J Neurosurg Pediatr. 2015 Dec;16(6):736-47. doi: 10.3171/2015.4.PEDS158.

Leksell Gamma Knife for pediatric and adolescent cerebral arteriovenous malformations: results of 100 cases followed up for at least 36 months

Antonio Nicolato, Michele Longhi, Nicola Tommasi, Giuseppe Kenneth Ricciardi, Roberto Spinelli, Roberto Israel Foroni, Emanuele Zivelonghi, Simone Zironi, Stefano Dall’oglio, Alberto Beltramello, Mario Meglio.


Istituto di Neurochirurgia, DAI di Neuroscienze, AOUI Verona, Italy.

Centro Interdipartimentale di Documentazione Economica (C.I.D.E.), University of Verona, Italy. Neuroradiologia d.O., DAI Patologia e Diagnostica, AOUI Verona, Italy.

Servizio Fisica Ospedaliera, DAI Patologia e Diagnostica, AOUI Verona, Italy.

Radioterapia Ospedaliera, DAI di Chirurgia ed Oncologia, AOUI Verona, Italy.



Objectives: The goal of this study was to evaluate advantages, risks, and failures of Gamma Knife radiosurgery (GKRS) in a large series of pediatric and adolescent patients with cerebral arteriovenous malformations (cAVMs) who were followed up for at least 36 months.

Methods: Since February 1993, 100 pediatric and adolescent patients (≤ 18 years of age) with cAVMs have undergone GKRS at the authors’ institution and were followed up for at least 36 months. Forty-six patients were boys and 54 were girls; the mean age was 12.8 years (range 3–18 years). Hemorrhage, either alone or combined with seizure, was the clinical onset in 70% of cases. The mean pre-GK cAVM volume was 2.8 ml; 92% of cAVMs were Spetzler-Martin (S-M) Grades I–III. Most lesions (94%) were in eloquent or deep-seated brain regions, according to S-M classification. The parameters for mean and range in treatment planning were prescription isodose 53.8% (40%–90%); prescription dose (PD) 20.2 Gy (9.0–26.4 Gy); maximal dose (MD) 37.8 Gy (18–50 Gy); and number of shots 4.7 (1–17). On the day of GKRS, stereotactic CT or stereotactic MRI and digital subtraction angiography were used.

Results: Obliteration rate (OR) was angiographically documented in 75 of 84 cases (89.3%) after single-session GKRS, with actuarial ORs at 3 and 5 years of 68.0% and 88.1%, respectively. A repeat treatment was performed in 7 patients (6 with obliteration), and 16 patients with cAVMs underwent staged treatment (9 of them were angiographically cured). Thus, the overall OR was 90%, with actuarial ORs at 3, 5, and 8 years of 59.0%, 76.0%, and 85.0%, respectively. Permanent symptomatic GK-related complications were observed in 11% of cases, with surgical removal of enlarged mass seen on post-RS imaging needed in 5 cases. Hemorrhage during the latency period occurred in 9% of patients, but surgical evacuation of the hematoma was required in only 1 patient. One patient died due to rebleeding of a brainstem cAVM. Radiosurgery outcomes varied according to cAVM sizes and doses: volumes ≤ 10 ml and PDs > 16 Gy were significantly associated with higher ORs and lower rates of permanent complication and bleeding during the latency period.

Conclusions: The data from this study reinforce the conclusion that GKRS is a safe and effective treatment for pediatric and adolescent cAVMs, yielding a high OR with minimal permanent severe morbidity and no mortality. The very low frequency of severe hemorrhages during the latency period further encourages a widespread application of RS in such patients. Univariate analysis found that modified RS-based cAVM score, nidus volume, PD, integral dose, S-M grade, and preplanned treatment (the last 2 parameters were also confirmed on multivariate analysis) significantly influenced OR. Lower S-M grades and single-session planned treatments correlated with shorter treatment obliteration interval on univariate analysis. This statistical analysis suggests that a staged radiosurgical treatment should be planned when nidus volume > 10 ml and/or when the recommended PD is ≤ 16 Gy.

PMID: 26339954



Cerebral arteriovenous malformations (cAVMs) are the most frequent cause of intracranial hemorrhage in children after infancy, accounting for as many as half of all hemorrhagic strokes in the pediatric and adolescent population. These cAVMs more frequently present after hemorrhage than those in adults and clinical presentation as spontaneous hemorrhage is reported in up to 80-100% of pediatric cases. Furthermore, children have a higher risk of rebleeding (25% incidence of rebleeding within 5 years from initial hemorhage) than adults with cumulative mortality and morbidity presumably greater because of the prolonged risk period (1). These young patients are exposed to a 50% combined risk of permanent morbidity and mortality.

