Cardiac digital twins of the adult population: linking electrophysiological functions to anatomy

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): This project is supported by the Wellcome Trust and EPSRC Centre for Medical Engineering. Background QRS duration derived from 12-lead ECGs provides an estimate of the ventricular depolarization time and is an important metric to assess cardiac function. The QRS duration is determined in part by the combination of heart size and conduction velocity (CV). Differences in QRS duration in terms of gender, body mass index (BMI) and age have been well-documented, however, the relative impact of concurrent changes in heart size and CV on QRS duration is poorly understood. Cardiac digital twins, as in-silico replicas, provide a non-invasive physics and physiology-constrained framework to separate the impact of CV and heart anatomy on the QRS duration across populations. Purpose This study aims to identify the impact of CV and heart size on differences in QRS duration and quantify their difference in the general population in different groups of gender, BMI and age. Methods Patient-specific biventricular models (n=400) were constructed using long- and short-axis cardiac magnetic resonance (CMR) images from patients without self-reported cardiovascular diseases in the UK Biobank. Cardiac electrophysiology was simulated using a physiological-detailed model that includes fiber orientation, anisotropic conductivity, the His-Purkinje system activation, and fast endocardial conductivity[1]. 12-lead ECGs were reconstructed from the simulations and the QRS duration was computed based on corresponding vectorcardiogram (VCG) reconstructed from ECG data [2]. We calibrated patient-specific CVs to match the clinically measured QRS duration in ECGs at rest and then compare the heart size, QRS duration and estimated CVs for different patient groups. Results QRS duration was approx. 7% longer in males than females (median: 90ms vs 84ms, P<0.001). Consistent with a longer QRS duration in males had larger hearts (male: 165.5 mL vs 123 mL, P<0.001), while the median CVs were the same between genders (both: 0.594 m/s). QRS duration was longer for obese (BMI≥30) and overweight group compared to healthy (BMI<25) groups (88ms and 86 ms vs 84 ms, P=0.05 and 0.3). Again, the change in QRS duration with BMI was reflected with a corresponding increase in heart size (healthy: 127 mL, overweight: 147.6 mL, obese:160 mL), while CV remained unchanged (healthy: 0.58 m/s vs overweight and obese: 0.594 m/s, P=0.051 and 0.32). There was no significant difference in QRS duration, CV or myocardial volume for different age groups (medians: 86-88 ms and 0.573-0.594 m/s and 138.7-143 mL). Conclusions QRS duration is shorter in females and individuals with healthy weights. Most of the change in QRS duration can be explained by heart size, with the estimated CV remaining consistent across groups.