Numerical study on patient-specific haemodynamics subjected to embolisation and wall-distensibility

Computational fluid dynamics (CFD) simulations have been peformed to investigate the hemodynamics of patient-specific cerebral aneurysm treated with endovascular coils; and arteriovenous fistula (AVF) using Star-CCM+. Fluid-structure interactions (FSI) between the elastic vessel walls and the blood flow within were also taken into account to provide a more realistic environment and better understanding of haemodynamic effects on wall remodelling. The blood in both studies was modelled as non-Newtonian fluid and comprises of three phases to fully incorporate the effects of shear-thinning and distributions of blood cells, respectively. The use of a less invasive ultrasonic imaging texhnique for CFD simulations is shown to be a viable alternative to magnetic resonance imaging (MRI). This has proven to be beneficial especially for haemodialysis patients who require fistula check-up on a regular basis. Excessively enlarged sections of arteries, called aneurysms, are vulnerable to vessel wall degradation. When blood flows into a cerebral aneurysm, it causes abnormal haemodynamic changes, which increases the risk of aneurysm rupture and strokes. Patients diagnosed with a cerebral aneurysm are therefore treated by stenting the parent artery or aneurysmal coiling to achieve occlusion. Despite high coiling packing density, aneurysm may recanalise, which consequently leads to aneurysm recurrence. Our understanding of the relationship between coiling density and aneurysmal occlusion and aneurysm recurrence in a non-Newtonian environment are limited. The effects of coil packing density on aneurysmal haemodynamics and the mechanism behind aneurysmal recurrence are discussed in this thesis. In the present aneurysm study, the aneurysm dome was embolised with seven different coil configurations of different packing densities. A time-dependent passive scalar was added to the multiphase blood inflow to represent medical dyes which allows for the visualisation of blood flow penetrating into the coils. The observed relationship between passive scalar visualisations, white blood cells distribution, and hemodynamic quantities will be beneficial for clinical evaluation of aneurysm occlusion. It is shown that a packing density of 31% (7 coils) is the optimal coil density that can supress the aneurysmal volume-averaged velocity and wall shear stress. Furthermore, the temporal variation in streamwise velocity inside the aneurysm dome does not nescessarily decrease with coiling packing density during peak systole. Local packing density, distribution of red and white blood cells, and wall compliance have been correlated with aneurysm recurrence. Circumferential wall shear stress, radial wall displacement, adhesion of white blood cells on the wall, and whole blood velocity magnitude in the six and seven coils cases are compared. These two coiling cases are chosen to represent an event of aneurysm recurrence as unexpected increase in the mean inflow into the aneurysm is observed despite higher coil packing densities. To the best of the author’s knowledge, the present aneurysm study is the first to investigate the effects of coil packing densities and blood cells distribution on non-Newtonian aneurysmal flow reduction and aneurysm recurrence in both rigid and compliant cerebral aneurysms. Additionally, the effects of aneurysmal haemodynamics on coil movement and vice versa are investigated in a study featuring a compliant coil in a rigid aneurysm. Over-estimation in aneurysmal velocity and wall shear stress at multiple locations by the rigid coil model has been observed. An arteriovenous fistula (AVF) is a connection between a brachial artery and vein that is surgically created to provide haemodialysis patients with matured vascular access points. AVF maturation failure, however, often occurs and its underlying mechanisms still remain controversial. The present AVF study investigates the effects of the compliant wall and non-Newtonian blood viscosity in an end-to-end AVF. Four simulations were performed to compare Newtonian and non-Newtonian haemodynamics in both rigid and wall-compliant fistulas. Different ranges of wall shear stress parameters corresponding to certain endothelial changes are compared among the four cases. It is found that the effects of wall compliance is more significant than that of non-Newtonian rheology. Furthermore, non-Newtonian effects are more clear when the AVF walls are compliant. Volumetric quantities like flow recirculations and helicity, which are related to abnormal endothelial changes, are also found to be overestimated by the rigid wall assumption. The study also investigates the effects of multiphase haemodynamics on inward wall remodelling and thus AVF maturation failure. Low and oscillating wall shear stress index is introduced as a tool for predicting the risk of maturation failure. Wall shear stresses, both directional and magnitude, red blood cell viscosity, flow recirculations are correlated with wall remodelling and endothelial damages derived from von-Mises stresses.