Stereotactic radiosurgery for intracranial arteriovenous malformations: A review
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.170102
Source of Support: None, Conflict of Interest: None
Stereotactic radiosurgery (SRS) has proven to be an effective strategy in the management of intracranial arteriovenous malformations (AVMs) in children and adults over the past three decades. Its application has resulted in lowering the morbidity and mortality associated with treatment of deep-seated AVMs. SRS has been used as a primary modality of therapy as well as in conjunction with embolization and microsurgery in the management of AVMs. The obliteration rate after SRS has been reported to range from 35% to 92%. Smaller AVMs receiving higher marginal doses have obliteration rates of 70% and more. The median follow-up reported in most series is approximately 36–40 months. The median time to obliteration has been reported to be approximately 24–36 months in most series. Radiation-induced neurological complications have been reported in less than 10% of patients, with a 1.5%–6% risk of developing a new permanent neurological deficit. The bleeding rate during the latency to obliteration has been reported to be approximately 5%. This review describes the experience reported in literature with respect to the indications, dosage, factors affecting obliteration rate of AVMs, and complications after SRS.
Keywords: Arteriovenous malformations; brain AVM; gamma knife; hemorrhage; intracranial; linear accelerator; stereotactic radiosurgery
Intracranial arteriovenous malformations (AVMs) are developmental anomalies that present with parenchymal or intraventricular bleeds in children and adults, resulting in considerable morbidity exceeding 25%.,,,, Children are probably at a higher lifetime risk of hemorrhage from an AVM as compared with adults. They are more likely to present with hemorrhage from an AVM than adults. The annual rate of hemorrhage from intracranial AVMs has been reported to be between 2 and 4%, with the yearly risk being possibly higher among children., The primary goal of management in these patients is to prevent rebleed. Although endovascular embolization has been used to treat AVMs, its role as a sole treatment modality in achieving complete obliteration is limited., While microsurgery remains the treatment of choice for patients with small superficial AVMs that have bled, stereotactic radiosurgery (SRS) has a definite role in the management of these AVMs as well as deep-seated AVMs that are associated with a higher surgical morbidity.,,,,,,,,,,
A literature search was performed from PubMed using a combination of the following keywords: "brain AVM," "intracranial AVM," "stereotactic radiosurgery," "haemorrhage," "complications of radiosurgery", "repeat radiosurgery", "hypofractionated stereotactic radiotherapy"; and articles published in English language retrieved. While the data related to the results in children were retrieved from articles dating back to 2000, the overall data incorporating findings amongst all age-groups was summarized from articles published over the past three years.
SRS, traditionally, refers to the technique of treating a lesion in the brain with a single session of radiation therapy, administering a dose of radiation to the target with rapid dose fall-off, thereby minimizing the dose delivered to the adjacent normal brain. In 2006, the American Association of Neurological Surgeons, Congress of Neurological Surgeons, and American Society for Therapeutic Radiology and Oncology modified the definition to include delivery of radiation in up to five doses. It has been in vogue for nearly four decades for the management of several pathologies in the brain. The mode of delivery of SRS has been the linear accelerator (LINAC), gamma knife (GK), or proton beam.,,
LINAC uses high-energy X-rays (photons) to deliver the radiation, while GK uses gamma rays from cobalt-60 isotope to deliver the radiation. Proton beam radiosurgery involves the use of high-energy protons as the source of radiation. A variation of LINAC-based planning and radiation delivery is robotic radiosurgery (CyberKnife), in which the patient need not be immobilized during radiation delivery.,, The software accounts for movements of the patient, and inverse planning is used, applying constraints to the dose delivered to eloquent structures in the vicinity of the AVM. With the currently available dose delivery techniques, LINAC and GK systems have comparable results. Dose distribution and delivery in plans comparing CyberKnife and micromultileaf collimator have also been found to be not significantly different from each other.
