Childs Nerv Syst DOI 10.1007/s00381-014-2469-5
ORIGINAL PAPER
Gamma Knife radiosurgery for arteriovenous malformations in pediatric patients Alp Özgün Börcek & Hakan Emmez & Koray M. Akkan & Özgür Öcal & Gökhan Kurt & Şükrü Aykol & Eray Karahacioğli & Kemali M. Baykaner
Received: 1 March 2014 / Accepted: 12 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Objective The authors present the results of Gamma Knife stereotactic radiosurgery performed in a series of children with arteriovenous malformations (AVMs). Methods Between June 2005 and January 2014, 75 patients 18 years old or younger received Gamma Knife radiosurgery for AVMs. Of these, 58 patients were eligible for further analysis. The median age of the population was 12 years; 41 % presented with hemorrhage, 34 % with neurological insult, and 24 % patients were diagnosed incidentally. The median AVM volume was 3.5 cm3. The median radiosurgerybased AVM score (RSBAVMS) was 0.86. The median followup period was 32 months. Results Single session Gamma Knife radiosurgery resulted in complete AVM obliteration in 40 (68.9 %) patients. There were 35 (60.3 %) excellent outcome (complete obliteration with no new deficits) in this series. During the follow-up period, nine (15.51 %) patients experienced new deficits and three (5.1 %) patients experienced intracranial hemorrhage. The annual rate of developing new deficits and hemorrhage was calculated as 5.45 and 1.8 %, respectively. Volume, gender, RSBAVMS, and nidus type factor were factors associated with excellent outcome.
A. Ö. Börcek (*) : H. Emmez : Ö. Öcal : G. Kurt : Ş. Aykol : K. M. Baykaner Department of Neurosurgery, Division of Pediatric Neurosurgery, Faculty of Medicine, Gazi University, Ankara 06500, Turkey e-mail: [email protected] K. M. Akkan Department of Radiology, Faculty of Medicine, Gazi University, Ankara, Turkey E. Karahacioğli Department of Radiation Oncology, Faculty of Medicine, Gazi University, Ankara, Turkey
Conclusions Radiosurgery was successful in majority of patients with minimal morbidity. Gamma Knife radiosurgery for AVMs can be a safe and successful method in pediatric patients. Keywords AVM . Gamma knife . Pediatric neurosurgery . Stereotactic radiosurgery Arteriovenous malformations (AVM) of the brain are the most common factor underlying intracranial hemorrhages in children [6, 9]. It is estimated that the incidence of AVM in pediatric population is 1 per 100,000, and of these, 12–18 % will be symptomatic during the childhood [7]. Pediatric AVMs seem to have different clinical, radiological, and outcome characteristics when compared to adult counterparts [10, 11]. The risk of rupture of an AVM is 2–4 % per year [16]. The cumulative hemorrhage risk of a child is higher than adults, and this makes definitive management mandatory whenever possible. Although microsurgery is accepted as the method of choice for cure, stereotactic radiosurgery was also proven for many cases, particularly for those associated with high surgical risks. This study aims to present the results of Gamma Knife radiosurgery in a pediatric population treated in a tertiary referral center from Ankara, Turkey.
Material and methods Patient characteristics Evaluation of prospectively recorded data between June 2005 and January 2014 for AVMs revealed a total of 75 pediatric patients that received Gamma Knife radiosurgery in our department. Patients requiring fractionated treatment, patient
Childs Nerv Syst
that have a history of previous radiotherapy or radiosurgery, and patients with less than 12 months of follow-up were excluded leaving a total number of 58 patients. The median age of the patient population was 12 (mean, 12.41±3.63; range, 4–18). There were 24 (41.4) boys and 34 (58.6 %) girls. Diagnosis was established after intracranial hemorrhage in 24 patients (41.4 %), after seizures in 17 patients (29.3 %), after headaches in 10 (17.2 %), and after progressive deficits in 3 (5.2 %). AVMs were diagnosed as an incidental finding in four (6.8 %) patients. Thirty-nine (61.9 %) patients had no history of related treatment whereas
8 (1.7 %) had hematoma surgery, 5 had (7.9 %) AVM surgery, and 11 (17.5 %) had embolization history. According to Spetzler Martin Grading (SM) system, there were 7 (12.1 %) grade 1, 19 (32.8 %) grade 2, 16 (27.6 %) grade 3, 8 (13.8 %) grade 4, and 8 (13.8 %) grade 5 AVMs in this cohort [19]. The median AVM volume was 3.5 cc (mean, 4.99±4.53; range, 0.42–23 cc) radiosurgery-based AVM score (RBAVMS) [14] and the later proposed modified radiosurgery-based AVM scores (mRBAVMS) [15] were calculated for patients. Used formulations for those scores were as follows:
RBAVMS ¼ ð0:1Þ ðvolume; mLÞ þ ð0:02Þ ðage; yearÞ þ ð0:3Þ location; frontal=temporal ¼ 0; parietal=occipital=corpus callosum=cerebellar ¼ 1; basal ganglia=thalamus=brainstem ¼ 2 mRBAVMS ¼ ð0:1Þ ðvolume; mLÞ þ ð0:02Þ ðage; yearÞ þ ð0:5Þ location; hemispheric=corpus callosum=cerebellar ¼ 0; basal ganglia=thalamus=brainstem ¼ 1
The median RBAVMS was 0.86 (mean, 1.02±0.57; range, 0.21–3.22) while the median mRBAVMS was 0.68 (mean, 0.85±0.57; range: 0.21–3.12). The main characteristics of the patients and the AVMs are demonstrated in Table 1.
