An Outcomes-based Analysis of Patients Treated for Arteriovenous Malformation by Frameless LINAC-based Stereotactic Radiosurgery Using 3-Dimensional Rotational Angiography

Abstract

Arteriovenous Malformation (AVM) nidus localization remains a significant radiosurgical challenge. Previous work at our institution utilizing phantom models demonstrated superior nidus visualization and geometric accuracy with use of 3-dimensional rotational angiography (3DRA) compared to digital angiography orthogonal image pairs. Incorporation of 3DRA into the radiosurgery (RS) planning process in 2009 has since allowed for a frameless LINAC-based approach to treatment. We evaluated clinical outcomes for patients treated in the time since adopting this technique. Clinical data were queried for all patients treated at our institution for AVM by single-fraction RS from 2010 to present. Post-RS surveillance for nidus obliteration and radiation-induced changes (RIC) consisted of an MRI every six months for two years, then annually. Beyond two years, confirmatory cerebral angiography was recommended for any patient with MRI suggestive of nidus obliteration. Patients were excluded if they lacked sufficient post-RS radiographic assessment, defined as either MRI, CTA, or cerebral angiography at least two years after RS. Maximum diameter (MD) and volume were measured for each lesion and used to calculate a modified Radiosurgery-based AVM Score (mRBAS). RICs were identified and classified as radiologic, symptomatic, or permanent. Excellent outcome was defined as nidus obliteration without intracranial hemorrhage (ICH) or symptomatic/permanent RIC. Clinical predictors of study outcomes were identified through univariate and multivariate (MVA) logistic regression and backwards elimination was used to optimize a predictive model for each outcome. A total of 65 AVMs in 62 patients were included with a median follow-up of 67 months. Median age at treatment was 31 years. Of these lesions, 58% presented with ICH. Prior intervention took place in 23% of cases. Median MD, volume, and mRBAS were 2.3cm, 1.93cc, and 1.16, respectively. 3DRA was utilized during treatment planning in 78% of cases. Treatments were delivered to a median isocenter and periphery dose of 2000 cGy and 1600 cGy, respectively. Obliteration was observed in 63% of cases with 88% confirmed angiographically. Median time-to-obliteration was 32 months. Radiologic, symptomatic, and permanent RICs developed in 40%, 8%, and 3% of cases, respectively. Post-RS ICH occurred in four patients with one ICH-related death. MVA showed that younger age and smaller MD predicted for obliteration. Larger MD predicted for radiologic RIC. Female gender and lower mRBAS predicted for excellent outcome. Incorporation of 3DRA in the RS planning process has allowed for a frameless approach to AVM treatment. Since instituting this technique, we have achieved obliteration and excellent outcome rates comparable to frame-based approaches with minimal treatment-related morbidity.