Moreover, cAVMs are more frequently deep-seated (in the basal ganglia, thalamus, corpus callosum or brain stem) or located in critical areas of the brain (i.e., in the motor strip, speech area, or visual cortex) in pediatric and adolescent patients  (up to 70%–90% of the time) than in adults (2). Particularly as concerns deep-seated cAVMs, hemorrhagic onset is described in 83% to 100% of cases. Variuos explanations have been offered for the greater frequency of bleeding in deep-seated cAVMs conpared with cAVMs in other locations: higher perfusion pressures in cAVMs with a high flow rate, the higher incidence of associated arterial and venous aneurysms, and a smaller number of venous drains. However, the explanation most commonly cited in the literature is venous hypertension within the angioma nidus. Deep-seated cAVMs usually drain into a single vein, the internal cerebral vein, and this constitutes an important risk factor for hemorrhage. Stenosis at the level of the venous drain readily develops because the high blood flow from the angioma nidus encounters a sharp bend and a narrowing of the vessel at the point where the vein of Galen enters the straight sinus. So, increased pressure and resistance within the internal cerebral vein are transmitted back to the nidus, facilitating hemorrhages (3).

Always referring to deep-seated and critically located cAVMs, given the functional significance of their location and a pattern of vascularisation without anastomoses, this may have devastating clinical and neurological consequences in case of bleeding: serious deficits of sensory-motor system, speech, visual field, cognition and memory, leading to permanent handicaps in large numbers of these young patients (3).

For all these reasons, complete surgical removal, with the aim of removing the risk of further, devastating hemorrhagic strokes, should be the treatment of choice for cAVMs, above all for pediatric and adolescent patients. But, in these cases, craniotomy turns out to be extremely dangerous and hazardous because of its invasive nature. Indeed, microsurgical treatment for pediatric and adolescent cAVMs is hindered by difficulties of access, deep-seated location, the complexity of vascular architecture and the close anatomical relationships between cAVMs and cerebral areas having enourmous functional importance in several cases. These problems may represent serious obstacles to complete microsurgical excision in a single operation, so that a second intervention is required for radical surgical removal in 17-25% of reported cases. The cAVM remnant leaves operated patients with a continuing risk of post-microsurgical bleeding/ rebleeding, an occurrence which is reported 4.5-12.5% of cases (3). Finally, microsurgical series report higher rates of permanent neurological damage and operative mortality in children than in adults.

Therefore, endovascular embolization (EE) has been proposed for the treatment of these patients. Nevertheless, EE still carries the risks of an invasive technique and it is uncommon that complete obliteration is achieved when used as a single treatment modality (1). Most relevant papers dealing with predictors of unfavorable outcome identified separately several factors for the different treatment modalities. In the case of EE, adverse factor linked to neurological complications were identified by means of morphological cAVM characteristics such as deep venous drainage and the size of the cAVM, basal ganglia location and eloquent areas (4). There is also the concern that revascularisation of obliterated vessels, despite its rarity, may occur more frequently in children than in adults (1). Thus, EE alone is often unsatisfactory as a single therapeutic strategy for the majority of inoperable cAVMs.

Therefore, taking into account the complicated and challenging management of pediatric and adolescent patients with cAVMs, the use of a noninvasive treatment tool such as radiosurgery (RS) has become more and more prevalent worldwide. Since the early 1970s, the application of RS in the treatment of cAVMs has become increasingly widespread, with thousands patients treated worldwide, including patients in pediatric and adolescent ages (1). In the last 2 decades, several studies on the radiosurgical treatment of pediatric and adolescent cAVMs have reported extremely favourable results: 74%-86% obliteration rates (ORs) and 1.1%-5.0% permanent complication rates (1,2,5). In particular, higher actuarial OR at 3 years and shorter treatment oblitaration interval (TOI) in younger than in older patients have been reported (5). Furthermore, these differences between the two patient populations were statistically significant. This phenomenon might be related to greater radiation sensitivity in younger patients, although no biologic evidence emerged to support this theory until recently. Histopathologic, immunohistochemical, and electromicroscopic examination of cAVM specimens obtained postoperatively 10–60 months after Gamma Knife (GK) RS demonstrated that the endothelial damage caused by irradiation induces the proliferation of smooth muscle cells and the production of extracellular collagen by these cells, which leads to progressive stenosis and obliteration of the cAVM nidus (6). In addition, the contractile activity of these gamma ray–activated, spindle-shaped smooth muscle cells and the transformation of the resting cells into an activated form after irradiation might be relevant to the shrinking process and eventual occlusion of cAVMs after RS (7). More recently, Hashimoto et al. (8) identified nonresting endothelial cells by using immunohistochemistry for the Ki-67 antigen from surgical samples of human cAVMs. They showed that the mean Ki-67 index was higher for cAVM vessels than for control brain cortical vessels (0.7% ± 0.6% vs. 0.1% ± 0.2%; p = 0.005), with an approximately sevenfold difference between the number of nonresting endothelial cells in the two samples. In the cAVM group, there was a trend for younger patients to have a wider variation and a higher Ki-67 index than older patients; no trend was evident in the control group. The greater number of nonresting cells found in young patients’ vessels might allow an earlier activation response in case of radiation treatment, inducing more rapid cAVM nidus obliteration in children/adolescents than in adults. The shorter TOI in children/adolescents than in adults has the clinical advantage of a lower risk of bleeding during the latency period in younger patients.

