The effects of geometric variation from OCT-derived 3D reconstructions on wall shear stress in a patient-specific coronary artery

Computational fluid dynamics (CFD) can be used to model the blood flow in patient-specific arterial geometries and predict atherosclerotic plaque progression, as the site specificity of plaques has been shown to depend on the wall shear stress (WSS) experienced by an artery’s endothelial layer. The level of artery wall detail in CFD models is expected to underpin predictions of plaque deposition, as it influences the computation of WSS. In this study, the sensitivity of WSS computation to geometric variation in a (proximally) stented left anterior descending (LAD) coronary artery geometry is investigated. The geometry is reconstructed from intravascular optical coherence tomography (OCT) images, which are registered and merged with computed tomography (CT) to include surrounding arteries. The WSS is modelled for low-, medium- and high-resolution geometries; created by altering the point resolution of the contour algorithm used to trace the (OCT imaged) artery lumen. The results show that areas of low WSS (and thus high thrombotic susceptibility) in the stented portion of the artery depended greatly on the resolution of the geometric reconstruction. Compared with the high-resolution geometry, the low- and medium-resolutions failed to accurately determine the areas of low WSS in the stented region, as minor geometric features were inadequately represented. The sensitivity of WSS to geometric variation is also relevant to analyses using coronary artery geometries derived from other imaging modalities (particularly those of lower resolution than OCT).