Pressure Mapping and Hemodynamic Assessment of Intracranial Dural Sinuses and Dural Arteriovenous Fistulas with 4D Flow MRI

SUMMARY: The feasibility of 4D flow MR imaging to visualize flow patterns and generate relative pressure maps in the dural venous sinus in healthy subjects (n = 60) and patients with dural arteriovenous fistulas (n = 7) was investigated. Dural venous drainage was classified based on torcular Herophili anatomy by using 4D flow MR imaging–derived angiograms and magnitude images. Subjects were scanned in a 3T clinical MR imaging system. 4D flow MR imaging enabled noninvasive characterization of dural sinus anatomy and mapping of relative pressure differences.

V enous hypertension is thought to be implicated in dural arteriovenous fistulas (DAVFs) with aggressive presentation. 1,2 We investigated the use of 4D flow MR imaging for the noninvasive assessment of DAVFs 3 through measuring the vascular velocity vector field. These data can be processed for the analysis of the spatial and temporal distributions of flow and pressure gradients, [4][5][6] including dural venous sinus pressure and flow patterns. The purpose of this study was to compare 4D flow MR imagingderived hemodynamics and relative pressure maps in the dural venous drainage of patients with DAVF and healthy subjects. In addition, we report on the anatomic variations found in the dural venous drainage.

MATERIALS AND METHODS
Seven patients (age range, 33-72 years; mean age, 52 years; 2 women) diagnosed with unilateral DAVFs affecting the transverse/sigmoid sinus and 60 healthy subjects (age range, 45-75 years; mean age, 64 years; 32 women) were scanned by using a 3T clinical MR imaging system (MR750; GE Healthcare, Milwaukee, Wisconsin) with an 8-channel head coil. Phase-contrast MR imaging data with 3-directional velocity encoding were acquired with a radially undersampled 4D flow MR imaging sequence, phase contrast with vastly undersampled isotropic projection (PC-VIPR), 7,8 and the following imaging parameters: velocity encoding ϭ 80 cm/s for control patients (TR, 7.4 ms; TE, 2.7 ms) and 100 cm/s for patients with DAVFs (TR, 7.8 ms; TE, 2.5 ms); FOV ϭ 22 ϫ 22 ϫ 16 cm 3 ; 0.7 mm isotropic resolution; 14,000 projection angles; scan time ϭ ϳ7 minutes; flip angle ␣ ϭ 10°; and receiver bandwidth ϭ Ϯ83 kHz. Magnitude, velocity data, and PC angiograms were all generated with off-line reconstruction. Background phase correction of velocities was performed in Matlab (MathWorks, Natick, Massachusetts), and vessel segmentation was performed semiautomatically (Mimics; Materialise, Leuven, Belgium) from the phase-contrast angiograms. Pressure maps were also calculated in Matlab by using previously validated methods. 5,6 This approach solves the Navier-Stokes equation by assuming blood is an incompressible Newtonian fluid (density ϭ 1060 kg/m 3 ; viscosity ϭ 3.2 cP). Visualization of pressure differences and flow quantification were carried out in EnSight (CEI, Apex, North Carolina). Flow and area were quantified from cut planes in each transverse sinus (TS), left and right, placed 25 mm from the torcular Herophili, and pressure gradients were quantified as the pressure difference at 25 mm and 50 mm from the torcular Herophili. These measurements were per-formed by a Ph.D. candidate with 5 years of MR imaging postprocessing experience (L.A.R.-R.). The anatomy of the dural sinuses was classified according to Gökçe et al 9 based on torcular Herophili anatomy. A senior neuroradiologist (P.A.T., Ͼ30 years of experience) inspected the images to determine anatomic classification.
Statistics were assessed by using ANOVA followed by post hoc analysis with the Tukey-Kramer method. Data normality was assessed with quantile-quantile plots. Analysis was performed in Matlab. P Ͻ .05 was set as the threshold for statistical significance.

RESULTS
Representative color pressure maps in a healthy subject and a patient with DAVF are shown in Fig 1. For the healthy subject ( Fig  1A, -D), blood pressure decreased downstream with lower pressure in the TS than in the superior sagittal sinus. However, the patient with DAVF (Fig 1C, -F) shows a pressure increase in the TS affected by the fistula and a decrease on the TS contralateral to the DAVF. Another patient pressure map and flow streamlines can be found in On-line Figs 1 and 2. The streamlines move retrograde from the site affected by the fistula into the right TS.
The anatomic variations in the dural sinus drainage were catalogued as follows: for healthy subjects, there were 10 type I, 35 type II, and 15 type III variations (On-line Fig 3 and On-line Table  1). For patients with DAVF, there were 5 type I and 2 type II variations. Flow and cross-sectional area measurements in healthy subjects are summarized in On-line Table 2.
Flow and pressure drops in the TS for all subjects are shown in Fig 2. For patients with DAVF, pressure increased along the TS ipsilateral to the fistula with a median of 0.055 Ϯ 0.130 mm Hg, whereas pressure decreased in the TS contralateral to the fistula with a median of Ϫ0.367 Ϯ 0.205 mm Hg (P ϭ .109). In healthy subjects, the pressure decreased very similarly in both TSs, with values of Ϫ0.105 Ϯ 0.033 mm Hg and Ϫ0.101 Ϯ 0.043 mm Hg, respectively (P ϭ .994). In patients with DAVF, flow in the TS affected by the fistula was 40 Ϯ 129 mL/min, and flow in the TS contralateral to the fistula was 328 Ϯ 134 mL/min (P ϭ .046). In healthy subjects, independent of anatomy type, flow was significantly larger in the right TS (262 Ϯ 83 mL/min) compared with the left TS (121 Ϯ 56 mL/min; P Ͻ .001).

DISCUSSION
Initially, we characterized dural venous sinus pressure in a population of 60 healthy subjects, establishing a reference frame for the interpretation of dural sinus pathology. The main findings of this study are an increase in pressure in the TS ipsilateral to the DAVF compared