- Email: [email protected]
Stereotactic irradiation for intracranial arteriovenous malformation using stereotactic radiosurgery or hypofractionated stereotactic radiotherapy
Int. J. Radiation Oncology Biol. Phys., Vol. 60, No. 3, pp. 861– 870, 2004 Copyright © 2004 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/04/$–see front matter
doi:10.1016/j.ijrobp.2004.04.041
CLINICAL INVESTIGATION
Brain
STEREOTACTIC IRRADIATION FOR INTRACRANIAL ARTERIOVENOUS MALFORMATION USING STEREOTACTIC RADIOSURGERY OR HYPOFRACTIONATED STEREOTACTIC RADIOTHERAPY TA-CHEN CHANG, M.D.,* HIROKI SHIRATO, M.D.,* HIDEFUMI AOYAMA, M.D.,* SATOSHI USHIKOSHI, M.D.,* NORIO KATO, M.D.,* SATOSHI KURODA, M.D.,† TATSUYA ISHIKAWA, M.D.,† KIYOHIRO HOUKIN, M.D.,‡ YOSHINOBU IWASAKI, M.D.,† AND KAZUO MIYASAKA, M.D.* Departments of *Radiology and †Neurosurgery, Hokkaido University Hospital, Sapporo, Japan; ‡Department of Neurosurgery, Sapporo Medical University, Sapporo, Japan Purpose: To investigate the appropriateness of the treatment policy of stereotactic irradiation using both hypofractionated stereotactic radiotherapy (HSRT) and stereotactic radiosurgery (SRS) for arteriovenous malformations (AVMs) located in an eloquent region or for large AVMs and using SRS alone for the other AVMs. Methods and Materials: Included in this study were 75 AVMs in 72 patients, with a mean follow-up of 52 months. Of the 75 AVMs, 33 were located in eloquent regions or were >2.5 cm in maximal diameter and were given 25–35 Gy (mean, 32.4 Gy) in four daily fractions at a single isocenter if the patient agreed to prolonged wearing of the stereotactic frame for 5 days. The other 42 AVMs were treated with SRS at a dose of 15–25 Gy (mean, 24.1 Gy) at the isocenter. The 75 AVMs were classified according to the Spetzler-Martin grading system; 21, 23, 28, 2, and 1 AVM were Grade I, II, III, IV, V, and VI, respectively. Results: The overall actuarial rate of obliteration was 43% (95% confidence interval [CI], 30 –56%) at 3 years, 72% (95% CI, 58 – 86%) at 5 years, and 78% (95% CI, 63–93%) at 6 years. The actuarial obliteration rate at 5 years was 79% for the 42 AVMs <2.0 cm and 66% for the 33 AVMs >2 cm. The 5- and 6-year actuarial obliteration rate was 61% (95% CI, 39 – 83%) and 71% (95% CI, 47–95%), respectively, after HSRT and 81% (95% CI, 66 –96%) and 81% (95% CI, 66 –96%), respectively, after SRS; the difference was not statistically significant. Radiation-induced necrosis was observed in 4 subjects in the SRS group and 1 subject in the HSRT group. Cyst formation occurred in 3 patients in the SRS group and no patient in the HSRT group. A permanent symptomatic complication was observed in 3 cases (4.2%), and 1 of the 3 was fatal. All 3 patients were in the SRS group. The annual intracranial hemorrhage rate was 5.5–5.6% for all patients. Conclusion: Our treatment policy using SRS and HSRT was as effective as the policy involving SRS alone. The HSRT schedule was suggested to have a lower frequency of radiation necrosis and cyst formation than the high-dose SRS schedule. The benefit of HSRT compared with lower dose SRS has not yet been determined. © 2004 Elsevier Inc. Stereotactic irradiation, Arteriovenous malformation, Hypofractionation.
INTRODUCTION
AVMs without causing radiation damage to the normal tissue (1, 2). We have used hypofractionated stereotactic radiotherapy (HSRT) as an alternative treatment for AVMs in eloquent regions or large AVMs. Single-fraction SRS was used as a basic treatment method for AVMs located in noneloquent regions or small ones. Our preliminary results have already been published, suggesting that HSRT was as effective as
Single-fraction stereotactic radiosurgery (SRS) is an effective treatment for patients with intracranial arteriovenous malformations (AVMs) with a diameter of about ⱕ2.5 cm. However, AVMs ⬎2.5 cm in mean diameter, or around 10 cm3 in volume, and AVMs in eloquent regions are often not cured, because we cannot give an efficient dose to the
Japan. Acknowledgments—We appreciate the help of the staff of the Departments of Neurosurgery and Radiology, Hokkaido University Hospital. We also appreciate Drs. Takeshi Nishioka and Rikiya Onimaru, for their help in preparing this paper. Received Sep 29, 2003, and in revised form Mar 8, 2004. Accepted for publication Apr 12, 2004.
Reprint requests to: Hiroki Shirato, M.D., Department of Radiology, Hokkaido University School of Medicine, North-15 West-7, Sapporo 060-8638, Japan. Tel: (⫹81) 11-706-7876; Fax: (⫹81) 11-706-5977; E-mail: [email protected] Parts of this paper were presented at the ISRS 2003 in Kyoto, Japan. Partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, and Technology of 861
862
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 3, 2004
Fig. 1. Arteriovenous malformation (AVM) treatment protocol selection between stereotactic radiosurgery (SRS) and hypofractionated stereotactic radiotherapy (HSRT).
SRS, with possibly a lower complication rate (3). However, the study included only 27 SRS and 26 HSRT patients (3). The present study was not a randomized study involving SRS and HSRT. The efficacy of HSRT was investigated, taking account the selection bias of the treatment protocol. The appropriateness of our treatment policy of stereotactic irradiation (STI) using both SRS and HSRT is discussed. METHODS AND MATERIALS Irradiation schedule Stereotactic irradiation was SRS or HSRT. The hypofractionation schedule in this study was determined using the linear-quadratic formula for late complications without correcting for slow repair (4). The ␣/ ratio of the normal brain was estimated to be 2.0, which is widely accepted with regard to late radiation effects against brain parenchyma (5). A dose of 35 Gy in four fractions and 28 Gy in four fractions would have the same effect as 18.4 Gy and 14.2 Gy, respectively, in a single exposure of radiation, assuming an ␣/ ratio of 2.0. An isocenter dose of 35 Gy and a peripheral dose of 28 Gy delivered in four daily fractions was prescribed as the standard dose in this study. Patients This study included 72 consecutive patients with angiography-proven AVMs who had agreed to undergo STI between January 1991 and December 2000. Patients initially presented with one or more of the following symptoms: episode of intracranial hemorrhage (n ⫽ 37, 51%), epileptic seizure (n ⫽ 12, 17%), and neurologic deficits (n ⫽ 46,
64%). In 19 patients, the AVM was discovered incidentally without any symptoms (n ⫽ 13), with symptoms unrelated to AVM (n ⫽ 1), or with symptoms from other disease states (subarachnoid hemorrhage in 1, facial hemangioma in 1, traffic accident in 2, and brain infarction in 1). Before STI, 11 patients had undergone surgical resection of the AVM. Embolization was performed in 7 patients before STI. Of 11 patients with coincidental intracranial aneurysms, 3 were treated by clipping the aneurysm before STI. All patients underwent digital subtraction angiography and CT. MRI and magnetic resonance angiography were used in the patients treated since the late 1990s. The AVMs were classified according to the Spetzler-Martin grading system (6). Treatment protocol—selection between SRS and HSRT The recursive tree used for the selection of the treatment method in the 75 AVMs is shown in Fig. 1. If the AVM nidus was primarily located in eloquent regions, such as the sensory-motor region, language region, and visual cortex; hypothalamus and thalamus; internal capsule; brainstem; cerebellar peduncles, or deep cerebellar nuclei, HSRT was used. In addition, when the nidus was located in a noneloquent region and was ⬎2.5 cm in maximal diameter as measured by angiography, HSRT was used. However, of those with AVMs that fulfilled the above criteria and who were eligible for HSRT, 5 patients were judged not to be suited for prolonged head-frame wearing because of physical instability or mental deficiency, and 8 patients did not agree to the prolonged head-frame wearing. These 13 patients were treated with SRS instead of HSRT. Conse-
STI for AVMs
Table 1. Patient characteristics Characteristic Age (y) ⬍20 ⱖ20 Gender Male Female History of hemorrhage Yes No Prior resection Yes No Prior embolization Yes No Diameter (cm) ⱕ2.5 ⬎2.5 ⱕ2.0 ⬎2.0 Location Eloquent Noneloquent Drainage vein Deep Superior Spetzler-Martin grade I ⫹ II III ⫹ IV ⫹ V
HSRT (n)
SRS (n)
2 31
7 35
20 13
23 19
15 18
22 20
6 27
5 37
3 30
4 38
18 15 13 20
36 6 29 13
20 13
14 28
19 14
24 18
16 17
28 14
quently, 33 lesions were treated with HSRT, and 42 lesions with SRS. All 6 patients with AVMs in an eloquent area with a maximal diameter ⬎2.5 cm were treated with HSRT. Table 1 shows the patient and AVM distribution in each treatment group. Radiosurgery technique All patients were treated with an invasive stereotactic head frame (RADFRAME, Mizuho Ika Kogyo, Tokyo, Japan) for use in performing CT, MRI, angiography, and RT. The RADFRAME was developed in our institution to be suitable for prolonged wearing while maintaining 1.0-mm accuracy. When SRS was used, the patients wore the stereotactic frame for 1–2 days. When HSRT was used, patients wore the stereotactic frame for 5 days. The procedure was performed with i.v. sedation when necessary. Between 1991 and April 2003, the AVM nidus was defined using biplane angiography, and the target was displayed on CT. After 1998, MRI was also used as a complementary tool for delineation (7, 8). A single isocenter and three-dimensional RT planning system (Focus, CMS, St. Louis, MO) were used. An isodose contour conforming to the entire AVM nidus was selected as the prescription isodose, and 6- or 10-mV X-ray beams from a linear accelerator were used. In general, a multileaf collimator was used for AVMs ⬎1.5 cm, and a cone-shape collimator was used for smaller
● T.-C. CHANG et al.
