Relationship between ventricular morphology and aqueductal cerebrospinal fluid flow in healthy and communicating hydrocephalus

Abstract

Objectives: Differences in the magnitude of cerebrospinal fluid (CSF) volumetric flow through the cerebral aqueduct between healthy and hydrocephalic patients have been previously reported. However it is not clear whether this is directly related to the pathophysiology or secondary to altered ventricular morphology and hydrodynamics. This work aims to determine the role of anatomic and hydrodynamic factors in modulating the magnitude of CSF flow through the aqueduct.

Materials and methods: Twenty subjects (10 healthy and 10 patients with communicating hydrocephalus of different causes) were studied by MRI. Scans included T1-weighted 3D anatomic imaging and velocity-encoded cine phase-contrast scans of transcranial blood and CSF flows as well as CSF flow through the aqueduct. Anatomic MR data were used for quantitation of ventricular volumes, third ventricular width, and gray and white brain tissue volumes. Velocity-encoded imaging was used for quantitation of aqueductal and cervical CSF stroke volumes (SV), aqueductal lumen area, and systolic maximal intracranial volume change. Because data from normal and hydrocephalic patients were aggregated, a battery of statistical methods that accounted for the group effects were used. Partial correlation was used to determine which of these parameters were most significantly associated with aqueductal stroke volume (ASV). Multiple linear regression analyses were employed to identify anatomic and hydrodynamic models with the least amount of variables that are significant predictors of ASV. Finally, the association between the magnitude of ASV and the aqueductal lumen area, and its implication on the CSF flow dynamic characteristics and aqueductal pressure difference was established.

Results: Using partial correlations, 5 of the 6 anatomic parameters and none of the hydrodynamic parameters and brain tissue volume were found to be statistically significant. The highest partial correlations were with the total ventricular volume (r = 0.838) and third ventricle width (r = 0.811). These parameters were also found to be significant predictors of ASV in the multiple linear regression analyses with third ventricle volume and group effects as insignificant predictors (F = 28.08, P < 0.0001, R = 0.85). On the other hand, both cervical CSF SV and maximal ICVC were found to be weak predictors of ASV with group effects as the only significant variable of the hydrodynamic model (F = 4.18, P = 0.023, R = 0.33). A combined anatomic-hydrodynamic model including the predictive variables of the anatomic model and the ICVC provides the strongest coefficient of determination (R = 0.873). Pearson correlation analysis revealed a very strong relationship between ASV and the aqueductal lumen area (r = 0.947).

Conclusions: Aqueductal CSF flow is strongly correlated with ventricular morphology, especially with the total ventricular volume and the third ventricle width, but not with the tested hydrodynamic parameters. In addition, ASV is linearly correlated with aqueductal lumen area, suggesting that the aqueductal CSF flow characteristics can be explained by oscillating pressure differences on the order of less than 0.01 mmHg. These findings may explain why a standalone ASV is a poor diagnostic marker and an insensitive indicator of shunt outcome in idiopathic normal pressure hydrocephalus.