Radiosurgical technique In our clinic, AVMs are managed according to the decision of the multidisciplinary team of experts including radiation oncologist, interventional radiologist, and pediatric neurosurgeons. Patients are offered radiosurgery based on their AVM size, location, and clinical status. Eligible lesions are offered surgical resection in the first place. However, patients’ requirements are also strongly considered in treatment planning. In case of a large AVM (generally >25 cc of volume), fractionated sessions are considered 3–4 months apart. The main determinant for this decision is the presence of critical nearby structures such as optic pathways and brain stem that will be exposed to significant doses in a single session. Additionally, in case of a large AVM, embolization is also considered prior to Gamma Knife surgery (GKS) in order to decrease the required dose; however, after some reports [4, 18] on negative influence of prior embolization on obliteration and complication rates, we prefer to avoid embolization in the first place whenever it is possible. All of the procedures were performed under general anesthesia for uncooperative children and under conscious sedation for older children. All children underwent regional scalp infiltration with local anesthetics. Stereotactic radiosurgery
procedure started with fixation of Leksell stereotactic frame (Elekta Instruments, Atlanta, GA) to the head of the patient. After frame application, each patient underwent both high-resolution, contrast-enhanced magnetic resonance imaging (MRI) and biplanar digital subtraction angiography (DSA). Radiosurgical target was defined using both imaging modalities, and a conformal plan was obtained with multiple isocenters. Dosing and planning were performed by a neurosurgeon, a radiation oncologist, and a medical physicist. The treatment plan was then transferred to the Leksell Gamma Knife Unit (Model 4C, Elekta Instruments, GA, USA), and patients received their treatment. The frame was removed after the treatment. All patients were observed for at least one night as inpatient and then discharged. Median prescription dose for the cohort was 22 Gy (mean, 20.77±1.94; range, 15–24). Median number of isocenters was 10 (mean, 10.22±5.55; range, 1–30). Other data regarding treatment plans are demonstrated in Table 2.
Conformity To determine the conformity of the treatment plans formulation described by Paddick [12] was used. Briefly, this formulation takes both the target volume covered by the prescription isodose volume and radiated volume outside the target into account in calculating conformity and enables to demonstrate more precise results regarding the treatment. It is formulated as follows:
Childs Nerv Syst Table 1 Main characteristics of the patient population and AVM-treated Number Age Mean Median Range Gender Male Female Presentation Hemorrhage Seizure Headache Neurological deficit Incidental Previous intervention None Hematoma evac. AVM Surgery Embolization AVM location Frontal Parietal Occipital Temporal Thalamic Basal gg. Pineal Pontine Cerebellar Corpus callosum Spetzler-Martin grade Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 AVM volume 10 cc Venous drain Deep Superficial Both Arterial feeders Single Multiple Nidus type Diffuse Compact Pial surface relation Yes No Radiosurgery-based AVM score 3
12.41 12 4–18 24 34
41.4 58.6
24 17 10 3 4
41.4 29.3 17.2 5.2 6.8
39 8 15 11
61.9 1.7 7.9 17.5
10 13 7 8 6 4 2 1 6 1
17.2 22.4 12.1 13.8 10.3 6.9 3.4 1.7 10.3 1.7
7 19 16 8 8
12.1 32.8 27.6 13.8 13.8
25 27 6
43.1 46.6 10.3
23 31 4
39.6 53.4 6.8
44 14
75.9 24.1
19 39
32.8 67.2
37 21
63.8 36.2
36 13
62.1 22.4
Number
Percent
5 4
8.6 6.9
TVPIV 2 100 TV PIV
ð1Þ
In this equation, TVPIV stands for the volume of the target covered by prescription isodose, TV for target volume, and PIV for prescription isodose volume. Additionally, we documented how much of the target volume was radiated by the prescribed isodose volume to demonstrate how much of the target volume was radiated. It is formulated as follows: TVPIV TV
ð2Þ
Follow-up Follow-up imaging with MRI was requested at 3 months, 6 months, and then annually. As soon as an MRI suggests obliteration of the nidus in a patient, a control DSA was requested to demonstrate complete obliteration. In case of a patent AVM with reduction in size compared to pre-GKS size after 36 months of follow-up, a second session was offered to the patients. During the follow-up period, all events related to GKS were recorded such as new hemorrhages and new or worsening deficits. The primary outcome was set as complete obliteration with no new deficits (excellent outcome). All other possibilities were graded according to both AVM obliteration status and neurological deficit status. The explanation of those criteria is demonstrated in Table 3. The median clinical follow-up duration of the entire cohort was 32 months in this series (mean, 35.37±17.97; range, 12– 100 months).
Table 2 Radiosurgery planning data of the patients
Prescription dose Min dose Max dose Conformity TVPIV/TV (TVPIV 2/TV x PIV)
Mean
Median
Range
20.77 16.19 42.09
22 16.95 44
15–24 5.4–21.70 31.40–48.50
98.59 79.20
100 80
92.68–100 62.37–95.97
Childs Nerv Syst Table 3 Outcome distribution among cases Description Excellent
Complete obliteration and no new deficit Good Complete obliteration and minor new deficit Fair Compete obliteration and major new deficit Unchanged Patent nidus and no new deficit Poor Patent nidus and new deficit
Number Percent 35
60.3
3
5.2
2
3.4
14 4
24.1 6.9
dard deviation (SD), percentages, and range. Continuous variables were compared with t test and categorical ones with chisquare tests. Kaplan-Meier method and log-rank analysis were used for actuarial rates of outcome and related factors. All analyses were performed using commercially available statistical software (IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.). Probability values less than 0.05 were considered significant.