863
AVMs. The arc technique was used in three lesions, and the static noncoplanar multiport technique was used for 72 lesions. The number of arcs or planes ranged from three to four, and the median number of ports was 8 (range, 3–11). Prescription The standard total dose for SRS was 25 Gy at the isocenter of the treatment target volume, and the nidus was enclosed by an 80% isodose surface, with a minimal dose of 20 Gy. HSRT was primarily performed using 35 Gy at the isocenter, with a minimal dose of 28 Gy in four daily fractions. The total doses used for SRS and HSRT were modified depending on the size and location of the AVM, as well as patient age, with patients ⬍20 years old receiving 90% of the dose described above. Thus, the total dose at the isocenter ranged from 15 to 25 Gy (mean, 24.1 Gy) for SRS and 25 to 35 Gy (mean, 32.4 Gy) for HSRT. In SRS, 29 AVMs ⬍2 cm were treated with 25 Gy in one fraction. The other 13 AVMs that were ⬎2 cm were given a lower dose to reduce the possibility of severe complications according to the isoeffect line of Kjellberg et al. (1) and Flickinger et al. (2). In HSRT, 18 AVMs received 35 Gy in four fractions, and 15 patients were treated with different doses and fractions in the early period or because of the proximity of the nidus to specifically critical structures such as the optic chiasm (4 AVMs, 25 Gy in four fractions; 1 AVM, 28 Gy in four fractions; 5 AVMs, 30 Gy in four fractions; 1 AVM, 32 Gy in four fractions; 1 AVM, 32.5 Gy in four fractions; 1 AVM, 30 Gy in eight fractions; and 2 AVMs, 35 Gy in eight fractions). The minimal dose ranged from 12 to 20 Gy (mean, 19.3 Gy) for SRS and 20 to 28 Gy (mean, 25.9 Gy) for HSRT. Follow-up evaluation For most of the patients who lived close enough to our institution, clinical examination and serial imaging studies (MRI or CT when MRI was contraindicated) were performed at 3- or 4-month intervals for the first year and every 6 months until complete obliteration was angiographically demonstrated. Some patients who lived far from our institution underwent the imaging studies near their homes and were evaluated by the referring physicians. Angiography was planned only when MRI demonstrated complete obliteration or when otherwise clinically indicated. The term “complete angiographic obliteration” was defined as normal blood flow, complete absence of pathologic vessels at the site of the nidus, and the normalization of flow in the efferent veins. The date of obliteration was defined as the date of angiography demonstrating complete obliteration. Clinical complications from STI were classified as new or aggravated neurologic symptoms or signs; complications requiring steroids, shunt placement, or surgical intervention; an imaging change suggesting radiation-induced edema (increased with a T2-weighted signal change), radiation-induced necrosis, or cyst formation; or treatment-related death.
864
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 3, 2004
Fig. 2. Actuarial obliteration-free curves according to arteriovenous malformation (AVM) diameter.
Statistical analysis For statistical analysis, we used a computer software statistical package (StatView for Windows, version 5, SAS Institute, Cary, NC). The actuarial rates of complete obliteration or T2-weighted signal changes or postradiosurgical hemorrhage were calculated by the Kaplan-Meier method. The overall actuarial rate of obliteration was calculated using the total number of patients at risk for the denominator and the number of patients with angiographically demonstrated complete obliteration for the numerator. The log– rank test was used to compare the outcomes of the different groups. Univariate analysis of the clinical and angiographic AVM factors was performed using the log–rank test. The rates of post-STI hemorrhage were calculated from the time of STI using the Kaplan-Meier method, and patients were considered to be at risk of hemorrhage until the date of their last follow-up. RESULTS Follow-up All 72 patients had clinical follow-up after STI (mean, 52 months; median, 48 months; range, 0 –138 months). One patient did not appear for follow-up at all but was counted as a patient at risk of complete obliteration and was also used for the analysis of acute complications. One died of an unknown cause after angiographic confirmation of AVM obliteration and was counted as having angiographic obliteration. After angiographic confirmation of no obliteration, 1 patient died of bile duct adenocarcinoma and 1 patient
died of treatment-related complications (details given below). These 2 patients were counted as angiographic failures in the analysis of obliteration. Four died before angiographic examination (rebleeding in 1, brain infarction in 1, and unknown causes in 2) and were counted as having been withdrawn before angiographic obliteration. Obliteration After STI, 57 of 75 AVMs were examined using angiography, and complete obliteration of the nidus was confirmed in 39 AVMs between 12 and 90 months (mean, 36 months; median, 32 months). Patients in whom the last angiographic study demonstrated partial obliteration or a residual presence of the AVM were considered to have treatment failure. The overall actuarial rate of obliteration was 43% (95% confidence interval [CI], 30 –56%) at 3 years, 72% (95% CI, 58 – 86%) at 5 years, and 78% (95% CI, 63–93%) at 6 years. The actuarial obliteration rate for the 42 AVMs ⱕ2.0 cm in maximal diameter (66%) was 50% (95% CI, 32– 68%) at 3 years and 79% (95% CI, 62–98%) at 5 years. For the 33 AVMs with a maximal diameter ⬎2 cm (44%), the actuarial complete obliteration rate was 44% (95% CI, 26 – 62%) at 3 years and 66% (95% CI, 45– 87%) at 5 years (p ⫽ 0.2674; Fig. 2). Similarly, the 3-year and 5-year complete obliteration rate was 39% (95% CI, 45– 87%) and 57% (95% CI, 45– 87%), respectively, for AVMs ⬎2.5 cm. Figure 2 shows that the larger AVMs tended to become obliterated later. The 3-year and 5-year complete obliteration rate of AVMs in eloquent regions was 51% (95% CI, 32–70%) and 86% (95% CI, 70 –100%), respectively.
STI for AVMs
● T.-C. CHANG et al.
865
Fig. 3. Actuarial obliteration-free curves according to treatment schedule.
The 3-year, 5-year, and 6-year actuarial complete obliteration rate was 32% (95% CI, 19 – 49%), 61% (95% CI, 39 – 83%), and 71% (95% CI, 47–95%), respectively, for patients treated with HSRT and followed angiographically for 21–90 months (mean, 37 months; median, 27 months) and was 52% (95% CI, 35– 69%), 81% (95% CI, 66 –96%), and 81% (95% CI, 66 –96%), respectively, for patients treated with SRS and followed angiographically for 11– 60 months (mean, 32 months; median, 31 months). The complete obliteration rates were not significantly different statistically between the HSRT and SRS groups (p ⫽ 0.1401; Fig. 3). The 3-year and 5-year complete obliteration rate was 26% (95% CI, 0 –52%) and 53% (95% CI, 18 – 88%), respectively, for AVMs ⬎2.5 cm treated with HSRT. The 3-year and 5-year complete obliteration rate of AVMs in eloquent regions treated with HSRT was 35% (95% CI, 12–58%) and 80% (95% CI, 56 –100%), respectively. Of the 16 AVMs with incomplete obliteration or residual AVM confirmed by angiography, 1 patient in the SRS group and 7 patients in the HSRT group were retreated with STI. For the second treatment, we used SRS with varying doses depending on the initial radiosurgical dose and the maximal tolerance dose of the brain. Of these 8 patients, 3 (1 in the SRS and 2 in the HSRT group) experienced obliteration. If we assume that the retreatment was not a failure, the obliteration rate at 3, 5, and 6 years was 43% (95% CI, 30 – 56%), 70% (95% CI, 57– 83%), and 76% (95% CI, 63– 99%), respectively. Univariate analysis of the possible prognostic factors for obliteration showed that the variables of treatment method (SRS, HSRT), age, gender, history of hemorrhage, prior
treatment (resection, embolization), and AVM characteristics (location, diameter, drainage vein, Spetzler-Martin grade) did not reach statistically significant levels (Table 2). Complications Of the 67 patients periodically examined using MRI, 39 (58%) exhibited a newly increased high signal surrounding their AVM site on T2-weighted MRI. The 3-year actuarial rate for developing post-STI T2-weighted signal changes was 64% (95% CI, 45–73%) for HSRT and 59% (95% CI, 42–76%) for SRS (Fig. 4). The cumulative actuarial rates for post-STI T2-weighted signal changes were not significantly different statistically between the HSRT and SRS groups. Of the 72 patients, 9 (9.7%) experienced transient symptomatic adverse effects after STI, 7 of which were related to the T2-weighted signal changes. All 9 patients recovered without a symptomatic deficit. Of the 72 patients, 3 (4.2%) had permanent radiationinduced symptomatic adverse effects. One (1.4%) of the three permanent adverse effects was fatal. This patient had status epilepticus after surgery for a large AVM and had maintained a good performance status after STI for 47.6 months. However, the epilepsy appeared again after cyst formation due to the STI. She died of status epilepticus owing to the expansion of a radiation-related cyst 1 week before the planned surgery for the cyst. The other 2 patients had chronic neurologic deficits (double vision in one and slight paresis of the left leg in the other). In 71 patients with MRI and/or CT follow-up, 5 (7.0%) were diagnosed with radiation necrosis. In the SRS group, 4 patients experienced radiation-related necrosis, 3 of which
866
I. J. Radiation Oncology
● Biology ● Physics
Table 2. Various parameters for obliteration rates: univariate analysis Variable Age (y) ⬍20 ⱖ20 Gender Male Female History of hemorrhage Yes No Prior resection Yes No Prior embolization Yes No Treatment SRS HSRT Diameter (cm) ⱕ2.0 ⬎2.0 ⱕ2.5 ⬎2.5 Location Eloquent Noneloquent Drainage vein Deep Superior Spetzler-Martin grade I ⫹ II III ⫹ IV ⫹ V
n
5-y Obliteration rate (95% CI)
p
9 66
56 ⫾ 33 73 ⫾ 14
0.5698
43 32
75 ⫾ 18 69 ⫾ 20
0.73
37 38
91 ⫾ 16 67 ⫾ 19
0.7712
11 64
72 ⫾ 30 73 ⫾ 15
0.8463
7 68
54 ⫾ 44 75 ⫾ 14
0.0775
42 33
81 ⫾ 15 61 ⫾ 22
0.1401
42 33 54 21
79 ⫾ 17 66 ⫾ 21 77 ⫾ 15 57 ⫾ 17
34 41
86 ⫾ 16 59 ⫾ 20
0.1381
43 32
71 ⫾ 17 75 ⫾ 22
0.8721
44 31
72 ⫾ 18 73 ⫾ 27
0.963
0.2674 0.2324
Abbreviation: CI ⫽ confidence interval; other abbreviations as in Table 2.
