Haemodynamic Assessment of Coronary Flow with CFD and Phase-Contrast MRI

Abstract

Coronary artery atherosclerosis is a major cause of morbidity and mortality worldwide. Some anatomical characteristics of coronaries may increase the likelihood of disease development, and stenting, a common treatment, has a high probability of restenosis and in-stent thrombosis. Coronary flow patterns, determined by vessel geometry and stent design and deployment, are thought to be related to adverse clinical outcomes. This thesis represents a comprehensive assessment of the relationships between flow and major coronary artery shape, and stent design and deployment. It further demonstrates the feasibility of assessing dynamically scaled coronary phantom flow with phase-contrast magnetic resonance imaging (PC-MRI). A range of idealised and patient-specific coronary artery geometries, with and without stents, were generated based on a statistical shape database from computed tomography (CT) angiograms, and micro-CT scanned benchtop deployed stents, using a computer aided drawing program. Transient flow was simulated using computational fluid dynamics (CFD). The geometries were also up-scaled, 3D printed, and incorporated into a dynamically scaled blood-analogue flow circuit for velocity acquisition with PC-MRI. The CFD results revealed little flow effect for bifurcation angle when varied in non-stented and stented idealised and patient-specific bifurcations. Flow was also computed in four patient geometries with wide shape variation, where a statistical shear stress analysis suggested that the vessel tortuosity may influence the effect of bifurcation angle. Similarly complex flow alterations were observed when systematically analysing the flow effects of major stent design features. The haemodynamic interrelations provided insights for the development of design guidelines for stent manufacture. Computational findings were validated by means of a new approach using dynamically scaled phantom flow and PC-MRI acquisition for non-invasive imaging with higher effective resolution (7-fold up-scaled geometries). The co-registered flow fields yielded good agreement (R2>0.8, magnitude error <8%) after using the PC-MRI measured inlet flow to inform the computational boundary conditions. A complementary approach of CFD and dynamically scaled PC-MRI can deliver further insights into the important haemodynamic effects of vessel shape and stent design and deployment. The results represent a potential clinical assessment tool, new considerations for stent design guidelines and deployment strategies, and a novel approach for coronary flow assessment for patient-profiling and pre-surgical planning.