ECCOMAS 2024

In-Silico Modeling of Atherosclerosis: Hybrid Approach of Convection-Diffusion-Reaction Model and Agent-Based Model

  • Caballero, Ricardo (Aragón Institute of Engineering Research (I3A)
  • Martínez, Miguel Ángel (Aragón Institute of Engineering Research (I3A)
  • Peña, Estefanía (Aragón Institute of Engineering Research (I3A)

Please login to view abstract download link

Atherosclerosis, a primary factor in cardiovascular diseases1, is a major challenge in biomedical research. In this study, we present an in-silico model that combines a hybrid approach of continuous convection-diffusion-reaction model and agent-based model to predict atheroma plaque growth in coronary arteries. Our hybrid model has three coupled modules: computational fluid dynamics (CFD), mass transport, and agent-based modeling (ABM). The CFD module starts from a 3D coronary artery geometry, in which hemodynamics are calculated using the Navier-Stokes equations. Then, several 2D cross-sectional surfaces are selected in the zone of least wall shear stress (WSS) of the 3D model, and the plasma filtration process is reproduced using Darcy's Law. The mass transport module models the transport of substances using the convection-diffusion-reaction equations. Low-density lipoprotein (LDL) leakage through the endothelium is evaluated using the three-pore model2, which considers three pore sizes: small, normal, and large. The ABM module imports both the 2D geometry and the WSS and LDL concentration values from the CFD and mass transport modules. This module is responsible for predicting arterial wall remodeling using conditional and stochastic behavioral rules, which calculate the probability of occurrence for each cellular process. The cellular processes included: mitosis and apoptosis of cells; production and degradation of extracellular matrix; production, phagocytosis, and necrosis of macrophages, becoming foam cells (FCs) due to excess LDL in the wall; and the change of phenotype of smooth muscle cells (SMCs) from contractile to synthetic. To couple both models, a segmentation process is performed on the ABM output image, identifying the different layers that form the arterial wall. Then, with this information, the 3D geometry is reconstructed and the process is repeated. Our agent-based model (ABM) demonstrates promising capabilities in realistically predicting the growth of atheroma plaques, particularly in areas with lower wall shear stress (WSS). It is important to note that these findings are preliminary and require further validation. This work represents an early exploration of the complex dynamics involved in cardiovascular disease progression.