The Role of Biological Effective Dose in Predicting Obliteration After Stereotactic Radiosurgery of Cerebral Arteriovenous Malformations

Objective: To determine whether biological effective dose (BED) was predictive of obliteration after stereotactic radiosurgery (SRS) for cerebral arteriovenous malformations (AVMs). Patients and Methods: We studied patients undergoing single-session AVM SRS between January 1, 1990, and December 31, 2014, with at least 2 years of imaging follow-up. Excluded were patients with syndromic AVM, previous SRS or embolization, and patients treated with volume-staged SRS. Biological effective dose was calculated using a mono-exponential model described by Jones and Hopewell. The primary outcome was likelihood of total obliteration defined by digital subtraction angiography or magnetic resonance imaging (MRI). Variables were analyzed as continuous and dichotomous variables based on the maximum value of (sensitivity–[1–specificity]). Results: This study included 352 patients (360 AVM, median follow-up, 5.9 years). The median margin dose prescribed was 18.75 Gy (interquartile range [IQR]: 18 to 20 Gy). Two hundred fifty-nine patients (71.9%) had obliteration shown by angiography (n=176) or MRI (n=83) at a median of 36 months after SRS (IQR: 26 to 44 months). Higher BED was associated with increased likelihood of obliteration in univariate Cox regression analyses, when treated as either a dichotomous (≥133 Gy; hazard ratio [HR],1.52; 95% confidence interval [CI], 1.19 to 1.95; P<.001) or continuous variable (HR, 1.00, 95% CI, 1.0002 to 1.005; P=.04). In multivariable analyses including dichotomized BED and location, BED remained associated with obliteration (P=.001). Conclusion: Biological effective dose ≥133 Gy was predictive of AVM obliteration after single-session SRS within the prescribed margin dose range 15 to 25 Gy. Further study is warranted to determine whether BED optimization should be considered as well as treatment dose for AVM SRS planning.