Statistical analysis
Results
Various factors that may affect the outcome were analyzed. Descriptive statistics is demonstrated as mean, median, stan-
Single-session Gamma Knife radiosurgery resulted in complete AVM obliteration in 40 (68.9 %) patients after a median
Fig. 1 Kaplan-Meier plot demonstrating the obliteration rates of the patient population throughout the follow-up period
Childs Nerv Syst Table 4 Factors associated with excellent outcome
Factor
p value
Volume
0.018
(3 cc) Age
0.584
(12) Gender Spetzler-Martin grade
0.049 0.098
(I, II, III vs IV, V) Prescription dose
0.305
(22) AVM score
0.018
(1) mAVM score
0.214
(1) Previous invasive intervention
0.495
(surgery and/or embolization) Previous embolization Hematoma presentation Nidus type
0.543 0.049 0.012
(compact vs diffuse) Pial surface relation
0.154
(yes vs no)
follow-up time of 21 months (mean, 26.87±10.15; range, 11– 49). The actuarial obliteration rate was calculated as 52 % at 24 months, 73.49 % at 36 months, and 82.54 % at 48 months (Fig. 1). Obliteration was diagnosed with a DSA in 25 (62.5 %) patients and with MRI in 15 (37.5 %) patients. Of these, 35 patients (60.3 %) had complete obliteration with no new deficits (excellent outcome), 3 (5.2 %) experienced minor deficits during obliteration, and 2 (3.4 %) experienced major deficits. Of the 18 (31 %) patients that did not have complete obliteration, 14 (24.1 %) had no new deficits while 4 (6.9 %) experienced new deficits (Table 3). During the follow-up period, three (5.1 %) patients experienced intracranial hemorrhage 16, 18, and 20 months after index treatment. Two of them recovered with severe neurological deficits and the other with minor deficits. Two of these patients had a SM grade 5 AVM while the other had a SM grade 2 AVM. No additional intervention was performed for these patients and their deficits improved; however, none of those AVMs obliterated until their last follow-up, and two of them were offered an additional session. The annual rate of hemorrhage after the GKS treatment is calculated as 1.8 %. Univariate analysis demonstrated no associated factors with new deficits. Regardless of obliteration during the follow-up period, nine (15.51 %) patients experienced new deficits related to treatment after a median follow-up time of 8 months (mean, 10.2± 6.03; range, 4–20 months). The actuarial new deficit rate was calculated as 10 % at 12 months and 16 % at 24 months. The annual rate of developing new deficits was calculated as
5.45 %. Univariate analysis demonstrated no associated factors with new deficits. In general, there were nine (15 %) patients who experienced complications related to the management (presentation with new deficit and/or hematoma) with an annual complication rate of 5.45 %. We performed univariate analysis to demonstrate any association between excellent outcome and following factors: age, sex, AVM volume, prescription dose, any intervention history, previous embolization, hematoma presentation, Spetzler Martin grade, radiosurgery-based AVM score, modified radiosurgery-based AVM score, nidus type, and pial surface relation. The analyses revealed that sex (p=0.049), AVM volume (p=0.018), radiosurgery-based AVM score (p= 0.018), and nidus type (p=0.012) were significantly associated with excellent outcome (Table 4). Results on conformity The analyses of our data revealed that median conformity index calculated by Eq. 1 for our patient population was 80 % (mean, 79.2±6.85; range, 62.37–95.97 %). The conformity index of patients with an excellent outcome was not different from the others (p=0.137). According to Eq. 2, the median conformity of the plans was 100 % (mean, 98.59 %±2.09; range, 93–100 %) Further treatment At the end of the follow-up period, a new Gamma Knife treatment decision was made for five (8.6 %) patients after a median follow-up time of 45 months (mean, 43.8±5.06; range, 36–50 months). All other patients are still followed up for a median time of 30 months (mean, 34.58±18.56; range, 12–45 months) for obliteration and complications.
Discussion AVMs pose significant difficulties in management with mortality and morbidity rates as high as 18 and 53 %, respectively [8]. In their recent meta-analysis on natural history of AVMs, Gross and Du have demonstrated that the annual hemorrhage rate of an AVM was 3 %. They also found that prior hemorrhage, deep location, deep venous drainage, and associated aneurysms were significant risk factors for hemorrhage [3]. Although the developing brain of a child is more vulnerable to radiation-induced damage than adults [2], radiosurgery is a good option for managing pediatric patients with AVMs. The gold standard method for obliteration confirmation is DSA. In our series, more than half of the obliterations were confirmed with a DSA; however, there are also patients that
Childs Nerv Syst
Fig. 2 Pre-GKS (left superior) and post-GKS (right superior) DSA and the treatment plan of a patient treated in our clinic. The conformity ratio of this patient was calculated as 0.74 according to Eq. 1 revealing an overtreatment
Childs Nerv Syst
decided on MRI. Although this may cause biases in calculating obliteration rates and durations, literature has reported similar problems and related solutions. Kano et al. have studied AVM obliteration diagnosis methods and demonstrated that “only 3 of 100 patients with MRI-defined obliteration have patent residuals demonstrated in follow-up DSA” [5]. Due to the absence of DSA-defined obliteration in all of the cases, our results should be considered as the “best case scenario” for our patients. During the follow-up period, patients are scanned usually at 1 year intervals; however, due to patient-related problems, some obliterations were determined in longer intervals, and the exact time of obliteration is not known. This may negatively affect time-to-closure analysis. Some suggestions have been made to solve—at least alleviate—this problem [5]; however, we did not perform such calculation and reported the results as obtained at the last event-documented follow-up. Buis et al. have documented results on pediatric patients harboring AVMs treated with Gamma Knife radiosurgery of different centers in their review [1]. In this report, obliteration rates after Gamma Knife application range between 33 and 95 %. Our raw obliteration rate is 68.9 %, and this is in accordance with the current literature. Besides, our series have achieved this ratio after a relatively short time period of 21 months. This is one of the important points in our study. It is well known that the typical latency period after radiosurgery is 3 years, and any evaluation before this threshold may interfere with the results. However, the aim of this study was to demonstrate the closure rate in a group of pediatric patients, and although we have put our threshold point at 12 months, we have demonstrated a relatively higher incidence of closure (probable underlying factor of these results is discussed below). On the other hand, we are aware of the potential bias in calculating complication rates in our study. It is obvious that longer follow-up periods may prove higher complication rates. Univariate analyses demonstrated that volume, gender, RSBAVMS, and nidus type were the factors correlating with excellent outcome. Due to the construct of the proposed formulation, the age of the patient is an important factor in calculating RSBAVMS. It can be speculated that implementing this formulation in pediatric population may provide a bias in favor of lower scores; however, as in many other studies [1, 2, 6], this work also confirms that RSBAVMS is a valid instrument in predicting outcome even in pediatric patients. The calculated conformity index according to the formulation suggested by Paddick is 79.20 % in this series, and it is surprisingly low. However, we could not demonstrate any associations between, for example, excellent outcome and the conformity calculated with Eq. 1. On the other hand, when we look at how much of the target volume is radiated (Eq. 2), we see that 98.59 % of the target was irradiated. These two
calculations suggest that there is an overtreatment in our series. In other words, while covering the target volume with the prescribed isodose, we also irradiated normal tissue outside the target needlessly (Fig. 2). Conformity is an important issue not only in stereotactic radiosurgery but also in other radiation techniques. Although increasing technology facilitates dose planning, still, it could be a problem to achieve higher conformity for targets with complex morphology. AVMs are perhaps among the most challenging lesions regarding their complex shapes [13]. Additionally, previous embolizations make finding the exact border of the residual lesion even harder. The major difference between two conformity indexes in our series can be attributed to the morphological complexity of the pathology dealt with. This situation may have both favorable and hazardous effects for patients. These patients may have increased risk for radiation-induced pathologies such as secondary malignancies or asymptomatic MR changes, and longer follow-up is mandatory for these patients. On the other hand, angiographically undetectable immature vessels in the periphery of the nidus that are missed in the first planning may benefit from such a generous treatment and decrease recurrence [2, 17]. The relatively shorter time to reach obliteration demonstrated in this series may also be the result of this “overtreatment.” For the cases presented in this series, this overtreatment does not seem to result in severe complications, but strict follow-up is necessary to make certain judgments. We suggest evaluation of relationship between conformity and outcome to find the answers to these questions. Technological improvements in both stereotactic radiosurgery methods and angiographic demonstration of the pathology may help to solve conformity problem encountered in targets with complex morphology.
