Towards the Development of Virtual Heart Technology for Creating Digital Twins of Cardiac Electrophysiology

Introduction: Virtual heart technologies (VHTs) promise to offer an detailed view on cardiac electrophysiology (EP). However, only very few VHTs simulating EP of four-chamber hearts have been reported so far and widespread use is limited due to labor intense workflows and the difficulty of ascertaining a close match between simulations and observations. This motivates the development of efficient workflows for both critical stages of creating a digital twin of cardiac EP: i) the generation of patient-specific anatomical models, and ii) the calibration of EP models to match clinical observations. Methods: A workflow for the generation of anatomically accurate models of the heart, with emphasis on the atria.from imaging data was developed. This includes, automated creation of multi-label segmentations using a convolutional neural network (CNN), a meshing step, annotation of tissue regions, and assignment of fibers. To speed up functional twinning, a prototype to run close to real-time EP simulations that can be used to find a good initial guess for model calibration methods that fit physiologically more accurate mono- or bidomain EP models to clinical data has been designed. Results: A semi-automated workflow to create atrial models at scale was developed, implemented and applied to a first clinical data set. A GUI based tool to plan and run whole heart EP simulations was prototyped.