| Type I Chiari malformation (CMI) is a complex disorder of the craniospinal system that is estimated to effect as many as 1:1000 adults in the US. In CMI, the fluid space near the craniovertebral junction becomes compressed and the tonsils of the cerebellum become stretched downward through the foramen magnum (FM), creating a partial blockage to the motion of cerebrospinal fluid (CSF) in the subarachnoid spaces. Diagnosis of the disorder is difficult because the most commonly used diagnostic criterion, cerebellar tonsil descent (CTD) >3-5 mm past the foramen magnum does not correlate well with patient symptom severity.;Because of the hydrodynamic component of the disorder, subject-specific computational modeling of cerebrospinal fluid dynamics may have utility in identifying biomechanical parameters to (1) quantify the severity of CMI and (2) quantify the changes to the CMI-affected spinal canal created by corrective surgery. In the first study presented in this work, magnetic resonance image based computational models of the cervical spinal canal were used of CSF dynamics showed that longitudinal impedance (LI) to CSF motion was elevated in a group of CMI patients compared to healthy volunteers and did not correlate well with CTD.;In the second study, a similar modeling methodology was used to compare pre-surgery CMI patients, post-surgery CMI patients, and healthy volunteers. LI was compared against CTD, cross-sectional area, hydraulic diameter, gradients at peak systole, peak diastole, and peakto- peak pressure gradient. LI was found to decrease significantly post-surgery on average, but remained significantly higher than in healthy volunteers. LI was found to be sensitive to small average changes in the cross-sectional geometry of the SSS, but only correlated weakly with 14 changes to CTD. However, because of the preponderance of female patients in the study, presurgery and post-surgery LI data may have been biased low.;Finally, velocity fields at the foramen magnum and C2 level of the spinal canal were compared to validate the subject-specific models against physical measurements from 2D phasecontrast MRI. Peak velocities and through-plane velocity profiles compared poorly between CFD and 2D pcMRI. CFD showed a trend of underestimating the peak velocities observed in 2D pcMRI at both the FM and C2 levels of the spinal canal. CFD velocity profiles at C2 lacked the anterior flow dominance and heterogeneous velocity distributions typically seen in 2D pcMRI measurements. Thus, CFD velocity fields could not be validated using 2D pcMRI. |