Volume 60, Number 3, 2004
were in patients with permanent radiation-induced symptomatic adverse effects. Radiation-related cyst formation was observed 2–3 years later in 3 patients with radiation necrosis after SRS. All 3 patients were to be treated in the HSRT group according to our decision tree but underwent SRS instead of HSRT (Table 3). In 33 patients who were treated with HSRT, 1 case of necrosis was observed that was symptomatically reversible. Seven patients (18%) of the 39 with T2-weighted signal changes and 4 patients (12%) of the 33 without T2-weighted signal changes experienced radiation-induced symptomatic complications.
Hemorrhage Post-STI hemorrhage was observed in 10 patients: 4 in the SRS group at a mean of 20 months (range, 1–21 months) and 6 in the HSRT group at a mean of 33 months (range 2–93). Of these 10 patients, 3 experienced neurologic deterioration owing to bleeding of the AVM after STI. One patient experienced hemiparesis after SRS. One experienced seizure and left homonymous hemianopsia after HSRT. One patient died just after AVM bleeding at 14 months after HSRT. The 5-year actuarial accumulative intracranial hemorrhage rate after STI was 15.2% (95% CI, 5.4 –25.0% ) for all patients combined, 22.3% (95% CI, 5.8 –38.8% ) for HSRT patients, and 7.8% (95% CI, 0 –16.2%) for SRS patients. The annual intracranial hemorrhage rate was 5.6% during the first year and 5.5% during the second year for all patients combined. It was 9.2% and 3.1% for HSRT patients at 1 and 2 years after STI and 3.6% and 4.4% for SRS patients, respectively. No statistically significant difference was found between the HSRT and SRS groups in the postradiosurgical hemorrhage rate (p ⫽ 0.2442). The actuarial accumulative intracranial hemor-
Fig. 4. T2-weighted signal change curves according to treatment schedule.
STI for AVMs
● T.-C. CHANG et al.
867
Table 3. Results of treating with SRS instead of HSRT Age Pt. (y)/ Prior medical No. Gender KPS history
Prior treatments (n)
Location Maximal (Spetzlerdiameter Martin grade) (cm)
1
53/M
90
Headache (incident)
None
Right basal ganglia (III)
2.2
2
16/F
90
None
39/M
90
Left frontal (I) Left thalamus (III)
2.7
3
Headache, nausea Headache, left limb numbness
4
19/F
60
Hemorrhage, coma
Surgery
Right frontal (II)
3.5
5
30/M
80
None (incident)
None
6
29/F
80
Hemorrhage
7
29/F
90
Right thalamus (III) Embolization Left thalamus (2) (III) None Right frontal (motor area) (III)
8
41/M
70
Headache, right internal carotid cave aneurysm (unruptured) Hemorrhage, Embolization hemiparesis (8)
9
16/F
90
Hemorrhage, coma
10
55/M
90
11
46/M
70
12
17/F
90
13
78/M
80
None None (incident), BuddChiari syndrome Hemorrhage, Ventricular seizure, drainage coma Hemorrhage, Ventricular seizure, drainage, coma VP shunt None Multiple brain infarction, Parkinson disease, hypertension, viral hepatitis
None
Ventricular drainage
0.9
Reason for Dose Angiographic SRS (cGy) obliteration Disagreement 2500
Child
Residual
2000
Complete
Disagreement 2500
Complete
Symptomatic complications and treatment Radiation necrosis, cyst formation, steroids, glycerol infusions, mild headache None
2250
Residual
2.3
Disagreement 2000
Residual
Radiation necrosis, cyst formation, steroids, glycerol infusions, hyperbaric oxygen therapy, tremor, seizure, persistent hemiparesis Radiation necrosis, cyst formation, steroids, glycerol infusions, seizure, death from status epilepticus None
2.0
Disagreement 2500
Complete
None
2.0
Disagreement 2500
Complete
None
Left corpus callosum, frontal (III) Left paraventricle (III) Right internal caudate head (III)
3.4
Disagreement 1500
Residual
None
2250
Complete
None
3.5
Disagreement 2500
Complete
None
Left thalamus (III)
1.3
Disagreement 2000
—
None
Left basal ganglia (III) Left basal ganglia (III)
1.0
Disagreement 2500
Complete
None
0.5
Physical 2500 instability mental deficiency
—
None
2.0
Physical instability
Child
Abbreviations: SRS ⫽ stereotactic radiosurgery; HSRT ⫽ hypofractionated stereotactic radiotherapy; Pt. No. ⫽ patient number; M ⫽ male; F ⫽ female.
868
I. J. Radiation Oncology
● Biology ● Physics
rhage rate at 5 years for AVMs ⱕ2.0 cm was 12.4% (95% CI, 0.6 –23.2%) and for AVMs ⬎2.0 cm was 17.7% (95% CI, 4 –30.4%). The difference did not reach statistical significance. Of the 13 patients who had been treated with SRS instead of HSRT, 7 (53%) had confirmed complete obliteration and 4 had partial obliteration or residual AVM by angiography (Table 3). Radiation necrosis and cyst formation was observed in 3 patients each (23%). Two patients experienced intracranial hemorrhage after treatment. DISCUSSION The primary subject of our study was to examine the appropriateness of our treatment policy compared with previous reports in which SRS alone was used. Particularly for AVMs ⬎2.0 –2.5 cm (44% ⬎ 2.0 cm and 29% ⬎ 2.5 cm in our series) and AVMs in eloquent areas (45% of our series), making a comparison with the previous report is quite important. However, we found that it was quite difficult to compare the results of STI for AVMs among different institutions. The crude obliteration rates of AVMs after SRS vary significantly from 40% to 85% at 3 years and 33% to 93% at 5 years (9 –24). If we adopt these crude obliteration rates using the number of patients who underwent angiography as the denominator, the crude complete obliteration rates for our treatment were 92% (12 of 13) at 3 years for 33 AVMs with a maximal diameter ⬎2 cm. This crude obliteration rate was better than that observed in previous papers but obviously was dependent on the selection bias and was far from a determination of the actual outcome of treatment. Because recent improvements in MRI and magnetic resonance angiography have helped to preclude unnecessary confirmatory angiography for patients with patent AVMs, the numerator can be very close to the denominator without an improvement in the treatment method. The number of reports using the Kaplan-Meier method for the analysis of the obliteration rate is increasing. Miyawaki et al. (23) have used the Kaplan-Meier method and reported a 33% 5-year actuarial rate of angiographic obliteration, assuming MRI/magnetic resonance angiography obliteration and retreatment to be failures, as we did. Touboul et al. (25) have reported 62.5% ⫾ 7% as the 5-year actuarial obliteration rate for 100 AVM patients. Pan et al. (20) reported a 75% actuarial complete obliteration rate at 40 months. Our results showed a 72% (95% CI, 58 – 86%) overall actuarial obliteration rate. Considering that 44% of the patients had AVMs ⬎2.0 cm in our series, the treatment outcome of our study is at least comparable with those of the previous studies. The decision tree in our treatment policy of STI, including HSRT, was not detrimental for cerebral AVMs. The treatment method was not a prognostic factor in this analysis, even though larger AVMs were more frequent in HSRT. The diameter of AVM was not a statistically significant prognostic factor for angiographic obliteration in this
Volume 60, Number 3, 2004
series. This latter result was not seen in our preliminary study (17), in which size (ⱕ2 cm or not) was a statistically significant favorable factor (p ⫽ 0.05). The finding that diameter was not a statistically significant factor may have been because the number of patients with occluded large AVMs increased in accordance with the longer follow-up period. Younger age was suggested to be a statistically significant factor in our previous study, but age also disappeared as a statistically significant factor in the present study. This result may also have been because the patients with AVMs ⬎2.0 cm who experienced complete obliteration with HSRT were older, with a median age of 45 years (range, 18 –74 years). Late radiation-induced symptomatic injury occurred in 3–18% of the patients with AVMs in the previous reports (12, 15, 16, 22, 23, 27, 28). Symptomatic complications appeared in the patients with larger AVMs and those who received higher doses of SRS as predicted by the radiobiologic volume effect models of the central nervous system (29). Steinberg et al. reported that 50% (10 of 20) had major or minor complications with doses ⬎18 Gy and volumes ⬎13 cm3 (24). Miyawaki et al. (23) reported that radiation necrosis occurred when minimal doses of 18 Gy and maximal doses of 34 Gy were administered to treatment volumes ⱖ14 cm3 (mean, 27.8 cm3). Flickinger et al. (26) found that eloquent location and volume irradiated to ⱖ12 Gy were significant risk factors for symptomatic radiation necrosis. However, in our series, symptomatic radiation injury was not more frequent in patients with larger AVMs or in patients with AVMs in eloquent regions. Radiation necrosis and cyst formation were observed in 4 of 42 in the SRS group and 1 of 33 in the HSRT group, the latter of which included larger AVMs. All three cyst formations were seen in patients who should have been treated by HSRT in our protocol but were treated with SRS instead. We decreased the SRS dose for these patients, in principle, but the dose reduction might have been too small. However, if we had used a lower dose, the obliteration rate in these patients might have been lower. We have used the same hypofractionated schedule for 140 brain metastases with a mean diameter of 6.7 cm (range, 0.006 – 48.3 cm) in 75 patients. We found that the schedule was as safe as we have seen in the present study with a radiation necrosis rate of 2.7% (2 patients) (30). These brain metastasis results and the results of the present AVM study suggest that our hypofractionation schedule is safe and close to the maximal tolerance dose. The maximal tolerance dose is consistent with the prediction made by the radiobiologic model of Brenner et al. (4), assuming an ␣/ ratio of 2.0 (3). The annual bleeding rate after STI was 5.5–5.6% in this series, within the range of the natural history of untreated AVMs, which runs from 2% to 17% annually (31–33). The number of AVMs that bled after STI was greater in the HSRT group (6 of 33) than in the SRS group (4 of 42) in this series. Also, the bleeding rate in the first year for HSRT was three times the rate of the SRS patients, although the
STI for AVMs
difference did not reach statistical significance. Stefani et al. (34), in prospective analyses of follow-up data from 390 patients with brain AVMs, found that large and deep-seated AVMs were the most important and significant factors associated with a greater risk of future hemorrhagic events. The greater bleeding rate in the first year in the HSRT group may have been related to the larger size of the AVMs. Steiner et al. (24) indicated that patients treated with radiosurgery may be at a decreased risk of hemorrhage once obliteration has effectively occurred. In our series, the actuarial obliteration rates for the HSRT group increased after 2 years. This might have been in relation to the decrease in the hemorrhage rate from 9.2% to 3.1% during the first 2 years of follow-up. The small number of patients prevented us from speculating more about the difference. For large AVMs, staged SRS has been shown to be an effective approach (35). In this approach, SRS is repeated during an interval of 5–36 months. Because we have also experienced good obliteration rates after repeated treatment of patient AVMs, we agree that staged STI may be one of the treatment options for large AVMs. This does not preclude the potential importance of HSRT as the initial treatment for large AVMs. Each approach should be considered one of several options for treating this difficult disease.
