Abstract
Background. Proton beam radiotherapy of arteriovenous malformations (AVM) in the brain has been performed in Uppsala since 1991. An earlier study based on the first 26 patients concluded that proton beam can be used for treating large and medium sized AVMs that were considered difficult to treat with photons due to the risk of side effects. In the present study we analyzed the result from treating the subsequent 65 patients.
Material and methods. A retrospective review of the patients’ medical records, treatment protocols and radiological results was done. Information about gender, age, presenting symptoms, clinical course, the size of AVM nidus and rate of occlusion was collected. Outcome parameters were the occlusion of the AVM, clinical outcome and side effects.
Results. The rate of total occlusion was overall 68%. For target volume 0–2cm3 it was 77%, for 3–10 cm3 80%, for 11–15 cm3 50% and for 16–51 cm3 20%. Those with total regress of the AVM had significantly smaller target volumes (p < 0.009) higher fraction dose (p < 0.001) as well as total dose (p < 0.004) compared to the rest. The target volume was an independent predictor of total occlusion (p = 0.03). There was no difference between those with and without total occlusion regarding mean age, gender distribution or symptoms at diagnosis. Forty-one patients developed a mild radiation-induced brain edema and this was more common in those that had total occlusion of the AVM. Two patients had brain hemorrhages after treatment. One of these had no effect and the other only partial occlusion from proton beams. Two thirds of those presenting with seizures reported an improved seizure situation after treatment.
Conclusion. Our observations agree with earlier results and show that proton beam irradiation is a treatment alternative for brain AVMs since it has a high occlusion rate even in larger AVMs.
Brain arteriovenous malformations (AVM) are the most common source of intracranial hemorrhage in the younger population, causing one third of the bleedings in patients younger than 40 years [Citation1]. The malformations may also cause epilepsy, headache and various other neurological symptoms. Today many asymptomatic AVMs are found since radiology is more used.
The available treatments are surgical resection, endovascular intervention and radiotherapy with photons or proton beams. They have various advantages and risks, and are often combined to get the best result. Neurovascular centers have their own strategies based on their individual experiences and available techniques. Patients, especially with incidental AVMs, often prefer radiotherapy as it is non-invasive. It is known though that there can be serious long-term adverse effects after irradiation. Development of cysts, persistent edema and radiation-induced tumors have been described, e.g. [Citation2]. The three main radiotherapy techniques, proton beams and photon beams delivered with gamma knife or linear accelerator (LINAC), have comparable biological effects on the malformation. A difference between proton beams and photon beams (gamma knife/LINAC) is that proton beams have a finite range with a sharp dose fall beyond which no dose is given, while the photon beams deliver dose along their path through the patient's head. Proton beams are therefore capable of giving a homogenous dose to the target and a substantially lower dose to surrounding normal brain tissues compared to photon beams. The proton beam treatment is therefore regarded to be more advantageous than photon beams when treating large AVMs and also when treating children since the irradiation of the normal brain tissue is minimized. One disadvantage is that the proton beam treatment can be elaborate and calls for an expensive cyclotron to generate the high energy proton beams needed for therapy.
The “The Svedberg” Laboratory (TSL), (named after Nobel Prize winner in Chemistry Theodor Svedberg) is a cyclotron facility offering proton beams with energies suitable for radiotherapy. The proton beam therapy at this facility is a joint project of Uppsala University and Uppsala University Hospital. The first proton beam treatments of AVMs at TSL were performed in 1991. Between 1991 and 1997 26 patients were treated. The results [Citation3] showed that proton beam treatment was successful in treating large and medium sized AVMs. Since then another 80 patients have been treated. In 65 of these patients treatment and follow-up are completed while the rest still are under follow-up. In the present paper we report the results and experiences from the 65 patients. Primary outcome was the rate of total occlusion of the AVM. Secondary outcomes were adverse effects and clinical results.
Methods
Patients
Sixty-five patients with brain AVM, treated at Uppsala University Hospital with proton beam radiotherapy between November 1997 and November 2008, i.e. after the earlier report [Citation3], were included in this material. For a description of the patients see . Sixty-three patients were treated once and two patients twice, resulting in 67 proton beam treatments altogether. Thirty-two AVMs (48%) were treated with proton beams only, 18 (27%) with proton beams after one endovascular procedure and 17 (25%) with proton beams after several endovascular, surgical or radiotherapy procedures.
Table I. Summary of the initial clinical data on the 65 patients treated with proton beam therapy.
