Effect of Prior Hemorrhage on Intracranial Arteriovenous Malformation Radiosurgery Outcomes

Abstract

Background: Intracerebral hemorrhage is simultaneously the most frequent and most debilitating manifestation of intracranial arteriovenous malformations (AVM), but its impact on success and complications of radiosurgery has not been rigorously assessed. In this case-control study, we define the effect of prior hemorrhage on AVM radiosurgery outcomes. Methods: From a prospective, institutional database of 1,400 AVM patients treated with Gamma Knife radiosurgery, unruptured and ruptured AVMs were matched in a 1:1 fashion, blinded to outcome, based on patient demographics, prior embolization (26.6% of each cohort), AVM size (mean volume of unruptured AVMs 3.7 cm3 versus ruptured AVMs 3.5 cm3, p = 0.195), Spetzler-Martin grade (Grade I 17.0%, Grade II 37.8%, Grade III 34.8%, Grade IV 10.4% for each cohort), and radiosurgical treatment parameters (mean prescription dose for unruptured AVMs 20.9 Gy versus ruptured AVMs 21.0 Gy, p = 0.837). There were 270 patients in each cohort. Matched statistical analyses were used to compare the baseline characteristics, obliteration rates, post-radiosurgery latency period hemorrhage risks, and incidences of radiation-induced changes (RIC) between the two cohorts. Results: The actuarial obliteration rates of the two cohorts were similar (unruptured AVMs: 38, 58, and 76% at 3, 5, 10 years, respectively; ruptured AVMs: 40, 60, and 73% at 3, 5, 10 years, respectively; p = 0.592). However, for embolized AVMs, complete obliteration was more likely to be achieved in unruptured lesions (unruptured AVMs: 25, 32, and 54% at 3, 5, 10 years, respectively; ruptured AVMs: 18, 27, and 42% at 3, 5, 10 years, respectively; p = 0.038). Prior AVM rupture resulted in a higher annual risk of post-radiosurgery latency period hemorrhage (ruptured AVMs 2.3% versus unruptured AVMs 1.1%, p = 0.025) but a lower rate of cumulative and symptomatic RIC (cumulative RIC: ruptured AVMs 30.4% versus unruptured AVMs 48.9%, p < 0.0001; symptomatic RIC: ruptured AVMs 7.0% versus unruptured AVMs 12.2%, p = 0.041, respectively). The rates of permanent RIC were similar between the unruptured (2.2%) and ruptured (1.9%) AVM cohorts (p = 0.761). The mean time interval to onset of RIC (unruptured AVMs 13.3 months versus ruptured AVMs 12.1 months, p = 0.783), and the mean duration of RIC (unruptured AVMs 22.0 months versus ruptured AVMs 21.7 months, p = 0.599) were not significantly different between the two cohorts. Conclusions: Prior AVM rupture significantly alters the risk of latency period hemorrhage and RIC following radiosurgery. These effects should be taken into consideration with the multidisciplinary management of AVM patients. Radiosurgery does not significantly alter the natural history of the hemorrhage risks of unruptured and ruptured AVMs unless obliteration is achieved.