Noninvasive characterization of cranial dural arteriovenous fistulas possible

MR selective flow-tracking cartography allows for noninvasive characterization of cranial dural arterivenous fistulas (DAVFs), according to a study published in the January 2014 issue of Radiology.

Though DAVFs are rare, they are a major contributor to long-term morbidity and mortality. DAVFs are generally diagnosed by digital subtraction angiography (DSA), an invasive tool with possible severe complications and ionizing radiation exposure. While noninvasive imaging methods exist, they are incapable of correctly classifying DAVFs by their bleeding risk on the basis of the Cognard classification. These techniques cannot combine optimized temporal and spatial resolution, necessary to define a DAVF’s many characteristics.

However, lead author Myriam Edjlali, MD, of the Université Paris Descartes, and colleagues wrote, “Accelerated four-dimensional (4D) flow imaging techniques based on phase-contrast MR imaging sequences open the possibility of capturing quantitative flow dynamics with high isotropic spatial resolution in a clinically acceptable acquisition time.” Edjlali and colleagues created a study that assessed the feasibility of a selective flow-tracking cartographic procedure being applied to 4D flow imaging and portrayed its usefulness in DAVF characterization.

Eight patients with a total of nine DAVFs underwent 3.0-T MRI and DSA in the study. Four-dimensional flow MRI was performed with the use of a 4D radial phase-contrast undersampled isotropic projection reconstruction pulse sequence with an isotropic spatial resolution of 0.86 millimeters. Images from the MR flow-tracking cartography were independently reviewed by two radiologists who reported the location of arterial feeder vessels, the venous drainage type, and classified DAVFs by the risk of rupture with the Cognard classification system. The findings were then compared with those at DSA. Intermodality agreement was tested for the identification of arterial feeder vessels, draining veins, and Cognard classification.

The study’s results revealed that interreader agreement for shunt location on MR images was perfect. Good-to-excellent interreader agreement was found for arterial feeder vessel identification and matched all cases with shunt location that was defined at DSA. Good-to-excellent agreement was seen between MR cartography and DSA in the definition of the main feeding arteries, presence of retrograde flow in dural sinuses, presence of retrograde cortical venous drainage, presence of venous ectasia, and final Cognard classification of DAVFs.

“The good-to-excellent agreement between findings at MR cartography and DSA may, if confirmed in a larger study, favor the substitution of DSA by noninvasive selective flow-tracking cartography for the purpose of initial grading of cranial DAVFs and for the follow-up of untreated patients,” concluded the study’s authors.