The follow-up protocol after radiotherapy included magnetic resonance imaging (MRI) after 6, 12, 24 and 36 months followed by digital subtraction angiography (DSA) if there was no obvious AVM left. Otherwise the MRI controls continued.
Not all of the patients followed this protocol for various reasons, e.g. other diseases and reluctance to do DSA just to confirm the MRI findings. The last radiology in the present study was done median 49 (6–140) months after treatment.
The AVMs were also assessed using the Spetzler-Martin scale [Citation4]. This assessment was done based on the original extent of the malformation. Post- irradiation brain edema was considered to be present when described in the radiology reports done by neuroradiologists who reviewed MRIs and computed tomography (CT) scans for clinical purposes.
Treatment planning and delivery of proton beams
The procedure in Uppsala regarding treatment planning and proton beam treatment was described in detail earlier [Citation3]. For defining the target, DSA and CT were done using a stereotactic system (Leksell® Coordinate Frame G, Elekta AB, Stockholm Sweden). The stereotactic system was used to define the position of the target in relation to fiducial markers consisting of small titanium screws implanted into the skull. The target, i.e. the AVM nidus as defined from DSA, was transferred to the CT study which formed the basis for the dose plan. The calculation of the nidus volume and the treatment planning was carried out with a commercial treatment planning system (Helax, Uppsala, Sweden) (for references see [Citation3]). In the treatment planning system, minor modifications of the target volume (the DSA defined nidus shape) were made manually, in order to exclude normal tissue. The geometrical relationship between the target and the fiducials was then utilized for accurate positioning of the patient during treatment [Citation5].
The treatment was carried out using either two or three coplanar proton beam directions. Each beam was optimized with respect to: 1) lateral field extension with an individually machined brass collimator; 2) depth penetration with an individually shaped range filter (bolus); and 3) Bragg peak modulation over the target, with a rotating, stepped absorber which resulted in a homogeneous dose distribution over the target volume [Citation6]. The beam directions were chosen to minimize the dose to healthy tissues outside the target volume.
The proton beam treatment was carried out at TSL using a fixed horizontal beam. The energy of the proton beam was 172 MeV when entering the patient, which corresponds to a depth penetration in soft tissue of 205 mm. With a maximum beam diameter of 98 mm most intracranial targets could be irradiated. The patients were treated in a sitting position with the head fixed to the treatment chair by an individually formed helmet and a bite-block. With a fixed horizontal beam the only way to change the beam direction was to rotate the treatment chair with the patient. The chair could be precisely positioned in relation to the beam and to the calculated target volume which was guided by x-ray imaging of the implanted fiducials. Corrections of the position were calculated by a separate algorithm [Citation5].The error between repeated sessions was < 1 mm [Citation5].
Doses and fractions
In this paper we report the proton beam treatment as the total dose, which was delivered in fractions. The total dose was in 64 treatments 18–25 physical Gy, delivered in two fractions separated by a 24-hour interval. In the remaining three treatments 35 Gy was delivered in 5 equal fractions (see Supplementary Table I, available online at: http://informahealthcare.com/doi/abs/10.3109/0284186X.2015.1043023).
Statistics
Descriptive and analytic statistics were performed using Statistica 12 (Stat soft Inc., Tulsa, OK, USA). The patient group with total regress of the AVM after proton beam treatment was compared to a second group including patients with partial regress and no effect. Comparison between groups was done using Mann-Whitney U-test and Fisher exact. Multiple regression was used in order to evaluate independent predictive values. A difference was considered statistically significant when p < 0.05.
Ethical considerations
Uppsala regional ethical committee for human research approved the study.
Results
Occlusion rate
One patient deceased from rapidly progressive amyotrophic lateral sclerosis before any follow-up could be done. In the remaining group 45 of 66 (68%) proton beam treatments led to a complete occlusion of the AVM. Twelve (18%) resulted in a partial occlusion and nine (14%) had no effect according to the available radiology.
We compared the characteristics of treatments leading to total regress to those with partial regress and no effect as one group. The AVM that became completely occluded had significantly smaller target volumes and received significantly higher fraction doses as well as total doses compared to the rest (). As the doses depend on the target volume this was further analyzed with multiple regression. The results from that showed that the target volume was an independent predictor for total occlusion of the AVM (p = 0.03). We further observed that AVM Spetzler-Martin grade I–II was more common in the total occlusion group. There was no difference in age or gender distribution between the two groups. Total occlusion was achieved in 78% of AVMs treated with proton beam only compared to 76% after one previous endovascular treatment.
