Abbreviations
AVM — arteriovenous malformation
PCF — posterior cranial fossa
MRI — magnetic resonance imaging
MS — microsurgery
RS — radiosurgery
S—M — Spetzler-Martin
CAG — cerebral angiography
GOS — Glasgow outcome scale
EV — endovascular
Introduction
Arteriovenous malformations (AVM) is a cerebrovascular malformation occurring between the third and eighth weeks of intrauterine development. This disease is usually diagnosed in young patients (mean age — 31.2 years). AVM is commonly diagnosed after hemorrhage or the first epileptic seizure [1, 2]. AVMs are often revealed in children due to improved diagnostic methods and their wider availability. Mortality after AVM-associated hemorrhage in children can reach 25% [3].
Considering high cumulative risk of AVM-associated hemorrhage and better recovery opportunities in children, treatment strategy is more aggressive in these patients compared with adults [4—6].
The main objective of the treatment of AVM in children is prevention of primary and recurrent intracerebral hemorrhage.
There are three main approaches to the management of cerebral AVMs: microsurgical, endovascular repair and radiotherapy.
The choice of cerebral AVM treatment in children is complex and depends on various clinical and X-ray factors.
Objective. To estimate the risk of hemorrhage in children with AVM before and after treatment, to evaluate the outcomes of various methods of AVM management in children.
Material and methods
A retrospective study was performed.
Inclusion criteria:
1. Patient's age 0—17 years at admission.
2. Cerebral AVM confirmed by angiography, CT-angiography or MR-angiography.
Exclusion criteria: AVM of vein of Galen; Conservative treatment of cerebral AVM.
There were 376 patients for the period 2008—2017 who were treated at the Burdenko National Research Centre for Neurosurgery. Repeated hospitalizations were not considered as separate cases. Mean age of patients was 11.1 years. There were 214 (57%) male and 162 (43%) female patients.
Angiomatous AVM were diagnosed in 345 (91,8%) patients. These malformations consisted of 3 components: afferent arteries, tangle, drainage veins. Fistulous AVMs were diagnosed in 31 (8,2%) cases. These malformations consisted of 2 components: afferent arteries and drainage veins (direct shunt).
Localization of AVMs and clinical symptoms are shown in Table 1.
Favorable treatment outcomes were considered GOS score 5 and 4, unfavorable outcomes — (severe disability) — score 3.
Quality of resection, embolization and post-irradiation obliteration was carried out using angiography data. Complete occlusion of AVM was characterized by the absence of AVM contrast enhancement, subtotal occlusion — insignificant contrast enhancement (1—30% of initial volume), partial occlusion — contrast enhancement of 31—99% of initial volume.
Choice of treatment approach for cerebral AVM in study patients
Microsurgical procedures were considered for small and middle AVMs (S—M grade I—III).
One-stage endovascular embolization was preferred in patients with favorable angioarchitectonic features (single afferent vessel, fistulous type of AVM, etc.).
Radiosurgery was scheduled for compact deep AVMs (subcortical nodes, internal capsule) or brain stem AVM followed by hemorrhage.
Large AVMs (S—M grade III—IV) with hemorrhagic type of course and multiple afferent vessels were scheduled for microsurgical resection after preliminary embolization or staged endovascular embolization.
Giant AVMs (S—M grade IV—V) extending to several lobes or functionally important and vital (brain stem) zones required individual approach. Symptomatic therapy was administered in case of isolated headache as a symptom of AVM. Anticonvulsant treatment was administered in patients with epilepsy.
Treatment of AVM in children depending on clinical manifestations, S—M grade and localization is summarized in Tables 2—4.
Results
Recurrent hemorrhages before treatment
Features of recurrent hemorrhage before treatment are shown in Table 5.
Case 1
A 5-year-old patient T. suffered parenchymal-ventricular hemorrhage on May 18, 2009. Bleeding was verified by CT of the brain (Fig. 1a). Severe general cerebral symptoms prevailed. Considering MRI data (Fig. 1b), a giant aneurysm of anterior communicating artery was suspected. Clinical deterioration with impairment of consciousness was noted 2 days later (May 20, 2009). CT revealed recurrent hemorrhage (Fig. 1c). The patient was hospitalized. Cerebral angiography (Fig. 1d, e) revealed fistulous AVM in the pool of the right anterior cerebral artery. Considering massive ventricular hemorrhage into posterior horns of the lateral ventricles, 2 external drainage tubes were installed for a period of 7 days (Fig. 1f). High-flow fistula was occluded by spirals (Fig. 1g, h). The patient satisfactorily recovered. CT of the brain (Fig. 1i) after 1 year (May 30, 2010) revealed moderate enlargement of ventricular system without need for bypass surgery.
