In!. _I. Radiatmn Oncology Biol. Phys.. Vol. Printed in the U.S.A. All rights reserved.
0360-3016/90 $3.00 + .oO Copyright 0 1990 Pergamon Press plc
19, pp. 775-782
0 Technical Innovations and Notes RADIOSURGERY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS WITH THE DYNAMIC STEREOTACTIC IRRADIATION LUIS SOUHAMI, ERVIN
B. PODGORSAK,
M.D.,* ANDRE PH.D., AND
OLIVIER,
F.C.C.P.M.,*
M.D.,
MARINA
PH.D.,
F.R.C.S.(C),+
PLA, M.Sc.,
M.C.C.P.M.*
G. BRUCE PIKE, M.E.E.+
McGill University, Montrkal, Canada From December 1986 through December 1988, 33 patients with inoperable arteriovenous malformation (AVM) were treated in our center with the dynamic stereotactic radiosurgery, which uses a standard 10 MV isocentric
linear accelerator. There were 18 females and 15 males with a median age of 26 years (range: 9-69) and a median follow-up time of 16 months (range: 7-32). The arteriovenous malformation volumes treated ranged from 0.2 to 42 cm3. The prescribed doses at the isocenter varied from 50 to 55 Gy and were given as a single fraction in the majority of the patients (31/33). Late complications consisting of intracranial bleeding and/or hemiparesis were observed in three patients. To date, 21 patients underwent repeat angiographic studies at 1 year post-treatment. A complete obliteration of the lesion was achieved in 38% of these patients. For the patients whose arteriovenous malformation nidus was covered by a minimum dose of 25 Gy, the total obliteration rate was 61.5% (g/13), whereas none of the patients who had received less than 25 Gy at the edge of the nidus obtained a total obliteration. Our preliminary analysis at 1 year post-radiosurgery reveals results comparable to those previously reported for other radiosurgical techniques for the same follow-up period. Stereotactic radiosurgery, Radiosurgery, Arteriovenous malformation, Dynamic rotation, Radiosurgery with linear accelerator. INTRODUCXION
Since the pioneering work of Leksell in 195 1 ( 15), radiosurgery has become an attractive form of therapy for intracranial arteriovenous malformations (AVM) (29, 30) and other selected intracranial lesions (6, 11, 16). Initially, radiosurgery was performed with heavy charged particle beams from cyclotrons (9, 10, 13) and later by the Gamma unit* (32). The well proven efficacy of radiosurgery has in recent years led to the development of linear accelerator (linac) based radiosurgical techniques (1, 4, 7, 8, 18, 26), making radiosurgery potentially available to most major radiation oncology centers. Several centers have developed or are developing linacbased techniques. It is estimated that by early 1990, approximately 30-40 linac-based facilities will be operational in the U.S. (12). The linac-based radiosurgical techniques
Presented at the 3 1st Annual Meeting of the American Society for Therapeutic Radiology and Oncology, San Francisco, CA, 1-6 October, 1989. * Dept. of Radiation Oncology. + Dept. of Neurosurgery. Reprint requests to: L. Souhami, M.D., Department of Radiation Oncology, Montreal General Hospital, 1650 Cedar Ave., Montreal, QuCbec, H3G IA4 Canada.
reported so far are: (a) single plane rotation (8), (b) multiple non-coplanar converging arcs ( 1, 4, 7, 1S), and (c) dynamic rotation (25, 26). A comparison study among the high energy photon beam radiosurgical techniques has been published recently (27). Several important criteria should be fulfilled before any new radiosurgical technique is used in clinical practice, including high spatial and numerical accuracy of dose delivery to the target, steep dose fall-off outside the target volume, and knowledge of dose distribution within the target volume. Recently, Podgorsak et al. (28) have studied the adequacy of linacs in performing radiosurgery. They have shown that two of the existing linac-based radiosurgical techniques (multiple non-coplanar converging arcs and dynamic rotation) meet these criteria and can be considered an adequate and less expensive option to the Gamma unit and/or proton beam radiosurgery. In
Acknowledgements-The authors wish to thank Drs. G. Bertrand and R. Leblanc for helpful discussions and referral of some of the patients and Ms. Jennifer Manal for dedicated secretarial work. Accepted for publication 29 March 1990. * Leksell Gamma unit, Electa Instrument AB, Stockholm, Sweden.
