Editorial: Radiosurgery and deep arteriovenous malformations

Abstract

Cheng et al.1 present the results of Gamma Knife surgery, collected over 3 decades, in 182 patients with basal ganglia (85) and thalamic (97) arteriovenous malformations (AVMs) followed up for more than 2 years. The mean nidus volume was 3.4 cm3 (range <1 to 29 cm3), and the mean margin dose was 21 Gy (range 10–28 Gy). Half of the AVMs were angiographically obliterated after a single GKS treatment, and roughly one-third of those retreated were angiographically obliterated, giving an overall definitive cure rate of 58%. If one includes MR imaging–suspected obliterations, the rate is about 66%, and if one includes lesions with only a residual venous shunt, the rate is close to 75%, but the authors rightly caution against doing so. These results were achieved at the cost of a 10% incidence of symptomatic radiationrelated necrosis and a 5% incidence of permanent new deficits related to radiation effect. As expected, predictors of angiographically confirmed obliteration were as follows: small lesion size, favorable shape (a surrogate for fewer isocenters), high margin dose, and no prior embolization. In a multivariate analysis, the only predictor of permanent deficits was unruptured status. Despite the fact that 76% of the lesions had previously bled and more than 98% were symptomatic, only 2 hemorrhages were noted in the posttreatment period (both cases fatal), and only 3 patients had symptoms progression. If one adds these patients, the overall morbidity rate was 7.5% and the mortality rate was 1.0% (8.5% combined). Given that 23% of the lesions were Spetzler-Martin Grade IV or V, 93% were Grade III or worse, and 31% had failed prior surgery or embolization, the results seem quite laudable. The authors conclude that in young patients with ruptured basal ganglia/thalamic AVMs in whom microsurgery is thought to pose a high risk by an experienced surgeon, radiosurgery is a reasonable alternative. I would concur, but ask a simple question: what is the definition of “experienced” when it comes to microsurgical treatment of basal ganglia and thalamic AVMs? I suspect that even busy surgeons in large cerebrovascular referral centers will not operate on more than 1 or 2 of these lesions a year and that surgeons associated with even the busiest radiosurgical units may operate on no more than 5 in a lifetime. Therefore, I suspect the truth, in most practice settings, is actually the reverse of what Cheng and colleagues propose. The first question asked is not, “Is this basal ganglia/thalamic AVM good for an ‘experienced microsurgeon?’” but rather, “Is it small enough and round enough that we’ll be able to get 20 Gy on the margin without causing a major radiation-induced deficit, and, if not, can we shrink the lesion with embolization to get it to that size and potentially obliterate the bleed site at the same time?” The data presented here do not necessarily invalidate this real-world approach, but they do not give us as much guidance as we’d like. For example, while there was a lower obliteration rate in the 38 patients treated with preradiosurgical embolic agents, the hemorrhage rate was far lower than what natural history studies would predict, given that 42% of patients were incompletely protected. Future prospective multicenter efforts will be needed if we are to ever really understand what is the best treatment for lesions so heterogeneous and rare, especially when the treatments are so dependent on user skill and experience and ongoing technological improvements. (http://thejns.org/doi/abs/10.3171/2011.10.JNS111419)