● T.-C. CHANG et al.
869
CONCLUSION Our treatment policy using SRS and HSRT was effective in producing an obliteration rate and complication rate compatible with those found in previous SRS studies. HSRT may have a wider therapeutic window in the treatment of AVMs ⱖ2.5 cm or AVMs in eloquent areas, because the selection bias in our treatment decision tree could favorably affect SRS but not HSRT. Some might conclude that the SRS dose in this series (minimal dose range, 12–20 Gy; mean, 19.3), was too high and that HSRT possibly has a lower obliteration rate than SRS. The current study does not answer the question of whether SRS using lower doses such as 15–18 Gy at the periphery of the AVM can achieve results similar to HSRT. We also cannot answer the question of whether HSRT has a role in decreasing late radiation necrosis for small AVMs and AVMs in noneloquent areas. Only a prospective randomized trial will answer these questions. An optimal hypofractionation schedule may be worth additional investigation, although the optimal dose distribution and rigid fixation continue to be the critical requirements for STI of AVMs in the central nervous system (36).
REFERENCES 1. Kjellberg RN, Hanamura T, Davis KR, et al. Bragg-peak proton-beam therapy for arteriovenous malformations of the brain. N Engl J Med 1983;309:269–274. 2. Flickinger JC, Schell MC, Larson DA. Estimation of complications for linear accelerator radiosurgery with the integrated logistic formula. Int J Radiat Oncol Biol Phys 1990;19:143– 148. 3. Aoyama H, Shirato H, Nishioka T, et al. Treatment outcome of single or hypofractionated single-isocentric stereotactic irradiation (STI) using a linear accelerator for intracranial arteriovenous malformation. Radiother Oncol 2001;59:323–328. 4. Brenner DJ, Martel MK, Hall EJ. Fractionated regimens for stereotactic radiotherapy of recurrent tumors in the brain. Int J Radiat Oncol Biol Phys 1991;21:819–824. 5. Bentzen SM, Baumann M. The linear-quadratic model in clinical practice. In: Steel GG, editor. Basic clinical radiobiology. London: Arnold; 2002. p. 134–146. 6. Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg 1986;65:476–483. 7. Kondziolka D, Dempsey PK, Lunsford LD, et al. A comparison between magnetic resonance imaging and computed tomography for stereotactic coordinate determination. Neurosurgery 1992;30:402–406. 8. Aoyama H, Shirato H, Nishioka T, et al. Magnetic resonance imaging system for three-dimensional conformal radiotherapy and its impact on gross tumor volume delineation of central nervous system tumors. Int J Radiat Oncol Biol Phys 2001; 50:821–827. 9. Yamamoto M, Jimbo M, Hara M, et al. Gamma knife radiosurgery for arteriovenous malformations: Long-term follow-up results focusing on complications occurring more than 5 years after irradiation. Neurosurgery 1996;38:906–914. 10. Yamamoto Y, Coffey RJ, Nichols DA, et al. Interim report on the radiosurgical treatment of cerebral arteriovenous malfor-
11.
12.
13.
14.
15.
16.
17.
18.
19. 20.
mations: The influence of size, dose, time, and technical factors on obliteration rate. J Neurosurg 1995;83:832–837. Shin M, Kawamoto S, Kurita H, et al. Retrospective analysis of a 10-year experience of stereotactic radio surgery for arteriovenous malformations in children and adolescents. J Neurosurg 2002;9:779–784. Colombo F, Pozza F, Chierego G, et al. Linear accelerator radiosurgery of cerebral arteriovenous malformations: An update. Neurosurgery 1994;34:14–20. 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;46:1135–1142. Hadjipanayis CG, Levy EI, Niranjan A, et al. Stereotactic radiosurgery for motor cortex region arteriovenous malformations. Neurosurgery 2001;48:70–76. Pollock BE, Gorman DA, Coffey RJ. Patient outcomes after arteriovenous malformation radiosurgical management: Results based on a 5- to 14-year follow-up study. Neurosurgery 2003;52:1291–1297. Steinberg GK, Fabrikant JI, Marks MP, et al. Stereotactic heavy-charged-particle Bragg-peak radiation for intracranial arteriovenous malformations. N Engl J Med 1990;323:96– 101. Aoyama H, Shirato H, Nishioka T, et al. Treatment outcome of single or hypofractionated single-isocentric stereotactic irradiation (STI) using a linear accelerator for intracranial arteriovenous malformation. Radiother Oncol 2001;59:323–328. Friedman WA, Bova FJ, Mendenhall WM. Linear accelerator radiosurgery for arteriovenous malformations: The relationship of size to outcome. J Neurosurg 1995;82:180–189. Friedman WA, Bova FJ. Linear accelerator radiosurgery for arteriovenous malformations. J Neurosurg 1992;77:832–841. Pan DH, Guo WY, Chung WY, et al. Gamma knife radiosurgery as a single treatment modality for large cerebral arterio-
870
21. 22.
23.
24. 25.
26.
27.
I. J. Radiation Oncology
● Biology ● Physics
venous malformations. J Neurosurg 2000;93(Suppl. 3):113– 119. Lunsford LD, Kondziolka D, Flickinger JC, et al. Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg 1991;75:512–524. Engenhart R, Wowra B, Debus J, et al. The role of high-dose single-fraction irradiation in small and large intracranial arteriovenous malformations. Int J Radiat Oncol Biol Phys. 1994; 15:30:521–529. 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: 1089–1106. Steiner L, Lindquist C, Adler JR, et al. Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 1992;77:1–8. 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;41:855–861. for the Arteriovenous Malformation Radiosurgery Study Group, Flickinger JC, Kondziolka D, Lunsford LD, et al. Development of a model to predict permanent symptomatic postradiosurgery injury for arteriovenous malformation patients. Int J Radiat Oncol Biol Phys 2000;46:1143–1148. Flickinger JC, Lunsford LD, Kondziolka D, et al. Radiosurgery and brain tolerance: An analysis of neurodiagnostic imaging changes after gamma knife radiosurgery for arteriovenous malformations. Int J Radiat Oncol Biol Phys 1992;23: 19–26.
Volume 60, Number 3, 2004
28. Yamamoto M, Hara M, Ide M, et al. Radiation-related adverse effects observed on neuro-imaging several years after radiosurgery for cerebral arteriovenous malformations. Surg Neurol 1998;49:385–398. 29. Shirato H, Mizuta M, Miyasaka K. A mathematical model of the volume effect which postulates cell migration from unirradiated tissues. Radiother Oncol 1995;35:227–231. 30. Aoyama H, Shirato H, Onimaru R, et al. Hypofractionated stereotactic radiotherapy alone without whole-brain irradiation for patients with solitary and oligo brain metastasis using noninvasive fixation of the skull. Int J Radiat Oncol Biol Phys 2003;56:793–800. 31. Itoyama Y, Uemura S, Ushio Y, et al. Natural course of unoperated intracranial arteriovenous malformations: Study of 50 cases. J Neurosurg 1989;71:805–809. 32. Fults D, Kelly DL, Jr. Natural history of arteriovenous malformations of the brain: A clinical study. Neurosurgery 1984; 15:658–662. 33. Brown RD, Jr, Wiebers DO, Forbes G, et al. The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 1988;68:352–357. 34. Stefani MA, Porter PJ, terBrugge KG, et al. Large and deep brain arteriovenous malformations are associated with risk of future hemorrhage. Stroke 2002;33:1220–1224. 35. Pendl G, Unger F, Papaefthymiou G, et al. Staged radiosurgical treatment for large benign cerebral lesions. J Neurosurg 2000;93(Suppl. 3):107–112. 36. Hida K, Shirato H, Isu T, et al. Focal fractionated radiotherapy for intramedullary spinal arteriovenous malformations: 10year experience. J Neurosurg 2003;99(1 Suppl.):34 –38.
