Efficacy of Surgical Resection for Tentorial Dural Arteriovenous Fistulae. Presented on Behalf of the Consortium for Dural Arteriovenous Fistula Outcomes Research (CONDOR)

Introduction: Dural arteriovenous fistulae (dAVF) are rare cerebrovascular anomalies. One important subgroup of these lesions are tentorial dAVFs. This type of dAVF is more likely to have aggressive features, making them prone to symptomatic presentations and hemorrhage. Here, we utilize an international, multi-center registry to describe the presentation of tentorial dAVFs, quantify outcomes following resection, and identify predictors for surgical outcome.

Methods: Patients were obtained from the CONDOR database. Variables analyzed included age at treatment, sex, comorbidities, clinical presentation, arterial anatomy, venous anatomy, and Borden grade. To identify factors predicting surgical outcome, a multivariable logistic regression for each endpoint was performed using stepwise backwards selection in SAS studio v3.8. The primary efficacy endpoint in our study was defined as complete surgical resection.

Results: A total of 1241 dAVFs were available for analysis in the CONDOR database, including 68 tentorial dAVFs. This included 44 male patients and 24 female patients. Presentation was due to hemorrhage in 59% (40/68) of cases, due to non-hemorrhagic neurologic deficit in 31% (21/68) of cases, and due to flow-related symptoms in 24% (16/68) of cases. The most common arterial feeder was the middle meningeal artery in 66% (43/65) and the most common venous drainage was via the deep cerebral vein 44% (28/63). Preoperative embolization was utilized in 62% (42/68) of cases. Extent of resection data was unavailable for two patients. For those with available data, complete resection was achieved in 82% (54/66) of patients. In our multivariate analysis, we found that prior ischemic stroke decre ased probability of complete resection (95% CI, 0.002-0.873).

Conclusion : Complete surgical resection of tentorial dAVFs can be accomplished in a high number of these lesions. While prior ischemic stroke was a predictor for incomplete resection, it is unclear what caused this relationship. Further work is necessary to identify optimal surgical approaches for these lesions.