The K index—calculated as the prescribed minimum dose of radiation delivered x (AVM volume)1/3—has been proposed to guide the dose of radiation delivered. However, its use may be limited to SRS for small AVMs, with obliteration rates increasing linearly up to a value of 27. In one series, the prescribed dose was reported to be lower than the calculated K index, and this modification was done owing to smaller volume of the AVM and did not result in a lower obliteration rate. Thus, the prescribed dose of radiation to the AVM is a function of the AVM volume and location. SRS planning ensures that only a small volume of normal tissue adjacent to the target receives radiation. If required, several separate targets can be chosen to treat a lesion with a large volume, and each target is treated as a separate isocenter. Radiation may be delivered via circular collimators through noncoplanar arcs or as fixed shots (in GK), or via micromultileaf collimator through fixed beams.,,,,
The prescribed dose of radiation is with reference to the isodose that would cover the periphery of the AVM. In effect, the areas within this isodose would receive higher than the prescribed dose, while the adjacent tissue outside this isodose would receive far lower doses of radiation. Treatment planning in SRS differs from that in conventional radiotherapy techniques in that it ensures that the dose delivered to the normal brain tissue is minimal, and this is achieved by conforming the dose delivered to the desired volume of the target.
Radiobiological effects of SRS
Stereotactic definition of the target restricts the radiobiological effects of SRS to the AVM, with only a small volume of the surrounding normal tissue receiving radiation. The immediate effect of SRS is damage to the endothelial cells of the vessels in the nidus, probably mediated by release of tissue-specific cytokines. This is followed by initiation of a chronic inflammatory process, with formation of granulation tissue that has fibroblasts and new capillaries. Myofibroblasts, which are actin-producing fibroblasts, have been detected in the region of radiation, and these have been postulated to exert contractile properties and facilitate AVM obliteration. The ensuing radiation-induced vasculopathy results in progressive occlusion of vessels within the AVM nidus. This process takes from 1 to 3 years.
AVM obliteration after SRS
MRI has been used as an effective screening tool to document nidus obliteration after SRS., MRI has been shown to have a reliability of 97% in documenting AVM obliteration. Most authors recommend an yearly follow-up with MRI after SRS, and DSA may be used to confirm its obliteration once the MRI shows evidence of obliteration. MRI also aids in assessing radiation-induced changes in the vicinity of the nidus.
Obliteration of AVMs after SRS has been reported to range from 35% to 92%, with the obliteration rate exceeding 70% in most series [Table 1].,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, In a series of unruptured AVMs in children, SRS resulted in an obliteration rate of 58%. In another series reporting on the results of radiosurgery for incidental AVMs, the obliteration rate was 61%. The obliteration rate with small AVMs has exceeded 80% in most series., In several of these series, the obliteration rate of AVM after initial radiosurgery has been further improved upon by subjecting patients with residual lesions to repeat SRS., For large AVMs and Spetzler–Martin grade 4 AVMs, volume-staged SRS has been reported to increase the obliteration rate.,
The interval-to-obliteration after SRS could be from 1 to 4 years or even longer. In one series, the median time to obliteration of AVMs was as delayed as 4.2 years. The highest possibility of AVM obliteration is within the first 3 years after SRS, with AVMs in noneloquent regions having a shorter latency period before obliteration. This shorter latency period may be a function of the higher marginal dose of radiation that was delivered to the AVM located in a noneloquent region. A minimum of 3- to 4-year follow-up of the AVM is required, before SRS can be deemed as a failure.
Factors affecting AVM obliteration
Smaller AVM volume, higher marginal dose of radiation, smaller maximal diameter, smaller number of isocenters, radiosurgery-based AVM score <1, lower modified Pollock–Flickinger score, lower Spetzler–Martin grade, younger age, and absence of a history of embolization have all been documented to have correlation with higher obliteration rates in series reporting on SRS for AVMs.,,,,,,[22,,,,,,,,,,,,,,,,, No difference in the AVM obliteration rate was observed among the patients >60 years of age as compared with adults <60 years.