Microsurgical resection of AVM
Resection of AVM was performed in 157 (41.8%) patients. Combined approach including preliminary endovascular embolization followed by microsurgical resection was applied in 22 (14%) out of these patients. Both procedures were performed within the same day.
Early (~10 days) clinical outcomes of AVM resection in children are shown in Table 6.
There were no patients in vegetative status. Postoperative mortality was absent.
Postoperative neurological complications were associated with surgical trauma as a rule. In one case (0.6%), deterioration was caused by hematoma in the bed of resected AVM that required debridement.
Hemianopsia was the most common symptom in 38 patients with postoperative neurological deficiency (n=21, 53.3%). Hemianopsia usually occurred (n=19) after resection of occipital AVM and AVM at the junction of parietal and occipital lobes. Hemiparesis was noted in 9 (5.7%) patients.
Selective cerebral angiography in early postoperative period was performed in 68 patients. These were predominantly patients with AVM S—M grade III and IV. Outpatient CT was recommended in most children with AVM grade I and II.
Considering outpatient CT angiography, quality of resection was assessed in 89 patients. Total resection was achieved in 84 (94.4%) patients, subtotal — in 4 (4.5%), partial resection — in 1 (1.1%) case. It is noteworthy that combined approach was followed by complete resection of AVM in all cases (Fig. 2).
Case 2
A 6-year-old patient with AVM of the left frontal lobe. Parenchymal hemorrhage (cold period). Cerebral angiography before embolization (Fig. 2a, b), after partial embolization with histoacryl (Fig. 2c, d) and after total microsurgical resection (Fig. 2e, f).
Residual contrast enhancement of AVM was observed in 5 patients. Redo microsurgical operations were performed in 3 cases. Complete resection was confirmed by angiography. Considering staged procedures, quality of resection was increased up to 97.8%. Radiosurgery was recommended in 2 patients with residual small deep AVMs.
Follow-up data for the period from 1 month to 10 years were collected in 76 patients. Favorable outcomes were observed in most patients (n=72; 94.7%). Severe disability (GOS score 3) persisted in 4 (5.3%) patients. No hemorrhage was observed in any patient. Postoperative seizures were absent in 9 out of 10 patients with previous epileptic seizures and only one of these patients takes anticonvulsants. In one patient with hemorrhage, epilepsy occurred after surgery.
Postoperative headache was not observed in 7 out of 10 patients with previous cephalalgia. Partial regression of cephalgic syndrome occurred in 2 patients. The same incidence and severity of headache were noted in 1 patient.
Endovascular treatment of AVM
Endovascular embolization of AVM was performed in 79 (21%) patients. The number of embolization stages varied from 1 to 8 (mean 1.6). There were 24 (30.4%) patients who required staged embolization.
Early (~10 days) clinical outcomes of endovascular embolization of AVM are shown in Table 7.
In one patient with small basal temporo-occipital left-sided AVM, unfavorable outcome after embolization was caused by massive ventricular hemorrhage in early postoperative period.
Thus, postoperative mortality after endovascular operations was 1.3%, after surgery for angiomatous AVM — 1.9%.
Overall quality of treatment after one or several stages of endovascular procedures is shown in Table 8.
Radiosurgery was recommended in those cases when total embolization of AVM was not achieved within one or more endovascular surgeries.
Clinical follow-up within 1—10 years after endovascular treatment was available in 31 patients. Neurological results were favorable in 27 (87.1%) cases, severe disability persisted in 4 (12.9%) patients. Recurrent hemorrhages between the stages occurred in 5 patients. There was no mortality associated with hemorrhage after embolization. Long-term postoperative seizures were absent in 2 out of 3 patients with preoperative epileptic seizures. In one patient with preoperative hemorrhage, convulsive syndrome appeared after embolization.
Radiotherapy of AVM
RS as an independent method of treatment was performed in 140 patients. RS after partial embolization, partial resection or partial post-radiation obliteration is not analyzed in this article.
The feature of treatment in children is anesthetic management during frame installation. All patients underwent MRI and cerebral angiography for preoperative scheduling.
