Object: Little information is available on follow-up MR imaging after treatment of spinal dural arteriovenous fistulas (DAVFs). The authors studied MR imaging findings in treated spinal DAVFs in relation to clinical outcomes.
Methods: A retrospective review of patients with spinal DAVFs who had undergone both pre- and postoperative spinal MR imaging was conducted. Postoperative MR images were obtained as routine follow-up studies or because of subjective or objective clinical deterioration. Several pre- and posttreatment MR imaging characteristics were evaluated by 2 neuroradiologists blinded to the clinical outcome. Clinical outcomes of motor, sensory, and urinary function (in relation to the patient's preoperative status) at the time of the postoperative MR imaging were obtained from the clinical record. The chi-square, Fisher exact, and rank-sum tests were performed to correlate imaging findings and changes with clinical outcomes.
Results: Thirty-four patients met inclusion criteria. Treatment was surgical in 33 patients and endovascular in 1 patient. Follow-up MR imaging was performed at a mean 168 ± 107 days after treatment. Twenty-seven patients (79.4%) were either clinically stable or improved, and 7 (20.6%) experienced worsening in one or more clinical outcomes. Most patients were found to have improvement of MR imaging changes. However, some degree of persistent spinal cord signal abnormality, enhancement, and swelling was observed in 31 (91.2%), 29 (85.3%), and 18 (52.3%) patients, respectively. Changes in these MR imaging characteristics compared with preoperative MR imaging did not correlate with clinical outcomes (p > 0.05), with the one exception of a significant correlation between change in urinary function and extent of spinal cord contrast enhancement (p = 0.026), a correlation of uncertain importance. Ten of the 34 patients underwent posttreatment digital subtraction angiography, and 3 of these patients had recurrent/residual DAVFs. Worsening of motor function significantly correlated with recurrent/residual DAVF (p = 0.053).
Conclusions: Spinal cord abnormalities persist on postoperative MR imaging studies in patients with treated spinal DAVFs, and although they tend to mildly improve with time, these changes may not correlate with clinical outcomes. However, regardless of imaging findings, worsening motor function may correlate with a recurrent or residual DAVF.