Ischaemic stroke is a major cause of death and disability worldwide. Cerebral autoregulation, which can be impaired during acute stroke, and collateral flow to brain tissue through the circle of Willis, both play a role in preventing tissue infarction. The configuration of the arterial circle varies between individuals. Thus, personalised modelling of the cerebral arterial network, to determine the potential for collateral flow, can be of significant value in the clinical context of stroke. The interaction between autoregulation and collateral flow remains poorly understood. In this study, steady-state physiological models of the cerebral arterial network, including several common variants of the circle of Willis, were coupled to a spatially variable mathematical representation of cerebral autoregulation. The resulting model was used to simulate various arterial occlusions, as well as bilateral and unilateral impairment of autoregulation, in each structural variant. The work identified few circle of Willis variants that present either particularly high-risk or particularly low-risk of cerebral ischaemia. Instead it was found that most variants are dependent upon the bilateral function of autoregulation to facilitate collateral flow and preserve cerebral blood flows. When autoregulation was impaired unilaterally, downstream of an occlusion, blood flows in the contralateral hemisphere were preserved at the expense of the ipsilateral tissue at risk. Arterial network models have in the past been personalised using structural, rather than functional, angiography measurements. This thesis presents a novel model-based method for absolute blood volume flow rate quantification in short arterial segments using dynamic magnetic resonance angiography data. The work also investigated the additional information that can be obtained from such functional angiography. The flow quantification technique was found to accurately estimate flows in shorter arterial segments than an existing technique. However, improvements to noise performance, and strategies for rejection of contaminating signals from overlapping vessels within the imaging plane, are required before the technique can be applied to personalised cerebral arterial network modelling.

Objectif : Déterminer la capacité de la séquence IRM 4D flow à mesurer la forme et le flot (débit, vélocité) de l’artère hépatique et de ses branches en trois dimensions. Méthodologie : Un fantôme de l’artère hépatique réaliste qui imite le flux sanguin et les mouvements respiratoires ainsi que 20 volontaires ont été imagés. La précision du 4D flow Cartésien avec navigateur et remplissage de l’espace-k selon la position respiratoire était déterminée in-vitro à quatre résolutions spatiales (0,5 à 1,0 mm isotropique) et fenêtres d’acceptation du navigateur (± 8 et ± 2 mm) avec un scanner IRM à 3T. Deux séquences centrées sur les branches hépatiques et gastroduodénales étaient évaluées in-vivo et comparés au contraste de phase 2D. Résultats : In vitro, l’augmentation de la résolution spatiale diminuait plus l’erreur qu’une fenêtre d’acceptation plus étroite (30.5 à -4.67% vs -6.64 à -4.67% pour le débit). In vivo, les artèreshépatiques et gastroduodénales étaient mieux visualisées avec la séquence de haute résolution (90 vs 71%). Malgré un accord interobservateur similaire (κ = 0.660 et 0.704), la séquence à plus haute résolution avait moins de variabilité pour l’aire, le débit, et la vélocité moyenne. Le 4D flow avait une meilleure cohérence interne entre l’afflux et l’efflux à la bifurcation de l’artère hépatique (1.03 ± 5.05% et 15.69 ± 6.14%) que le contraste de phase 2D (28.77 ± 21.01%). Conclusion : Le 4D flow à haute résolution peut évaluer l’anatomie et l’hémodynamie de l’artère hépatique avec une meilleure précision, visibilité, moindre variabilité et meilleure concordance interne. / Objectives: To assess the ability of four-dimensional (4D) flow, an MRI sequence that captures the form and flow of vessels in three dimensions, to measure hepatic arterial hemodynamics. Methods: A dynamic hepatic artery phantom and 20 consecutive volunteers were scanned. The accuracies of Cartesian 4D flow sequences with k-space reordering and navigator gating at four spatial resolutions (0.5- to 1-mm isotropic) and navigator acceptance windows (± 8 to ± 2 mm) were assessed in vitro at 3 T. Two sequences centered on gastroduodenal and hepatic artery branches were assessed in vivo for intra - and interobserver agreement and compared to 2D phase-contrast (0.5-mm in -plane). Results In vitro, higher spatial resolution led to a greater decrease in error than narrower navigator window (30.5 to −4.67% vs−6.64 to −4.67% for flow). In vivo, hepatic and gastroduodenal arteries were visualized more frequently with the higher resolution sequence (90 vs 71%). Despite similar interobserver agreement (κ = 0.660 and 0.704), the higher resolution sequence had lower variability for area, flow, and average velocity. 4D flow had lower differences between inflow and outflow at the hepatic artery bifurcation (11.03 ± 5.05% and 15.69 ± 6.14%) than 2D phase-contrast (28.77 ± 21.01%). Conclusion: High-resolution 4D flow can assess hepatic artery anatomy and hemodynamics with improved accuracy, greater vessel visibility, better interobserver reliability, and internal consistency.