisciplinary experts spanning engineering and medicine.