doi:10.1016/j.ijrobp.2004.04.041
CLINICAL INVESTIGATION
Brain
STEREOTACTIC IRRADIATION FOR INTRACRANIAL ARTERIOVENOUS MALFORMATION USING STEREOTACTIC RADIOSURGERY OR HYPOFRACTIONATED STEREOTACTIC RADIOTHERAPY TA-CHEN CHANG, M.D.,* HIROKI SHIRATO, M.D.,* HIDEFUMI AOYAMA, M.D.,* SATOSHI USHIKOSHI, M.D.,* NORIO KATO, M.D.,* SATOSHI KURODA, M.D.,† TATSUYA ISHIKAWA, M.D.,† KIYOHIRO HOUKIN, M.D.,‡ YOSHINOBU IWASAKI, M.D.,† AND KAZUO MIYASAKA, M.D.* Departments of *Radiology and †Neurosurgery, Hokkaido University Hospital, Sapporo, Japan; ‡Department of Neurosurgery, Sapporo Medical University, Sapporo, Japan Purpose: To investigate the appropriateness of the treatment policy of stereotactic irradiation using both hypofractionated stereotactic radiotherapy (HSRT) and stereotactic radiosurgery (SRS) for arteriovenous malformations (AVMs) located in an eloquent region or for large AVMs and using SRS alone for the other AVMs. Methods and Materials: Included in this study were 75 AVMs in 72 patients, with a mean follow-up of 52 months. Of the 75 AVMs, 33 were located in eloquent regions or were >2.5 cm in maximal diameter and were given 25–35 Gy (mean, 32.4 Gy) in four daily fractions at a single isocenter if the patient agreed to prolonged wearing of the stereotactic frame for 5 days. The other 42 AVMs were treated with SRS at a dose of 15–25 Gy (mean, 24.1 Gy) at the isocenter. The 75 AVMs were classified according to the Spetzler-Martin grading system; 21, 23, 28, 2, and 1 AVM were Grade I, II, III, IV, V, and VI, respectively. Results: The overall actuarial rate of obliteration was 43% (95% confidence interval [CI], 30 –56%) at 3 years, 72% (95% CI, 58 – 86%) at 5 years, and 78% (95% CI, 63–93%) at 6 years. The actuarial obliteration rate at 5 years was 79% for the 42 AVMs <2.0 cm and 66% for the 33 AVMs >2 cm. The 5- and 6-year actuarial obliteration rate was 61% (95% CI, 39 – 83%) and 71% (95% CI, 47–95%), respectively, after HSRT and 81% (95% CI, 66 –96%) and 81% (95% CI, 66 –96%), respectively, after SRS; the difference was not statistically significant. Radiation-induced necrosis was observed in 4 subjects in the SRS group and 1 subject in the HSRT group. Cyst formation occurred in 3 patients in the SRS group and no patient in the HSRT group. A permanent symptomatic complication was observed in 3 cases (4.2%), and 1 of the 3 was fatal. All 3 patients were in the SRS group. The annual intracranial hemorrhage rate was 5.5–5.6% for all patients. Conclusion: Our treatment policy using SRS and HSRT was as effective as the policy involving SRS alone. The HSRT schedule was suggested to have a lower frequency of radiation necrosis and cyst formation than the high-dose SRS schedule. The benefit of HSRT compared with lower dose SRS has not yet been determined. © 2004 Elsevier Inc. Stereotactic irradiation, Arteriovenous malformation, Hypofractionation.
INTRODUCTION
AVMs without causing radiation damage to the normal tissue (1, 2). We have used hypofractionated stereotactic radiotherapy (HSRT) as an alternative treatment for AVMs in eloquent regions or large AVMs. Single-fraction SRS was used as a basic treatment method for AVMs located in noneloquent regions or small ones. Our preliminary results have already been published, suggesting that HSRT was as effective as
Single-fraction stereotactic radiosurgery (SRS) is an effective treatment for patients with intracranial arteriovenous malformations (AVMs) with a diameter of about ⱕ2.5 cm. However, AVMs ⬎2.5 cm in mean diameter, or around 10 cm3 in volume, and AVMs in eloquent regions are often not cured, because we cannot give an efficient dose to the
Japan. Acknowledgments—We appreciate the help of the staff of the Departments of Neurosurgery and Radiology, Hokkaido University Hospital. We also appreciate Drs. Takeshi Nishioka and Rikiya Onimaru, for their help in preparing this paper. Received Sep 29, 2003, and in revised form Mar 8, 2004. Accepted for publication Apr 12, 2004.
Reprint requests to: Hiroki Shirato, M.D., Department of Radiology, Hokkaido University School of Medicine, North-15 West-7, Sapporo 060-8638, Japan. Tel: (⫹81) 11-706-7876; Fax: (⫹81) 11-706-5977; E-mail: [email protected] Parts of this paper were presented at the ISRS 2003 in Kyoto, Japan. Partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, and Technology of 861
862
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 3, 2004
Fig. 1. Arteriovenous malformation (AVM) treatment protocol selection between stereotactic radiosurgery (SRS) and hypofractionated stereotactic radiotherapy (HSRT).
SRS, with possibly a lower complication rate (3). However, the study included only 27 SRS and 26 HSRT patients (3). The present study was not a randomized study involving SRS and HSRT. The efficacy of HSRT was investigated, taking account the selection bias of the treatment protocol. The appropriateness of our treatment policy of stereotactic irradiation (STI) using both SRS and HSRT is discussed. METHODS AND MATERIALS Irradiation schedule Stereotactic irradiation was SRS or HSRT. The hypofractionation schedule in this study was determined using the linear-quadratic formula for late complications without correcting for slow repair (4). The ␣/ ratio of the normal brain was estimated to be 2.0, which is widely accepted with regard to late radiation effects against brain parenchyma (5). A dose of 35 Gy in four fractions and 28 Gy in four fractions would have the same effect as 18.4 Gy and 14.2 Gy, respectively, in a single exposure of radiation, assuming an ␣/ ratio of 2.0. An isocenter dose of 35 Gy and a peripheral dose of 28 Gy delivered in four daily fractions was prescribed as the standard dose in this study. Patients This study included 72 consecutive patients with angiography-proven AVMs who had agreed to undergo STI between January 1991 and December 2000. Patients initially presented with one or more of the following symptoms: episode of intracranial hemorrhage (n ⫽ 37, 51%), epileptic seizure (n ⫽ 12, 17%), and neurologic deficits (n ⫽ 46,
64%). In 19 patients, the AVM was discovered incidentally without any symptoms (n ⫽ 13), with symptoms unrelated to AVM (n ⫽ 1), or with symptoms from other disease states (subarachnoid hemorrhage in 1, facial hemangioma in 1, traffic accident in 2, and brain infarction in 1). Before STI, 11 patients had undergone surgical resection of the AVM. Embolization was performed in 7 patients before STI. Of 11 patients with coincidental intracranial aneurysms, 3 were treated by clipping the aneurysm before STI. All patients underwent digital subtraction angiography and CT. MRI and magnetic resonance angiography were used in the patients treated since the late 1990s. The AVMs were classified according to the Spetzler-Martin grading system (6). Treatment protocol—selection between SRS and HSRT The recursive tree used for the selection of the treatment method in the 75 AVMs is shown in Fig. 1. If the AVM nidus was primarily located in eloquent regions, such as the sensory-motor region, language region, and visual cortex; hypothalamus and thalamus; internal capsule; brainstem; cerebellar peduncles, or deep cerebellar nuclei, HSRT was used. In addition, when the nidus was located in a noneloquent region and was ⬎2.5 cm in maximal diameter as measured by angiography, HSRT was used. However, of those with AVMs that fulfilled the above criteria and who were eligible for HSRT, 5 patients were judged not to be suited for prolonged head-frame wearing because of physical instability or mental deficiency, and 8 patients did not agree to the prolonged head-frame wearing. These 13 patients were treated with SRS instead of HSRT. Conse-
STI for AVMs
Table 1. Patient characteristics Characteristic Age (y) ⬍20 ⱖ20 Gender Male Female History of hemorrhage Yes No Prior resection Yes No Prior embolization Yes No Diameter (cm) ⱕ2.5 ⬎2.5 ⱕ2.0 ⬎2.0 Location Eloquent Noneloquent Drainage vein Deep Superior Spetzler-Martin grade I ⫹ II III ⫹ IV ⫹ V
HSRT (n)
SRS (n)
2 31
7 35
20 13
23 19
15 18
22 20
6 27
5 37
3 30
4 38
18 15 13 20
36 6 29 13
20 13
14 28
19 14
24 18
16 17
28 14
quently, 33 lesions were treated with HSRT, and 42 lesions with SRS. All 6 patients with AVMs in an eloquent area with a maximal diameter ⬎2.5 cm were treated with HSRT. Table 1 shows the patient and AVM distribution in each treatment group. Radiosurgery technique All patients were treated with an invasive stereotactic head frame (RADFRAME, Mizuho Ika Kogyo, Tokyo, Japan) for use in performing CT, MRI, angiography, and RT. The RADFRAME was developed in our institution to be suitable for prolonged wearing while maintaining 1.0-mm accuracy. When SRS was used, the patients wore the stereotactic frame for 1–2 days. When HSRT was used, patients wore the stereotactic frame for 5 days. The procedure was performed with i.v. sedation when necessary. Between 1991 and April 2003, the AVM nidus was defined using biplane angiography, and the target was displayed on CT. After 1998, MRI was also used as a complementary tool for delineation (7, 8). A single isocenter and three-dimensional RT planning system (Focus, CMS, St. Louis, MO) were used. An isodose contour conforming to the entire AVM nidus was selected as the prescription isodose, and 6- or 10-mV X-ray beams from a linear accelerator were used. In general, a multileaf collimator was used for AVMs ⬎1.5 cm, and a cone-shape collimator was used for smaller
● T.-C. CHANG et al.
