E-103 angiographic classification of non-traumatic spinal epidural arteriovenous fistulas (seavf)

Abstract

Purpose To propose a SEAVF classification scheme based on their angiographic angioarchitecture and hemodynamic characteristics. Materials and methods The proposed classification is based on the review of 53 consecutive SEAVFs angiographically diagnosed between 1998 and 2014. Each SEAVF type is documented by several angiographic observations. Results Non-traumatic SEAVFs can be divided into two main categories: Type 1 SEAVFs are congenital high-flow arteriovenous shunts typically seen in pediatric patients without concomitant venous thrombosis; they often present with cord compression from dilated epidural venous pouches in children of all ages or high-output cardiac failure in infants. Type 2 SEAVFs are acquired slow-flow arteriovenous shunts associated with extensive venous thrombotic changes. Type 2 SEAVFs are further subdivided according to their connection with the perimedullary venous system (PVS) and their drainage pattern. Type 2a SEAVFs (Figure 1A) are generally benign lesions not connected to the PVS. While Type 2b SEAVFs (Figure 1B) are connected to the PVS, a competent anti-reflux mechanism prevents retrograde drainage; increased perimedullary venous pressure starts developing in order to maintain a drainage gradient towards the epidural venous plexus. Failure of the anti-reflux mechanism in Type 2c (Figure 1C) results in the development of retrograde perimedullary drainage. Type 2d (Figure 1D) is characterized by an isolated venous pouch with a radiculomedullary vein as the sole or highly dominant drainage pathway, leading to severe perimedullary venous hypertension. This type of SEAVFs is morphologically close to and often mistaken for a spinal dural arteriovenous fistula. Conclusion We propose an angiographically-based classification system for non-traumatic SEAVFs that takes into account the morphology of the arteriovenous shunts and their drainage pathways. This classification scheme has in our experience demonstrated its clinical value, showing excellent correlations with the clinical presentation and pathogeny of the SEAVFs. Disclosures D. Sorte: None. L. Gregg: None. E. Wyse: None. P. Gailloud: None.