CLINICAL STUDIES THE ROLE OF NON-INVASIVE COMPUTED TOMOGRAPHY IN PATIENTS WITH SUSPECTED DURAL FISTULAS WITH SPINAL DRAINAGE Peter Zampakis, M.B.B.S., Ph.D. Department of Neuroradiology, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland Celestine Santosh, M.B.B.S. Department of Neuroradiology, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland William Taylor, M.B.B.S. Department of Neurosurgery, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland Evelyn Teasdale, M.R.C.P. Department of Neuroradiology, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland Reprint request: Evelyn Teasdale, M.R.C.P., Department of Neuroradiology, Institute of Neurological Sciences, Southern General Hospital, 1345 Govan Road, Glasgow G51 4TF Scotland. E-mail: evelyn.teasdale @sgh.scot.ns.uk Received, February 4, 2005. Accepted, September 7, 2005. OBJECTIVE: Accurate diagnosis and demonstration of the angioarchitecture and lo- calization of a spinal dural arteriovenous fistula is of crucial importance before treatment. Selective spinal angiography (DSA) has been considered the standard technique, but is invasive, time-consuming, and may be falsely negative. This report evaluates the use of noninvasive vascular imaging (computed tomographic and mag- netic resonance angiography [MRA]) in patients suspected to have a dural fistula with spinal drainage. METHOD: Ten consecutive patients had DSA and multidetector computed tomo- graphic angiography (MDCTA), eight also had MRA. Nine were men with an average age of 67 years. In nine patients, the diagnosis was confirmed at surgery or intravas- cular treatment. Eight were proven to have a spinal dural fistula. In two, the fistula was within the cervical cranial dura. In all patients, the venous drainage involved only the spinal venous plexus. RESULTS: MDCTA identified the level of the feeding artery in nine patients. In two cases, selective DSA failed to show the abnormality found on MDCTA, but both were confirmed at surgery. MRA was diagnostic in a case in which the lesion was not accurately depicted by either MDCTA or DSA. MRA was less accurate than MDCTA in determining the level of the feeding artery. CONCLUSION: MDCTA and MRA can direct and focus DSA. MDCTA gives addi- tional useful three-dimensional bone detail and localization information for the sur- geon. It may replace DSA if surgery is the planned treatment. KEY WORDS: Dural fistula, Multidetector computed tomographic angiography, Magnetic resonance angiography, Vasculitis Neurosurgery 58:686-694, 2006 DOI: 10.1227/01.NEU00199163.10539.56 www.neurosurgery-online.com D ural arteriovenous fistulas (DAVF) commonly present in middle-aged people with a progressive spastic paraparesis, and are the most common spinal vascular malformation to present in this way. The fistula results in cord ischemia, which, if left untreated, causes permanent neurological deficits. The diagnosis is most often suspected on the basis of the clinical history and the typical findings in magnetic resonance imag- ing (MRI) scans: high signal within the cord (ischemia) on T2-weighted sequences and flow voids of abnormal veins in the adjacent subarachnoid space (Fig. 1A). Selective spinal angiography (DSA) has been the definitive examination in the diagno- sis and planning of treatment (3), but it is invasive and time-consuming. It also requires specialized expertise, but rarely causes any serious complication (2). Surgery has been shown to be twice as effective as interven- tional catheter treatment in curing a DAVF (21) and requires only demonstration of the side and level of the fistula and the origin of the arterialized vein. Consequently, if nonin- vasive angiography could provide this infor- mation, it could replace DSA in the investiga- tion of patients undergoing surgery. There is, however, little information about the use of noninvasive angiographic methods in the assessment of spinal DAVF. Magnetic resonance angiography (MRA) has been re- 686 | VOLUME 58 | NUMBER 4 | APRIL 2006 www.neurosurgery-online.com