Background: Cardiac resynchronization therapy (CRT) has become an important treatment option in patients with typical left bundle branch block (LBBB). However, approximately 30% of patients undergoing CRT show little or no improvement. Although echocardiography could bring essential information concerning LV dyssynchrony, the interpretation of echocardiographic strains could be difficult due the complexity of mechanisms involved in cardiac contraction.
Objective: The aim of this study is to propose a patient-specific model-based approach in order to assist the analysis and to improve the interpretability of myocardial strains.
Methods: We propose a model of the cardiovascular system that integrates 4 main sub-models: i) cardiac electrical system, ii) right and left atrium, iii) multi-segment right and left ventricles and iv) systemic and pulmonary circulations. The left ventricle (LV) was divided into 16 segments and the right ventricle (RV) into 3 segments in order to evaluate concurrently different regions during the ventricular contraction process. For each LV segment, some parameters, associated with active and passive components of the cardiac muscle and electrical depolarization time, were identified using Evolutionnary Algorithms. The proposed approach was evaluated on data obtained from 3 LBBB patients, including ischemic (n=2) and non-ischemic (n=1) cardiomyopathy.
Results: A close match was observed between experimental and simulated myocardial strain curves for all the subjects. The RMSE, over all LV segments, is equal to 2.87 (±1.00), 2.49 (±0.55) and 3.63 (± 0.81) for the anterior ischemic, the lateral ischemic and the non-ischemic LBBB patients, respectively. Bull's-eye representations of electrical activation delay and contractility levels could be estimated from the model-based approach. Contractility bull's-eye results allow ischemic and non-ischemic cases distinction, where low levels of contractility could be associated with damaged tissues.
Conclusion: The proposed model is able to simulate successfully patient-specific myocardial strains and to estimate electrical conduction and mechanical activation delays.