Patient-Specific Panoramic Mapping of Ventricular Tachycardia Mechanisms
This PhD project aims to develop and validate ANOIA (ACIF/2021/205) , an interpretable deep learning-based tool that integrates patient-specific electrical and structural cardiac data into a single panoramic map. This tool will help identify key mechanisms sustaining ventricular tachycardia (VT) and guide personalized ablation strategies.
To achieve this, we will explore the value of non-invasive electrocardiographic imaging (ECGI) alone or in combination with invasive mapping and late gadolinium-enhanced cardiac MRI. Our work involves:
- Building a multimodal clinical database, collecting imaging and ablation outcome data from VT patients.
- Generating a virtual population of personalized electrophysiological heart models to simulate various ablation strategies—enabling regional importance assessment without needing to test every strategy in each patient.
- Developing ANOIA, an interpretable AI framework that scores each cardiac region by its role in arrhythmia maintenance—leveraging both clinical and simulated data to reveal not only electrical but broader mechanistic insights.
- Validating the methodology by using ANOIA to guide ablation in a subsequent VT patient cohort and evaluating outcomes.
Ultimately, ANOIA will deliver personalized, data-driven panoramic maps to support clinical decision-making in complex VT cases.
This PhD is being carried out by Clara Herrero Martín
