Abstract DP173: Predictors of Failure of Endovascular Treatment for Dural Arterio-Venous Fistulas with Cortical Venous Drainage

Introduction: Dural arteriovenous fistulas (dAVFs) with cortical venous drainage (CVD) are arteriovenous shunts with significant risk of intracranial hemorrhage. Endovascular treatment is considered first-line treatment, yet it can be associated with lower complete obliteration and higher recurrence rates when compared to surgery. This study aims to identify predictors of failed endovascular-only treatment for dAVFs with CVD using the Consortium for Dural Arteriovenous Fistula Outcomes Research (CONDOR) database. Methods: The CONDOR registry was queried for patients with dAVFs with presence of angiographically-confirmed CVD who underwent endovascular embolization as first-line therapy. Successful endovascular treatment was defined as durable complete obliteration of the fistula or disconnection of the CVD (regardless of the number of embolization sessions) without the need for additional treatment modalities. Patients who required surgery or radiosurgery following embolization due to persistent CVD, or demonstrated recurrence after initially successful embolization requiring additional treatment, were considered as failure of primary endovascular treatment. Results: A total of 507 patients with dAVFs with CVD that underwent first-line endovascular treatment were included. Of these, 332 (65.5%) had successful initial endovascular embolization and 175 (34.5%) had unsuccessful embolization failing to achieve either complete obliteration or CVD disconnection. Of the 332 initially successful embolizations, 5 had a recurrence of their dAVF requiring surgery or radiosurgery. Endovascular treatment success was thereby achieved in 327 (64.5%) patients, while treatment failure was observed in 180 (35.5%). On multivariable logistic regression, embolization of the middle meningeal artery lowered the odds of failed endovascular treatment (OR: 0.558, p=0.010). The presence of intracranial feeders (p=1.868, p=0.007) and venous drainage into the deep venous system (OR 1.859, p=0.013) increased the odds of failed endovascular treatment. Conclusions: Intracranial feeders and deep venous drainage to straight sinus and vein of Galen were predictors of failed primary endovascular treatment in patients with dAVFs with CVD, while arterial supply by the MMA was predictive of successful endovascular treatment. Predicting which dAVFs are more likely to fail first-line endovascular treatment may help improve patient selection for endovascular-only versus multimodal therapy.