Volume of AVM and marginal dose
In several series, smaller-volume AVMs have been reported to have higher obliteration rates than the overall obliteration rate reported.,,,,,, Smaller AVM volume allows prescription of a higher dose of radiation, and hence, it would be expected to have a higher chance of obliteration. Recently, several authors have confirmed that the most important predictor of obliteration was a higher marginal dose of radiation.,,,,,,,,,, In LINAC-based SRS for AVMs, the obliteration rate ranged from 55% to 89.2%, with a mean marginal dose ranging from 18 to 25 Gy.,,,, Potts et al., reported a 52% rate of obliteration with a dose of >18 Gy, while only 16% of AVMs that received <18 Gy were obliterated in a series reporting on the results of GK-based SRS for AVMs.
Spetzler–Martin grading and AVM radiosurgery score
The Spetzler–Martin grading was designed to assess the risk of morbidity related to excision of AVMs. After SRS, Spetzler–Martin grade I/II AVMs have been reported to have a higher obliteration rate as compared with grade III–V AVMs.,,, Complications after SRS were also reported to be lower among Spetzler–Martin grade I/II AVMs. The radiosurgery-based AVM score developed at Pittsburgh and its subsequent modifications (Pollock–Flickinger score) have taken into account the role of AVM volume and location that are indirect determinants of the dose of radiation that can be optimally prescribed to an AVM., Hence, this modified AVM radiosurgery score is a better predictor of outcome after SRS as compared with the Spetzler–Martin grading. The AVM radiosurgery score has been validated in various series of AVMs in adults and children treated with LINAC as well as GK, where it was found to be a significant determinant of outcome.,,,,,,, Children with modified AVM radiosurgery score of <1 had a higher obliteration rate associated with a lower rate of radiation-induced complications, with up to 88% chance of an excellent outcome., Kano et al., reported a higher obliteration rate with lower symptomatic adverse radiation effects in children with a lower Pollock–Flickinger score. Raffa et al., reported excellent outcomes in 70% of patients with an AVM radiosurgery score of <1 as compared with only 11% of patients with an AVM radiosurgery score of >2.5. A scoring system developed in the University of Virginia has also been validated among adults and children, with lower scores being associated with increased chance of AVM obliteration and low post-SRS morbidity.,,, However, some authors have reported no correlation between AVM obliteration and the AVM radiosurgery score or the Spetzler–Martin grading.,,,
Embolization of AVMs is often not curative but could decrease the chances of rebleed in the interim period while awaiting SRS., Several reports as well as a meta-analysis have suggested that previous embolization could result in a suboptimal definition of the AVM target for SRS, and this could result in the AVMs not being completely obliterated.,,,,,, It is also possible that the nidus that was not visualized after embolization could recanalize at a later stage and hence would contribute to lower obliteration rates. Moreover, it is likely that AVMs that are subjected to partial embolization will be larger in volume compared with those that can be completely embolized, and this would be a confounding factor while assessing outcomes comparing embolization and SRS.
However, some recent reports suggest that previous embolization may not be associated with a lower obliteration rate after SRS.,,, For large AVMs, the strategy of staged embolization to reduce the volume of lesion up to 33% followed by SRS for the residual AVM has resulted in obliteration rates up to 45%.
The angioarchitecture of AVMs has shown poor correlation with AVM obliteration in the pediatric age-group. A compact nidus and a lower number of draining veins have been reported to be favorable factors towards achieving AVM obliteration., Others have not reported any relationship between the angioarchitecture of the AVM and their obliteration rates. Target definition is easier and more likely to be complete with a compact nidus, and it has been suggested that one of the main reasons for failure of radiosurgery is suboptimal delineation of the AVM. In various series reporting the outcomes of SRS in children and adult population, the presence of a single draining vein has been a predictor of obliteration.,
Complications of SRS in AVMs
New-onset neurological deficits after SRS have been reported in 0–17.6% patients in different series of LINAC- and GK-based SRS in AVMs.,,,,,,,,,,, Permanent neurological deficits after SRS have been reported to occur in 1.5–6% of patients.,,,,,, A higher incidence of radiation-induced complications has been reported after LINAC-/GK-based SRS in children with larger volumes of AVM as well as with Spetzler–Martin grade IV and V AVMs.,,,, AVMs located in the brainstem, thalamus, or basal ganglia as well as those with a higher Pollock–Flickinger score have been reported to have a higher risk of developing post-SRS neurological deficits.