863
AVMs. The arc technique was used in three lesions, and the static noncoplanar multiport technique was used for 72 lesions. The number of arcs or planes ranged from three to four, and the median number of ports was 8 (range, 3–11). Prescription The standard total dose for SRS was 25 Gy at the isocenter of the treatment target volume, and the nidus was enclosed by an 80% isodose surface, with a minimal dose of 20 Gy. HSRT was primarily performed using 35 Gy at the isocenter, with a minimal dose of 28 Gy in four daily fractions. The total doses used for SRS and HSRT were modified depending on the size and location of the AVM, as well as patient age, with patients ⬍20 years old receiving 90% of the dose described above. Thus, the total dose at the isocenter ranged from 15 to 25 Gy (mean, 24.1 Gy) for SRS and 25 to 35 Gy (mean, 32.4 Gy) for HSRT. In SRS, 29 AVMs ⬍2 cm were treated with 25 Gy in one fraction. The other 13 AVMs that were ⬎2 cm were given a lower dose to reduce the possibility of severe complications according to the isoeffect line of Kjellberg et al. (1) and Flickinger et al. (2). In HSRT, 18 AVMs received 35 Gy in four fractions, and 15 patients were treated with different doses and fractions in the early period or because of the proximity of the nidus to specifically critical structures such as the optic chiasm (4 AVMs, 25 Gy in four fractions; 1 AVM, 28 Gy in four fractions; 5 AVMs, 30 Gy in four fractions; 1 AVM, 32 Gy in four fractions; 1 AVM, 32.5 Gy in four fractions; 1 AVM, 30 Gy in eight fractions; and 2 AVMs, 35 Gy in eight fractions). The minimal dose ranged from 12 to 20 Gy (mean, 19.3 Gy) for SRS and 20 to 28 Gy (mean, 25.9 Gy) for HSRT. Follow-up evaluation For most of the patients who lived close enough to our institution, clinical examination and serial imaging studies (MRI or CT when MRI was contraindicated) were performed at 3- or 4-month intervals for the first year and every 6 months until complete obliteration was angiographically demonstrated. Some patients who lived far from our institution underwent the imaging studies near their homes and were evaluated by the referring physicians. Angiography was planned only when MRI demonstrated complete obliteration or when otherwise clinically indicated. The term “complete angiographic obliteration” was defined as normal blood flow, complete absence of pathologic vessels at the site of the nidus, and the normalization of flow in the efferent veins. The date of obliteration was defined as the date of angiography demonstrating complete obliteration. Clinical complications from STI were classified as new or aggravated neurologic symptoms or signs; complications requiring steroids, shunt placement, or surgical intervention; an imaging change suggesting radiation-induced edema (increased with a T2-weighted signal change), radiation-induced necrosis, or cyst formation; or treatment-related death.
864
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 3, 2004
Fig. 2. Actuarial obliteration-free curves according to arteriovenous malformation (AVM) diameter.
Statistical analysis For statistical analysis, we used a computer software statistical package (StatView for Windows, version 5, SAS Institute, Cary, NC). The actuarial rates of complete obliteration or T2-weighted signal changes or postradiosurgical hemorrhage were calculated by the Kaplan-Meier method. The overall actuarial rate of obliteration was calculated using the total number of patients at risk for the denominator and the number of patients with angiographically demonstrated complete obliteration for the numerator. The log– rank test was used to compare the outcomes of the different groups. Univariate analysis of the clinical and angiographic AVM factors was performed using the log–rank test. The rates of post-STI hemorrhage were calculated from the time of STI using the Kaplan-Meier method, and patients were considered to be at risk of hemorrhage until the date of their last follow-up. RESULTS Follow-up All 72 patients had clinical follow-up after STI (mean, 52 months; median, 48 months; range, 0 –138 months). One patient did not appear for follow-up at all but was counted as a patient at risk of complete obliteration and was also used for the analysis of acute complications. One died of an unknown cause after angiographic confirmation of AVM obliteration and was counted as having angiographic obliteration. After angiographic confirmation of no obliteration, 1 patient died of bile duct adenocarcinoma and 1 patient
died of treatment-related complications (details given below). These 2 patients were counted as angiographic failures in the analysis of obliteration. Four died before angiographic examination (rebleeding in 1, brain infarction in 1, and unknown causes in 2) and were counted as having been withdrawn before angiographic obliteration. Obliteration After STI, 57 of 75 AVMs were examined using angiography, and complete obliteration of the nidus was confirmed in 39 AVMs between 12 and 90 months (mean, 36 months; median, 32 months). Patients in whom the last angiographic study demonstrated partial obliteration or a residual presence of the AVM were considered to have treatment failure. The overall actuarial rate of obliteration was 43% (95% confidence interval [CI], 30 –56%) at 3 years, 72% (95% CI, 58 – 86%) at 5 years, and 78% (95% CI, 63–93%) at 6 years. The actuarial obliteration rate for the 42 AVMs ⱕ2.0 cm in maximal diameter (66%) was 50% (95% CI, 32– 68%) at 3 years and 79% (95% CI, 62–98%) at 5 years. For the 33 AVMs with a maximal diameter ⬎2 cm (44%), the actuarial complete obliteration rate was 44% (95% CI, 26 – 62%) at 3 years and 66% (95% CI, 45– 87%) at 5 years (p ⫽ 0.2674; Fig. 2). Similarly, the 3-year and 5-year complete obliteration rate was 39% (95% CI, 45– 87%) and 57% (95% CI, 45– 87%), respectively, for AVMs ⬎2.5 cm. Figure 2 shows that the larger AVMs tended to become obliterated later. The 3-year and 5-year complete obliteration rate of AVMs in eloquent regions was 51% (95% CI, 32–70%) and 86% (95% CI, 70 –100%), respectively.
STI for AVMs
● T.-C. CHANG et al.
865
Fig. 3. Actuarial obliteration-free curves according to treatment schedule.
The 3-year, 5-year, and 6-year actuarial complete obliteration rate was 32% (95% CI, 19 – 49%), 61% (95% CI, 39 – 83%), and 71% (95% CI, 47–95%), respectively, for patients treated with HSRT and followed angiographically for 21–90 months (mean, 37 months; median, 27 months) and was 52% (95% CI, 35– 69%), 81% (95% CI, 66 –96%), and 81% (95% CI, 66 –96%), respectively, for patients treated with SRS and followed angiographically for 11– 60 months (mean, 32 months; median, 31 months). The complete obliteration rates were not significantly different statistically between the HSRT and SRS groups (p ⫽ 0.1401; Fig. 3). The 3-year and 5-year complete obliteration rate was 26% (95% CI, 0 –52%) and 53% (95% CI, 18 – 88%), respectively, for AVMs ⬎2.5 cm treated with HSRT. The 3-year and 5-year complete obliteration rate of AVMs in eloquent regions treated with HSRT was 35% (95% CI, 12–58%) and 80% (95% CI, 56 –100%), respectively. Of the 16 AVMs with incomplete obliteration or residual AVM confirmed by angiography, 1 patient in the SRS group and 7 patients in the HSRT group were retreated with STI. For the second treatment, we used SRS with varying doses depending on the initial radiosurgical dose and the maximal tolerance dose of the brain. Of these 8 patients, 3 (1 in the SRS and 2 in the HSRT group) experienced obliteration. If we assume that the retreatment was not a failure, the obliteration rate at 3, 5, and 6 years was 43% (95% CI, 30 – 56%), 70% (95% CI, 57– 83%), and 76% (95% CI, 63– 99%), respectively. Univariate analysis of the possible prognostic factors for obliteration showed that the variables of treatment method (SRS, HSRT), age, gender, history of hemorrhage, prior
treatment (resection, embolization), and AVM characteristics (location, diameter, drainage vein, Spetzler-Martin grade) did not reach statistically significant levels (Table 2). Complications Of the 67 patients periodically examined using MRI, 39 (58%) exhibited a newly increased high signal surrounding their AVM site on T2-weighted MRI. The 3-year actuarial rate for developing post-STI T2-weighted signal changes was 64% (95% CI, 45–73%) for HSRT and 59% (95% CI, 42–76%) for SRS (Fig. 4). The cumulative actuarial rates for post-STI T2-weighted signal changes were not significantly different statistically between the HSRT and SRS groups. Of the 72 patients, 9 (9.7%) experienced transient symptomatic adverse effects after STI, 7 of which were related to the T2-weighted signal changes. All 9 patients recovered without a symptomatic deficit. Of the 72 patients, 3 (4.2%) had permanent radiationinduced symptomatic adverse effects. One (1.4%) of the three permanent adverse effects was fatal. This patient had status epilepticus after surgery for a large AVM and had maintained a good performance status after STI for 47.6 months. However, the epilepsy appeared again after cyst formation due to the STI. She died of status epilepticus owing to the expansion of a radiation-related cyst 1 week before the planned surgery for the cyst. The other 2 patients had chronic neurologic deficits (double vision in one and slight paresis of the left leg in the other). In 71 patients with MRI and/or CT follow-up, 5 (7.0%) were diagnosed with radiation necrosis. In the SRS group, 4 patients experienced radiation-related necrosis, 3 of which
866
I. J. Radiation Oncology
● Biology ● Physics
Table 2. Various parameters for obliteration rates: univariate analysis Variable Age (y) ⬍20 ⱖ20 Gender Male Female History of hemorrhage Yes No Prior resection Yes No Prior embolization Yes No Treatment SRS HSRT Diameter (cm) ⱕ2.0 ⬎2.0 ⱕ2.5 ⬎2.5 Location Eloquent Noneloquent Drainage vein Deep Superior Spetzler-Martin grade I ⫹ II III ⫹ IV ⫹ V
n
5-y Obliteration rate (95% CI)
p
9 66
56 ⫾ 33 73 ⫾ 14
0.5698
43 32
75 ⫾ 18 69 ⫾ 20
0.73
37 38
91 ⫾ 16 67 ⫾ 19
0.7712
11 64
72 ⫾ 30 73 ⫾ 15
0.8463
7 68
54 ⫾ 44 75 ⫾ 14
0.0775
42 33
81 ⫾ 15 61 ⫾ 22
0.1401
42 33 54 21
79 ⫾ 17 66 ⫾ 21 77 ⫾ 15 57 ⫾ 17
34 41
86 ⫾ 16 59 ⫾ 20
0.1381
43 32
71 ⫾ 17 75 ⫾ 22
0.8721
44 31
72 ⫾ 18 73 ⫾ 27
0.963
0.2674 0.2324
Abbreviation: CI ⫽ confidence interval; other abbreviations as in Table 2.
