Improving Computational Fluid Dynamics Simulations of Coiled Aneurysms Using Finite Element Modeling

Cerebral aneurysms are a serious clinical challenge, with ∼half resulting in death or disability. Treatment via endovascular coiling significantly reduces the chances of rupture, but the technique has failure rates between 25-40%. This presents a pressing need to develop a method for determining optimal coil deployment strategies. Quantification of aneurysm hemodynamics through computational fluid dynamics (CFD) has the potential to significantly improve the understanding of the mechanics of aneurysm coiling and improve treatment outcomes, but accurately representing the coil mass in CFD simulations remains a challenge. We have used the Finite Element Method (FEM) for simulating patient-specific coil deployment based on mechanical properties and coil geometries provided by the device manufacturer for n=4 ICA aneurysms for which 3D printedin vitromodels were also generated, coiled, and scanned using ultra-high resolution synchrotron micro-CT. The physical and virtual coil geometries were voxelized onto a binary structured grid and porosity maps were generated for geometric comparison. The average binary accuracy score is 0.836 and the average error in porosity map is 6.3%. We then conduct patient-specific CFD simulations of the aneurysm hemodynamics using virtual coils geometries, micro-CT generated oil geometries, and using the porous medium method to represent the coil mass. Hemodynamic parameters of interest including were calculated for each of the CFD simulations. The average error across hemodynamic parameters of interest is ∼19%, a 58% reduction from the average error of the porous media simulations, demonstrating a marked improvement in the accuracy of CFD simulations using FEM generated coil geometries.