Modeling the geometry, hemodynamics and tissue mechanics of cerebral aneurysms

Cerebral aneurysms are abnormal dilatations of the intracranial arteries that often cause death or significant morbidity once ruptured. The current project is a comprehensive modeling effort from geometrical, hemodynamic and solid mechanical perspectives using anatomically realistic lesion geometry to improve upon the current understanding and management of the disease.;Twenty-eight saccular cerebral aneurysms and the connected arterial branches were reconstructed based on computed tomography angiography data from patients. A set of indices for the global geometrical (size and shape) features of the isolated aneurysm sac were developed based on differential and computational geometry techniques. Sensitivity analysis showed that these indices are not sensitive to random noise during the reconstruction process. Comparison between the ruptured and unruptured groups of aneurysms revealed the relative importance of different indices as predictors of aneurysm rupture, and demonstrated that lesion shape is better correlated to rupture potential than lesion size.;Steady and pulsatile blood flows in the reconstructed vasculature were simulated based on computational fluid dynamics (CFD) techniques. Hemodynamic indices for the isolated aneurysm sac, i.e. pressure differential, wall shear stress, and particle residence time were obtained. The Fung-type hyperelasticity material model was assumed for the aneurysm wall and implemented in finite element analysis using ABAQUS. Effective material fiber directions on the aneurysm surface were estimated based on surface curvature. Nonlinear, anisotropic, static deformation analyses were performed for all aneurysms. The computed wall stresses were not able to predict the rupture of aneurysms. Wall thickness variation was correlated with surface curvature, and variable wall thickness was found to homogenize the stress field as compared with constant wall thickness.;Pearson's correlation coefficients between the computed geometrical indices and biomechanical indices were obtained. The results showed that aneurysm geometry correlated better with wall stress than with hemodynamic indices, and wall stress correlated better with size indices than with shape indices.;These results are still preliminary due to the small sample size. However, the same technique of aneurysm analysis can be implemented in a larger patient population in the future to give clinically valuable insights.