Conclusion Gamma Knife radiosurgery is an effective and safe treatment modality for pediatric patients with AVMs. While deciding for the method of treatment, RBAVMS can be used as a marker for future success of radiosurgery.
References 1. Buis DR, Dirven CM, Lagerwaard FJ, Mandl ES, Lycklama ANGJ, Eshghi DS, van den Berg R, Baayen JC, Meijer OW, Slotman BJ, Vandertop WP (2008) Radiosurgery of brain arteriovenous malformations in children. J Neurol 255:551–560 2. Cohen-Gadol AA, Pollock BE (2006) Radiosurgery for arteriovenous malformations in children. J Neurosurg 104:388–391 3. Gross BA, Du R (2013) Natural history of cerebral arteriovenous malformations: a meta-analysis. J Neurosurg 118:437–443
Childs Nerv Syst 4. Kano H, Kondziolka D, Flickinger JC, Park KJ, Iyer A, Yang HC, Liu X, Monaco EA 3rd, Niranjan A, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations after embolization: a case–control study. J Neurosurg 117:265–275 5. Kano H, Kondziolka D, Flickinger JC, Yang HC, Flannery TJ, Awan NR, Niranjan A, Novotny J Jr, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations, part 3: outcome predictors and risks after repeat radiosurgery. J Neurosurg 116:21–32 6. Kano H, Kondziolka D, Flickinger JC, Yang HC, Flannery TJ, Awan NR, Niranjan A, Novotny J, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations, part 2: management of pediatric patients. J Neurosurg Pediatr 9:1–10 7. Menovsky T, van Overbeeke JJ (1997) Cerebral arteriovenous malformations in childhood: state of the art with special reference to treatment. Eur J Pediatr 156:741–746 8. Nataraj A, Mohamed MB, Gholkar A, Vivar R, Watkins L, Aspoas R, Gregson B, Mitchell P, Mendelow AD (2013) Multimodality treatment of cerebral arteriovenous malformations. World neurosurgery 9. Nicolato A, Foroni R, Seghedoni A, Martines V, Lupidi F, Zampieri P, Sandri MF, Ricci U, Mazza C, Beltramello A, Gerosa M, Bricolo A (2005) Leksell gamma knife radiosurgery for cerebral arteriovenous malformations in pediatric patients. Child’s Nerv Syst : ChNS : Off J Int Soc Pediatr Neurosurg 21:301–307, discussion 308 10. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, Maluta S, Beltramello A, Gerosa M (2006) Gamma knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part I: differences in epidemiologic, morphologic, and clinical characteristics, permanent complications, and bleeding in the latency period. Int J Radiat Oncol Biol Phys 64:904–913
11. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, Pasqualin A, Beltramello A, Gerosa M (2006) Gamma Knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part II: differences in obliteration rates, treatment-obliteration intervals, and prognostic factors. Int J Radiat Oncol Biol Phys 64:914–921 12. Paddick I (2000) A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 93(Suppl 3):219–222 13. Paddick I, Motti E (2013) Targeting and conformality in arteriovenous malformation radiosurgery. Prog Neurol Surg 27:35–48 14. Pollock BE, Flickinger JC (2002) A proposed radiosurgery-based grading system for arteriovenous malformations. J Neurosurg 96:79– 85 15. Pollock BE, Flickinger JC (2008) Modification of the radiosurgerybased arteriovenous malformation grading system. Neurosurgery 63: 239–243, discussion 243 16. Reyns N, Blond S, Gauvrit JY, Touzet G, Coche B, Pruvo JP, Dhellemmes P (2007) Role of radiosurgery in the management of cerebral arteriovenous malformations in the pediatric age group: data from a 100-patient series. Neurosurgery 60:268– 276, discussion 276 17. Rodriguez-Arias C, Martinez R, Rey G, Bravo G (2000) Recurrence in a different location of a cerebral arteriovenous malformation in a child after radiosurgery. Child’s Nerv Syst : ChNS : Off J Int Soc Pediatr Neurosurg 16:363–365 18. Schwyzer L, Yen CP, Evans A, Zavoian S, Steiner L (2012) Longterm results of gamma knife surgery for partially embolized arteriovenous malformations. Neurosurgery 71:1139–1147, discussion 1147–1138 19. Spetzler RF, Martin NA (1986) A proposed grading system for arteriovenous malformations. J Neurosurg 65:476–483
ORIGINAL PAPER
Gamma Knife radiosurgery for arteriovenous malformations in pediatric patients Alp Özgün Börcek & Hakan Emmez & Koray M. Akkan & Özgür Öcal & Gökhan Kurt & Şükrü Aykol & Eray Karahacioğli & Kemali M. Baykaner
Received: 1 March 2014 / Accepted: 12 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Objective The authors present the results of Gamma Knife stereotactic radiosurgery performed in a series of children with arteriovenous malformations (AVMs). Methods Between June 2005 and January 2014, 75 patients 18 years old or younger received Gamma Knife radiosurgery for AVMs. Of these, 58 patients were eligible for further analysis. The median age of the population was 12 years; 41 % presented with hemorrhage, 34 % with neurological insult, and 24 % patients were diagnosed incidentally. The median AVM volume was 3.5 cm3. The median radiosurgerybased AVM score (RSBAVMS) was 0.86. The median followup period was 32 months. Results Single session Gamma Knife radiosurgery resulted in complete AVM obliteration in 40 (68.9 %) patients. There were 35 (60.3 %) excellent outcome (complete obliteration with no new deficits) in this series. During the follow-up period, nine (15.51 %) patients experienced new deficits and three (5.1 %) patients experienced intracranial hemorrhage. The annual rate of developing new deficits and hemorrhage was calculated as 5.45 and 1.8 %, respectively. Volume, gender, RSBAVMS, and nidus type factor were factors associated with excellent outcome.
