| Summary: | <p>A clot in a cerebral aneurysm can either accelerate the road to rupture, through inflammatory processes and furthering vascular wall degradation, or stabilise the situation by occluding the aneurysm, and thus prevent rupture. A three-dimensional computational model of clotting in patient-derived cerebral aneurysm geometries is presented. The model accounts for the biochemical reactions that make up the clotting process, for realistic three-dimensional haemodynamics in image-derived vasculature representations and for the growing clot's interaction with and impact on the flow field. The flow is accounted for by the Navier Stokes equations and the transport equation describes the changes in biochemical species concentrations. Level Set methods are used to track the surface of the growing clot in the three-dimensional geometries studied. The influence of the thrombosed region on the haemodynamics is accounted for by modifying the local porosity and permeability, to reflect the fibrous and permeable nature of the clot. The model is first developed, examined and parameterised for a physiological model of clotting in two dimensions and is then extended and demonstrated for the pathological case in three dimensions.</p> <p>The framework developed is used to examine various aspects of clotting. The two-dimensional model is used to investigate the effects of critical thrombin concentration, tissue factor and the underlying biochemical cascade. The critical thrombin concentration at which coagulation transitions from the initiation to the propagation phase was found to be [TH] = 1 – 10nM. The inclusion of blood-borne tissue factor was found to reflect a more realistic thrombin production curve and an increase in initial tissue factor concentration led to a decrease in thrombin production lag time. The different underlying biochemical cascades produce similar results.</p> <p>The model is then extended to three dimensions and is used to investigate clot propagation and initiation in patient-derived aneurysms. The propagation velocity is linked to mechanical factors and biochemical species concentrations. An inverse relationship between strain rate and propagation velocity showed realistic clot growth. Realistic growth was also observed for a direct relationship with thrombin concentration and this seemed to be the most suitable approximation. The ways in which tissue factor, strain rate threshold and location of endothelial damage affect initiation are also examined. A strain rate of 500s<sup>-1</sup> was found to be the highest strain rate at which fibrin controlled clot initiation took place. Above that value, no clotting was observed. The location of endothelial damage affected clot growth as evidenced by the reproduction of the clots observed under Lawton's classification scheme.</p> <p>The three-dimensional model is then applied to patient-derived geometries and is used to examine the efficacy of flow diverter treatment. For a given geometry, clot growth is predicted for the case with and without a flow diverter. In some cases, clotting is a positive outcome while in other cases, the clot occludes the parent vessel.</p> <p>The unique contribution of this thesis is the combination of computational fluid dynamics, biochemistry and Level Set methods in complex, realistic, three-dimensional aneurysm geometries for clot prediction. The impact of the clot on the flow field is modelled by altering the porosity and permeability values of the clot region. </p>
|
|---|