The hypothesis of this project is rooted in the development of a tool that, by combining ventricular electrical activity, the relationship with signals measured on the patient’s torso, and its correlation with the ECGi solution, yields an approximation of the patient’s personalized electrophysiological properties that cannot be non-invasively measured to date. By detecting areas of very low amplitude and/or areas of very low conduction velocities through the ECGi solution and applying artificial intelligence, estimators will be created for those regions in the myocardium that do not contain functional tissue and are associated with multiple cardiac conditions.

This project aims to use cardiac digital twins to train an artificial intelligence model that enables the use of ECGi to obtain characteristics of electrical remodeling, as well as the location and properties of structural remodeling in a non-invasive and timely manner, providing the physician with information to tailor a treatment that suits the patient.

To achieve the main objective, the following specific objectives have been identified:

i) Generate 1000 ventricular digital twins that encompass anatomical variability between men and women and different degrees of electrical and structural remodeling based on available research group data.

ii) Train and validate an artificial intelligence model using digital twin data.

iii) Generate a proof of concept with patient data to validate the proposed model in collaboration with the Clinic Hospital of Barcelona.

CIAICO/2022/020  was funded by Generalitat Valenciana