Table II. In this table is the group with total regress of the AVM compared to those with partial or no effect. The univariate analysis showed that the group with total occlusion had lower Spetzler-Martin grade, smaller target volume and higher fraction as well as total dose compared to those with partial or no effect. Multivariate analysis showed that the target volume was independent predictor of total occlusion (p = 0.03). The group with total regress also had significantly more patients with transient brain edema after the irradiation.
shows a plot of target volumes and occlusion. Two of the volumes are retreated AVMs. We divided in volume intervals and found for target volume 0–2 cm3 occlusion rate of 77%, for 3–10 cm3 80%, for 11–15 cm3 50% and for 16–51 cm3 20%. We also calculated the cumulative numbers of patients over time with a complete occlusion of the AVM (). The figure shows that if there will be a complete occlusion, it seems that it will be within the first five years.
Irradiation-induced side effects
The development of irradiation-induced side effects was in most cases monitored with MRI and in a few cases with CT. Forty-one (62%) developed edema around the AVM, visible on MRI or CT. We compared patients with and without edema and found that those with edema had significantly higher rate of total occlusion than those without (80% vs. 50%, p = 0.01). Otherwise, there was no difference in target size, fraction dose, total dose, Spetzler-Martin grade, mean age or gender distribution between the two groups.
In most cases the edema was mild and transient but in 11 cases this was symptomatic. Two patients developed permanent neurological deficits. One of them had an AVM nidus of 51 cm3 that decreased to 45 cm3 after the first proton beam treatment. He was then asymptomatic but after a second treatment with proton beams he developed permanent cognitive and motor deficits. The other patient had a nidus of 1 cm3 that was treated once with proton beam. The nidus was located in a cerebellar peduncle and the patient developed edema there that spread to the brain stem giving permanent cranial nerve and cerebellar symptoms.
Clinical outcome
Two patients had new bleedings after the proton beam treatment. The first was a patient in whom the treatment had no radiological effect on the AVM. She was operated on in the 1950s in an attempt to remove the AVM. She also had epilepsy for many years. Three years after the radiotherapy she had an intraventricular bleeding. The other patient had an AVM that had been treated with five endovascular procedures first. This was followed by radiotherapy but with no obvious effect on the AVM. Four years later, however, he had a new hemorrhage. Six months after that he had a second proton beam treatment and since then there have been no more hemorrhages. Unfortunately he has developed cognitive and motor deficits (see above).
Twenty-two patients had seizures as the first symptom at diagnosis. Fifteen of them (68%) reported an improved seizure situation after radiotherapy. One patient had a worse situation and one patient that was diagnosed with seizures had an intracranial hemorrhage (see above).
Development of a cavernoma
One patient was treated for an AVM in the left temporal lobe that was an incidental finding. The target volume was 1 cm3 and it was treated 2 fractions of 12 Gy each. In the following MRI controls the AVM gradually disappeared without any apparent complications. DSA three years after irradiation showed that the AVM was completely occluded. Five years after treatment the patient developed quite severe headaches. MRI showed a new and prominent edema in the left temporal lobe. The edema surrounded a small contrast enhancing object. It was considered to be a late adverse reaction to irradiation and was initially just followed. The problems with headaches increased as did the size of the edema and the contrast enhancing object. Eventually the patient developed signs of increased intracranial pressure and the contrast enhancing object was removed with microneurosurgery. The histopathology suggested a cavernoma. MRI control two years after surgery showed no recurrence. The intracranial pressure slowly but spontaneously normalized ().
Discussion
We had an overall occlusion rate of 68% which is comparable to our own earlier results and in range of reports in the literature on proton beam treatment of brain AVM. Hattangadi et al. [Citation7] report the experiences from treating 248 patients with mean nidus of 3.4 cm3. They used a single fraction with the most common dose 15 Gy. With this strategy they achieved total occlusion in 64.6% of the whole material. Our results are also in line with those using the gamma knife. In recent studies [Citation2,Citation8] 69% obliteration after one treatment was found.
It seems that one thing proton beam and photon treatment have in common is that the size of the target volume and the radiation dose are the most important factors predicting the occlusion rate. This was supported by our results which showed that the size of the target volume was an independent predictor for total occlusion. We also found that larger total dose as well as fraction dose was delivered to those AVMs that occluded. Paul et al. [Citation8] had similar results using the gamma knife, with positive predictors of occlusion including smaller AVM, compact nidus and higher margin dose.