Leksell Gamma Knife and Cyber Knife were used for radiosurgery. Physical parameters of irradiation are shown in Table 9.
Clinical follow-up within 1—10 years was available in 75 patients after radiosurgery. Results were favorable in 64 (85.3%) patients, severe disability persisted in 10 (13.3%) cases. In one patient, clinical deterioration was caused by subcortical radiation-induced necrosis. Recurrent hemorrhages were observed in 6 (8%) patients. Hemorrhage was followed by clinical deterioration in all cases. One (1.3%) patient died. Preoperative convulsive syndrome was detected in 2 patients. No positive dynamics was observed throughout the follow-up period. In addition, convulsive syndrome appeared after treatment in 5 patients. Among 8 patients with AVM followed by headache, complete regression was noted in 5 cases. Positive changes were not observed in 3 patients. In 7 patients, headache arose after treatment.
Angiographic control within 6 months—12 years was available in 73 patients. Total obliteration was achieved in 47 (64.4%) patients (Fig. 3),
Correlation of post-radiation obliteration and localization of AVM is shown in Table 10.
It is noteworthy that obliteration was worse in patients with AVM of subcortical nuclei and thalamus: complete obliteration was achieved only in 5 (35.7%) out of 14 cases. This was explained by smaller radiation dose (20—22 Gy) in this area to avoid neurological complications.
Case 3
A 5-year-old patient E. Ventricular hemorrhage developed on January 7, 2010. The disease was manifested by general cerebral symptoms. MRI and cerebral angiography (Fig. 3a, b) revealed left-sided AVM within subcortical nuclei. Radiosurgery was performed on December 9, 2010 (CyberKnife, dose — 22 Gy). Clinical state was favorable. There was no subsequent deterioration. Control angiography after 3 years (March 14, 2014) revealed no residual contrast enhancement of cerebral AVM (Fig. 3c, d).
Among all study patients available for follow-up (n=182), favorable long-term results were achieved in 163 (89.6%) cases. Recurrent hemorrhages were observed in 11 (6%) patients. One patient (0.6%) died from recurrent hemorrhage.
Discussion
Microsurgical resection of AVM in children is the most common approach [4, 6–11]. According to the literature, disability is observed in 5–25% of children with AVM who underwent microsurgical resection [4, 10, 11]. According to our data, neurological deterioration was observed in 24.2% of patients after resection of AVM. However, these disorders were moderate as a rule and most children (89.8%) were discharged in satisfactory condition. In our series, there were no deaths after microsurgical resection. At the same time, mortality of 3.7—5% is reported in the literature [4, 10].
It should be also noted that combined surgical approach in patients with complex AVMs (S—M grade III—IV) was followed by favorable neurological results.
Various authors reported positive aspects of combined treatment of AVM [4, 5, 6, 8, 9]. Firstly, preliminary endovascular embolization significantly reduces the risk of blood loss and time of surgery that is especially important for younger children. Secondly, a glue in abnormal vessels of AVM allows clearer demarcation between AVM stroma and intact cerebral parenchyma with normal vessels.
According to the literature, complete microsurgical resection of cerebral AVM in children is achieved in 67—89% of cases [4, 5, 10], in some series — 97—100% [6, 11, 12].
In our series, complete microsurgical resection was performed in 97.8% of cases.
According to several large studies, complete resection of cerebral AVM confirmed by cerebral angiography is not followed by recurrent hemorrhage [1, 13]. In our group, we also did not observe hemorrhages after complete resection of AVM in children.
At the same time, there are literature data on residual AVMs in children after angiographic confirmation of complete occlusion of AVM in early postoperative period [14, 15]. Bristol R. E. et al. [4] reported 5.6% of children with residual AVMs in long-term postoperative period (over 11 months) while postoperative cerebral angiography did not reveal contrast enhancement of abnormal vessels.
There are several hypotheses explaining development of AVM de novo in children. Some authors suppose that small immature vessels remaining in surgical field after resection of AVM transform into a new AVM along with growth of a child [15, 16]. Others believe that stimulation of angiogenesis can be triggered by surgical trauma and inflammation [17, 18] or increased expression of humoral growth factors especially vascular endothelial growth factor [19—22].
