A multicenter retrospective study of frameless robotic radiosurgery for intracranial arteriovenous malformation.

Introduction: CT-guided, frameless radiosurgery (SRS) is an alternative treatment to traditional catheter-angiography targeted, frame-based methods for intracranial arteriovenous malformations (AVMs). Despite the widespread use of frameless radiosurgery for treating intracranial tumors, its use for treating AVM is not well described. Methods: Patients who completed a course of single fraction SRS at The University of North Carolina or Georgetown University between 4/1/2005 – 4/1/2011 with single fraction SRS and received at least one follow-up imaging study were included. All patients received pre-treatment planning with CTA ± MRA and were treated on the CyberKnife (Accuray) radiosurgery system. Patients were evaluated for changes in clinical symptoms and radiographic changes evaluated with MRI/MRA and catheter-angiography. Results: 26 patients, 15 male and 11 female, were included in the present study at a median age of 41 years (IQR, 26-55 years). The Spetzler Martin grades of the AVMs included seven Grade I, twelve Grade II, six Grade III, and one Grade IV with fourteen (54%) of the patients having a pre-treatment hemorrhage. Median AVM nidal volume was 1.62cm3 (IQR, 0.57-8.26 cm3) and was treated with a median of 1900 cGy to the 80% isodose line. At median follow-up of 25 months (IQR, 19-36 months), 15 patients had a complete closure of their AVM, 6 patients had a partial closure, and 5 patients were stable. Time since treatment was a significant predictor of response, with patients experience complete closure having on average 11 months more follow-up than patients with partial or no closure (p = 0.03). One patient experienced a post-treatment hemorrage at 22 months. Conclusions: Frameless SRS can be targeted with non-invasive MRI/MRA and CTA imaging. Despite the difficulty of treating AVM without catheter angiography, early results with frameless, CT-guided SRS suggest that it can achieve similar results to frame-based methods at these time points.