Personalization of Biomechanical Models for Early Detection of Disease

Nick Van Osta1, Feddo Kirkels2, Aurore Lyon1, Tammo Delhaas1, Maarten-Jan Cramer2, Arco Teske2, Joost Lumens1
1Maastricht University, Cardiovascular Research Institute Maastricht (CARIM), 2University Medical Center Utrecht, Dept. of Cardiology


Abstract

Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited cardiomyopathy clinically characterized by ventricular arrhythmias and predominantly right ventricular (RV) dysfunction. Disease expression in ARVC mutation carriers varies from sudden cardiac death in young individuals to lifelong absence of any phenotype. Current electrocardiographic and structural imaging methods for screening of genotype positive ARVC family members fail to detect early-stage ARVC-related disease of the myocardium. Here, we propose a cardiac imaging-based personalized modelling approach to identify and characterize right ventricular (RV) electro-mechanical tissue abnormalities in early-stage ARVC mutation carriers. We obtained multi-modality cardiac imaging data in a population of 10 subclinical stage ARVC mutation carriers without any of the conventional electrocardiographic and structural abnormalities. The reduced-order CircAdapt model of the human heart and circulation was used to perform organ-scale patient-specific ARCV simulations for all 10 subjects. A fitting algorithm was developed that minimizes the errors between simulated and measured RV volumes (end-diastolic and end-systolic), regional RV strain patterns (apical, mid-ventricular, and subtricuspid), and times of pulmonary and tricuspid valve opening and closure. The output of this algorithm was a limited set of model parameters describing the active and passive myocardial tissue behavior, represented by a three-element Hill muscle model, and regional delay of electromechanical activation. Patient-specific simulations revealed significant regional heterogeneities in active and passive tissue properties, mostly a reduced ability of active force generation typically in the subtricuspid region. None of the patient-specific simulations revealed significant delays of regional RV electromechanical activation. In conclusion, we developed a fast reduced-order patient-specific modeling algorithm that enables to detect early-stage electro-mechanical disease substrates in asymptomatic ARVC mutation carriers who are at risk for sudden cardiac death, but without conventional electrocardiographic and structural disease criteria. Future studies should evaluate whether these early-stage disease substrates are predictive for arrhythmic event or ARVC disease progression.