Search

MOTIVATION Cerebral aneurysms have presented considerable challenges to effective treatment since their first clinical observation. For some 100 years the only clinical solution was dangerous and often debilitating open-surgery in the form of a craniotomy. The safety and efficiency of cerebral aneurysm treatment was greatly improved in the early- 1990s with the introduction of minimally invasive procedures such as endovascular coiling. The development in the mid-2000s of low-porosity stents, known as flow- diverters, presented a more sophisticated treatment rationale that focussed on vessel reconstruction, and addressed cases that would otherwise be problematic to treat by other endovascular means. However, current designs of flow-diverter (and all known designs in development) are limited to near-identical products that risk under-exploiting the potential benefits of the treatment method. These devices are of a cylindrical, woven construction that necessitates a fixed mesh design, which means that any variation in design focusses on device porosity alone. As such, the precise hemodynamic mechanisms, and the resulting quality of aneurysm treatment by flow-diverter, are under-developed and poorly understood. There is also concern in the clinical community that the cylindrical nature of current flow-diverter devices may lead to daughter vessel sacrifice when used to treat bifurcation aneurysms. This thesis specifically focusses on the effect of flow-diverter mesh topology on the complex hemodynamic environment within the treated aneurysm, and the resulting implications for treatment success. METHODOLOGY In this study, a number of both novel and commercially available flow-diverter designs are virtually implanted into idealised and patient-specific aneurysm geometries. These devices are then modelled under realistic blood flow conditions with computational fluid dynamics (CFD) techniques. The devices chosen encapsulate a range of features that allow the effects of device porosity and device mesh design to be controlled for and evaluated independently. Conventional cylindrical devices are also compared to a novel, non-cylindrical device for the dedicated treatment of bifurcation aneurysms. The aneurysm hemodynamic environment both pre- and post-treatment is quantified with a number of in-silico measures, such as flow rate, velocity, pressure, and wall shear stress. Finally, a model of angiographic contrast agent transport is developed in-silico to begin to verify the complex flow patterns predicted by CFD simulations with the corresponding in-vivo behaviour seen clinically. A number of inaccuracies of previous 'virtual contrast' models, as presented in the literature, are addressed and quantified by comparison to a gold-standard in-vivo porcine model. RESULTS In a number of aneurysm geometries, novel flow-diverter designs are shown to out-perform current commercially available devices. The strong influence of device porosity on aneurysm inflow reduction is confirmed with a 10% decrease in device porosity correlating with around 10-20% reduction in inflow. Variation attributed to device design alone is also seen; in a number of geometries, differences in flow-diverter mesh design, at constant porosity, are shown to introduce approximately 10-30% variation in aneurysm inflow reduction. Such variation is significant and potentially very important to the development of future endovascular treatment devices. The greatest variation in inflow reduction seen across the device designs also corresponds to changes in vorticity and secondary flow effects. The flow reduction achieved by the novel bifurcation aneurysm device discussed compares favourably to conventional devices in half of the geometries simulated, achieving a flow reduction of approximately 30-70%. Across all bifurcation aneurysms studied, conventional devices produce a relatively uniform flow reduction of 50-70%. Little change in flow rates for vessels jailed by a conventional flow diverter suggest that concerns of daughter vessel sacrifice appear unfounded, at least immediately following device deployment and prior to any neointimal formation. A number of unusual flow phenomena, which are almost entirely unreported in the literature, are observed in several aneurysm geometries. In particular, the emergence of stable and unstable laminar flow patterns under stationary conditions lead to oscillatory behaviour in one aneurysm geometry. Pseudo-stable behaviour is also observed in another aneurysm geometry whereby steady state and transient simulations conducted under identical conditions converge to solutions with radically different flow patterns. A good in-silico prediction of contrast residence is possible for a pre-treatment aneurysm geometry, where mixing is prominent. Aneurysm flow patterns after device placement reveal large spatial-variation in contrast infiltration that challenge the validity of quantifying contrast residence with a spatially-averaged decay curve. The most significant source of error between aneurysm contrast decay rate predicted in-silico and the in-vivo porcine model is found to be the variation in parent vessel flow rate, and not the density and viscosity changes of the contrast-blood mixture. CONCLUSIONS In addition to porosity, flow-diverter device design has a substantial effect on aneurysm inflow reduction and aneurysm flow pattern. When coupled with improved knowledge of aneurysm thrombosis mechanisms, appropriate changes in flow-diverter mesh design have the potential to improve the success rate of flow-diverter treatment. However, the clinical implications of these effects are currently unknown due to the lack of design variation in commercially available devices. In cases of bifurcation aneurysms, conventional flow-diverter devices have been shown to equal or outperform a novel device designed exclusively for bifurcation cases. Although no immediate complications were predicted from daughter vessel occlusion, the long-term integrity of treatment with conventional cylindrical devices is uncertain, and dedicated devices designed to preserve daughter vessel patency may become necessary. The complexity of the flow environment within a cerebral aneurysm has been reinforced, and a number of effects have been observed that demand further investigation. In particular, the role that the observed flow instabilities and the resulting changes in aneurysm wall shear stress may have in aneurysm rupture mechanisms is an emerging area of the literature. Verification of CFD simulations with in-vivo results remains a significant challenge. The virtual modelling of angiographic contrast residence currently offers the best route to such verification; although contrast transport may be integrated into current CFD simulations with relative ease, washout behaviour appears highly sensitive to parent vessel flow rates. As such, any contrast-based validation requires knowledge of patient-specific flow rates, which begins to challenge how representative flow conditions seen during angiography may be.