The Impact of Magnetic Resonance Imaging Findings in Predicting Neurological Status Pre- and Post-Treatment of Spinal Dural Arteriovenous Fistulas: A 22-Year Experience in a Neurovascular and Spine Center

Abstract

Background: Successful treatment of spinal dural arteriovenous fistulas (SDAVF) requires prompt diagnosis with definitive fistula localization and non-delayed treatment. Magnetic resonance imaging (MRI) is used for the screening and follow-up of SDAVF, although the value of MRI signs such as myelopathy and flow voids is controversial. Therefore, we investigated the predictive value of MRI signs pre- and post-treatment and their correlation with the neurological status of SDAVF patients. Methods: We retrospectively analyzed the clinical records of 81 patients who underwent surgical or endovascular treatment for SDAVF at our hospital between 2002 and 2023. A total of 41 SDAVF patients with follow-up MRI of 4.6 [2.9–6.5] months (median [interquartile range]) post-treatment and clinical follow-up of 3, 6, and 12 months were included. Results: The extent of pretreatment myelopathy was seven [6–8] vertebral levels, with follow-up MRI showing no myelopathy in 70.7% of cases. The pretreatment flow voids extended over seven [4.5–10] vertebral levels and completely disappeared on follow-up MRI in 100% of cases. The modified Aminoff–Logue scale of disability (mALS) was four [2–7] pretreatment and two [0–4.5] at the third follow-up, with improvement in 65.9% of patients. The American Spinal Injury Association motor score (ASIA-MS) was 97 [88–100] pretreatment and 100 [95–100] at the third follow-up assessment, with 78% of patients improving. Pretreatment ASIA-MS correlated with the extent of myelopathy at admission (R2: 0.179; 95% CI: −0.185, −0.033; p = 0.006) but not with flow voids at admission, while pretreatment mALS showed no correlation with either MRI signs. The improvement in ASIA-MS and mALS between admission and the last follow-up showed no correlation with the extent of pretreatment myelopathy and flow voids or with pos-treatment MRI changes. The diagnostic sensitivity of magnetic resonance angiography (MRA) for localization of the fistula was 68.3% (28/41). Conclusions: The severity of the clinical condition in SDAVF patients has a multifactorial cause, whereby the ASIA-MS correlates with the extent of myelopathy pretreatment. MRI changes after treatment showed no correlation with the clinical outcome and cannot be used as a prognostic factor.

Keywords

• SDAVF
• spinal arteriovenous fistula
• myelopathy
• endovascular treatment
• surgical treatment
• spinal cord edema