In patients presenting with seizures before SRS, radiosurgery was associated with persistence of seizures as compared with microsurgery. AVM obliteration was associated with seizure freedom. The risk of new-onset seizures after treatment was higher with surgery than with radiosurgery. Another series has corroborated the fact that no new seizures developed after SRS in patients who had been seizure-free before treatment. Patients presenting with simple partial or generalized seizures had better seizure outcomes after SRS, and AVM obliteration was strongly associated with withdrawal of antiepileptic drug therapy. Contrary to this, seizure control after SRS did not correlate with AVM obliteration in a series reporting on only temporal lobe AVMs. New-onset seizures have been reported after proton beam radiosurgery that persisted in up to 9.2% of patients on a long-term follow-up.
Radiation-induced imaging changes
Asymptomatic signal changes on MRI have been reported in up to 30% of patients from 2 to 7 years after SRS., In a large series, 33.8% of 1426 patients who underwent SRS were reported to have radiation-induced imaging changes at a median duration of 13 months after SRS. The imaging changes were symptomatic in 8.6% patients, and there was resolution of these changes at a follow-up of almost 2 years. Large size of the nidus, presence of a single draining vein, absence of previous hemorrhage, presence of diabetes mellitus, a higher Virginia radiosurgery AVM scale score, and absence of surgical intervention before SRS, have all been associated with the risk of eliciting radiation-induced imaging changes., 53, ,, In a smaller series of 10 patients, the incidence of signal changes on MRI after SRS and hypofractionated SRT (HSRT) was as high as 60%. Segmental narrowing in normal arteries, which is clinically silent, has been documented in children on long-term follow-up angiograms. Dysplastic changes in feeding arteries and early venous thrombosis with resultant venous infarction have been reported after SRS in one series.
Rebleed or hemorrhage from AVM
Almost 60% of patients presented with hemorrhage from AVM before SRS in most series., Hence, the main goal of SRS was to prevent rebleeding of the AVM. The inherent risk of rebleeding remains up to the time AVM is obliterated. Rebleeding could be due to the residual nidus as the AVM undergoes obliteration. It has not yet been established whether partial obliteration of AVM would minimize the risk of a bleed. Rebleeding has also been associated with development of de novo arterial pseudoaneurysms and new venous varix after radiosurgery. Rebleeding from previously ruptured AVMs or hemorrhage from AVMs after SRS has been reported in up to 25% of patients by Potts et al. Other series have reported rates between 1.3% and 9%.,,,,,,, The risk of hemorrhage has been observed to be the highest during the first year after treatment. Low marginal dose of radiation, presence of multiple draining veins, and periventricular location of the AVM have been associated with the risk of hemorrhage in the latency period between SRS and AVM obliteration., Pregnancy during the latency period before AVM obliteration may be a risk factor for AVM hemorrhage and hence patients may be advised to defer pregnancy till obliteration has been achieved.
Cyst formation within the brain parenchyma can occur as a delayed complication after SRS in 1.5%–3.4% of patients, the median duration of cyst formation being reported from 4 years to up to 10 years after SRS.,,, These cysts could causes a mass effect necessitating surgical decompression. It has been postulated that the cysts form owing to radiation-induced necrosis and expand as a result of repeated hemorrhages into it from an enhancing nodule (probably radiation-induced cavernous angioma or noncavernous angioma-like granulation tissue with fragile capillaries).
