Computational Fluid Dynamics in Patient-Specific Models of Normal and Chiari I Geometries

Abstract

Chiari I malformation characterizes by cerebellar tonsils descending below foramen magnum, leading to obstruction of subarachnoid space (SAS) that alters the cerebrospinal fluid (CSF) flow. The condition leads to diverse symptoms in patients and is in some cases accompanied by a syrinx in the spinal cord. In order to improve the understanding of effects of SAS geometry on the CSF flow characteristics, numerical simulations of the CSF flow were conducted on a a series of personalized models of Chiari I patients, healthy volunteers and post-operative patients. The models were generated with the help of VMTK, which is a software tool for reconstruction of anatomical structures based on segmentation of medical images. To simulate the CSF flow, we applied the Navier Stokes equations for incompressible Newtonian fluid, which were solved numerically by applying the Finite Element Method (FEM) to the Incremental Pressure Correction Scheme (IPCS). The simulations were conducted by means of a pre-provided code for computational fluid dynamics (CFD) based on the FEniCS software library. To assess the validity of our results, the computed velocities were compared with MRI-measurements of the studied patients. If the computations were significantly different from the velocity measurements, the inflow and outflow conditions were changed accordingly and new simulations were conducted. Chiari I models have shown more evident flow complexities, greater peak velocities, higher flux-values and larger magnitude of bidirectional flow compared to models of healthy volunteers. We observed that to achieve more realistic results in future simulations, setting pressure and velocity conditions based on MRI measurements is recommended.