Endovascular surgery is a common approach in children for the treatment of AVM [8, 23—26]. In general, postoperative aggravation of focal neurological disorders was less common compared with microsurgical operations. At the same time, according to our data, disorders were more significant in complicated patients (severe disability in 20.2% of cases, mortality — 1.3%). According to the literature, mortality after endovascular operations in children with AVM is 0—4.9% [23—25].
It was noted that endovascular occlusion of AVM in children is characterized by worse quality if this approach is used in isolated fashion. According to various authors [23–26], incidence of total occlusion of AVM in children after embolization is 12–27.5% (including several surgical stages).
According to our data, incidence of total embolization in children was 29.1%. This value was higher in patients with fistulous AVM.
Risk of hemorrhage should be considered in patients with incomplete embolization of AVM. This aspect is not discussed in numerous trials since long-term results are still being studied. Among our patients, recurrent hemorrhages after embolization were observed in 6.3% of cases.
Radiosurgery is also common method for treatment of cerebral AVM in children. Total radiation dose in pediatric group is 14–40 Gray [27–31].
Incidence of AVM obliteration in children is 43—72% within 6—36 months after radiosurgery [28, 29, 31, 32, 33]. Several courses of RS are followed by complete occlusion of 80—95% of AVMs in long-term post-radiation period (4—5 years) [27, 33—36]. In our sample, total obliteration after a single exposure was achieved in 64.4% of cases. Better obliteration is observed in children under 12 years old, AVM with a diameter of less than 3 cm and volume of less than 3.8 cm3 [34, 36]. Post-radiation reactions followed by neurological deterioration are observed in 3—11.4% of cases [31, 32, 35], mortality rate — 1% [36]. Intracranial hemorrhage may also be a cause of deterioration and even mortality after irradiation of AVM. According to the literature, overall incidence of hemorrhages after radiosurgical treatment of AVM in children is 4—7.5% [31, 34, 36, 37]. In our series, hemorrhages after irradiation were observed in 8% of patients. Annual risk of hemorrhage after irradiation of AVM in children is determined within 1.5–2.6% [28, 29, 31, 36]. It is noteworthy that in rare cases hemorrhages occurred after angiographic confirmation of complete radiation-induced obliteration of AVMs [36].
Timing of treatment of cerebral AVM in children is important especially in patients with previous hemorrhage. According to our data, early (within the first month) recurrent hemorrhages from angiomatous AVMs in children are rare (1.1%) and more typical for deep AVMs. Therefore, as in adults [38], we prefer microsurgery in long-term post-hemorrhagic period in children in the absence of a large intracerebral hematoma because brain tissue changes and cerebral edema associated with hemorrhage significantly complicate surgery. At the same time, there are no restrictions on the timing of endovascular surgery in children with cerebral AVM if severe edema is absent.
In our opinion, it is not advisable to postpone radiotherapy in children with AVM of subcortical nuclei and thalamus due to high risk of recurrent hemorrhage. It is important to remember that the process of radiation-induced obliteration of AVM is quite long.
An interesting question is epileptic syndrome after treatment of cerebral AVM in children.
According to the literature, microsurgical resection is followed by relatively good results regarding relief of epileptic syndrome. Thus, Heros R. C. et al. [13] reported regression of epilepsy in more than 50% of patients after microsurgical resection. Kondziolka D. et al. [3] revealed that total resection of AVM in children ensures complete relief of epilepsy in 73% of patients. According to our data, seizures in long-term postoperative period were not observed in 9 (90%) out of 10 children with preoperative epilepsy. There is also evidence that embolization of AVM reduces incidence of seizures. Krasnova M.A. [7] analyzed material of the Institute of Neurosurgery in earlier years. The authors found out that endovascular surgery was followed by reduced number of seizures in 12 (48%) out of 25 children with drug-resistant seizures and complete regression in 4 (16%) cases.
Data on regression of seizures after radiosurgery in children with epileptic syndrome are unclear. Gerszten P. et al. [39] found cessation of seizures without the use of anticonvulsants in 1 year after radiosurgery in 11 (85%) out of 13 children. Among our patients, there was no positive dynamics regarding epilepsy after irradiation. Moreover, seizures de novo appeared after irradiation in 6.6% of patients.
Conclusion
Microsurgical resection is preferred in children with angiomatous AVM because this approach ensures higher quality of treatment and prevents recurrent hemorrhages. Combination of embolization and microsurgical resection is advisable for large AVM and characterized by low risk of postoperative complications.