Radiation-induced neoplasms such as meningiomas and glioblastomas have been reported to occur in children after SRS for AVMs.,, However, the reported incidence of occurrence of these neoplasms is not higher than that reported in the general population. Pollock et al., have reported an incidence of radiation-induced tumors of 0.007% at a long-term follow-up after radiosurgery. In a cohort of patients with long-term follow-up, the chance of developing a radiation-induced neoplasm was 0.64%, 10 years after undergoing SRS for an AVM, and the cumulative rate among 78 patients who had a 15-year follow-up was 3.4%. These reports do underscore the importance of long-term follow-up of patients, particularly children, with AVMs treated with SRS.
Recurrence of AVM
Lindqvist et al., reported the recurrence of an AVM with hemorrhage in children who had documented complete obliteration of the AVM after SRS. This has been reported in adult patients too, and the recurrence may be at the site of the previous nidus or a new nidus may develop adjacent to the previous one. Children with AVMs undergoing SRS are at a lifetime risk of symptomatic recurrence, albeit very rare, and hence need a long-term follow-up.
Although no cognitive or neurobehavioral sequelae were reported in a small number of patients by Riva et al., Yeon et al., reported a decline in the scholastic performance in 44% of children who underwent SRS for AVMs. Long-term studies on larger numbers of patients are needed before it can be ascertained that SRS in children will have no bearing on cognition. In a series reporting data on 10 patients who underwent hypo-fractionated stereotactic radiotherapy (HSRT) for AVMs, defects were observed in several domains tested before therapy. At 6 weeks after completion of HSRT, there was worsening of semantic processing that returned to baseline after 6 months. At longer-term assessment 2.5–3.5 years after HSRT, there was improvement in memory function, probably as a result of restoration of normal hemodynamic pattern of blood flow to previously hypoperfused regions.
Repeat radiosurgery is an option when patients have persistent or residual nidus 3 or more years after the initial SRS.,, The dose–response curve for obliterating a previously treated AVM has been reported to be similar to that of an untreated AVM. When there is reduction in the nidus volume after the initial SRS, it may be possible to administer a higher marginal radiation dose, facilitating obliteration., The marginal dose of radiation delivered during the repeat SRS may have to be reduced if the nidus volume has increased (as occurs rarely after recanalization of partially embolized AVMs) or if the patient has had an adverse radiation effect after the initial SRS. Several series have improved upon previous obliteration rates by repeating SRS of the residual nidus. Kano et al., reported an 80% actuarial obliteration rate at 10 years after repeat SRS in 105 patients. Reduction in the volume of an AVM by 50% or more and a smaller target volume were associated with a higher chance of total obliteration. Reduction in target volume by less than 50% after the initial SRS and the presence of hemorrhages before repeat SRS were associated with increased risk of hemorrhage after repeat SRS. In this series, the rate of symptomatic radiation-related side effects was 10.5% after repeat SRS, while it was 4.8% after initial SRS. Other series have reported obliteration rates of up to 65% after repeat SRS, with similar rates of complications.,,, In a meta-analysis of 733 patients who underwent repeat SRS for incompletely obliterated AVMs, the mean obliteration rate was 61% (95% confidence interval, 51.9%–71.7%). The interval to complete obliteration after repeat SRS ranged from 21 to 40.8 months. Overall, a hemorrhage risk of 7.6% and radiation-induced changes among 7.4% of cases were reported after repeat SRS. The results are summarized in [Table 2].
Radiosurgery strategies for large AVMs
While small AVMs and SM grade I/II AVMs have a good response to SRS, management of larger AVMs require individualized strategies that are often multimodal.,,,,,,,,,, Large AVMs, defined in most series as those with a volume more than 10 cm 3, can be treated with volume-staged SRS or dose-staged SRS. In each of these strategies, the underlying principle is to reduce the volume being treated in each radiosurgery session, so that an optimal dose of radiation can be administered. This may be achieved by dividing the AVM into smaller volumes and administering SRS to each volume in a staged manner, a few months apart (volume-staged SRS).,,, The other strategy is to treat the entire volume with a single-session RS and then to repeat the same after 3 years, so that the initial SRS reduces the volume of the AVM to a size that would have a higher chance of obliteration with repeat SRS (dose staging).,,
Apart from repeat SRS, HSRT offers an alternative dose-staging strategy to treat large AVMs by delivering the dose in 3–6 fractions (5–7 Gy per fraction) and reducing the dose delivered per fraction. Between each fraction, an interval of up to a week is given for the normal tissue to recover from radiation-related acute injury.,,,
With the aforementioned strategies, obliteration rates more than 50% have been achieved. However, as summarized in [Table 3], the post-SRS hemorrhage risks and radiation-related complications appear to be slightly higher in this subgroup of patients, as compared with that reported with single-session SRS for smaller AVMs., 63, ,,,,,,,, In one series, the paradigm of reducing large AVMs, considered inoperable, to smaller volumes with multiple sessions of SRS followed by surgery for the residue has been reported. It has also been proposed that after SRS, large-sized AVMs could be partially embolized with a view to achieve size reduction and partial cure as well as to reduce the risk of hemorrhage., As any intervention in large AVMs is associated with lesser chance of AVM obliteration and higher percentage of risks of therapy, Spetzler–Martin grade IV/V AVMs may be observed without any intervention unless they have bled or have produced progressive neurological deficits.
Radiosurgery for unruptured AVMs
Results of the ARUBA trial (wherein recruitment was prematurely halted) showed a clear superiority of medical management for adult patients with unruptured AVMs as compared with any intervention, with risk of death or stroke being higher in the intervention group. However, in this randomized trial of medical management versus intervention for unruptured AVMs, 31 patients in the intervention group underwent radiotherapy alone, and the mortality/morbidity rates in this subgroup have not been reported separately. It is possible that the mortality and morbidity may have been skewed by combining patients with multimodality intervention as well as surgery alone with patients who received radiotherapy alone. SRS for AVMs have been reported to be safe among unruptured AVMs treated in children and in adults.,, However, SRS may currently be reserved for unruptured AVMs of lower Spetzler–Martin grade and may not be recommended for unruptured grade IV/V AVMs. In patients with unruptured AVMs presenting with fixed neurological deficits or well-controlled seizures, it is important to discuss the management option of intervention versus follow-up with the patient. In such instances, management will need to be individualized based on patient preference and the experience of the treating neurosurgeon. A proposed radiosurgery-based algorithm for management of AVMs is summarized in [Figure 1].
Radiosurgery in childhood versus adult AVMs
Although children who undergo radiosurgery are more likely than adults to have a hemorrhagic presentation, bleeding rates during the latency period to obliteration have been reported to be lower among children, though not statistically significant. It has also been observed that AVMs take a shorter time to obliterate in children than in adults after SRS. Pan et al., observed that the obliteration rate of medium-sized AVMs was lower in children compared with adults. Although there is a concern regarding side effects secondary to the delivery of high dose of radiation to the developing brain, complication rates among children have been reported to be comparable to that seen in adults, if not lower. Hence, no attempt has been made to reduce the radiation dose in children vis-à-vis adults and radiosurgery treatment paradigms report the same dose range in children and adults.,, Attempts at lowering the marginal dose have been correlated with lower obliteration rates.
SRS has been proven to be effective in the management of ruptured as well as unruptured AVMs. A higher marginal dose of radiation is the most important factor in predicting AVM obliteration after SRS. Obliteration and complication rates reported in literature suggest that there is no difference in the efficacy and safety of different delivery systems. While it is most effective in the management of small AVMs, treatment paradigms for larger AVMs include multiple-session SRS or HSRT. Even after obliteration has been achieved, these patients need a long-term follow-up to determine the cognitive sequelae, delayed complications, and rarely, AVM recurrence.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3]