But, despite all these favorable evidences, considering the young age of such patients associated with a long-life expectancy and the absolutely benign nature of cAVMs, the question is: does radiosurgical treatment expose children/adolescents with cAVMs to higher risks of revascularization, long-term adverse radiation effect, postradiosurgery bleeding and permanent complication rates ? The evaluation of large numbers of observations with very long follow-up times are needed to give appropriate answers to these questions. Thus, we conducted a retrospective study which describes advantages, risks, and failures of GKRS on a large series of pediatric and adolescent patients with cAVMs (100 cases) who were followed up for a long period (at least 36 months). RS shows the great advantage of a non-invasive tool and the treatment needs a 36-hour-hospitalization, only. Briefly, the procedure foresees the applicationn of a MRI-compatible Leksell Model G stereotactic frame (Elekta Instruments) to the patient’s head. Then, neuroradiological localization was routinely performed using stereotactic 2D cerebral angiography or high-resolution magnification subtraction stereotactic angiograms. More recently, 3D rotational stereotactic angiography (with evaluation of the early arterial to late venous phases) has also been used to define the cAVM nidus (gross target volume) and to determine target coordinates. Angiographic examination was supplemented with stereotactic CT/MRI, with specific algorithms and sequences, to obtain additional information about the 3D shape of the cAVM and the surrounding normal brain structure. Postcontrast 2-mm-thick coronal and axial images and, more recently, 1-mm-isovoxel volumetric images with gadolinium enhancement, 2-mm-thick T2-weighted MR images, and MR angiography sequences were acquired. Radiosurgical procedures were performed with a model C 201-source Co60 Leksell Gamma Unit and, since June 2008, with GK Perfexion (both from Elekta Instruments). Three-dimensional treatment planning was developed using commercially available softwares. The neurological surgeon, radiation oncologist, and medical physicist created highly conformal dose planning using multiple collimators and performed the dose selection. General endotracheal anesthesia was usually induced in patients under 14 years of age and in selected cases in older patients. Patients were discharged from the hospital on the day after treatment (1,2,9).

At the end of our study, 99 patients were still alive (the only death in our series was due to rebleeding of a brainstem cAVM) with a mean and median follow-up period of 92.0 and 82.2 months, respectively. Our experience showed that an aggressive therapeutic approach allows to achieve a very high cure rate in such patients. Indeed, the reported OR following the first treatment was 89.3% which raised to 96.4% when a second radiosurgical treatment was performed on the cAVM residual nidus.

Furthermore, in patients with particularly critically located cAVMs (deep brainstem and basal ganglia) and/or with nidus volume exceeding 10 ml, a radiosurgically staged treatment – either dose staged (the whole cAVM nidus volume treated twice or three times with lower radiation doses than usual at 6 month-interval) or volume staged (different parts of the whole cAVM nidus volume treated with usual radiation doses at 6 month-interval, as well) was scheduled. This aggressive approch allows to achieve a 56.3% cAVM cure in patients otherwise untreatable.

At the same time, this aggressive approach does not entail an increasd risk of RS-related permanent complications. We registered a permanent symptomatic adverse radiation effect (ARE) in 11 out of 100 cases but 8 of them showed a Grade I/II Rankin Handicap Scale (10), that is minor symptoms which no interferes with lifestyle (Grade I) or minor handicap with some restriction in lifestyle but still totally independent (Grade II).

Bleeding during the latency period occurred in 9 out 100 cases (9%). In 1 of them, it dealt with a fatal rebleeding of a brainstem cAVM leading to the death of the patient (the only death in our series). Among the other 8 patients, hemorrhage resulted in a permanent neurological worsening in just 1 case.