A. Ö. Börcek (*) : H. Emmez : Ö. Öcal : G. Kurt : Ş. Aykol : K. M. Baykaner Department of Neurosurgery, Division of Pediatric Neurosurgery, Faculty of Medicine, Gazi University, Ankara 06500, Turkey e-mail: [email protected] K. M. Akkan Department of Radiology, Faculty of Medicine, Gazi University, Ankara, Turkey E. Karahacioğli Department of Radiation Oncology, Faculty of Medicine, Gazi University, Ankara, Turkey
Conclusions Radiosurgery was successful in majority of patients with minimal morbidity. Gamma Knife radiosurgery for AVMs can be a safe and successful method in pediatric patients. Keywords AVM . Gamma knife . Pediatric neurosurgery . Stereotactic radiosurgery Arteriovenous malformations (AVM) of the brain are the most common factor underlying intracranial hemorrhages in children [6, 9]. It is estimated that the incidence of AVM in pediatric population is 1 per 100,000, and of these, 12–18 % will be symptomatic during the childhood [7]. Pediatric AVMs seem to have different clinical, radiological, and outcome characteristics when compared to adult counterparts [10, 11]. The risk of rupture of an AVM is 2–4 % per year [16]. The cumulative hemorrhage risk of a child is higher than adults, and this makes definitive management mandatory whenever possible. Although microsurgery is accepted as the method of choice for cure, stereotactic radiosurgery was also proven for many cases, particularly for those associated with high surgical risks. This study aims to present the results of Gamma Knife radiosurgery in a pediatric population treated in a tertiary referral center from Ankara, Turkey.
Material and methods Patient characteristics Evaluation of prospectively recorded data between June 2005 and January 2014 for AVMs revealed a total of 75 pediatric patients that received Gamma Knife radiosurgery in our department. Patients requiring fractionated treatment, patient
Childs Nerv Syst
that have a history of previous radiotherapy or radiosurgery, and patients with less than 12 months of follow-up were excluded leaving a total number of 58 patients. The median age of the patient population was 12 (mean, 12.41±3.63; range, 4–18). There were 24 (41.4) boys and 34 (58.6 %) girls. Diagnosis was established after intracranial hemorrhage in 24 patients (41.4 %), after seizures in 17 patients (29.3 %), after headaches in 10 (17.2 %), and after progressive deficits in 3 (5.2 %). AVMs were diagnosed as an incidental finding in four (6.8 %) patients. Thirty-nine (61.9 %) patients had no history of related treatment whereas
8 (1.7 %) had hematoma surgery, 5 had (7.9 %) AVM surgery, and 11 (17.5 %) had embolization history. According to Spetzler Martin Grading (SM) system, there were 7 (12.1 %) grade 1, 19 (32.8 %) grade 2, 16 (27.6 %) grade 3, 8 (13.8 %) grade 4, and 8 (13.8 %) grade 5 AVMs in this cohort [19]. The median AVM volume was 3.5 cc (mean, 4.99±4.53; range, 0.42–23 cc) radiosurgery-based AVM score (RBAVMS) [14] and the later proposed modified radiosurgery-based AVM scores (mRBAVMS) [15] were calculated for patients. Used formulations for those scores were as follows:
RBAVMS ¼ ð0:1Þ ðvolume; mLÞ þ ð0:02Þ ðage; yearÞ þ ð0:3Þ location; frontal=temporal ¼ 0; parietal=occipital=corpus callosum=cerebellar ¼ 1; basal ganglia=thalamus=brainstem ¼ 2 mRBAVMS ¼ ð0:1Þ ðvolume; mLÞ þ ð0:02Þ ðage; yearÞ þ ð0:5Þ location; hemispheric=corpus callosum=cerebellar ¼ 0; basal ganglia=thalamus=brainstem ¼ 1
The median RBAVMS was 0.86 (mean, 1.02±0.57; range, 0.21–3.22) while the median mRBAVMS was 0.68 (mean, 0.85±0.57; range: 0.21–3.12). The main characteristics of the patients and the AVMs are demonstrated in Table 1.
