Development of a computational hemodynamic framework for the simulation of coronary arteries wall motion

Cardiovascular diseases (CVD) are one of the leading causes of death and disabilities in western countries. Coronary arteries are one of the sites most likely to be affected by atherosclerosis (AS), one of the most common forms of CVD. In this field, hemodynamic computational fluid dynamics (CFD) played a key role in describing the impact of wall shear stress (WSS) patterns in atheromatous plaque genesis and progression. Moreover, coronary arteries do not present a fixed geometry, as they are bounded to the heart, and they are constantly subjected to bending and torsional forces through the entire duration of the cardiac cycle. The continuous change in shape of arterial lumen may cause alterations in the internal blood flow. Incorporating this aspect in CFD simulations may help to recreate more comprehensive and accurate models, which can provide a more realistic assessment of hemodynamic quantities related to the predictions of the progression of atherosclerotic disease. This thesis aims to implement a computational framework to impose the arterial vessel wall motion as a boundary condition for computational hemodynamic simulations. The tested wall motions were a planar bending of the coronary geometry (with no variation of diameter), a dilatation of the coronary lumen, and a combination of bending and dilatation, to study the effect of a composite wall deformation coronary hemodynamic. A 3D idealized geometry based on left anterior descending (LAD) artery with a reconstructed 56% eccentric stenosis was used. Three methods were developed to impose cyclic wall motion, assigning displacements for each node of the lumen wall. The first method involves directly changing the vessel mean curvature (SC), while the other two use the coherent point drift (CPD) algorithm to calculate the displacements that transform the initial geometry into the target one: in the first case CPD was used directly on the wall surface nodes (CPD-wall), while in the latter displacements were calculated on the vessel centerline (CPD-cl), and then transferred to the wall nodes. The wall complex deformation was created by combining a 3D cyclic bending with a cyclic diameter variation, with a maximum of 5.1% dilatation. The simulations were performed on the fluid domain alone, based on an interface-tracking Arbitrary Lagrangian-Eulerian (ALE) approach. Blood was modelled as an incompressible, homogeneous, and Newtonian fluid, and the imposed boundary conditions were unsteady flow and pressure waveforms, respectively at the inlet and at the outlet cross sections. The results were compared with CFD rigid wall simulations where the wall did not move during the cardiac cycle. The comparison was carried out in terms of WSS-based and bulk-flow descriptors. From the comparison between the three displacements calculation methods, the average nodal spatial discrepancy taken at the maximum deformation between SC method and CPD-wall and CPD-cl were respectively 0.46 mm and 0.04 mm. In general, the results demonstrate modest differences with the rigid-wall CFD model. Mean time-averaged WSS (TAWSS) values obtained in the stenosis region were respectively 9.601, 9.635, 9.568 and 9.602 Pa for rigid wall, dilatation, bending, and the combination of dilatation and bending. The differences in TAWSS values can be ascribed to the fact that the wall motion including dilatation leads to an average increase in flow rate. The next development will be transitioning this method to patient-specific models, implementing realistic wall motion.