Session S74.5

A Cardiovascular Model for the Analysis of Pacing Configurations in Cardiac Resynchronization Therapy

K Tse Ve Koon, V Le Rolle, G Carrault*, AI Hernandez

Université de Rennes
Rennes, France

Among patients selected for cardiac resynchronization therapy (CRT), around 30% are non-responders. A possible reason for this is that the atrioventricular (AVD) and interventricular (VVD) pacing delays are not correctly optimized. The aim of this work is to propose a mathematical model of the cardiovascular system reproducing the main haemodynamic effects resulting from changes in AVD and VVD in CRT. This model may then be useful to assist the AVD and VVD optimization procedure. Clinical studies have shown that the variation of systolic blood pressure (SBP) for different AVD and VVD is quasi-parabolic, with the former being more curved. These studies show an increased SBP(AVD) dependence at higher heart rates (HR).
In order to reproduce these effects, a cardiovascular model made of 3 interacting modules is proposed: i) an electrical heart model, made of a set of coupled cellular automata; ii) an elastance-based mechanical model of the heart, which is segmented in five sub-systems (one for each atrium, one for each ventricle and one for the interventricular septum) and iii) the circulatory system. This model is able to reproduce the quasi-parabolic nature of the dependence of SBP on AVD and VVD. Results from sensitivity analyses of the model to four different model parameters related to ventricular contractility, pulmonary and systemic vasodilatation and venous regulation are firstly presented. A first step towards a patient-specific approach (identification of the parameters of this model on a set of interpolated clinical data) is also proposed. An evolutionary algorithm is used for this purpose. For each pacing configuration, simulated SBP outputs are obtained and a relative mean squared error (rMSE) is computed between simulated data and a set of parabolic models, derived from clinical data. Three optimal AVD/VVD configurations: 100/0, 90/20 and 110/-30 have been tested. For these three different optimal configurations, the best set of parameters has been derived to represent a patient-specific model with rMSE respectively equal to 0.24, 0.23 and 0.10.

(Abstract Control Number: 122)