Radiosurgical technique In our clinic, AVMs are managed according to the decision of the multidisciplinary team of experts including radiation oncologist, interventional radiologist, and pediatric neurosurgeons. Patients are offered radiosurgery based on their AVM size, location, and clinical status. Eligible lesions are offered surgical resection in the first place. However, patients’ requirements are also strongly considered in treatment planning. In case of a large AVM (generally >25 cc of volume), fractionated sessions are considered 3–4 months apart. The main determinant for this decision is the presence of critical nearby structures such as optic pathways and brain stem that will be exposed to significant doses in a single session. Additionally, in case of a large AVM, embolization is also considered prior to Gamma Knife surgery (GKS) in order to decrease the required dose; however, after some reports [4, 18] on negative influence of prior embolization on obliteration and complication rates, we prefer to avoid embolization in the first place whenever it is possible. All of the procedures were performed under general anesthesia for uncooperative children and under conscious sedation for older children. All children underwent regional scalp infiltration with local anesthetics. Stereotactic radiosurgery
procedure started with fixation of Leksell stereotactic frame (Elekta Instruments, Atlanta, GA) to the head of the patient. After frame application, each patient underwent both high-resolution, contrast-enhanced magnetic resonance imaging (MRI) and biplanar digital subtraction angiography (DSA). Radiosurgical target was defined using both imaging modalities, and a conformal plan was obtained with multiple isocenters. Dosing and planning were performed by a neurosurgeon, a radiation oncologist, and a medical physicist. The treatment plan was then transferred to the Leksell Gamma Knife Unit (Model 4C, Elekta Instruments, GA, USA), and patients received their treatment. The frame was removed after the treatment. All patients were observed for at least one night as inpatient and then discharged. Median prescription dose for the cohort was 22 Gy (mean, 20.77±1.94; range, 15–24). Median number of isocenters was 10 (mean, 10.22±5.55; range, 1–30). Other data regarding treatment plans are demonstrated in Table 2.
Conformity To determine the conformity of the treatment plans formulation described by Paddick [12] was used. Briefly, this formulation takes both the target volume covered by the prescription isodose volume and radiated volume outside the target into account in calculating conformity and enables to demonstrate more precise results regarding the treatment. It is formulated as follows:
Childs Nerv Syst Table 1 Main characteristics of the patient population and AVM-treated Number Age Mean Median Range Gender Male Female Presentation Hemorrhage Seizure Headache Neurological deficit Incidental Previous intervention None Hematoma evac. AVM Surgery Embolization AVM location Frontal Parietal Occipital Temporal Thalamic Basal gg. Pineal Pontine Cerebellar Corpus callosum Spetzler-Martin grade Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 AVM volume 10 cc Venous drain Deep Superficial Both Arterial feeders Single Multiple Nidus type Diffuse Compact Pial surface relation Yes No Radiosurgery-based AVM score 3
12.41 12 4–18 24 34
41.4 58.6
24 17 10 3 4
41.4 29.3 17.2 5.2 6.8
39 8 15 11
61.9 1.7 7.9 17.5
10 13 7 8 6 4 2 1 6 1
17.2 22.4 12.1 13.8 10.3 6.9 3.4 1.7 10.3 1.7
7 19 16 8 8
12.1 32.8 27.6 13.8 13.8
25 27 6
43.1 46.6 10.3
23 31 4
39.6 53.4 6.8
44 14
75.9 24.1
19 39
32.8 67.2
37 21
63.8 36.2
36 13
62.1 22.4
Number
Percent
5 4
8.6 6.9
TVPIV 2 100 TV PIV
ð1Þ
In this equation, TVPIV stands for the volume of the target covered by prescription isodose, TV for target volume, and PIV for prescription isodose volume. Additionally, we documented how much of the target volume was radiated by the prescribed isodose volume to demonstrate how much of the target volume was radiated. It is formulated as follows: TVPIV TV
ð2Þ
Follow-up Follow-up imaging with MRI was requested at 3 months, 6 months, and then annually. As soon as an MRI suggests obliteration of the nidus in a patient, a control DSA was requested to demonstrate complete obliteration. In case of a patent AVM with reduction in size compared to pre-GKS size after 36 months of follow-up, a second session was offered to the patients. During the follow-up period, all events related to GKS were recorded such as new hemorrhages and new or worsening deficits. The primary outcome was set as complete obliteration with no new deficits (excellent outcome). All other possibilities were graded according to both AVM obliteration status and neurological deficit status. The explanation of those criteria is demonstrated in Table 3. The median clinical follow-up duration of the entire cohort was 32 months in this series (mean, 35.37±17.97; range, 12– 100 months).
Table 2 Radiosurgery planning data of the patients
Prescription dose Min dose Max dose Conformity TVPIV/TV (TVPIV 2/TV x PIV)
Mean
Median
Range
20.77 16.19 42.09
22 16.95 44
15–24 5.4–21.70 31.40–48.50
98.59 79.20
100 80
92.68–100 62.37–95.97
Childs Nerv Syst Table 3 Outcome distribution among cases Description Excellent
Complete obliteration and no new deficit Good Complete obliteration and minor new deficit Fair Compete obliteration and major new deficit Unchanged Patent nidus and no new deficit Poor Patent nidus and new deficit
Number Percent 35
60.3
3
5.2
2
3.4
14 4
24.1 6.9
dard deviation (SD), percentages, and range. Continuous variables were compared with t test and categorical ones with chisquare tests. Kaplan-Meier method and log-rank analysis were used for actuarial rates of outcome and related factors. All analyses were performed using commercially available statistical software (IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.). Probability values less than 0.05 were considered significant.
