Computational modelling of atherosclerotic coronary vessels with multi-layer material properties

Coronary arteries are a prominent site for atherosclerosis (AS), a chronic and progressive inflammatory condition characterized by arterial wall structural alterations. AS is a leading cause of fatalities and disabilities connected with cardiovascular diseases (CVD). Computational tools in coronary biomechanics investigate aspects not readily visible through imaging, such as wall shear stress (WSS) patterns and stress within coronary vasculature. While computational fluid dynamics (CFD) simulations have elucidated the role of WSS in the genesis and progression of coronary artery disease (CAD), they are constrained by the rigid-wall assumption, which is physiologically inaccurate. Conversely, fluid-structure interaction (FSI) methods account for vascular distensibility allowing for a more accurate representation of the healthy and/or altered mechanical properties of coronary wall. Since the mechanical properties of solid domain are determinants for the computational assessment of stress distribution, this work focuses on coronary vessel wall materials, modeled with two levels of complexity. On the one hand, a multi-component approach has been adopted to highlight marked mechanical heterogeneity across the wall thickness. This includes the co-existence of the three tunics and a lipid-rich necrotic core (LRNC) to explore interactions between healthy and pathological components of the arterial wall. On the other hand, a non-linear hyperelastic stress-strain behavior has been assigned to each layer. The Holzapfel-Gasser-Ogden (HGO) material model captures media and adventitia anisotropy, treating each tunica as a fiber-reinforced solid, where an isotropic ground matrix hosts two families of collagen fibers. An isotropic Neo-Hookean (NH) model was adopted for the plaque and the LRNC. An idealized left anterior descending (LAD) coronary geometry with eccentric stenosis was modeled from realistic morphological data. The cap over the LRNC was modeled to have a thickness of 50 μm. Material parameters of HGO model have been fitted on layer-specific quasi-static uniaxial tension tests on coronaries. End-diastolic steady rigid-wall CFD simulations were performed on the fluid domain to initialize the subsequent prestress, indicating a max percentual difference of 0.13% in the pressure maps when comparing results from a Newtonian rheological model (standard solver) and a non-Newtonian Carreau-Yasuda model (FSI fluid solver). Two-way fully coupled FSI simulations based on an Arbitrary Lagrangian-Eulerian approach (ALE) were first executed with a single-layer approach for human coronary arteries, i.e. considering only one component for the arterial wall, to assess its appropriateness for local stress analyses. In the prestressed configuration of the single-layer model, the maximum principal stress (MPS) peak (0.0014 MPa) was found in the region upstream of the central cross-section of the plaque, while the MPS is one order of magnitude higher (0.026 MPa) for the multilayer model. Following FSI, WSS-based and intravascular hemodynamic descriptors, as well as principal stress and strain for the wall, were calculated and a general comparative analysis between the FSI and the corresponding rigid-wall unsteady simulation was carried out. In this work, a computational framework for the advanced modeling of coronary vascular tissue was set. Once developed, such a benchmark will be useful for evaluating the reliability of single- and multi-component approaches in FSI compared to rigid-wall CFD simulations.