Volume 60, Number 3, 2004
were in patients with permanent radiation-induced symptomatic adverse effects. Radiation-related cyst formation was observed 2–3 years later in 3 patients with radiation necrosis after SRS. All 3 patients were to be treated in the HSRT group according to our decision tree but underwent SRS instead of HSRT (Table 3). In 33 patients who were treated with HSRT, 1 case of necrosis was observed that was symptomatically reversible. Seven patients (18%) of the 39 with T2-weighted signal changes and 4 patients (12%) of the 33 without T2-weighted signal changes experienced radiation-induced symptomatic complications.
Hemorrhage Post-STI hemorrhage was observed in 10 patients: 4 in the SRS group at a mean of 20 months (range, 1–21 months) and 6 in the HSRT group at a mean of 33 months (range 2–93). Of these 10 patients, 3 experienced neurologic deterioration owing to bleeding of the AVM after STI. One patient experienced hemiparesis after SRS. One experienced seizure and left homonymous hemianopsia after HSRT. One patient died just after AVM bleeding at 14 months after HSRT. The 5-year actuarial accumulative intracranial hemorrhage rate after STI was 15.2% (95% CI, 5.4 –25.0% ) for all patients combined, 22.3% (95% CI, 5.8 –38.8% ) for HSRT patients, and 7.8% (95% CI, 0 –16.2%) for SRS patients. The annual intracranial hemorrhage rate was 5.6% during the first year and 5.5% during the second year for all patients combined. It was 9.2% and 3.1% for HSRT patients at 1 and 2 years after STI and 3.6% and 4.4% for SRS patients, respectively. No statistically significant difference was found between the HSRT and SRS groups in the postradiosurgical hemorrhage rate (p ⫽ 0.2442). The actuarial accumulative intracranial hemor-
Fig. 4. T2-weighted signal change curves according to treatment schedule.
STI for AVMs
● T.-C. CHANG et al.
867
Table 3. Results of treating with SRS instead of HSRT Age Pt. (y)/ Prior medical No. Gender KPS history
Prior treatments (n)
Location Maximal (Spetzlerdiameter Martin grade) (cm)
1
53/M
90
Headache (incident)
None
Right basal ganglia (III)
2.2
2
16/F
90
None
39/M
90
Left frontal (I) Left thalamus (III)
2.7
3
Headache, nausea Headache, left limb numbness
4
19/F
60
Hemorrhage, coma
Surgery
Right frontal (II)
3.5
5
30/M
80
None (incident)
None
6
29/F
80
Hemorrhage
7
29/F
90
Right thalamus (III) Embolization Left thalamus (2) (III) None Right frontal (motor area) (III)
8
41/M
70
Headache, right internal carotid cave aneurysm (unruptured) Hemorrhage, Embolization hemiparesis (8)
9
16/F
90
Hemorrhage, coma
10
55/M
90
11
46/M
70
12
17/F
90
13
78/M
80
None None (incident), BuddChiari syndrome Hemorrhage, Ventricular seizure, drainage coma Hemorrhage, Ventricular seizure, drainage, coma VP shunt None Multiple brain infarction, Parkinson disease, hypertension, viral hepatitis
None
Ventricular drainage
0.9
Reason for Dose Angiographic SRS (cGy) obliteration Disagreement 2500
Child
Residual
2000
Complete
Disagreement 2500
Complete
Symptomatic complications and treatment Radiation necrosis, cyst formation, steroids, glycerol infusions, mild headache None
2250
Residual
2.3
Disagreement 2000
Residual
Radiation necrosis, cyst formation, steroids, glycerol infusions, hyperbaric oxygen therapy, tremor, seizure, persistent hemiparesis Radiation necrosis, cyst formation, steroids, glycerol infusions, seizure, death from status epilepticus None
2.0
Disagreement 2500
Complete
None
2.0
Disagreement 2500
Complete
None
Left corpus callosum, frontal (III) Left paraventricle (III) Right internal caudate head (III)
3.4
Disagreement 1500
Residual
None
2250
Complete
None
3.5
Disagreement 2500
Complete
None
Left thalamus (III)
1.3
Disagreement 2000
—
None
Left basal ganglia (III) Left basal ganglia (III)
1.0
Disagreement 2500
Complete
None
0.5
Physical 2500 instability mental deficiency
—
None
2.0
Physical instability
Child
Abbreviations: SRS ⫽ stereotactic radiosurgery; HSRT ⫽ hypofractionated stereotactic radiotherapy; Pt. No. ⫽ patient number; M ⫽ male; F ⫽ female.
868
I. J. Radiation Oncology
● Biology ● Physics
rhage rate at 5 years for AVMs ⱕ2.0 cm was 12.4% (95% CI, 0.6 –23.2%) and for AVMs ⬎2.0 cm was 17.7% (95% CI, 4 –30.4%). The difference did not reach statistical significance. Of the 13 patients who had been treated with SRS instead of HSRT, 7 (53%) had confirmed complete obliteration and 4 had partial obliteration or residual AVM by angiography (Table 3). Radiation necrosis and cyst formation was observed in 3 patients each (23%). Two patients experienced intracranial hemorrhage after treatment. DISCUSSION The primary subject of our study was to examine the appropriateness of our treatment policy compared with previous reports in which SRS alone was used. Particularly for AVMs ⬎2.0 –2.5 cm (44% ⬎ 2.0 cm and 29% ⬎ 2.5 cm in our series) and AVMs in eloquent areas (45% of our series), making a comparison with the previous report is quite important. However, we found that it was quite difficult to compare the results of STI for AVMs among different institutions. The crude obliteration rates of AVMs after SRS vary significantly from 40% to 85% at 3 years and 33% to 93% at 5 years (9 –24). If we adopt these crude obliteration rates using the number of patients who underwent angiography as the denominator, the crude complete obliteration rates for our treatment were 92% (12 of 13) at 3 years for 33 AVMs with a maximal diameter ⬎2 cm. This crude obliteration rate was better than that observed in previous papers but obviously was dependent on the selection bias and was far from a determination of the actual outcome of treatment. Because recent improvements in MRI and magnetic resonance angiography have helped to preclude unnecessary confirmatory angiography for patients with patent AVMs, the numerator can be very close to the denominator without an improvement in the treatment method. The number of reports using the Kaplan-Meier method for the analysis of the obliteration rate is increasing. Miyawaki et al. (23) have used the Kaplan-Meier method and reported a 33% 5-year actuarial rate of angiographic obliteration, assuming MRI/magnetic resonance angiography obliteration and retreatment to be failures, as we did. Touboul et al. (25) have reported 62.5% ⫾ 7% as the 5-year actuarial obliteration rate for 100 AVM patients. Pan et al. (20) reported a 75% actuarial complete obliteration rate at 40 months. Our results showed a 72% (95% CI, 58 – 86%) overall actuarial obliteration rate. Considering that 44% of the patients had AVMs ⬎2.0 cm in our series, the treatment outcome of our study is at least comparable with those of the previous studies. The decision tree in our treatment policy of STI, including HSRT, was not detrimental for cerebral AVMs. The treatment method was not a prognostic factor in this analysis, even though larger AVMs were more frequent in HSRT. The diameter of AVM was not a statistically significant prognostic factor for angiographic obliteration in this
Volume 60, Number 3, 2004
series. This latter result was not seen in our preliminary study (17), in which size (ⱕ2 cm or not) was a statistically significant favorable factor (p ⫽ 0.05). The finding that diameter was not a statistically significant factor may have been because the number of patients with occluded large AVMs increased in accordance with the longer follow-up period. Younger age was suggested to be a statistically significant factor in our previous study, but age also disappeared as a statistically significant factor in the present study. This result may also have been because the patients with AVMs ⬎2.0 cm who experienced complete obliteration with HSRT were older, with a median age of 45 years (range, 18 –74 years). Late radiation-induced symptomatic injury occurred in 3–18% of the patients with AVMs in the previous reports (12, 15, 16, 22, 23, 27, 28). Symptomatic complications appeared in the patients with larger AVMs and those who received higher doses of SRS as predicted by the radiobiologic volume effect models of the central nervous system (29). Steinberg et al. reported that 50% (10 of 20) had major or minor complications with doses ⬎18 Gy and volumes ⬎13 cm3 (24). Miyawaki et al. (23) reported that radiation necrosis occurred when minimal doses of 18 Gy and maximal doses of 34 Gy were administered to treatment volumes ⱖ14 cm3 (mean, 27.8 cm3). Flickinger et al. (26) found that eloquent location and volume irradiated to ⱖ12 Gy were significant risk factors for symptomatic radiation necrosis. However, in our series, symptomatic radiation injury was not more frequent in patients with larger AVMs or in patients with AVMs in eloquent regions. Radiation necrosis and cyst formation were observed in 4 of 42 in the SRS group and 1 of 33 in the HSRT group, the latter of which included larger AVMs. All three cyst formations were seen in patients who should have been treated by HSRT in our protocol but were treated with SRS instead. We decreased the SRS dose for these patients, in principle, but the dose reduction might have been too small. However, if we had used a lower dose, the obliteration rate in these patients might have been lower. We have used the same hypofractionated schedule for 140 brain metastases with a mean diameter of 6.7 cm (range, 0.006 – 48.3 cm) in 75 patients. We found that the schedule was as safe as we have seen in the present study with a radiation necrosis rate of 2.7% (2 patients) (30). These brain metastasis results and the results of the present AVM study suggest that our hypofractionation schedule is safe and close to the maximal tolerance dose. The maximal tolerance dose is consistent with the prediction made by the radiobiologic model of Brenner et al. (4), assuming an ␣/ ratio of 2.0 (3). The annual bleeding rate after STI was 5.5–5.6% in this series, within the range of the natural history of untreated AVMs, which runs from 2% to 17% annually (31–33). The number of AVMs that bled after STI was greater in the HSRT group (6 of 33) than in the SRS group (4 of 42) in this series. Also, the bleeding rate in the first year for HSRT was three times the rate of the SRS patients, although the
STI for AVMs
difference did not reach statistical significance. Stefani et al. (34), in prospective analyses of follow-up data from 390 patients with brain AVMs, found that large and deep-seated AVMs were the most important and significant factors associated with a greater risk of future hemorrhagic events. The greater bleeding rate in the first year in the HSRT group may have been related to the larger size of the AVMs. Steiner et al. (24) indicated that patients treated with radiosurgery may be at a decreased risk of hemorrhage once obliteration has effectively occurred. In our series, the actuarial obliteration rates for the HSRT group increased after 2 years. This might have been in relation to the decrease in the hemorrhage rate from 9.2% to 3.1% during the first 2 years of follow-up. The small number of patients prevented us from speculating more about the difference. For large AVMs, staged SRS has been shown to be an effective approach (35). In this approach, SRS is repeated during an interval of 5–36 months. Because we have also experienced good obliteration rates after repeated treatment of patient AVMs, we agree that staged STI may be one of the treatment options for large AVMs. This does not preclude the potential importance of HSRT as the initial treatment for large AVMs. Each approach should be considered one of several options for treating this difficult disease.
