Abstract
Background and purpose: To evaluate whether a radiosurgery-based arteriovenous malformation (AVM) scale (RBAS) could be used to predict obliteration of brain arteriovenous malformations (bAVMs) supposed for combined endovascular embolization (EMB) and gamma knife surgery (GKS) treatment. Methods: bAVM patients who underwent GKS with or without previous EMB from January 2011 to December 2016 at our institution were retrospectively reviewed. The patients were categorized into a combined treatment group and a GKS group. A 1:1 propensity score matching (PSM) was used to match the two groups. Pre-EMB and pre-GKS RBAS were assessed for every patient. Multivariate analysis was performed to find factors associated with complete obliteration in the combined treatment group. Survival analysis based on sub-groups according to RBAS was performed to compare obliteration rate and find cutoffs for appropriate treatment modalities. Results: A total of 96 patients were involved, and each group comprised 48 patients. There was no difference between the two groups in terms of obliteration rate (75.0 vs. 83.3%, p = 0.174). Pre-EMB RBAS (p = 0.010) and the number of feeding arteries (p = 0.014) were independent factors associated with obliteration rate in the combined treatment group. For the combined treatment patients, sub-group analysis according to pre-EMB RBAS (score <1.0, 1.0-1.5, and >1.5) showed statistical difference in obliteration rate (p = 0.002). Sub-group analysis according to RBAS between the two groups showed that the obliteration rate of the GKS group is significantly higher than the combined group when RBAS >1.5 (47.4 vs. 66.7%, p = 0.036). Conclusions: The RBAS is proposed to be efficient in predicting obliteration of bAVMs supposed to receive combined EMB and GKS treatment. Patients with RBAS >1.5 are inclined to be more suitable for GKS instead of the combined treatment.