PACE-VT: A Non-Invasive, Machine-Learning-Driven Digital Twin Approach for Precise Targeting of Ventricular Tachycardia in Cardiac Stereotactic Body Radiotherapy - Project Summary Ventricular tachycardia (VT) and ventricular fibrillation (VF) are major contributors to the approximately 300,000 sudden cardiac deaths occurring annually in the United States. VT, particularly when refractory to ablation and medication, poses a life-threatening risk that necessitates innovative treatment strategies. Cardiac stereotactic body radiotherapy (cSBRT) offers a promising non-invasive alternative; however, its success depends on precise localization of the VT substrate for targeted, high-dose radiation delivery. The PACE-VT (Personalized Automated VT Circuit and Exit Localization) represents a transformative step forward in the management of VT, particularly for patients who are refractory to traditional treatments. This innovative research aims to integrate patient-specific computational models derived from CT imaging with real-time ECG data to enhance the precision of non-invasive VT localization. This integration facilitates targeted, non-invasive interventions such as cSBRT, offering potential improvements in safety and efficacy over existing therapies. The primary objective of the proposed project is to validate the effectiveness of a novel non-invasive VT mapping methodology that combines detailed structural data from CT scans with functional data from ECGs. The project seeks to demonstrate that this PACE-VT approach can accurately identify VT substrate locations, improving the targeting of therapeutic interventions such as cSBRT. The project will leverage existing datasets from catheter ablation procedures to validate the accuracy of the PACE-VT approach against clinically-identified VT circuits and exits. Statistical analyses will assess the spatial concordance between predicted VT substrate and actual VT substrate locations, aiming to substantiate the model’s predictive power. The project also utilizes advanced image processing techniques, machine learning models, and personalized heart digital twins to analyze VT circuits and exit sites. Success in this endeavor could lead to broad adoption of more accurate and patient-friendly management strategies for individuals with challenging cardiac arrhythmias. This project aligns with the NHLBI’s mission to foster innovative research and has the potential to significantly improve clinical outcomes and quality of life for patients with life-threatening heart conditions.
