Towards Reduced Order Modelling of Cardiac Electroanatomical Mapping

Ventricular tachycardia (VT) is a lethal heart rhythm disorder responsible for 80% of sudden cardiac deaths.Radiofrequency ablation (RFA) of VT is minimally invasive and potentially curative but target identification using cardiac electroanatomical mapping (EAM) is challenging.Personalized computer heart models may provide deeper insight of patient electrophysiology (EP) but are often incompatible with clinical timeframes.This work is a proof of concept of applying personalized finite element EP modeling and machine learning (ML) to achieve near-instant prediction of patient EAM in sinus rhythm.We used one patient dataset to generate a personalized 3D left-ventricle model.A simulation batch with varying parameters was run and results were compared to patient EAMs.Simulated data was then used to generate a reduced order model (ROM) through ML.ROMs were capable of reproducing the simulation activation maps with < 1 ms accuracy using as low as 15 learning scenarios.