Personalized Heart Digital Twins Detect Substrate Abnormalities in Scar-Dependent Ventricular Tachycardia

Area of Science:

• Cardiovascular Electrophysiology
• Medical Imaging
• Computational Modeling

Background:

• Catheter ablation for scar-dependent ventricular tachycardia (VT) has limited success due to high recurrence rates.
• Personalized heart digital twin technology offers a potential noninvasive method for predicting VT substrate.
• The accuracy of digital twins in identifying critical VT substrate abnormalities requires investigation.

Purpose of the Study:

• To prospectively analyze the accuracy of heart digital twin technology in predicting critical substrate abnormalities for VT ablation.
• To compare electrogram characteristics at digital twin-predicted sites versus nonpredicted sites.
• To assess the ability of digital twins to identify regions of conduction slowing.

Main Methods:

• Heart digital twin models were created from 18 VT patients using contrast-enhanced MRI.
• Electrophysiological properties were applied to finite-element meshes, and VT circuits were defined using pacing protocols.
• Digital twin predictions were compared with invasive substrate mapping, including electrogram abnormalities and conduction slowing.

Main Results:

• Digital twin-predicted sites showed significantly higher rates of electrogram abnormalities (45.5% vs. 32.2%, P<0.001).
• Electrograms were significantly longer at predicted sites (82.0±25.9 ms vs. 69.7±22.3 ms, P<0.001).
• Digital twins correctly identified 80.8% of deceleration zones observed on late activation mapping.

Conclusions:

• Digital twin-predicted sites exhibit more frequent and prolonged electrogram abnormalities, indicating accurate substrate identification.
• Digital twin technology demonstrates potential for enhancing substrate-based VT ablation by improving target prediction.
• This study provides the first prospective analysis of digital twin accuracy in predicting VT substrate abnormalities.