The challenge, especially in large AVMs is to deliver a radiation dose that is large enough for occlusion but still not causing an injury to the surrounding brain. Different strategies of staged treatment has therefore been tried, mostly regarding the dose but also the volume. Hattangadi et al. [Citation9] used two fractions in treating 64 high risk large AVMs. The median total dose was 16 Gy and the nidus median 23 cm3. Occlusion rate was 15%. In a report by Kano et al. [Citation10] the results from gamma knife treatment of AVMs > 10 cm3 was presented. They used dose staged treatment which resulted in total occlusion of 11/47 (23%) of the AVMs.
Our strategy was also to use two fractions also in the smaller AVMs and 5 fractions in some of the larger malformations. The median total dose in our material was 23 Gy which could contribute to the larger degree of occlusion. Regarding results in volume intervals, we had 50% total occlusion for 11–15 cm3 and 20% for 16–51 cm3. Our results therefore support the view that proton beam therapy is an alternative in larger AVMs.
In another study on gamma knife treatment of AVM [Citation11] single dose was used with median margin dose of 18 Gy (13–24). They found an overall occlusion rate of 56%. No difference was seen in the occlusion rate between those < 10 cm3 and ≥ 10 cm3 and the cut-off for better results in their study was 3.5 cm3. However, they showed that treating large AVMs with gamma knife resulted in significantly more new neurological deficits compared to the smaller malformations. It is possible that the Bragg peak distribution offers a more advantageous dose distribution especially in large AVMs, protecting the surrounding brain.
One type of staged treatment is to reduce the nidus with endovascular techniques before the radiotherapy is done. However, some concerns have been raised considering incomplete endovascular treatment, and there are results from non-randomized studies indicating that there is a lower occlusion rate with radiosurgery following embolization. In a recently published report [Citation12] evidences for endovascular-induced angioneogenesis were reviewed. These include expression of vascular endothelial growth factors and inflammatory mechanisms (for references see [Citation12]). It has been observed that the change in morphology of the target after neurointervention can make the treatment planning for radiosurgery more difficult. Furthermore, it has been hypothesized that the embolization material could attenuate and deviate the irradiation, but presently there are no conclusive studies on these questions. Our results showed 78% total occlusion in AVMs treated with proton beam only compared to 76% after a proton beam treatment following one previous endovascular treatment. Although the number of patients is small, the result does not support the view that endovascular treatment before radiotherapy with proton beam is a problem.
AVMs are heterogeneous and can be described in many different ways regarding feeders, associated aneurysms, flow pattern, angioneogenesis, draining vessel, venous ectasies, venous pattern and venous rerouting. Reports have shown that this angio- architecture can be correlated to the occlusion rate (see, e.g. [Citation8,Citation13]). The size of this patient cohort did not allow an analysis of these numerous classes of AVMs, but our results in line with those of others support the conclusion that target volume is a robust predictor of success in occluding the nidus.
Radiation-induced side effects are defined and described in different ways in the literature, see e.g. Barker et al. [Citation14]. They defined complications as neurological deficits, either new or exacerbation of preexisting deficits. They analyzed their large database of 1250 proton beam treated AVMs and found, using multivariate analysis, that the risk for complications were related to a model consisting of age, location of AVM, volume of nidus and treatment. Parkhutik et al. [Citation15] studied 102 AVMs treated with gamma knife and defined radiological complications. Edema and blood-brain-barrier breakdown were common and necrosis occurred in 6%. The radiological complication was considered major in 19% and these patients had significantly increased risk for neurological deficits, increased intracranial pressure and brain hemorrhage.
We studied mainly the development of edema, radionecrosis, and possible tumors. Two thirds of the patients developed a transient edema, which was found to be significantly more common in the group with total occlusion of the AVM. We found no relation to the other parameters studied. Our results are supported by the report by Van der Bergh et al. [Citation16]. They studied brain AVMs treated with LINAC and quantified the post-irradiation edema. They found that those with more extensive edema had higher occlusion rate and also more often a single draining vein. They speculate on different mechanisms for the edemas, with the edema close to the nidus possibly being an irradiation effect on normal brain tissue, while the more extensive edema may be due to the venous occlusion. As the proton beam gives a sharp dose fall towards normal brain tissue it is possible that the edema seen in our material could be due to other factors involving the gradual changes in the angio-architecture. This has to be more studied. Eleven of our patients with edema had symptoms and two of these developed long lasting neurological deficits. One of the two patients with long lasting neurological deficits had two proton beam treatments.