Statistical analysis
Results
Various factors that may affect the outcome were analyzed. Descriptive statistics is demonstrated as mean, median, stan-
Single-session Gamma Knife radiosurgery resulted in complete AVM obliteration in 40 (68.9 %) patients after a median
Fig. 1 Kaplan-Meier plot demonstrating the obliteration rates of the patient population throughout the follow-up period
Childs Nerv Syst Table 4 Factors associated with excellent outcome
Factor
p value
Volume
0.018
(3 cc) Age
0.584
(12) Gender Spetzler-Martin grade
0.049 0.098
(I, II, III vs IV, V) Prescription dose
0.305
(22) AVM score
0.018
(1) mAVM score
0.214
(1) Previous invasive intervention
0.495
(surgery and/or embolization) Previous embolization Hematoma presentation Nidus type
0.543 0.049 0.012
(compact vs diffuse) Pial surface relation
0.154
(yes vs no)
follow-up time of 21 months (mean, 26.87±10.15; range, 11– 49). The actuarial obliteration rate was calculated as 52 % at 24 months, 73.49 % at 36 months, and 82.54 % at 48 months (Fig. 1). Obliteration was diagnosed with a DSA in 25 (62.5 %) patients and with MRI in 15 (37.5 %) patients. Of these, 35 patients (60.3 %) had complete obliteration with no new deficits (excellent outcome), 3 (5.2 %) experienced minor deficits during obliteration, and 2 (3.4 %) experienced major deficits. Of the 18 (31 %) patients that did not have complete obliteration, 14 (24.1 %) had no new deficits while 4 (6.9 %) experienced new deficits (Table 3). During the follow-up period, three (5.1 %) patients experienced intracranial hemorrhage 16, 18, and 20 months after index treatment. Two of them recovered with severe neurological deficits and the other with minor deficits. Two of these patients had a SM grade 5 AVM while the other had a SM grade 2 AVM. No additional intervention was performed for these patients and their deficits improved; however, none of those AVMs obliterated until their last follow-up, and two of them were offered an additional session. The annual rate of hemorrhage after the GKS treatment is calculated as 1.8 %. Univariate analysis demonstrated no associated factors with new deficits. Regardless of obliteration during the follow-up period, nine (15.51 %) patients experienced new deficits related to treatment after a median follow-up time of 8 months (mean, 10.2± 6.03; range, 4–20 months). The actuarial new deficit rate was calculated as 10 % at 12 months and 16 % at 24 months. The annual rate of developing new deficits was calculated as
5.45 %. Univariate analysis demonstrated no associated factors with new deficits. In general, there were nine (15 %) patients who experienced complications related to the management (presentation with new deficit and/or hematoma) with an annual complication rate of 5.45 %. We performed univariate analysis to demonstrate any association between excellent outcome and following factors: age, sex, AVM volume, prescription dose, any intervention history, previous embolization, hematoma presentation, Spetzler Martin grade, radiosurgery-based AVM score, modified radiosurgery-based AVM score, nidus type, and pial surface relation. The analyses revealed that sex (p=0.049), AVM volume (p=0.018), radiosurgery-based AVM score (p= 0.018), and nidus type (p=0.012) were significantly associated with excellent outcome (Table 4). Results on conformity The analyses of our data revealed that median conformity index calculated by Eq. 1 for our patient population was 80 % (mean, 79.2±6.85; range, 62.37–95.97 %). The conformity index of patients with an excellent outcome was not different from the others (p=0.137). According to Eq. 2, the median conformity of the plans was 100 % (mean, 98.59 %±2.09; range, 93–100 %) Further treatment At the end of the follow-up period, a new Gamma Knife treatment decision was made for five (8.6 %) patients after a median follow-up time of 45 months (mean, 43.8±5.06; range, 36–50 months). All other patients are still followed up for a median time of 30 months (mean, 34.58±18.56; range, 12–45 months) for obliteration and complications.
Discussion AVMs pose significant difficulties in management with mortality and morbidity rates as high as 18 and 53 %, respectively [8]. In their recent meta-analysis on natural history of AVMs, Gross and Du have demonstrated that the annual hemorrhage rate of an AVM was 3 %. They also found that prior hemorrhage, deep location, deep venous drainage, and associated aneurysms were significant risk factors for hemorrhage [3]. Although the developing brain of a child is more vulnerable to radiation-induced damage than adults [2], radiosurgery is a good option for managing pediatric patients with AVMs. The gold standard method for obliteration confirmation is DSA. In our series, more than half of the obliterations were confirmed with a DSA; however, there are also patients that
Childs Nerv Syst
Fig. 2 Pre-GKS (left superior) and post-GKS (right superior) DSA and the treatment plan of a patient treated in our clinic. The conformity ratio of this patient was calculated as 0.74 according to Eq. 1 revealing an overtreatment
Childs Nerv Syst
decided on MRI. Although this may cause biases in calculating obliteration rates and durations, literature has reported similar problems and related solutions. Kano et al. have studied AVM obliteration diagnosis methods and demonstrated that “only 3 of 100 patients with MRI-defined obliteration have patent residuals demonstrated in follow-up DSA” [5]. Due to the absence of DSA-defined obliteration in all of the cases, our results should be considered as the “best case scenario” for our patients. During the follow-up period, patients are scanned usually at 1 year intervals; however, due to patient-related problems, some obliterations were determined in longer intervals, and the exact time of obliteration is not known. This may negatively affect time-to-closure analysis. Some suggestions have been made to solve—at least alleviate—this problem [5]; however, we did not perform such calculation and reported the results as obtained at the last event-documented follow-up. Buis et al. have documented results on pediatric patients harboring AVMs treated with Gamma Knife radiosurgery of different centers in their review [1]. In this report, obliteration rates after Gamma Knife application range between 33 and 95 %. Our raw obliteration rate is 68.9 %, and this is in accordance with the current literature. Besides, our series have achieved this ratio after a relatively short time period of 21 months. This is one of the important points in our study. It is well known that the typical latency period after radiosurgery is 3 years, and any evaluation before this threshold may interfere with the results. However, the aim of this study was to demonstrate the closure rate in a group of pediatric patients, and although we have put our threshold point at 12 months, we have demonstrated a relatively higher incidence of closure (probable underlying factor of these results is discussed below). On the other hand, we are aware of the potential bias in calculating complication rates in our study. It is obvious that longer follow-up periods may prove higher complication rates. Univariate analyses demonstrated that volume, gender, RSBAVMS, and nidus type were the factors correlating with excellent outcome. Due to the construct of the proposed formulation, the age of the patient is an important factor in calculating RSBAVMS. It can be speculated that implementing this formulation in pediatric population may provide a bias in favor of lower scores; however, as in many other studies [1, 2, 6], this work also confirms that RSBAVMS is a valid instrument in predicting outcome even in pediatric patients. The calculated conformity index according to the formulation suggested by Paddick is 79.20 % in this series, and it is surprisingly low. However, we could not demonstrate any associations between, for example, excellent outcome and the conformity calculated with Eq. 1. On the other hand, when we look at how much of the target volume is radiated (Eq. 2), we see that 98.59 % of the target was irradiated. These two
calculations suggest that there is an overtreatment in our series. In other words, while covering the target volume with the prescribed isodose, we also irradiated normal tissue outside the target needlessly (Fig. 2). Conformity is an important issue not only in stereotactic radiosurgery but also in other radiation techniques. Although increasing technology facilitates dose planning, still, it could be a problem to achieve higher conformity for targets with complex morphology. AVMs are perhaps among the most challenging lesions regarding their complex shapes [13]. Additionally, previous embolizations make finding the exact border of the residual lesion even harder. The major difference between two conformity indexes in our series can be attributed to the morphological complexity of the pathology dealt with. This situation may have both favorable and hazardous effects for patients. These patients may have increased risk for radiation-induced pathologies such as secondary malignancies or asymptomatic MR changes, and longer follow-up is mandatory for these patients. On the other hand, angiographically undetectable immature vessels in the periphery of the nidus that are missed in the first planning may benefit from such a generous treatment and decrease recurrence [2, 17]. The relatively shorter time to reach obliteration demonstrated in this series may also be the result of this “overtreatment.” For the cases presented in this series, this overtreatment does not seem to result in severe complications, but strict follow-up is necessary to make certain judgments. We suggest evaluation of relationship between conformity and outcome to find the answers to these questions. Technological improvements in both stereotactic radiosurgery methods and angiographic demonstration of the pathology may help to solve conformity problem encountered in targets with complex morphology.
