Objective: The risk of intracranial dural arteriovenous fistula is linked to its pattern of venous drainage (Borden type), in particular the presence of cortical venous drainage. The purpose of this study was to assetss the accuracy of 3D pseudocontinuous arterial spin-labeling (ASL) MRI for noninvasive delineation of venous drainage.
Materials and methods: This retrospective study included 34 patients with a dural arteriovenous fistula who had undergone both digital subtraction angiography (DSA) and 3D pseudocontinuous ASL MRI. Two neuroradiologists blinded to the DSA results independently assessed ASL images for the presence of cortical vein hyperintensity (cortical venous drainage) and the distribution of venous hyperintensity (Borden type). DSA was used as the reference standard. The sensitivity and specificity of 3D pseudocontinuous ASL MRI for the detection of cortical venous drainage were determined. Intermodality and interobserver agreement for Borden type was determined by use of the weighted kappa statistic.
Results: Three-dimensional pseudocontinuous ASL MRI had high sensitivity (91%) and specificity (96%) for the detection of cortical venous drainage. Borden type was correctly identified with very good intermodality (weighted κ = 0.82) and interobserver (weighted κ = 0.85) agreement in 88% of patients.
Conclusion: Three-dimensional pseudocontinuous ASL MRI is highly accurate for the detection of cortical venous drainage and determination of Borden type. With this technique, high-risk fistulas requiring treatment can be reliably differentiated from low-risk lesions. Although it cannot replace DSA, incorporating 3D pseudocontinuous ASL into an MRI protocol for assessment of dural arteriovenous fistula can facilitate treatment planning.
Keywords: MRI; arterial spin labeling; cerebrovascular disease; dural arteriovenous fistula; noninvasive imaging.