Sex-specific Cardiac Digital Twins of Human Ventricular Electrophysiology using 12-Lead ECG and MRI

Julia Camps1, Maxx Holmes2, Ruben Doste3, Lucas Arantes Berg4, Zhinuo Jenny Wang4, Blanca Rodriguez4
1Universitat Pompeu Fabra, 2University of Leeds, 3Department of Computer Science, University of Oxford, 4University of Oxford


Abstract

A cardiac digital twin is a virtual tool representing a patient's heart that can simulate new events, such as evaluating therapies to inform clinical decision-making. We present a sex-specific digital twinning framework for personalising electrophysiological function based on routinely acquired magnetic resonance imaging (MRI) data and the standard 12-lead electrocardiogram (ECG). We generate the MRI-based biventricular geometry. Then, we use a sequential Monte Carlo approximate Bayesian computation algorithm to infer the heart's electrical activation and repolarisation properties from the QRS, and from the T wave, respectively. We use a dictionary of action potential models generated using sex-specific electrophysiology characteristics implemented on the ToR-ORd cellular model by calibrating the ionic conductances using relative mRNA expression ratios of ion channel markers from non-diseased adult human male and female ventricular myocytes. These new sex-specific models, combined with the MRI-derived geometries for the heart and torso, enabled personalising to the sex of the subject considered for the study. We applied our framework to a female subject, and the inferred population matched the clinical data's QRS and T wave with a Pearson's correlation coefficient of 0.89. These methodologies for cardiac digital twinning are a step towards personalised virtual therapy testing. These tools are available at github.com/juliacamps/Cardiac-Digital-Twin.