● T.-C. CHANG et al.
869
CONCLUSION Our treatment policy using SRS and HSRT was effective in producing an obliteration rate and complication rate compatible with those found in previous SRS studies. HSRT may have a wider therapeutic window in the treatment of AVMs ⱖ2.5 cm or AVMs in eloquent areas, because the selection bias in our treatment decision tree could favorably affect SRS but not HSRT. Some might conclude that the SRS dose in this series (minimal dose range, 12–20 Gy; mean, 19.3), was too high and that HSRT possibly has a lower obliteration rate than SRS. The current study does not answer the question of whether SRS using lower doses such as 15–18 Gy at the periphery of the AVM can achieve results similar to HSRT. We also cannot answer the question of whether HSRT has a role in decreasing late radiation necrosis for small AVMs and AVMs in noneloquent areas. Only a prospective randomized trial will answer these questions. An optimal hypofractionation schedule may be worth additional investigation, although the optimal dose distribution and rigid fixation continue to be the critical requirements for STI of AVMs in the central nervous system (36).
REFERENCES 1. Kjellberg RN, Hanamura T, Davis KR, et al. Bragg-peak proton-beam therapy for arteriovenous malformations of the brain. N Engl J Med 1983;309:269–274. 2. Flickinger JC, Schell MC, Larson DA. Estimation of complications for linear accelerator radiosurgery with the integrated logistic formula. Int J Radiat Oncol Biol Phys 1990;19:143– 148. 3. Aoyama H, Shirato H, Nishioka T, et al. Treatment outcome of single or hypofractionated single-isocentric stereotactic irradiation (STI) using a linear accelerator for intracranial arteriovenous malformation. Radiother Oncol 2001;59:323–328. 4. Brenner DJ, Martel MK, Hall EJ. Fractionated regimens for stereotactic radiotherapy of recurrent tumors in the brain. Int J Radiat Oncol Biol Phys 1991;21:819–824. 5. Bentzen SM, Baumann M. The linear-quadratic model in clinical practice. In: Steel GG, editor. Basic clinical radiobiology. London: Arnold; 2002. p. 134–146. 6. Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg 1986;65:476–483. 7. Kondziolka D, Dempsey PK, Lunsford LD, et al. A comparison between magnetic resonance imaging and computed tomography for stereotactic coordinate determination. Neurosurgery 1992;30:402–406. 8. Aoyama H, Shirato H, Nishioka T, et al. Magnetic resonance imaging system for three-dimensional conformal radiotherapy and its impact on gross tumor volume delineation of central nervous system tumors. Int J Radiat Oncol Biol Phys 2001; 50:821–827. 9. Yamamoto M, Jimbo M, Hara M, et al. Gamma knife radiosurgery for arteriovenous malformations: Long-term follow-up results focusing on complications occurring more than 5 years after irradiation. Neurosurgery 1996;38:906–914. 10. Yamamoto Y, Coffey RJ, Nichols DA, et al. Interim report on the radiosurgical treatment of cerebral arteriovenous malfor-
11.
12.
13.
14.
15.
16.
17.
18.
19. 20.
mations: The influence of size, dose, time, and technical factors on obliteration rate. J Neurosurg 1995;83:832–837. Shin M, Kawamoto S, Kurita H, et al. Retrospective analysis of a 10-year experience of stereotactic radio surgery for arteriovenous malformations in children and adolescents. J Neurosurg 2002;9:779–784. Colombo F, Pozza F, Chierego G, et al. Linear accelerator radiosurgery of cerebral arteriovenous malformations: An update. Neurosurgery 1994;34:14–20. 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;46:1135–1142. Hadjipanayis CG, Levy EI, Niranjan A, et al. Stereotactic radiosurgery for motor cortex region arteriovenous malformations. Neurosurgery 2001;48:70–76. Pollock BE, Gorman DA, Coffey RJ. Patient outcomes after arteriovenous malformation radiosurgical management: Results based on a 5- to 14-year follow-up study. Neurosurgery 2003;52:1291–1297. Steinberg GK, Fabrikant JI, Marks MP, et al. Stereotactic heavy-charged-particle Bragg-peak radiation for intracranial arteriovenous malformations. N Engl J Med 1990;323:96– 101. Aoyama H, Shirato H, Nishioka T, et al. Treatment outcome of single or hypofractionated single-isocentric stereotactic irradiation (STI) using a linear accelerator for intracranial arteriovenous malformation. Radiother Oncol 2001;59:323–328. Friedman WA, Bova FJ, Mendenhall WM. Linear accelerator radiosurgery for arteriovenous malformations: The relationship of size to outcome. J Neurosurg 1995;82:180–189. Friedman WA, Bova FJ. Linear accelerator radiosurgery for arteriovenous malformations. J Neurosurg 1992;77:832–841. Pan DH, Guo WY, Chung WY, et al. Gamma knife radiosurgery as a single treatment modality for large cerebral arterio-
870
21. 22.
23.
24. 25.
26.
27.
I. J. Radiation Oncology
● Biology ● Physics
venous malformations. J Neurosurg 2000;93(Suppl. 3):113– 119. Lunsford LD, Kondziolka D, Flickinger JC, et al. Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg 1991;75:512–524. Engenhart R, Wowra B, Debus J, et al. The role of high-dose single-fraction irradiation in small and large intracranial arteriovenous malformations. Int J Radiat Oncol Biol Phys. 1994; 15:30:521–529. 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: 1089–1106. Steiner L, Lindquist C, Adler JR, et al. Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 1992;77:1–8. 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;41:855–861. for the Arteriovenous Malformation Radiosurgery Study Group, Flickinger JC, Kondziolka D, Lunsford LD, et al. Development of a model to predict permanent symptomatic postradiosurgery injury for arteriovenous malformation patients. Int J Radiat Oncol Biol Phys 2000;46:1143–1148. Flickinger JC, Lunsford LD, Kondziolka D, et al. Radiosurgery and brain tolerance: An analysis of neurodiagnostic imaging changes after gamma knife radiosurgery for arteriovenous malformations. Int J Radiat Oncol Biol Phys 1992;23: 19–26.
Volume 60, Number 3, 2004
28. Yamamoto M, Hara M, Ide M, et al. Radiation-related adverse effects observed on neuro-imaging several years after radiosurgery for cerebral arteriovenous malformations. Surg Neurol 1998;49:385–398. 29. Shirato H, Mizuta M, Miyasaka K. A mathematical model of the volume effect which postulates cell migration from unirradiated tissues. Radiother Oncol 1995;35:227–231. 30. Aoyama H, Shirato H, Onimaru R, et al. Hypofractionated stereotactic radiotherapy alone without whole-brain irradiation for patients with solitary and oligo brain metastasis using noninvasive fixation of the skull. Int J Radiat Oncol Biol Phys 2003;56:793–800. 31. Itoyama Y, Uemura S, Ushio Y, et al. Natural course of unoperated intracranial arteriovenous malformations: Study of 50 cases. J Neurosurg 1989;71:805–809. 32. Fults D, Kelly DL, Jr. Natural history of arteriovenous malformations of the brain: A clinical study. Neurosurgery 1984; 15:658–662. 33. Brown RD, Jr, Wiebers DO, Forbes G, et al. The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 1988;68:352–357. 34. Stefani MA, Porter PJ, terBrugge KG, et al. Large and deep brain arteriovenous malformations are associated with risk of future hemorrhage. Stroke 2002;33:1220–1224. 35. Pendl G, Unger F, Papaefthymiou G, et al. Staged radiosurgical treatment for large benign cerebral lesions. J Neurosurg 2000;93(Suppl. 3):107–112. 36. Hida K, Shirato H, Isu T, et al. Focal fractionated radiotherapy for intramedullary spinal arteriovenous malformations: 10year experience. J Neurosurg 2003;99(1 Suppl.):34 –38.