Cardiac Electromechanical Model Sensitivity Analysis using Causal Discovery
Résumé
The emergence of advanced imaging technologies has considerably increased the availability of data to better characterise cardiac function, leading to improved accuracy, but at the cost of increasing complexity due to the need to integrate a greater number of parameters. Understanding the relationships among these parameters and their influence on model outputs has become a crucial yet challenging task. In this work, we employ causal discovery to perform global sensitivity analysis of an electromechanical model, and investigate the interrelationships among input parameters and the main outputs of interest, namely, ejection fraction (EF ) and pressure change in the left ventricular cavity (max(dP/dt)) that are typically used in cardiac resynchronization therapy. The obtained causal graph reveals and quantifies the strengths of the relationships among those variables. Notably, an augmented peak contractility is found to positively affect both biomarkers, while increasing the tissue’s stiffness has a negative impact on them. The parameters controlling the shape of the active stress were found to influence max(dP/dt), whereas leaving EF unchanged. Those results were further confirmed by independently performing clustering and classification, as well as classical local sensitivity analysis. Our method provides a precise identification of the key parameters to be focused on, and ensures stable results compared to classical global sensitivity analysis methods, despite a limited number of available simulations.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |