- Email: [email protected]
Linac-based Stereotactic Radiosurgery for Brain Arteriovenous Malformations
Clinical Oncology 23 (2011) 525e531 Contents lists available at ScienceDirect
Clinical Oncology journal homepage: www.elsevier.com/locate/clon
Original Article
Linac-based Stereotactic Radiosurgery for Brain Arteriovenous Malformations S. Blamek*, R. Tarnawski, L. Miszczyk Department of Radiotherapy, Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Gliwice, Poland Received 14 September 2010; received in revised form 11 January 2011; accepted 1 February 2011
Abstract Aims: Most papers dealing with radiosurgery for cerebral arteriovenous malformations (AVMs) present the results of gamma-knife treatment, whereas linac radiosurgery is becoming increasingly popular. Moreover, there is still much uncertainty about the rationale of combined endovascular and radiosurgical treatment. The aims of this study were to evaluate obliteration and rebleeding rates, and to determine factors influencing obliteration and adverse effects after linac-based stereotactic radiosurgery for cerebral AVMs. Materials and methods: Records of 62 consecutive patients were analysed. Thirty-one had partial embolisation, five surgery, 29 had no prior treatment. The mean follow-up was 28.4 months. The mean volume treated was 11.7 cm3 and included embolised portions of AVMs. Actuarial obliteration rates and annual bleeding hazard rates after radiosurgery were calculated using KaplaneMeier survival and life table analyses. Results: Actuarial obliteration rates after 1, 2 and 3 years of follow-up were 17, 36 and 40%, respectively. Prior embolisation did not influence the obliteration rate. In 77.3% of patients, obliteration occurred during the first 2 years of follow-up. Annual bleeding hazard rates after stereotactic radiosurgery were 3.4 and 1.1% during the first and second year of follow-up, respectively. Non-symptomatic imaging abnormalities were detected in 33.9% of patients after a median time of 8.8 months. The SpetzlereMartin grade, AVM score, radiation dose, volume and AVM nidus < 3 cm significantly influenced the probability of obliteration. A dose less than 15 Gy significantly reduced the probability of obliteration. Conclusion: At least a 3 year follow-up is required to accurately assess the outcome. The best effects of the treatment are achieved for small (<3 cm), low-grade lesions with a low AVM score. The bleeding risk after stereotactic radiosurgery gradually decreases. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Arteriovenous malformations (AVM); brain; obliteration; radiotherapy; stereotactic radiosurgery
Introduction Arteriovenous malformations (AVMs) of the brain can be the cause of epileptic seizures, neurological dysfunctions and intracranial haemorrhage. The annual bleeding risk from AVMs is estimated to be between 2 and 4% and the life risk of haemorrhage is usually calculated using the Brown formula: risk (%) ¼ 105 e age [1,2]. It is obvious that the cumulative risk of a life-threatening haemorrhage is substantial, especially in young patients and, thus, most AVMs require treatment that can eliminate or significantly minimise the risk of bleeding. The most effective treatment is microneurosurgery d total removal of the lesion guarantees cure. Stereotactic radiosurgery is a widely recognised Author for correspondence: S. Blamek, Department of Radiotherapy, Maria Sk1odowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, ul. Wybrzeze AK 15, 44-100 Gliwice, Poland. Tel: þ48322788052; Fax: þ48-322788001. E-mail address: [email protected] (S. Blamek).
method of treatment for AVMs of the brain that are not suitable for surgery. The reported percentage of obliteration is usually about 80e90% in selected cases of lesions smaller than 3 cm [3e6]. The results of stereotactic irradiation of larger AVMs are less encouraging and usually range between 43 and 70% [7e10]. Moreover, the influence of combined treatment involving endovascular procedures and stereotactic radiosurgery on treatment outcome is not fully elucidated. Available reports often contain conflicting data indicating either beneficial or deleterious effects of endovascular treatment on radiosurgery outcome [11e13]. Still more data are needed to establish reliable indications for embolisation before radiosurgery. Most papers dealing with stereotactic irradiation of brain AVMs report the results of gamma-knife radiosurgery. Linac-based radiosurgery for brain AVMs is less extensively described and still new data are needed to confirm the influence of selected clinical and treatment parameters on treatment outcome. The aim of the study was to evaluate the obliteration and rebleeding rates as well as the adverse
0936-6555/$36.00 Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.clon.2011.03.012
526
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
effects of linac-based stereotactic radiosurgery for brain AVMs of relatively large mean volume (11.7 cm3). We also attempted to identify clinical and treatment-related factors influencing obliteration rates after linac-based stereotactic radiosurgery.
Materials and Methods Patients The analysis was based on a group of 62 consecutive patients aged 8e76 years old (mean age 39.5 years), irradiated with a linac-based stereotactic technique for brain AVMs between October 2001 and December 2005. The proportion of males and females was almost equal d 32 and 30, respectively. Twenty-nine patients were referred to our institution without prior treatment, just after the diagnosis, whereas 33 patients were previously treated. Five patients were treated surgically. The surgical intervention was taken to evacuate the haematoma and ruptured malformation, but resulted in subtotal removal of AVM nidus. At least one embolisation before stereotactic radiosurgery was carried out in 31 patients. In most cases, staged embolisation was carried out; only 12 patients had a single procedure (Table 1). Intracranial haemorrhage was diagnosed in 31 (50%) patients before radiosurgery. The number of haemorrhages and other symptoms are summarised in Table 2. The mean and median volume treated was 11.7 and 6.64 cm3, respectively, and ranged between 0.14 and 63 cm3. In 13 (21%) patients, the volume of the nidus exceeded 14 cm3, which is the estimated volume of a spherical lesion of 3 cm diameter (Figure 1). All lesions were graded using the SpetzlereMartin grading system (Table 3). Grade VI (truly inoperable) was not used because its definition is relatively wide and AVMs of various size and volume can be classified as being grade VI lesions. Grade I lesions were irradiated only when the patient refused surgery or in the case of serious co-morbidities that made the risk of surgery unacceptable. The radiosurgery-based grading system proposed by Pollock and Flickinger [14] was also used and was calculated according to the following formula: AVM score ¼ 0.1 * AVM volume þ 0.02 * age þ 0.3 * AVM location. The parameter ‘AVM location’ has a value of 0e2 according to the site of the lesion [14]. Recent simplification of the formula did not
Table 2 Characteristics of clinical symptoms and frequency of intracranial haemorrhages
Intracranial haemorrhage 1 2 3 8 Symptom Headache Epileptic seizures Paresis Disturbed vision Vertigo Speech disturbances Tinnitus Vomiting Involuntary movements Transient ischaemic attacks
Number
Percentage
26 2 2 1
41.9 3.2 3.2 1.6
36 17 23 7 11 6 4 10 1 6
58.0 27.5 37.0 11.0 18.0 9.5 6.5 16.2 1.6 9.6
significantly change the potential of prediction of obliteration. Therefore, the classic formula was used to facilitate comparisons of our results with other reports [15]. There were nine lesions with an AVM score 1, 31 lesions with a score between 1 and 2 and 22 lesions with an AVM score above 2. Irradiation Technique All the patients were irradiated with a linear accelerator (Clinac 2300 C/D, Varian, Palo Alto, USA) equipped with a micro-multileaf collimator with a 3 mm centre leaf width at the isocentre (m3, BrainLab AG, Feldkirchen, Germany) with 6 MV photons. Patients were immobilised with thermoplastic masks with mouthbites used whenever possible to reduce positioning inaccuracies. Invasive fixation with head frames was not used. Radiosurgery plans were made with BrainLab software. Target definition was based on planning computed tomography, magnetic resonance
Table 1 Number of endovascular procedures carried out before stereotactic radiosurgery Number of embolisations
Number of patients
0 1 2 3 4 5
31 12 8 5 4 2
Fig. 1. Volume distribution of treated arteriovenous malformations.
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531 Table 3 Classification of cases according to the SpetzlereMartin grading system SpetzlereMartin grade
Number
I II III IV V
3 22 26 9 2
T1-weighted post-contrast, FLAIR and angio-magnetic resonance images; digital subtraction angiography (DSA) was not routinely used. A margin of 2 mm was added to account for intrafractional movement and positioning inaccuracies. The dose was specified at the isocentre and the whole nidus was defined as the target volume, including previously embolised portions. The prescribed dose varied according to the target volume and the proximity of organs at risk. In all cases, an attempt was made to follow ICRU rules for dose specification. However, in the case of AVMs involving or adjacent to organs at risk, such as the brain stem or optic chiasm, target coverage with a 90% isodose was accepted. The patients were irradiated with doses ranging between 8 and 28 Gy. The mean and median dose was 16 Gy. In 51 patients, a multifield conformal noncoplanar technique was applied with a micro-multileaf collimator used for field shaping. In 11 patients, intensity modulation was used due to the proximity of organs at risk or the complex shape of the nidus. Follow-up The mean and median follow-up were 28.4 and 26.8 months, respectively. About every 6 months the patients had control magnetic resonance imaging or computed tomography examinations, which also included magnetic resonance or computed tomography angiography. The first control examination was usually carried out 4e6 weeks after treatment; the next one was about 3 months after irradiation and then every 6 months. Once obliteration was diagnosed, control examinations were carried out yearly. Imaging examinations were carried out from the second control visit after treatment. Magnetic resonance or computed tomography angiography was used for the assessment of obliteration. DSA was not a standard tool for the assessment of obliteration in our protocol.
527
parameters that were significantly different and could influence the obliteration rate. KaplaneMeier survival analysis was used to establish the significance of selected parameters. All tests were carried out with the significance level a set to 0.05. All statistical calculations were carried out using Statistica 7.1 PL software.
Results AVMs were obliterated in 22 of 62 patients, which resulted in a crude obliteration rate of 35.5%. Actuarial obliteration rates after 1, 2 and 3 years of follow-up were 17, 36 and 40%, respectively. In 17 of 22 (77.3%) patients, obliteration was diagnosed during the first 2 years of follow-up (Figure 2). Nevertheless, five of 22 obliterations (22.7%) occurred after 2 years and three (13.6%) after 3 years of observation. The mean time to obliteration was 18.7 months (range 3e53 months). Among 13 patients with an AVM volume larger than 14 cm3 there was only one obliteration, whereas in the group of 49 patients with smaller AVMs, 21 (42.9%) were obliterated. The difference was statistically significant (P ¼ 0.0185, chi-square test). In the group of patients followed-up for at least 2 years, the crude obliteration rate was 43.2%. The crude obliteration rate of AVMs of maximum diameter <3 cm was 51.4%, 3e6 cm 15.6% and >6 cm 0%. Actuarial obliteration rates after 2 years of follow-up for AVMs grade IeIII and IVeV according to the SpetzlereMartin grading system were 38 and 10%, respectively (P ¼ 0.0076, F Cox test). Two of three grade I AVMs followed-up at least 1 year were obliterated within 2 years after treatment. Actuarial obliteration rates after 2 years for grade II, III and IV AVMs were 41, 30 and 12%, respectively, and after 3 years of follow-up 48, 37 and 12%, respectively. Two grade V AVMs did not obliterate after treatment. The obliteration rate of lesions with radiosurgery-based AVM score <1, 1e2 and >2 was 55.5, 48.3 and 9%, respectively. The total obliteration rate in the group with an AVM score 2 was 50% (20/40). The mean AVM
Statistics Actuarial obliteration rates were assessed using KaplaneMeier estimation. Annual hazard rates of haemorrhage were calculated using KaplaneMeier life tables. The patients were divided into two groups: patients with patent and obliterated AVMs. Chi-square, ManneWhitney U test and Student’s t-test for independent samples were used for comparisons according to variable type to identify the
Fig. 2. Obliteration kinetics in the group of 22 obliterated arteriovenous malformations.
528
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
score was 2.14 and within the group of patients with an AVM score below the mean the obliteration rate was 47.7% (21/44), whereas in the group with an AVM score equal or larger than the mean value it was only 5.5% (1/18). Three patients had intracranial haemorrhage after radiosurgery; 1, 5 and 46 months after treatment. In spite of bleeding, the neurological status of the patients did not change significantly and no new neurological deficits were diagnosed. All of them had bled before radiosurgery, in two patients haemorrhage was diagnosed 2 months before radiosurgery, in one 2 years before treatment. Two of them had partial embolisations before radiosurgery. The patients were irradiated with 15, 19 and 18 Gy, respectively. In two patients the lesion was in the central region of the brain (hypothalamus and corpus callosum), in one it involved the periventricular trigone and occipital lobe. In all patients AVMs drained to deep veins of the brain. In one case another attempt of embolisation was made after bleeding, but without success, and subsequently the patient was irradiated for the second time with 10 Gy in a single fraction. The annual bleeding hazard rates after stereotactic radiosurgery were 3.4% (standard error ¼ 2.4%) and 1.1% (standard error ¼ 1.5%) during the first and second year of follow-up, respectively. Because of a low incidence of bleeding after stereotactic radiosurgery, no reliable statistical analysis aiming to identify factors influencing the risk of bleeding could be carried out. The mean AVM score in the bleeders’ group was 2.76. whereas in non-bleeders it was 2.11; the difference was not statistically significant. Recanalisation was seen in one patient, 6 months after obliteration. The patient had two embolisations before stereotactic radiosurgery and was irradiated with a dose of 15 Gy. Revascularisation of the nidus was, however, diagnosed with magnetic resonance angiography as well as previous obliteration; neither obliteration nor recanalisation was confirmed by classic angiography. It is therefore possible that the obliteration could be misdiagnosed due to diagnostic inaccuracy of magnetic resonance angiography. Transient neurological deterioration was observed in two patients. They complained of headaches, dizziness and paraesthesiae, which resolved within 2 weeks after treatment. One patient worsened neurologically during follow-up, 9 months after treatment. The symptoms of hemiparesis and aphasia were, however, attributed to stroke involving the right frontoparietal region of the brain, which was confirmed with magnetic resonance imaging. The treated lesion was located in the cerebellum and irradiated with a dose of 12 Gy due to its large volume (63.07 cm3), but no cerebellar symptoms were diagnosed in spite of the presence of a T2 hyperintensity around the lesion on control imaging studies. Non-symptomatic imaging abnormalities were detected in 33.9% of patients after a median time of 8.8 months. Imaging abnormalities were defined as T1 hypointensity, T2 hyperintensity or regions of contrast enhancement around or adjacent to the nidus. Statistical analysis showed that SpetzlereMartin grade (P ¼ 0.0148), AVM score (P ¼ 0.0049), radiation dose (P ¼0.0107), minimum dose (P ¼ 0.0097), maximum dose (P ¼ 0.0189), AVM volume (P ¼ 0.0236) and maximum
diameter of AVM nidus <3 cm (P ¼ 0.0028) were significantly different in the group of patients with obliterated AVMs as compared with the remainder. KaplaneMeier estimation showed that application of a dose lower than 15 Gy significantly (P ¼ 0.0079) reduced the probability of obliteration (Figure 3). After 2 years of follow-up, the actuarial obliteration rate was 43% in the group irradiated with doses 15 Gy (46 patients) and only 17% in the group treated with lower doses. The history of embolisation or surgery did not influence the probability of obliteration. Occlusion of the nidus occurred in 10/31 patients after embolisation and in 9/31 patients who did not receive endovascular treatment (P ¼ 0.4855, chi-square test). Similarly, intracranial haemorrhage before treatment did not influence the probability of obliteration, AVMs obliterated in 11/31 bleeders and in 8/31 non-bleeders (P ¼ 0.4086, chi-square test).
Discussion Actuarial obliteration rates of 17, 36 and 40% after 1, 2 and 3 years of follow-up, respectively, are somewhat lower than usually reported in the literature and can be attributed to the relatively large mean volume of treated lesions, as well as to the moderate radiation doses applied. They are, however, comparable with obliteration rates reported by Zabel-du Bois et al. [16]. They reported a 50% actuarial 3 year obliteration rate in a group of 65 patients with a median time to obliteration of 22.4 months. The distribution of SpetzlereMartin grades was very similar: II 13 patients, III 39, IV 12 and V one, respectively [16]. Moreover, the mean volume of the nidus in their series was 5.2 cm3, which is about two times smaller than that reported here (11.7 cm3). The same team also reported obliteration rates of large AVMs treated with radiosurgery or hypofractionated radiotherapy. Actuarial 3 and 4 year obliteration rates after radiosurgery were 47 and 60%, respectively. The median target volume was 7.1 cm3, whereas in our series it was
Fig. 3. Obliteration rates in patients treated with doses <15 Gy (solid line) and 15 Gy (dashed line). The difference was statistically significant (P ¼ 0.0079).
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
6.64 cm3 [17]. It should be noted that the mean and median dose reported here was 16 Gy and ranged widely between 8 and 28 Gy, whereas in their series the median dose was 17 Gy and ranged between 15 and 19 Gy prescribed on the 80% (surrounding) isodose. The poorer results of treatment in our group could be the result of the administration of lower doses, especially on the periphery of the nidus, as we specified the dose at the isocentre. In our series, 22.7% of obliterations were diagnosed after 2 years of follow-up and 13.6% after 3 years of observation, which suggests that the latency period can be longer than the commonly accepted 2e3 years. It is especially true in the case of large AVMs. This observation is confirmed in several reports, indicating that the latency period can even last 4e5 years [9,10,17,18]. Touboul et al. [9] noticed that in a group of 100 patients, obliteration was diagnosed in 40% of patients after 3 years of observation, but after 5 years of follow-up it was 62%, which is a 50% rise after 2 additional years of follow-up [9]. Similar conclusions were drawn after analysis of the data reported by Chang and co-workers. They reported 43% obliterations 3 years, 72% 5 years and 78% 6 years after treatment [8]. These observations were also confirmed by Zabel-du Bois et al. The obliteration rate after 3 years of observation was 47% in their series, whereas after 4 years it rose to the level of 60% [17]. Bleeding after stereotactic radiosurgery was observed in only three patients in our series. The annual bleeding hazard rate of 3.4 and 1.1% in the first and second year of observation, respectively, is close to that reported in recent literature. The annual bleeding risk after radiosurgery reported by Zabel-du Bois and colleagues [16] was 4.7, 3.4 and 2.7% during the first, second and third year of observation, respectively. These values match our results and support the hypothesis that haemorrhage risk gradually decreases after treatment. The tendency of an annual bleeding hazard to decrease after stereotactic radiosurgery was also confirmed recently by Zabel-du Bois et al. [19] in the analysis of a group of patients irradiated after partial embolisation of the nidus. In that group, the annual bleeding risk was 7.9% after 1 year and dropped to 2.2% after the second year of follow-up [19]. According to other authors, the bleeding risk after stereotactic radiosurgery was usually estimated in the range of 0e3.6% [7,8,20]. Miyawaki and colleagues [20] found that the annual bleeding risk was 2.7% for AVMs with volumes less than 14 cm3, but for larger AVMs significantly increased to the value of 7.5% per annum. A large AVM volume was also reported as a risk factor of rupture by Hernesniemi et al. [21] in a recent study on the natural history of AVMs. The SpetzlereMartin grading system, which proved its usefulness in assessing the risk of AVM surgery, can also be a helpful tool in predicting radiosurgery outcome. Authors reporting the results of radiosurgery for brain AVMs have not always found the SpetzlereMartin grading system to be useful in predicting radiosurgery outcome [10]. However, our results of radiosurgical treatment of low-grade AVMs are significantly better than in the case of high-grade lesions. This observation was also confirmed in other reports [17,20,22,23].
529
Current analysis also supports the use of radiosurgerybased AVM score. The best results were obtained in patients with a low AVM score, whereas an AVM score of 2 and above predicts poor treatment outcome. According to Pollock and Flickinger [14], an AVM score of 2 gives a 50% chance of obliteration. Analysis of our material shows that in the group with an AVM score <2, the obliteration rate was 50% (20/40). If the AVM score exceeded a mean value of 2.14 calculated for the whole group of patients, the obliteration rate dropped to 9%. Similarly, in a group of 56 patients treated by Pollock and colleagues [24] for AVMs located in basal ganglia, thalamus or brainstem with a median AVM score of 1.83, 43% of lesions were obliterated after single irradiation and finally 57% after repeated procedures. The best results were obtained in the group with an AVM score <1.5 in which the obliteration rate was 67% [24]. AVM score was also identified to be useful in predicting the risk of intracranial haemorrhage after radiosurgery [16,25]. In our series, the incidence of bleeding after radiosurgery was low, which hindered reliable statistical analysis. Therefore, we could not show the influence of AVM score on bleeding risk as the mean AVM scores in non-bleeders and bleeders exceeded 2 and were 2.11 and 2.76, respectively. The analysis of the influence of previous embolisation on obliteration rates showed that irradiation results are similar irrespective of previous endovascular treatment. It stands in contrast with some reports suggesting that previous embolisation reduces the probability of obliteration. It can be explained by the fact that the whole nidus was our target volume, whereas most other groups defined the target volume according to angiographic appearance of the postembolisation nidus. This may lead to underdosage of previously embolised sections of the AVM and, consequently, to an increased risk of revascularisation [13,26]. Our policy of treating the whole AVM nidus is supported by recent reports indicating that previous embolisation reduces the obliteration rate if the embolised portion of the nidus is excluded from the target volume [27]. Endothelial cells of AVM vessels possess increased angiogenic activity and migration potential [28]. Moreover, it was shown that endothelial cells of AVM vessels may be stimulated by the process of embolisation and show even larger angiogenic activity as compared with untreated ones. In contrast, irradiated endothelial cells have the lowest angiogenic activity as compared with cells extracted from untreated or embolised nidi [29]. We suppose that irradiation of the whole nidus counterbalanced the proangiogenic effect of embolisation, which resulted in the same obliteration rates in patients with embolised and non-embolised AVMs in our series. KaplaneMeier analysis revealed that irradiation with a single dose below 15 Gy leads to a significant drop in the actuarial obliteration rate as compared with higher doses. According to Nataf et al. [25], irradiation with doses <15 Gy led to an obliteration rate of 44%, whereas irradiation with 15e20 Gy resulted in an obliteration rate of 89%. Miyawaki and colleagues [20] also noted that stereotactic radiosurgery doses 14 Gy resulted in a 9% obliteration rate, in contrast, doses of 18 Gy allowed for a 60% obliteration rate.
530
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
The tolerance of the treatment was excellent and no permanent neurological deficits were observed. Transient neurological symptoms in two patients can be attributed to acute effects of irradiation and resolved within 2 or 3 weeks. Imaging abnormalities after stereotactic radiosurgery were detected in 33.9% of cases [30]. This value is consistent with other reports estimating the incidence of imaging abnormalities at the level of about one-third up to about one-half of irradiated patients [31e36]. Limitations Estimation of blood flow through the nidus was made with computed tomography or magnetic resonance angiography. It may be considered as a drawback of the study because DSA is a commonly accepted ‘gold standard’ of AVM evaluation. However, DSA is an invasive procedure and can be associated with serious complications. For that reason it was not routinely used for AVM evaluation after treatment. Moreover, patients usually refused angiography proposed as a method of evaluation of treatment effect, preferring non-invasive computed tomography or magnetic resonance angiography. However, according to some reports, due to comparable results magnetic resonance or computed tomography angiography can be considered as a reliable alternative to DSA in assessing blood flow through the AVM nidus [37,38]. According to Pollock et al. [38], no patient with obliteration diagnosed with magnetic resonance angiography and not confirmed with DSA suffered from intracranial haemorrhage after stereotactic radiosurgery. Moreover, in 10 of 16 patients misdiagnosed with magnetic resonance angiography, angiography was repeated and eventually confirmed obliteration of the AVM in an additional seven patients [38]. For that reason, we decided not to include obligatory DSA examination as a method of confirmation of AVM obliteration in the follow-up protocol. Other factors influencing the quality of the study are the large heterogeneity of the group and moderate radiation doses resulting in a mean dose of 16 Gy, which may be considered too low to ensure an optimal treatment effect. The reasons for dose reduction were mostly large AVM volumes and the risk of adverse effects associated with the application of higher doses to large volumes of the brain and, finally, the proximity of dose-limiting organs at risk. However, the obliteration rate in the present study is comparable with results reported by other teams treating AVMs of similar volume. We suppose that the latency period after irradiation with moderate doses can be somewhat longer than after treatment with higher doses. The decreasing risk of haemorrhage indicates that our treatment influenced outcome in spite of a relatively low rate of radiologically confirmed obliteration. An excellent longterm study reported by Laakso et al. [39] showed that even partial but active treatment reduces mortality in patients with AVMs, which further supports our policy to treat even large AVMs or lesions close to organs at risk even at the cost of reduction of the dose. Moreover, partially obliterated AVMs, previously inoperable, may become suitable for surgery and previous irradiation can facilitate the surgical
procedure and reduce morbidity [40]. Longer follow-up will be crucial to a more reliable assessment of the influence of this treatment on bleeding risk, long-term obliteration rate and AVM-associated mortality.
Conclusions The large heterogeneity of the group, moderate doses of radiation used in some patients and the method of assessment of AVM obliteration are the disadvantages of the study that could influence our results. However, the current analysis shows that linac-based stereotactic radiosurgery proved to be a safe method of treatment for cerebral AVMs and within the range of applied doses no serious adverse effects were observed, even in the case of large AVMs of volumes exceeding 14 cm3. Adequate, at least 3 years, follow-up is required to accurately assess the results of stereotactic AVM irradiation. The best effects of the treatment can be expected for small (<3 cm), low-grade lesions with a low radiosurgery-based AVM score. The bleeding risk after stereotactic radiosurgery gradually decreases during follow-up. Inclusion of the obliterated portion of the AVM nidus into the target volume may prevent later recanalisation. This observation requires further studies because it may significantly influence the qualification criteria for endovascular procedures.
References [1] Brown Jr RD. Simple risk predictions for arteriovenous malformation hemorrhage. Neurosurgery 2000;46(4):1024. [2] Kondziolka D, McLaughlin MR, Kestle JRW. Simple risk predictions for arteriovenous malformation hemorrhage. Neurosurgery 1995;37:851e855. [3] Flickinger JC, Kondziolka D, Maitz AH, Lunsford LD. An analysis of the dose-response for arteriovenous malformation radiosurgery and other factors affecting obliteration. Radiother Oncol 2002;63:347e354. [4] Nicolato A, Lupidi F, Sandri MF, et al. Gamma knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part II: Differences in obliteration rates, treatmenteobliteration intervals, and prognostic factors. Int J Radiat Oncol Biol Phys 2006;64(3):914e921. [5] Sasaki T, Kurita H, Saito I, et al. Arteriovenous malformations in the basal ganglia and thalamus: management and results in 101 cases. J Neurosurg 1998;88:285e292. [6] Steiner L, Lindquist C, Adler JR, Torner JC, Alves W, Steiner M. Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 1992;77:1e8. [7] Bollet MA, Anxionnat R, Buchheit I, et al. Efficacy and morbidity of arc-therapy radiosurgery for cerebral arteriovenous malformations: a comparison with the natural history. Int J Radiat Oncol Biol Phys 2004;58(5):1353e1363. [8] Chang T-C, Shirato H, Aoyama H, et al. Stereotactic irradiation for intracranial arteriovenous malformation using stereotactic radiosurgery or hypofractionated stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2004;60(3):861e870. [9] Touboul E, Al Halabi A, Buffat L, et al. Single-fraction stereotactic radiotherapy: a dose-response analysis of arteriovenous malformation obliteration. Int J Radiat Oncol Biol Phys 1998;41(4):855e861.
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531 [10] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Pediatric cerebral arteriovenous malformations: the role of stereotactic linac-based radiosurgery. Int J Radiat Oncol Biol Phys 2006;65(4):1206e1211. [11] Henkes H, Nahser H-C, Berg-Dammer E, Weber W, Lange S, € hne D. Endovascular therapy of brain AVMs prior to radiKu osurgery. Neurol Res 1998;20:479e492. [12] Hurst RW, Berenstein A, Kupersmith MJ, Madrid M, Flamm ES. Deep central arteriovenous malformations of the brain: the role of endovascular treatment. J Neurosurg 1995;82:190e195. [13] Schlienger M, Atlan D, Lefkopoulos D, et al. Linac radiosurgery for cerebral arteriovenous malformations: results in 169 patients. Int J Radiat Oncol Biol Phys 2000;46(5):1135e1142. [14] Pollock BE, Flickinger JC. A proposed radiosurgery-based grading system for arteriovenous malformations. J Neurosurg 2002;96:79e85. [15] Pollock BE, Flickinger JC. Modification of the radiosurgerybased arteriovenous malformation grading system. Neurosurgery 2008;63(2):239e243. [16] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Stereotactic linac-based radiosurgery in the treatment of cerebral arteriovenous malformations located deep, involving corpus callosum, motor cortex or brainstem. Int J Radiat Oncol Biol Phys 2006;64(4):1044e1048. [17] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Linac-based radiosurgery or hypofractionated stereotactic radiotherapy in the treatment of large cerebral arteriovenous malformations. Int J Radiat Oncol Biol Phys 2006;64(4): 1049e1054. [18] Yamamoto M, Jimbo M, Kobayashi M, et al. Long-term results of radiosurgery for arteriovenous malformation: neuroradiodiagnostic imaging and histological studies of angiographically confirmed nidus obliteration. Surg Neurol 1992;37: 219e230. [19] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Risk of hemorrhage and obliteration rates of LINAC-based radiosurgery for cerebral arteriovenous malformations treated after prior partial embolization. Int J Radiat Oncol Biol Phys 2007;68(4):999e1003. [20] Miyawaki L, Dowd C, Wara W, et al. Five year results of LINAC radiosurgery for arteriovenous malformations: outcome for large AVMs. Int J Radiat Oncol Biol Phys 1999;44(5):1089e1106. €€ € M, [21] Hernesniemi JA, Dashti R, Juvela S, Va art K, Niemela Laakso A. Natural history of brain arteriovenous malformations: a long-term follow-up study of risk of hemorrhage in 238 patients. Neurosurgery 2008;63(5):823e829. [22] Ellis TL, Friedman WA, Bova FJ, Kubilis PS, Buatti JM. Analysis of treatment failure after radiosurgery for arteriovenous malformations. J Neurosurg 1998;89:104e110. [23] Meder JF, Oppenheim C, Blustajn J, et al. Cerebral arteriovenous malformations: the value of radiologic parameters in predicting response to radiosurgery. AJNR Am J Neuroradiol 1997;18:1473e1483. [24] Pollock BE, Gorman DA, Brown PD. Radiosurgery for arteriovenous malformations of the basal ganglia, thalamus and brainstem. J Neurosurg 2004;100:210e214. [25] Nataf F, Schlienger M, Lefkopoulos D, et al. Radiosurgery of cerebral arteriovenous malformations children: a series of 57 cases. Int J Radiat Oncol Biol Phys 2003;57(1):184e195.
531
[26] Gobin YP, Laurent A, Merienne L, et al. Treatment of brain arteriovenous malformations by embolization and radiosurgery. J Neurosurg 1996;85(1):19e28. [27] Andrade-Souza YM, Ramani M, Scora D, Tsao terBrugge MN, terBrugge K, Schwartz ML. Embolization before radiosurgery reduces the obliteration rate of arteriovenous malformations. Neurosurgery 2007;60(3):443e451. [28] Jabbour MN, Elder JB, Samuelson CG, et al. Aberrant angiogenic characteristics of human brain arteriovenous malformation endothelial cells. Neurosurgery 2009;64(1):139e146. [29] Akakin A, Ozkan A, Akgun E, et al. Endovascular treatment increases but gamma knife radiosurgery decreases angiogenic activity of arteriovenous malformations: an in vivo experimental study using a rat cornea model. Neurosurgery 2010;66(1):121e129. [30] Blamek S, Boba M, Larysz D, et al. The incidence of imaging abnormalities after stereotactic radiosurgery for cerebral arteriovenous and cavernous malformations. Acta Neurochir Suppl 2010;106:187e190. [31] Flickinger JC, Kondziolka D, Kalend AM, et al. Radiosurgeryrelated imaging changes in surrounding brain: multivariate analysis and model evaluation. Radiosurgery 1996;1:229e236. [32] Flickinger JC, Kondziolka D, Lunsford LD, et al. A multi-institutional analysis of complication outcomes after arteriovenous malformation radiosurgery. Int J Radiat Oncol Biol Phys 1999;44(1):67e74. [33] Flickinger JC, Kondziolka D, Maitz AH, Lunsford LD. Analysis of neurological sequelae from radiosurgery of arteriovenous malformations: how location affects outcome. Int J Radiat Oncol Biol Phys 1998;40(2):273e278. [34] Flickinger JC, Kondziolka D, Pollock BE, Maitz AH, Lunsford LD. Complications from arteriovenous malformation radiosurgery: multivariate analysis and risk modeling. Int J Radiat Oncol Biol Phys 1997;38(3):485e490. [35] Tomura N, Izumi J, Sakuma I, et al. Radiation-induced changes in the brain following stereotactic irradiation evaluated by sequential MRI. Comp Med Imag Graph 2004;28:73e79. [36] van den Berg R, Buis DR, Lagerwaard FJ, Lycklama A,
[37]
[38]
[39]
[40]
Nijeholt GJ, Vandertop WP. Extensive white matter changes after stereotactic radiosurgery for brain arteriovenous malformations: a prognostic sign for obliteration? Neurosurgery 2008;63(6):1064e1069. Guo WY, Pan HC, Chung WY, Wang LW, Teng MMH. Do we need conventional angiography? The role of magnetic resonance imaging in verifying obliteration of arteriovenous malformations after gamma knife surgery. Stereotact Funct Neurosurg 1996;66(Suppl. 1):71e84. Pollock BE, Kondziolka D, Flickinger JC, Patel AK, Bissonette DJ, Lunsford LD. Magnetic resonance imaging: an accurate method to evaluate arteriovenous malformations after stereotactic radiosurgery. J Neurosurg 1996;85: 1044e1049. €nen J, et al. Long-term excess Laakso A, Dashti R, Seppa mortality in 623 patients with brain arteriovenous malformations. Neurosurgery 2008;63(2):244e253. Sanchez-Mejia RO, McDermott MW, Tan J, Kim H, Young WL, Lawton MT. Radiosurgery facilitates resection of brain arteriovenous malformations and reduces surgical morbidity. Neurosurgery 2009;64(2):231e238.
Clinical Oncology journal homepage: www.elsevier.com/locate/clon
Original Article
Linac-based Stereotactic Radiosurgery for Brain Arteriovenous Malformations S. Blamek*, R. Tarnawski, L. Miszczyk Department of Radiotherapy, Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Gliwice, Poland Received 14 September 2010; received in revised form 11 January 2011; accepted 1 February 2011
Abstract Aims: Most papers dealing with radiosurgery for cerebral arteriovenous malformations (AVMs) present the results of gamma-knife treatment, whereas linac radiosurgery is becoming increasingly popular. Moreover, there is still much uncertainty about the rationale of combined endovascular and radiosurgical treatment. The aims of this study were to evaluate obliteration and rebleeding rates, and to determine factors influencing obliteration and adverse effects after linac-based stereotactic radiosurgery for cerebral AVMs. Materials and methods: Records of 62 consecutive patients were analysed. Thirty-one had partial embolisation, five surgery, 29 had no prior treatment. The mean follow-up was 28.4 months. The mean volume treated was 11.7 cm3 and included embolised portions of AVMs. Actuarial obliteration rates and annual bleeding hazard rates after radiosurgery were calculated using KaplaneMeier survival and life table analyses. Results: Actuarial obliteration rates after 1, 2 and 3 years of follow-up were 17, 36 and 40%, respectively. Prior embolisation did not influence the obliteration rate. In 77.3% of patients, obliteration occurred during the first 2 years of follow-up. Annual bleeding hazard rates after stereotactic radiosurgery were 3.4 and 1.1% during the first and second year of follow-up, respectively. Non-symptomatic imaging abnormalities were detected in 33.9% of patients after a median time of 8.8 months. The SpetzlereMartin grade, AVM score, radiation dose, volume and AVM nidus < 3 cm significantly influenced the probability of obliteration. A dose less than 15 Gy significantly reduced the probability of obliteration. Conclusion: At least a 3 year follow-up is required to accurately assess the outcome. The best effects of the treatment are achieved for small (<3 cm), low-grade lesions with a low AVM score. The bleeding risk after stereotactic radiosurgery gradually decreases. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Arteriovenous malformations (AVM); brain; obliteration; radiotherapy; stereotactic radiosurgery
Introduction Arteriovenous malformations (AVMs) of the brain can be the cause of epileptic seizures, neurological dysfunctions and intracranial haemorrhage. The annual bleeding risk from AVMs is estimated to be between 2 and 4% and the life risk of haemorrhage is usually calculated using the Brown formula: risk (%) ¼ 105 e age [1,2]. It is obvious that the cumulative risk of a life-threatening haemorrhage is substantial, especially in young patients and, thus, most AVMs require treatment that can eliminate or significantly minimise the risk of bleeding. The most effective treatment is microneurosurgery d total removal of the lesion guarantees cure. Stereotactic radiosurgery is a widely recognised Author for correspondence: S. Blamek, Department of Radiotherapy, Maria Sk1odowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, ul. Wybrzeze AK 15, 44-100 Gliwice, Poland. Tel: þ48322788052; Fax: þ48-322788001. E-mail address: [email protected] (S. Blamek).
method of treatment for AVMs of the brain that are not suitable for surgery. The reported percentage of obliteration is usually about 80e90% in selected cases of lesions smaller than 3 cm [3e6]. The results of stereotactic irradiation of larger AVMs are less encouraging and usually range between 43 and 70% [7e10]. Moreover, the influence of combined treatment involving endovascular procedures and stereotactic radiosurgery on treatment outcome is not fully elucidated. Available reports often contain conflicting data indicating either beneficial or deleterious effects of endovascular treatment on radiosurgery outcome [11e13]. Still more data are needed to establish reliable indications for embolisation before radiosurgery. Most papers dealing with stereotactic irradiation of brain AVMs report the results of gamma-knife radiosurgery. Linac-based radiosurgery for brain AVMs is less extensively described and still new data are needed to confirm the influence of selected clinical and treatment parameters on treatment outcome. The aim of the study was to evaluate the obliteration and rebleeding rates as well as the adverse
0936-6555/$36.00 Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.clon.2011.03.012
526
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
effects of linac-based stereotactic radiosurgery for brain AVMs of relatively large mean volume (11.7 cm3). We also attempted to identify clinical and treatment-related factors influencing obliteration rates after linac-based stereotactic radiosurgery.
Materials and Methods Patients The analysis was based on a group of 62 consecutive patients aged 8e76 years old (mean age 39.5 years), irradiated with a linac-based stereotactic technique for brain AVMs between October 2001 and December 2005. The proportion of males and females was almost equal d 32 and 30, respectively. Twenty-nine patients were referred to our institution without prior treatment, just after the diagnosis, whereas 33 patients were previously treated. Five patients were treated surgically. The surgical intervention was taken to evacuate the haematoma and ruptured malformation, but resulted in subtotal removal of AVM nidus. At least one embolisation before stereotactic radiosurgery was carried out in 31 patients. In most cases, staged embolisation was carried out; only 12 patients had a single procedure (Table 1). Intracranial haemorrhage was diagnosed in 31 (50%) patients before radiosurgery. The number of haemorrhages and other symptoms are summarised in Table 2. The mean and median volume treated was 11.7 and 6.64 cm3, respectively, and ranged between 0.14 and 63 cm3. In 13 (21%) patients, the volume of the nidus exceeded 14 cm3, which is the estimated volume of a spherical lesion of 3 cm diameter (Figure 1). All lesions were graded using the SpetzlereMartin grading system (Table 3). Grade VI (truly inoperable) was not used because its definition is relatively wide and AVMs of various size and volume can be classified as being grade VI lesions. Grade I lesions were irradiated only when the patient refused surgery or in the case of serious co-morbidities that made the risk of surgery unacceptable. The radiosurgery-based grading system proposed by Pollock and Flickinger [14] was also used and was calculated according to the following formula: AVM score ¼ 0.1 * AVM volume þ 0.02 * age þ 0.3 * AVM location. The parameter ‘AVM location’ has a value of 0e2 according to the site of the lesion [14]. Recent simplification of the formula did not
Table 2 Characteristics of clinical symptoms and frequency of intracranial haemorrhages
Intracranial haemorrhage 1 2 3 8 Symptom Headache Epileptic seizures Paresis Disturbed vision Vertigo Speech disturbances Tinnitus Vomiting Involuntary movements Transient ischaemic attacks
Number
Percentage
26 2 2 1
41.9 3.2 3.2 1.6
36 17 23 7 11 6 4 10 1 6
58.0 27.5 37.0 11.0 18.0 9.5 6.5 16.2 1.6 9.6
significantly change the potential of prediction of obliteration. Therefore, the classic formula was used to facilitate comparisons of our results with other reports [15]. There were nine lesions with an AVM score 1, 31 lesions with a score between 1 and 2 and 22 lesions with an AVM score above 2. Irradiation Technique All the patients were irradiated with a linear accelerator (Clinac 2300 C/D, Varian, Palo Alto, USA) equipped with a micro-multileaf collimator with a 3 mm centre leaf width at the isocentre (m3, BrainLab AG, Feldkirchen, Germany) with 6 MV photons. Patients were immobilised with thermoplastic masks with mouthbites used whenever possible to reduce positioning inaccuracies. Invasive fixation with head frames was not used. Radiosurgery plans were made with BrainLab software. Target definition was based on planning computed tomography, magnetic resonance
Table 1 Number of endovascular procedures carried out before stereotactic radiosurgery Number of embolisations
Number of patients
0 1 2 3 4 5
31 12 8 5 4 2
Fig. 1. Volume distribution of treated arteriovenous malformations.
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531 Table 3 Classification of cases according to the SpetzlereMartin grading system SpetzlereMartin grade
Number
I II III IV V
3 22 26 9 2
T1-weighted post-contrast, FLAIR and angio-magnetic resonance images; digital subtraction angiography (DSA) was not routinely used. A margin of 2 mm was added to account for intrafractional movement and positioning inaccuracies. The dose was specified at the isocentre and the whole nidus was defined as the target volume, including previously embolised portions. The prescribed dose varied according to the target volume and the proximity of organs at risk. In all cases, an attempt was made to follow ICRU rules for dose specification. However, in the case of AVMs involving or adjacent to organs at risk, such as the brain stem or optic chiasm, target coverage with a 90% isodose was accepted. The patients were irradiated with doses ranging between 8 and 28 Gy. The mean and median dose was 16 Gy. In 51 patients, a multifield conformal noncoplanar technique was applied with a micro-multileaf collimator used for field shaping. In 11 patients, intensity modulation was used due to the proximity of organs at risk or the complex shape of the nidus. Follow-up The mean and median follow-up were 28.4 and 26.8 months, respectively. About every 6 months the patients had control magnetic resonance imaging or computed tomography examinations, which also included magnetic resonance or computed tomography angiography. The first control examination was usually carried out 4e6 weeks after treatment; the next one was about 3 months after irradiation and then every 6 months. Once obliteration was diagnosed, control examinations were carried out yearly. Imaging examinations were carried out from the second control visit after treatment. Magnetic resonance or computed tomography angiography was used for the assessment of obliteration. DSA was not a standard tool for the assessment of obliteration in our protocol.
527
parameters that were significantly different and could influence the obliteration rate. KaplaneMeier survival analysis was used to establish the significance of selected parameters. All tests were carried out with the significance level a set to 0.05. All statistical calculations were carried out using Statistica 7.1 PL software.
Results AVMs were obliterated in 22 of 62 patients, which resulted in a crude obliteration rate of 35.5%. Actuarial obliteration rates after 1, 2 and 3 years of follow-up were 17, 36 and 40%, respectively. In 17 of 22 (77.3%) patients, obliteration was diagnosed during the first 2 years of follow-up (Figure 2). Nevertheless, five of 22 obliterations (22.7%) occurred after 2 years and three (13.6%) after 3 years of observation. The mean time to obliteration was 18.7 months (range 3e53 months). Among 13 patients with an AVM volume larger than 14 cm3 there was only one obliteration, whereas in the group of 49 patients with smaller AVMs, 21 (42.9%) were obliterated. The difference was statistically significant (P ¼ 0.0185, chi-square test). In the group of patients followed-up for at least 2 years, the crude obliteration rate was 43.2%. The crude obliteration rate of AVMs of maximum diameter <3 cm was 51.4%, 3e6 cm 15.6% and >6 cm 0%. Actuarial obliteration rates after 2 years of follow-up for AVMs grade IeIII and IVeV according to the SpetzlereMartin grading system were 38 and 10%, respectively (P ¼ 0.0076, F Cox test). Two of three grade I AVMs followed-up at least 1 year were obliterated within 2 years after treatment. Actuarial obliteration rates after 2 years for grade II, III and IV AVMs were 41, 30 and 12%, respectively, and after 3 years of follow-up 48, 37 and 12%, respectively. Two grade V AVMs did not obliterate after treatment. The obliteration rate of lesions with radiosurgery-based AVM score <1, 1e2 and >2 was 55.5, 48.3 and 9%, respectively. The total obliteration rate in the group with an AVM score 2 was 50% (20/40). The mean AVM
Statistics Actuarial obliteration rates were assessed using KaplaneMeier estimation. Annual hazard rates of haemorrhage were calculated using KaplaneMeier life tables. The patients were divided into two groups: patients with patent and obliterated AVMs. Chi-square, ManneWhitney U test and Student’s t-test for independent samples were used for comparisons according to variable type to identify the
Fig. 2. Obliteration kinetics in the group of 22 obliterated arteriovenous malformations.
528
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
score was 2.14 and within the group of patients with an AVM score below the mean the obliteration rate was 47.7% (21/44), whereas in the group with an AVM score equal or larger than the mean value it was only 5.5% (1/18). Three patients had intracranial haemorrhage after radiosurgery; 1, 5 and 46 months after treatment. In spite of bleeding, the neurological status of the patients did not change significantly and no new neurological deficits were diagnosed. All of them had bled before radiosurgery, in two patients haemorrhage was diagnosed 2 months before radiosurgery, in one 2 years before treatment. Two of them had partial embolisations before radiosurgery. The patients were irradiated with 15, 19 and 18 Gy, respectively. In two patients the lesion was in the central region of the brain (hypothalamus and corpus callosum), in one it involved the periventricular trigone and occipital lobe. In all patients AVMs drained to deep veins of the brain. In one case another attempt of embolisation was made after bleeding, but without success, and subsequently the patient was irradiated for the second time with 10 Gy in a single fraction. The annual bleeding hazard rates after stereotactic radiosurgery were 3.4% (standard error ¼ 2.4%) and 1.1% (standard error ¼ 1.5%) during the first and second year of follow-up, respectively. Because of a low incidence of bleeding after stereotactic radiosurgery, no reliable statistical analysis aiming to identify factors influencing the risk of bleeding could be carried out. The mean AVM score in the bleeders’ group was 2.76. whereas in non-bleeders it was 2.11; the difference was not statistically significant. Recanalisation was seen in one patient, 6 months after obliteration. The patient had two embolisations before stereotactic radiosurgery and was irradiated with a dose of 15 Gy. Revascularisation of the nidus was, however, diagnosed with magnetic resonance angiography as well as previous obliteration; neither obliteration nor recanalisation was confirmed by classic angiography. It is therefore possible that the obliteration could be misdiagnosed due to diagnostic inaccuracy of magnetic resonance angiography. Transient neurological deterioration was observed in two patients. They complained of headaches, dizziness and paraesthesiae, which resolved within 2 weeks after treatment. One patient worsened neurologically during follow-up, 9 months after treatment. The symptoms of hemiparesis and aphasia were, however, attributed to stroke involving the right frontoparietal region of the brain, which was confirmed with magnetic resonance imaging. The treated lesion was located in the cerebellum and irradiated with a dose of 12 Gy due to its large volume (63.07 cm3), but no cerebellar symptoms were diagnosed in spite of the presence of a T2 hyperintensity around the lesion on control imaging studies. Non-symptomatic imaging abnormalities were detected in 33.9% of patients after a median time of 8.8 months. Imaging abnormalities were defined as T1 hypointensity, T2 hyperintensity or regions of contrast enhancement around or adjacent to the nidus. Statistical analysis showed that SpetzlereMartin grade (P ¼ 0.0148), AVM score (P ¼ 0.0049), radiation dose (P ¼0.0107), minimum dose (P ¼ 0.0097), maximum dose (P ¼ 0.0189), AVM volume (P ¼ 0.0236) and maximum
diameter of AVM nidus <3 cm (P ¼ 0.0028) were significantly different in the group of patients with obliterated AVMs as compared with the remainder. KaplaneMeier estimation showed that application of a dose lower than 15 Gy significantly (P ¼ 0.0079) reduced the probability of obliteration (Figure 3). After 2 years of follow-up, the actuarial obliteration rate was 43% in the group irradiated with doses 15 Gy (46 patients) and only 17% in the group treated with lower doses. The history of embolisation or surgery did not influence the probability of obliteration. Occlusion of the nidus occurred in 10/31 patients after embolisation and in 9/31 patients who did not receive endovascular treatment (P ¼ 0.4855, chi-square test). Similarly, intracranial haemorrhage before treatment did not influence the probability of obliteration, AVMs obliterated in 11/31 bleeders and in 8/31 non-bleeders (P ¼ 0.4086, chi-square test).
Discussion Actuarial obliteration rates of 17, 36 and 40% after 1, 2 and 3 years of follow-up, respectively, are somewhat lower than usually reported in the literature and can be attributed to the relatively large mean volume of treated lesions, as well as to the moderate radiation doses applied. They are, however, comparable with obliteration rates reported by Zabel-du Bois et al. [16]. They reported a 50% actuarial 3 year obliteration rate in a group of 65 patients with a median time to obliteration of 22.4 months. The distribution of SpetzlereMartin grades was very similar: II 13 patients, III 39, IV 12 and V one, respectively [16]. Moreover, the mean volume of the nidus in their series was 5.2 cm3, which is about two times smaller than that reported here (11.7 cm3). The same team also reported obliteration rates of large AVMs treated with radiosurgery or hypofractionated radiotherapy. Actuarial 3 and 4 year obliteration rates after radiosurgery were 47 and 60%, respectively. The median target volume was 7.1 cm3, whereas in our series it was
Fig. 3. Obliteration rates in patients treated with doses <15 Gy (solid line) and 15 Gy (dashed line). The difference was statistically significant (P ¼ 0.0079).
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
6.64 cm3 [17]. It should be noted that the mean and median dose reported here was 16 Gy and ranged widely between 8 and 28 Gy, whereas in their series the median dose was 17 Gy and ranged between 15 and 19 Gy prescribed on the 80% (surrounding) isodose. The poorer results of treatment in our group could be the result of the administration of lower doses, especially on the periphery of the nidus, as we specified the dose at the isocentre. In our series, 22.7% of obliterations were diagnosed after 2 years of follow-up and 13.6% after 3 years of observation, which suggests that the latency period can be longer than the commonly accepted 2e3 years. It is especially true in the case of large AVMs. This observation is confirmed in several reports, indicating that the latency period can even last 4e5 years [9,10,17,18]. Touboul et al. [9] noticed that in a group of 100 patients, obliteration was diagnosed in 40% of patients after 3 years of observation, but after 5 years of follow-up it was 62%, which is a 50% rise after 2 additional years of follow-up [9]. Similar conclusions were drawn after analysis of the data reported by Chang and co-workers. They reported 43% obliterations 3 years, 72% 5 years and 78% 6 years after treatment [8]. These observations were also confirmed by Zabel-du Bois et al. The obliteration rate after 3 years of observation was 47% in their series, whereas after 4 years it rose to the level of 60% [17]. Bleeding after stereotactic radiosurgery was observed in only three patients in our series. The annual bleeding hazard rate of 3.4 and 1.1% in the first and second year of observation, respectively, is close to that reported in recent literature. The annual bleeding risk after radiosurgery reported by Zabel-du Bois and colleagues [16] was 4.7, 3.4 and 2.7% during the first, second and third year of observation, respectively. These values match our results and support the hypothesis that haemorrhage risk gradually decreases after treatment. The tendency of an annual bleeding hazard to decrease after stereotactic radiosurgery was also confirmed recently by Zabel-du Bois et al. [19] in the analysis of a group of patients irradiated after partial embolisation of the nidus. In that group, the annual bleeding risk was 7.9% after 1 year and dropped to 2.2% after the second year of follow-up [19]. According to other authors, the bleeding risk after stereotactic radiosurgery was usually estimated in the range of 0e3.6% [7,8,20]. Miyawaki and colleagues [20] found that the annual bleeding risk was 2.7% for AVMs with volumes less than 14 cm3, but for larger AVMs significantly increased to the value of 7.5% per annum. A large AVM volume was also reported as a risk factor of rupture by Hernesniemi et al. [21] in a recent study on the natural history of AVMs. The SpetzlereMartin grading system, which proved its usefulness in assessing the risk of AVM surgery, can also be a helpful tool in predicting radiosurgery outcome. Authors reporting the results of radiosurgery for brain AVMs have not always found the SpetzlereMartin grading system to be useful in predicting radiosurgery outcome [10]. However, our results of radiosurgical treatment of low-grade AVMs are significantly better than in the case of high-grade lesions. This observation was also confirmed in other reports [17,20,22,23].
529
Current analysis also supports the use of radiosurgerybased AVM score. The best results were obtained in patients with a low AVM score, whereas an AVM score of 2 and above predicts poor treatment outcome. According to Pollock and Flickinger [14], an AVM score of 2 gives a 50% chance of obliteration. Analysis of our material shows that in the group with an AVM score <2, the obliteration rate was 50% (20/40). If the AVM score exceeded a mean value of 2.14 calculated for the whole group of patients, the obliteration rate dropped to 9%. Similarly, in a group of 56 patients treated by Pollock and colleagues [24] for AVMs located in basal ganglia, thalamus or brainstem with a median AVM score of 1.83, 43% of lesions were obliterated after single irradiation and finally 57% after repeated procedures. The best results were obtained in the group with an AVM score <1.5 in which the obliteration rate was 67% [24]. AVM score was also identified to be useful in predicting the risk of intracranial haemorrhage after radiosurgery [16,25]. In our series, the incidence of bleeding after radiosurgery was low, which hindered reliable statistical analysis. Therefore, we could not show the influence of AVM score on bleeding risk as the mean AVM scores in non-bleeders and bleeders exceeded 2 and were 2.11 and 2.76, respectively. The analysis of the influence of previous embolisation on obliteration rates showed that irradiation results are similar irrespective of previous endovascular treatment. It stands in contrast with some reports suggesting that previous embolisation reduces the probability of obliteration. It can be explained by the fact that the whole nidus was our target volume, whereas most other groups defined the target volume according to angiographic appearance of the postembolisation nidus. This may lead to underdosage of previously embolised sections of the AVM and, consequently, to an increased risk of revascularisation [13,26]. Our policy of treating the whole AVM nidus is supported by recent reports indicating that previous embolisation reduces the obliteration rate if the embolised portion of the nidus is excluded from the target volume [27]. Endothelial cells of AVM vessels possess increased angiogenic activity and migration potential [28]. Moreover, it was shown that endothelial cells of AVM vessels may be stimulated by the process of embolisation and show even larger angiogenic activity as compared with untreated ones. In contrast, irradiated endothelial cells have the lowest angiogenic activity as compared with cells extracted from untreated or embolised nidi [29]. We suppose that irradiation of the whole nidus counterbalanced the proangiogenic effect of embolisation, which resulted in the same obliteration rates in patients with embolised and non-embolised AVMs in our series. KaplaneMeier analysis revealed that irradiation with a single dose below 15 Gy leads to a significant drop in the actuarial obliteration rate as compared with higher doses. According to Nataf et al. [25], irradiation with doses <15 Gy led to an obliteration rate of 44%, whereas irradiation with 15e20 Gy resulted in an obliteration rate of 89%. Miyawaki and colleagues [20] also noted that stereotactic radiosurgery doses 14 Gy resulted in a 9% obliteration rate, in contrast, doses of 18 Gy allowed for a 60% obliteration rate.
530
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531
The tolerance of the treatment was excellent and no permanent neurological deficits were observed. Transient neurological symptoms in two patients can be attributed to acute effects of irradiation and resolved within 2 or 3 weeks. Imaging abnormalities after stereotactic radiosurgery were detected in 33.9% of cases [30]. This value is consistent with other reports estimating the incidence of imaging abnormalities at the level of about one-third up to about one-half of irradiated patients [31e36]. Limitations Estimation of blood flow through the nidus was made with computed tomography or magnetic resonance angiography. It may be considered as a drawback of the study because DSA is a commonly accepted ‘gold standard’ of AVM evaluation. However, DSA is an invasive procedure and can be associated with serious complications. For that reason it was not routinely used for AVM evaluation after treatment. Moreover, patients usually refused angiography proposed as a method of evaluation of treatment effect, preferring non-invasive computed tomography or magnetic resonance angiography. However, according to some reports, due to comparable results magnetic resonance or computed tomography angiography can be considered as a reliable alternative to DSA in assessing blood flow through the AVM nidus [37,38]. According to Pollock et al. [38], no patient with obliteration diagnosed with magnetic resonance angiography and not confirmed with DSA suffered from intracranial haemorrhage after stereotactic radiosurgery. Moreover, in 10 of 16 patients misdiagnosed with magnetic resonance angiography, angiography was repeated and eventually confirmed obliteration of the AVM in an additional seven patients [38]. For that reason, we decided not to include obligatory DSA examination as a method of confirmation of AVM obliteration in the follow-up protocol. Other factors influencing the quality of the study are the large heterogeneity of the group and moderate radiation doses resulting in a mean dose of 16 Gy, which may be considered too low to ensure an optimal treatment effect. The reasons for dose reduction were mostly large AVM volumes and the risk of adverse effects associated with the application of higher doses to large volumes of the brain and, finally, the proximity of dose-limiting organs at risk. However, the obliteration rate in the present study is comparable with results reported by other teams treating AVMs of similar volume. We suppose that the latency period after irradiation with moderate doses can be somewhat longer than after treatment with higher doses. The decreasing risk of haemorrhage indicates that our treatment influenced outcome in spite of a relatively low rate of radiologically confirmed obliteration. An excellent longterm study reported by Laakso et al. [39] showed that even partial but active treatment reduces mortality in patients with AVMs, which further supports our policy to treat even large AVMs or lesions close to organs at risk even at the cost of reduction of the dose. Moreover, partially obliterated AVMs, previously inoperable, may become suitable for surgery and previous irradiation can facilitate the surgical
procedure and reduce morbidity [40]. Longer follow-up will be crucial to a more reliable assessment of the influence of this treatment on bleeding risk, long-term obliteration rate and AVM-associated mortality.
Conclusions The large heterogeneity of the group, moderate doses of radiation used in some patients and the method of assessment of AVM obliteration are the disadvantages of the study that could influence our results. However, the current analysis shows that linac-based stereotactic radiosurgery proved to be a safe method of treatment for cerebral AVMs and within the range of applied doses no serious adverse effects were observed, even in the case of large AVMs of volumes exceeding 14 cm3. Adequate, at least 3 years, follow-up is required to accurately assess the results of stereotactic AVM irradiation. The best effects of the treatment can be expected for small (<3 cm), low-grade lesions with a low radiosurgery-based AVM score. The bleeding risk after stereotactic radiosurgery gradually decreases during follow-up. Inclusion of the obliterated portion of the AVM nidus into the target volume may prevent later recanalisation. This observation requires further studies because it may significantly influence the qualification criteria for endovascular procedures.