The sharp dose fall and minimal irradiation of normal brain tissue makes proton beam radiosurgery especially beneficial in pediatric patients. Walcott et al. recently published their long-term results from treating 44 pediatric patients with AVM, where 40.9% had complete occlusion of the AVM after 3.5 years [Citation17].
Based on the experiences from irradiation to the brain in other diseases, there is a concern that patients treated with radiosurgery would develop intracranial tumors years after treatment. From what is known so far, the risk seems to be low. In a material of 1309 gamma knife treated AVMs three radiation-induced tumors were identified [Citation18]. In a review from 2003 only four cases of tumors after any type of radiosurgery were identified in the literature, together with two cases from the author's own experience [Citation19]. One patient in our material developed an intracranial mass five years after irradiation and two years after total regression of the AVMs was seen on DSA. The mass was removed and on histopathology it resembled a cavernoma. This has not been described before. Cavernomas are not true neoplastic masses, and its genesis is unclear. One can speculate that it either was occult at the time of radiosurgery or that it resulted from micro-bleedings and impaired venous draining. The latter possibility is especially likely since the patient developed signs of increased intracranial pressure that gradually resolved over a period of a few months after the mass was removed.
Twenty-two patients in our material had seizures as the first symptom at diagnosis. Fifteen of them (68%) reported an improved seizure situation after radiosurgery. The effect of radiosurgery on seizures has been observed before, see e.g. [Citation20], and the mechanisms are not clear. One obvious explanation is that the epilepsy resolves when the epileptogenic lesion is treated. This is supported by at meta- analysis in which seizure control after radiosurgery, microneurosurgery and neurointervention was compared. The highest overall seizure control was achieved by microneurosurgery, but the cases in which radiosurgery achieved complete occlusion had even better seizure control [Citation21]. It has been suggested that there are more direct anticonvulsive effects of the irradiation, i.e. effects on protein synthesis which in turn affect both the GABA-ergic as well as the excitatory transmitter systems (for references see [Citation20,Citation22]). Some promising results were seen in animal studies [Citation22,Citation23] but the experience in patients has been varying [Citation22]. A randomized controlled study comparing radiosurgery and open surgery is presently underway.
It is obviously difficult to compare the treatment methods of the studies described in the literature since patient groups, AVMs characteristics, follow-up protocols as well as radiotherapy techniques are heterogeneous. Another methodological issue is that the follow-up of the patients in our material varied due to a number of factors, e.g. the clinical condition of the patient, the patient's own preferences and available methods.
We report doses in physical Gy and prescribe them in that way as well. We prefer this method to the use of dose corrected for relative biological effectiveness (RBE), which is not a physical constant and may vary over the target volume. We are aware that many studies report RBE-corrected dose in Gy (RBE) but we understand that concept to be based on proliferating tumor cells and suspect that it may not apply to the capillaries and arteries in a growing arterio-venous malformation.
The overall complete occlusion of brain AVMs after proton beam radiosurgery was in our material 68% which is comparable with other patient series, independent of radiotherapy method. The total occlusion rate was for the target volumes 0–2 cm3 77%, for 3–10 cm3 80%, for 11–15 cm3 50% and for 16–51 cm3 20% and the occlusion seems to occur within five years. This agrees with other studies showing best results in smaller malformations, however our results also support that proton beam treatment is an alternative in the larger malformations.
Supplementary material available online
Supplementary Table I available online at: http://informahealthcare.com/doi/abs/10.3109/0284186X.2015.1043023)
ionc_a_1043023_sm5762.docx
Download MS Word (24.2 KB)Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Ruiz-Sandoval JL, Cantu C, Barinagarrementeria F. Intracerebral hemorrhage in young people: Analysis of risk factors, location, causes, and prognosis. Stroke 1999;30:537–41.
- Pollock BE, Gorman DA, Coffey RJ. Patient outcomes after arteriovenous malformation radiosurgical management: Results based on a 5- to 14-year follow-up study. Neurosurgery 2003;52:1291–6; discussion 6–7.
- Silander H, Pellettieri L, Enblad P, Montelius A, Grusell E, Vallhagen-Dahlgren C, et al. Fractionated, stereotactic proton beam treatment of cerebral arteriovenous malformations. Acta Neurol Scand 2004;109:85–90.
- Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg 1986;65:476–83.
- Grusell E, Montelius A, Russell KR, Blomquist E, Pellettieri L, Lilja A, et al. Patient positioning for fractionated precision radiation treatment of targets in the head using fiducial markers. Radiother Oncol 1994;33:68–72.