Conclusion Gamma Knife radiosurgery is an effective and safe treatment modality for pediatric patients with AVMs. While deciding for the method of treatment, RBAVMS can be used as a marker for future success of radiosurgery.
References 1. Buis DR, Dirven CM, Lagerwaard FJ, Mandl ES, Lycklama ANGJ, Eshghi DS, van den Berg R, Baayen JC, Meijer OW, Slotman BJ, Vandertop WP (2008) Radiosurgery of brain arteriovenous malformations in children. J Neurol 255:551–560 2. Cohen-Gadol AA, Pollock BE (2006) Radiosurgery for arteriovenous malformations in children. J Neurosurg 104:388–391 3. Gross BA, Du R (2013) Natural history of cerebral arteriovenous malformations: a meta-analysis. J Neurosurg 118:437–443
Childs Nerv Syst 4. Kano H, Kondziolka D, Flickinger JC, Park KJ, Iyer A, Yang HC, Liu X, Monaco EA 3rd, Niranjan A, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations after embolization: a case–control study. J Neurosurg 117:265–275 5. Kano H, Kondziolka D, Flickinger JC, Yang HC, Flannery TJ, Awan NR, Niranjan A, Novotny J Jr, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations, part 3: outcome predictors and risks after repeat radiosurgery. J Neurosurg 116:21–32 6. Kano H, Kondziolka D, Flickinger JC, Yang HC, Flannery TJ, Awan NR, Niranjan A, Novotny J, Lunsford LD (2012) Stereotactic radiosurgery for arteriovenous malformations, part 2: management of pediatric patients. J Neurosurg Pediatr 9:1–10 7. Menovsky T, van Overbeeke JJ (1997) Cerebral arteriovenous malformations in childhood: state of the art with special reference to treatment. Eur J Pediatr 156:741–746 8. Nataraj A, Mohamed MB, Gholkar A, Vivar R, Watkins L, Aspoas R, Gregson B, Mitchell P, Mendelow AD (2013) Multimodality treatment of cerebral arteriovenous malformations. World neurosurgery 9. Nicolato A, Foroni R, Seghedoni A, Martines V, Lupidi F, Zampieri P, Sandri MF, Ricci U, Mazza C, Beltramello A, Gerosa M, Bricolo A (2005) Leksell gamma knife radiosurgery for cerebral arteriovenous malformations in pediatric patients. Child’s Nerv Syst : ChNS : Off J Int Soc Pediatr Neurosurg 21:301–307, discussion 308 10. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, Maluta S, Beltramello A, Gerosa M (2006) Gamma knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part I: differences in epidemiologic, morphologic, and clinical characteristics, permanent complications, and bleeding in the latency period. Int J Radiat Oncol Biol Phys 64:904–913
11. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, Pasqualin A, Beltramello A, Gerosa M (2006) Gamma Knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part II: differences in obliteration rates, treatment-obliteration intervals, and prognostic factors. Int J Radiat Oncol Biol Phys 64:914–921 12. Paddick I (2000) A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 93(Suppl 3):219–222 13. Paddick I, Motti E (2013) Targeting and conformality in arteriovenous malformation radiosurgery. Prog Neurol Surg 27:35–48 14. Pollock BE, Flickinger JC (2002) A proposed radiosurgery-based grading system for arteriovenous malformations. J Neurosurg 96:79– 85 15. Pollock BE, Flickinger JC (2008) Modification of the radiosurgerybased arteriovenous malformation grading system. Neurosurgery 63: 239–243, discussion 243 16. Reyns N, Blond S, Gauvrit JY, Touzet G, Coche B, Pruvo JP, Dhellemmes P (2007) Role of radiosurgery in the management of cerebral arteriovenous malformations in the pediatric age group: data from a 100-patient series. Neurosurgery 60:268– 276, discussion 276 17. Rodriguez-Arias C, Martinez R, Rey G, Bravo G (2000) Recurrence in a different location of a cerebral arteriovenous malformation in a child after radiosurgery. Child’s Nerv Syst : ChNS : Off J Int Soc Pediatr Neurosurg 16:363–365 18. Schwyzer L, Yen CP, Evans A, Zavoian S, Steiner L (2012) Longterm results of gamma knife surgery for partially embolized arteriovenous malformations. Neurosurgery 71:1139–1147, discussion 1147–1138 19. Spetzler RF, Martin NA (1986) A proposed grading system for arteriovenous malformations. J Neurosurg 65:476–483