References [1] Brown Jr RD. Simple risk predictions for arteriovenous malformation hemorrhage. Neurosurgery 2000;46(4):1024. [2] Kondziolka D, McLaughlin MR, Kestle JRW. Simple risk predictions for arteriovenous malformation hemorrhage. Neurosurgery 1995;37:851e855. [3] Flickinger JC, Kondziolka D, Maitz AH, Lunsford LD. An analysis of the dose-response for arteriovenous malformation radiosurgery and other factors affecting obliteration. Radiother Oncol 2002;63:347e354. [4] Nicolato A, Lupidi F, Sandri MF, et al. Gamma knife radiosurgery for cerebral arteriovenous malformations in children/ adolescents and adults. Part II: Differences in obliteration rates, treatmenteobliteration intervals, and prognostic factors. Int J Radiat Oncol Biol Phys 2006;64(3):914e921. [5] Sasaki T, Kurita H, Saito I, et al. Arteriovenous malformations in the basal ganglia and thalamus: management and results in 101 cases. J Neurosurg 1998;88:285e292. [6] Steiner L, Lindquist C, Adler JR, Torner JC, Alves W, Steiner M. Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 1992;77:1e8. [7] Bollet MA, Anxionnat R, Buchheit I, et al. Efficacy and morbidity of arc-therapy radiosurgery for cerebral arteriovenous malformations: a comparison with the natural history. Int J Radiat Oncol Biol Phys 2004;58(5):1353e1363. [8] Chang T-C, Shirato H, Aoyama H, et al. Stereotactic irradiation for intracranial arteriovenous malformation using stereotactic radiosurgery or hypofractionated stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2004;60(3):861e870. [9] Touboul E, Al Halabi A, Buffat L, et al. Single-fraction stereotactic radiotherapy: a dose-response analysis of arteriovenous malformation obliteration. Int J Radiat Oncol Biol Phys 1998;41(4):855e861.
S. Blamek et al. / Clinical Oncology 23 (2011) 525e531 [10] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Pediatric cerebral arteriovenous malformations: the role of stereotactic linac-based radiosurgery. Int J Radiat Oncol Biol Phys 2006;65(4):1206e1211. [11] Henkes H, Nahser H-C, Berg-Dammer E, Weber W, Lange S, € hne D. Endovascular therapy of brain AVMs prior to radiKu osurgery. Neurol Res 1998;20:479e492. [12] Hurst RW, Berenstein A, Kupersmith MJ, Madrid M, Flamm ES. Deep central arteriovenous malformations of the brain: the role of endovascular treatment. J Neurosurg 1995;82:190e195. [13] Schlienger M, Atlan D, Lefkopoulos D, et al. Linac radiosurgery for cerebral arteriovenous malformations: results in 169 patients. Int J Radiat Oncol Biol Phys 2000;46(5):1135e1142. [14] Pollock BE, Flickinger JC. A proposed radiosurgery-based grading system for arteriovenous malformations. J Neurosurg 2002;96:79e85. [15] Pollock BE, Flickinger JC. Modification of the radiosurgerybased arteriovenous malformation grading system. Neurosurgery 2008;63(2):239e243. [16] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Stereotactic linac-based radiosurgery in the treatment of cerebral arteriovenous malformations located deep, involving corpus callosum, motor cortex or brainstem. Int J Radiat Oncol Biol Phys 2006;64(4):1044e1048. [17] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Linac-based radiosurgery or hypofractionated stereotactic radiotherapy in the treatment of large cerebral arteriovenous malformations. Int J Radiat Oncol Biol Phys 2006;64(4): 1049e1054. [18] Yamamoto M, Jimbo M, Kobayashi M, et al. Long-term results of radiosurgery for arteriovenous malformation: neuroradiodiagnostic imaging and histological studies of angiographically confirmed nidus obliteration. Surg Neurol 1992;37: 219e230. [19] Zabel-du Bois A, Milker-Zabel S, Huber P, Schlegel W, Debus J. Risk of hemorrhage and obliteration rates of LINAC-based radiosurgery for cerebral arteriovenous malformations treated after prior partial embolization. Int J Radiat Oncol Biol Phys 2007;68(4):999e1003. [20] Miyawaki L, Dowd C, Wara W, et al. Five year results of LINAC radiosurgery for arteriovenous malformations: outcome for large AVMs. Int J Radiat Oncol Biol Phys 1999;44(5):1089e1106. €€ € M, [21] Hernesniemi JA, Dashti R, Juvela S, Va art K, Niemela Laakso A. Natural history of brain arteriovenous malformations: a long-term follow-up study of risk of hemorrhage in 238 patients. Neurosurgery 2008;63(5):823e829. [22] Ellis TL, Friedman WA, Bova FJ, Kubilis PS, Buatti JM. Analysis of treatment failure after radiosurgery for arteriovenous malformations. J Neurosurg 1998;89:104e110. [23] Meder JF, Oppenheim C, Blustajn J, et al. Cerebral arteriovenous malformations: the value of radiologic parameters in predicting response to radiosurgery. AJNR Am J Neuroradiol 1997;18:1473e1483. [24] Pollock BE, Gorman DA, Brown PD. Radiosurgery for arteriovenous malformations of the basal ganglia, thalamus and brainstem. J Neurosurg 2004;100:210e214. [25] Nataf F, Schlienger M, Lefkopoulos D, et al. Radiosurgery of cerebral arteriovenous malformations children: a series of 57 cases. Int J Radiat Oncol Biol Phys 2003;57(1):184e195.
531
[26] Gobin YP, Laurent A, Merienne L, et al. Treatment of brain arteriovenous malformations by embolization and radiosurgery. J Neurosurg 1996;85(1):19e28. [27] Andrade-Souza YM, Ramani M, Scora D, Tsao terBrugge MN, terBrugge K, Schwartz ML. Embolization before radiosurgery reduces the obliteration rate of arteriovenous malformations. Neurosurgery 2007;60(3):443e451. [28] Jabbour MN, Elder JB, Samuelson CG, et al. Aberrant angiogenic characteristics of human brain arteriovenous malformation endothelial cells. Neurosurgery 2009;64(1):139e146. [29] Akakin A, Ozkan A, Akgun E, et al. Endovascular treatment increases but gamma knife radiosurgery decreases angiogenic activity of arteriovenous malformations: an in vivo experimental study using a rat cornea model. Neurosurgery 2010;66(1):121e129. [30] Blamek S, Boba M, Larysz D, et al. The incidence of imaging abnormalities after stereotactic radiosurgery for cerebral arteriovenous and cavernous malformations. Acta Neurochir Suppl 2010;106:187e190. [31] Flickinger JC, Kondziolka D, Kalend AM, et al. Radiosurgeryrelated imaging changes in surrounding brain: multivariate analysis and model evaluation. Radiosurgery 1996;1:229e236. [32] Flickinger JC, Kondziolka D, Lunsford LD, et al. A multi-institutional analysis of complication outcomes after arteriovenous malformation radiosurgery. Int J Radiat Oncol Biol Phys 1999;44(1):67e74. [33] Flickinger JC, Kondziolka D, Maitz AH, Lunsford LD. Analysis of neurological sequelae from radiosurgery of arteriovenous malformations: how location affects outcome. Int J Radiat Oncol Biol Phys 1998;40(2):273e278. [34] Flickinger JC, Kondziolka D, Pollock BE, Maitz AH, Lunsford LD. Complications from arteriovenous malformation radiosurgery: multivariate analysis and risk modeling. Int J Radiat Oncol Biol Phys 1997;38(3):485e490. [35] Tomura N, Izumi J, Sakuma I, et al. Radiation-induced changes in the brain following stereotactic irradiation evaluated by sequential MRI. Comp Med Imag Graph 2004;28:73e79. [36] van den Berg R, Buis DR, Lagerwaard FJ, Lycklama A,
[37]
[38]
[39]
[40]
Nijeholt GJ, Vandertop WP. Extensive white matter changes after stereotactic radiosurgery for brain arteriovenous malformations: a prognostic sign for obliteration? Neurosurgery 2008;63(6):1064e1069. Guo WY, Pan HC, Chung WY, Wang LW, Teng MMH. Do we need conventional angiography? The role of magnetic resonance imaging in verifying obliteration of arteriovenous malformations after gamma knife surgery. Stereotact Funct Neurosurg 1996;66(Suppl. 1):71e84. Pollock BE, Kondziolka D, Flickinger JC, Patel AK, Bissonette DJ, Lunsford LD. Magnetic resonance imaging: an accurate method to evaluate arteriovenous malformations after stereotactic radiosurgery. J Neurosurg 1996;85: 1044e1049. €nen J, et al. Long-term excess Laakso A, Dashti R, Seppa mortality in 623 patients with brain arteriovenous malformations. Neurosurgery 2008;63(2):244e253. Sanchez-Mejia RO, McDermott MW, Tan J, Kim H, Young WL, Lawton MT. Radiosurgery facilitates resection of brain arteriovenous malformations and reduces surgical morbidity. Neurosurgery 2009;64(2):231e238.