- Grusell E, Montelius A, Blomquist E. The proton therapy project at The Svedberg Laboratory in Uppsala. PSI-Bericht Wrenligen und Villigen Switzerland. 1991;111:43–5.
- Hattangadi-Gluth JA, Chapman PH, Kim D, Niemierko A, Bussiere MR, Stringham A, et al. Single-fraction proton beam stereotactic radiosurgery for cerebral arteriovenous malformations. Int J Radiat Oncol Biol Phys 2014;89:338–46.
- Paul L, Casasco A, Kusak ME, Martinez N, Rey G, Martinez R. Results for a series of 697 AVMs treated by gamma knife: Influence of angiographic features on the obliteration rate. Neurosurgery 2014;75:568–83; discussion 582–3.
- Hattangadi JA, Chapman PH, Bussiere MR, Niemierko A, Ogilvy CS, Rowell A, et al. Planned two-fraction proton beam stereotactic radiosurgery for high-risk inoperable cerebral arteriovenous malformations. Int J Radiat Oncol Biol Phys 2012;83:533–41.
- Kano H, Kondziolka D, Flickinger JC, Park KJ, Parry PV, Yang HC, et al. Stereotactic radiosurgery for arteriovenous malformations, Part 6: multistaged volumetric management of large arteriovenous malformations. J Neurosurg 2012;116:54–65.
- Murray G, Brau RH. A 10-year experience of radiosurgical treatment for cerebral arteriovenous malformations: A perspective from a series with large malformations. Clinical article. J Neurosurg 2011;115:337–46.
- Buell TJ, Ding D, Starke RM, Webster Crowley R, Liu KC. Embolization-induced angiogenesis in cerebral arteriovenous malformations. J Clin Neurosci 2014;21:1866–71.
- Taeshineetanakul P, Krings T, Geibprasert S, Menezes R, Agid R, Terbrugge KG, et al. Angioarchitecture determines obliteration rate after radiosurgery in brain arteriovenous malformations. Neurosurgery 2012;71:1071–8; discussion 9.
- Barker FG, 2nd, Butler WE, Lyons S, Cascio E, Ogilvy CS, Loeffler JS, et al. Dose-volume prediction of radiation- related complications after proton beam radiosurgery for cerebral arteriovenous malformations. J Neurosurg 2003; 99:254–63.
- Parkhutik V, Lago A, Aparici F, Vazquez JF, Tembl JI, Guillen L, et al. Late clinical and radiological complications of stereotactical radiosurgery of arteriovenous malformations of the brain. Neuroradiology 2013;55:405–12.
- van den Berg R, Buis DR, Lagerwaard FJ, Lycklama a Nijeholt GJ, Vandertop WP. Extensive white matter changes after stereotactic radiosurgery for brain arteriovenous malformations: A prognostic sign for obliteration? Neurosurgery 2008;63:1064–9; discussion 9–70.
- Walcott BP, Hattangadi-Gluth JA, Stapleton CJ, Ogilvy CS, Chapman PH, Loeffler JS. Proton beam stereotactic radiosurgery for pediatric cerebral arteriovenous malformations. Neurosurgery 2014;74:367–73; discussion 74.
- Starke RM, Yen CP, Chen CJ, Ding D, Mohila CA, Jensen ME, et al. An updated assessment of the risk of radiation-induced neoplasia after radiosurgery of arteriovenous malformations. World Neurosurg 2014;82:395–401.
- Loeffler JS, Niemierko A, Chapman PH. Second tumors after radiosurgery: Tip of the iceberg or a bump in the road? Neurosurgery 2003;52:1436–40; discussion 40–2.
- Yang SY, Kim DG, Chung HT, Paek SH. Radiosurgery for unruptured cerebral arteriovenous malformations: Long-term seizure outcome. Neurology 2012;78:1292–8.
- Baranoski JF, Grant RA, Hirsch LJ, Visintainer P, Gerrard JL, Gunel M, et al. Seizure control for intracranial arteriovenous malformations is directly related to treatment modality: A meta-analysis. J Neurointerv Surg 2014;6:684–90.
- Quigg M, Rolston J, Barbaro NM. Radiosurgery for epilepsy: Clinical experience and potential antiepileptic mechanisms. Epilepsia 2012;53:7–15.
- Ronne-Engström E, Kihlström L, Carlson H, Flink R, Blom S, Hillered L, et al. Gamma Knife surgery in epilepsy, an experimental rat model. Abstract 20th International Congres in Epilepsy, Oslo, Norway, July 1993.