Project Summary
Sudden cardiac arrest is one of the leading causes of death in developed countries, accounting for
approximately 350,000 deaths per year in the United States. The majority of those events are caused by
ventricular arrhythmias (VA). Implantable defibrillators reduce mortality in high-risk patients, but do not
prevent recurrent arrhythmias. Suppression of recurrent ventricular tachycardia (VT) can be accomplished
effectively with catheter ablation; more recently stereotactic body radiotherapy (SBRT) has been shown to
have a potential role. Accurate identification of the substrate responsible for the VA is key to the success of
either of these modalities and may be facilitated using the standard 12-lead ECG to noninvasively identify
the site from which a focal ventricular tachycardia (VT)/premature ventricular complex (PVC) arises or from
which a reentrant circuit exits the central isthmus to activate the “normal” myocardium. Currently, there is
not an automatic real-time non-invasive patient-specific approach that can be used to accurately identify the
site of origin (SoO) of VA using the 12-lead ECG.
Rapid 12-lead ECG interpretation to identify the SoO of VA requires expertise and could be facilitated
with a computerized method to automatically locate the VT exit/PVC origin site in real-time. The ability to
accurately identify the VT exit/PVC origin site enables the electrophysiologist to concentrate
mapping/targeting to a specific region. To tackle this problem, this research proposes to develop a novel
non-invasive 3D mapping technique that relies on the assembly of personalized ventricular surfaces from
CT/MRI scans in combination with a statistical estimate derived from a large clinical database to accurately
identify the VT exit/PVC origin site from an induced/recorded VT/PVC ECG in real time. The project is
interdisciplinary as it combines expertise in biomedical engineering, clinical cardiac electrophysiology, ECG
signal processing, image processing, and computational statistical modeling. To this end, the project will
include the following two activities: 1) to develop the proposed system in clinically usable software; 2) to
assess the accuracy of the proposed software in a prospective case-series study (with the goal of achieving
a mean localization error of less than 10 mm).
The proposed software delivered by this project will provide significant accuracy improvement in the VT
exit/PVC origin site localization, potentially decrease in the time of an invasive VA ablation procedure, and
would be helpful to accurately target VT for non-invasive cardiac SBRT. The proposed project is innovative
in proposing to bring computational statistical modeling that integrates structure data (CT/MRI imaging),
function data (ECG), and a large clinical dataset into the realm of contemporary patient care. At its core, the
project is of translational nature, with personalized computational statistical modeling being used for
guidance of clinical therapies.