Statistical analysis of our study showed that the probability of complete obliteration highly increases, and the risks of symptomatic permanent complications due to ARE and bleeding during the latency period highly decrease, when the nidus volume is less than 10 ml and a PD higher than 16 Gy can be delivered. Multivariate regression analysis confirmed a highly significant correlation for Sptzler-Martin (S-M) grading (11) and for type of scheduled treatment alone. In other words, patients with lower S-M grades (I-III) and preplanned single session treatment have an extremely high probability of cAVM cure after the first radiosurgical procedure.


Importance of the study. Our results show that:

  1. GKRS remains a safe and effective treatment for pediatric and adolescent patients with cAVMs, even after a long-standing observation period.
  2. High OR, minimal permanent severe morbidity, no treatment-related mortality, and very low frequency of severe hemorrhages during the latency period further encourages a widespread application of RS in treatment of pediatric and adolescent cAVMs.
  3. Statistical analysis revealed that S-M grade and preplanned treatment (single versus staged treatment) significantly influenced ORs. In particular, S-M Grades I, II, and III, and single-session planned treatment correlated with shorter TOI.
  4. Finally, our statistical analysis suggests that a staged radiosurgical treatment should be planned when nidus volume exceeds 10 ml and/or when the recommended PD should be inferiuor to 16 Gy with the aim of increasing the OR for “large” or critically located pediatric and adoloscent cAVMs while keeping the risk of permanent side effects very low.





Figure 1. Pre-GK magnetic resonance imaging (MRI) (A) and digital subtraction angiography (DSA) (B) showing a left temporal cAVM in a 13-year-old girl (gross target volume: 6 cc; prescription isodose: 50%; prescription dose: 19 Gy).





Figure 2. MRI (A) and DSA (B) follow-up at 25 months from GK tratement showing complete obliteration of the cAVM sparing normal brain parenchyma and vessels.




Table 1. Summary of clinical and neurological results. tab1

OR: obliteration rate. *: 8/11 Grade I-II according to the Rankin Handicap Scale; 3/11 Grade III according to the Rankin Handicap Scale. RN: radiation necrosis. §: due to rebleeding of a brainstem cAVM.



  1. Nicolato A, Foroni R, Seghedoni A, Martines V, Lupidi F, Zampieri P, et al. 2005 Leksell gamma knife radiosurgery for cerebral arteriovenous malformations in pediatric patients. Childs Nerv Syst 21:301–308.
  2. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, et al. 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.
  3. Nicolato A, Foroni R, Crocco A, Zampieri PG, AlessandriniF, Bricolo A, et al. 2002 Gamma knife radiosurgery in the management of arteriovenous malformations of the Basal Ganglia region of the brain. Minim Invasive Neurosurg 45:211–223.
  4. Beltramello A, Ricciardi GK, Piovan E, Zampieri P, Pasqualin A, Nicolato A, Foroni R, Sala F, Bassi L, Gerosa M. 2009 Operative classification of brain arteriovenous malformation. Part two: validation. Interventional Neuroradiology, 15:266-274.
  5. Nicolato A, Lupidi F, Sandri MF, Foroni R, Zampieri P, Mazza C, et al. 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.
  6. Schneider BF, Eberhard DA, Steiner LE. 1997 Histopathology of arteriovenous malformations after gamma knife radiosurgery. J Neurosurg 87:352–357.
  7. Szeifert GT, Kemeny AA, Timperley WR, et al. 1997 The potential role of myofibroblasts in the obliteration of arteriovenous malformations after radiosurgery. Neurosurgery 40:61–66.
  8. Hashimoto T, Mesa-Tejada R, Quick CM, et al. 2001 Evidence of increased endothelial cell turnover in brain arteriovenous malformations. Neurosurgery 49:124 –132.
  9. Nicolato A, Gerosa M, Ferraresi P, Piovan E, Pasoli A, Perini S, et al. 1997 Stereotactic radiosurgery for the treatment of arteriovenous malformations in childhood. J Neurosurg Sci 41:359–371.
  10. de Haan R, Limburg M, Bossuyt P, van der Meulen J, Aaronson N. 1995 The clinical meaning of Rankin ‘handicap’ grades after stroke. Stroke 26:2027–2030.
  11. Spetzler RF, Martin NA. 1986 A proposed grading system for arteriovenous malformations. J Neurosurg 65:476–483.
  12. Pollock BE, Flickinger JC. 2002 A proposed radiosurgery-based grading system for arteriovenous malformations. J Neurosurg 96:79–85.



Multiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier SchönmannMultiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier Schönmann