Hemorrhagic Outcome of Brainstem Cavernous Malformations following Radiosurgery: Dose-Response Relationship

Abstract

<b><i>Introduction:</i></b> This study aimed to assess the impact of gamma knife radiosurgery on brainstem cavernous malformations (CMs). <b><i>Methods:</i></b> A total of 85 patients (35 females; median age 41.0 years) who underwent gamma knife radiosurgery for brainstem CMs at our institute between 2006 and 2015 were enrolled in a prospective clinical observation trial. Risk factors for hemorrhagic outcomes were evaluated, and outcomes were compared across different margin doses. <b><i>Results:</i></b> The pre-radiosurgery annual hemorrhage rate (AHR) was 32.3% (44 hemorrhages during 136.2 patient-years). The median planning target volume was 1.292 cc. The median margin and maximum doses were 15.0 and 29.2 Gy, respectively, with a median isodose line of 50.0%. The post-radiosurgery AHR was 2.7% (21 hemorrhages during 769.9 patient-years), with a rate of 5.5% within the first 2 years and 2.0% thereafter. The post-radiosurgery AHR for patients with margin doses of ≤13.0 Gy (<i>n</i> = 15), 14.0–15.0 Gy (<i>n</i> = 50), and ≥16.0 Gy (<i>n</i> = 20) was 5.4, 2.7, and 0.6%, respectively. Correspondingly, transient adverse radiation effects were observed in 6.7 (1/15), 10.0 (5/50), and 30.0% (6/20) of cases, respectively. An increased margin dose per 1 Gy (hazard ratio: 0.530, 95% CI: 0.341–0.826, <i>p</i> = 0.005) was identified as an independent protective factor against post-radiosurgery hemorrhage. Margin doses of ≥16.0 Gy were associated with improved hemorrhagic outcomes (hazard ratio: 0.343, 95% confidence interval [CI]: 0.157–0.749, <i>p</i> = 0.007), but an increased risk of adverse radiation effects (odds ratio: 3.006, 95% CI: 1.041–8.677, <i>p</i> = 0.042). <b><i>Conclusion:</i></b> The AHR of brainstem CMs decreased following radiosurgery, and our study revealed a significant dose-response relationship. Margin doses of 14–15 Gy were recommended. Further studies are required to validate our findings.