Endovascular surgery as a primary treatment option is indicated in children with fistulous AVM. Early intervention is indicated in patients with previous hemorrhage from fistulous AVM since the risk of recurrent hemorrhage within the first month is higher compared with angiomatous AVM.
The results of radiosurgical treatment of AVM in children also look quite good. However, it is not worthwhile to postpone treatment of AVM for a long time in children with increased risk of recurrent hemorrhage considering relatively long post-radiation obliteration.
Long-term angiographic control is desirable in all patients undergoing treatment for cerebral AVM in childhood
The authors declare no conflicts of interest.
Commentary
Treatment of patients with cerebral arteriovenous malformations (AVM) is still one of the most difficult problems. Optimal solution of this issue has not yet been achieved in any country in the world.
The main problem is unclear understanding of the pathogenesis of AVM. Therefore, accurate prediction of natural course of disease is impossible while assessment of acceptable risks of treatment is complicated.
There are three competitive methods (microsurgery, embolization, radiosurgery) with certain advantages and disadvantages in each case. This situation introduces significant dissonance in formation of a single point of view on treatment strategy in patients with cerebral AVM.
No consensus on the treatment of malformations in adults is followed by even more unclear approaches to AVM management in children. Therefore, this study is extremely relevant.
There were 376 children with cerebral AVM who were treated at the Burdenko Neurosurgery Center for a 10-year period. Microsurgical resection of malformations was performed in 157 cases (42%), embolization in 79 (21%) patients, radiosurgery — in 140 (37%) cases.
The authors collected the largest sample of children undergoing microsurgical resection in Russia and one of the largest in the world. Therefore, a reliable assessment of the outcomes in subgroups and comparison with other methods are possible.
High surgical experience is confirmed by no mortality, complete resection in almost 98% of cases and discharge of 90% of patients in satisfactory condition. The authors can confidently defend their point of view on approaches to surgical treatment of these patients.
In this study, we can talk about the evidence base of effectiveness of surgical concept. This strategy implies microsurgical resection of AVMs Spetzler—Martin grade 1—3 in cold period and additional embolization in especially difficult cases.
In fact, this concept duplicates the approaches previously developed by the same authors for the treatment of AVM in adults. Now the authors have convincingly demonstrated effectiveness of this approach in pediatric patients.
Children undergoing radiosurgery were comparable by most parameters. This approach was characterized by somewhat worse results regarding quality of removal (64%) and risk of hemorrhage (8%). Nevertheless, these values were similar to literature data.
The progress achieved in endovascular technologies over the past 5—7 years is a reason for another intensification of competitive approaches to the treatment of AVM. Actually, even the possibility of treatment of AVM grade 4 (sometimes even grade 5) is one of the serious advantages of modern endovascular operations (1—4).
Unfortunately, the authors did not clearly indicate specific techniques used in endovascular treatment (embolization substrates, technique, vascular pool for approach, dependence of technique and embolization stage, etc.) that makes perception somewhat more difficult.
It is reasonable that the authors did not consider the results of combined treatment in this work. Perhaps, these data are still not sufficiently analyzed and ready for presentation. I hope that we will be able to familiarize with these data.
I would like to congratulate the authors with excellent work and express confidence that this study will be really needed by our neurosurgical society.
References
1. Mosimann PJ, Chapot R. Contemporary endovascular techniques for the curative treatment of cerebral arteriovenous malformations and review of neurointerventional outcomes. Journal of Neurosurgical Sciences. 2018;62(4):505-513.
2. Crowley RW, Ducruet AF, Kalani MY, Kim LJ, Albuquerque FC, McDougall CG. Neurological morbidity and mortality associated with the endovascular treatment of cerebral arteriovenous malformations before and during the Onyx era. Journal of Neurosurgery. 2015;122(6):1492-1497.
3. Iosif C1, Mendes GA, Saleme S, Ponomarjova S, Silveira EP, Caire F, Mounayer C. Endovascular transvenous cure for ruptured brain arteriovenous malformations in complex cases with high Spetzler-Martin grades. Journal of Neurosurgery. 2015;122(5):1229-1238.
4. Saatci I, Geyik S, Yavuz K, Cekirge HS. Endovascular treatment of brain arteriovenous malformations with prolonged intranidal Onyx injection technique: long-term results in 350 consecutive patients with completed endovasculartreatment course. Journal of Neurosurgery. 2011;115(1):78-88.
A.Yu. Ivanov (St. Petersburg, Russia)