776
I. J. Radiation Oncology 0 Biology 0 Physics
this paper we report our initial results in the first 33 patients with AVM treated with the dynamic rotation at McGill University in Montreal. METHODS
AND MATERIALS
From December 1986 through December 1988,33 patients with AVM underwent stereotactic radiosurgery in our center. All patients were treated with the dynamic rotation, described in detail previously (2526). A 10 MV isocentric linac* with a remotely controlled couch rotation is used as the source of radiation. The gantry rotates from 30” to 330”, while simultaneously the couch rotates from 75” to -75” (Fig. 1). The circular beam converges at the target volume, but because of the simultaneous couch and gantry rotations, the entrance beam never coincides with an exit beam, yielding a uniform dose distribution in the target volume and a reasonably sharp dose fall-off
September 1990, Volume 19, Number 3
outside the target volume. Target localization, treatment set up, and patient immobilization during the treatment are accomplished with a stereotactic frame developed locally (2 1, 22), and compatible with computerized tomography scan (CT), magnetic resonance imaging (MRI), and digital subtraction angiography (DSA). All patients underwent MRI and DSA studies before radiosurgery to define the location and size of the AVM as well as its relationship to sensitive anatomical structures. MRI exams were performed on a 1.5 Tesla whole body system and consisted of short repetition time spin echo (T 1-weighted) acquisitions in the transverse, sagittal, and coronal orientations. The body coil was used in all studies since the stereotactic frame does not easily fit within a standard head coil. However, with slice thicknesses of 7.5 mm, good quality images were obtained with one signal measurement. We have achieved a localization accuracy off 1 mm using DSA and transverse MR images
Fig. 1. The dynamic stereotactic radiosurgical procedure starts with the couch at 75” and the gantry at 30”, as shown in Fig. IA. During treatment, the couch rotates 150” from +75’ to -75”, while the gantry simultaneously rotates 300” from 30” to 330”. Thus, each degree of couch rotation corresponds to two degrees of gantry rotation. Several sucessive positions through which the couch and gantry move during the complete radiosurgical procedure are shown, starting (A) with the gantry and couch angles of 30” and +75’, respectively, through (B) 90” and +45”, respectively, (C) 180” and O”, respectively, (D) 270” and -45”, respectively, and (E) stopping at 330” .and -75”, respectively.
* Clinac- 18, Varian Associates,
Palo Alto, CA.
Dynamic stereotactic radiosurgery 0 L. SOUHAMI efal.
(in plane); however, our sagittal and coronal MR images have been shown to be geometrically less reliable (23). DSA provides a good definition of the dynamic compartments of the AVM, namely the arterial feeders, the nidus, and the draining veins. MRI gives the picture of the bulk of the AVM without differentiating between arteries, capillaries, or veins. It delineates, however, the topographic information concerning vital cerebral structures neighboring the AVM (20). Following treatment, MRI was repeated every 3 months and DSA was obtained after 12 months or sooner if the MRI study demonstrated a major change in the lesion signal. A dedicated 3-dimensional treatment planning system developed at McGill was used for the patient dose distribution calculation. The system was verified experimentally and described in detail elsewhere (24). It was originally implemented on a large VAX computer and, more recently, a PC-based version has been used and integrated within a complete stereotactic image analysis system, capable of processing stereotactic CT, MR, and DSA images. Figure 2 shows lateral DSA as well as coronal and transverse MR diagnostic images and typical 3-dimensional isodose distributions superimposed onto these images. Specially made 10 cm thick lead collimators are used to obtain the small circular fields (5 mm to 30 mm diameter at SAD = 100 cm). The field diameter has been arbitrarily defined as the separation at 90% on the stationary beam dose profile measured at the depth of dose maximum in a water phantom. Once the patient set up is completed, it takes about 25-30 minutes to deliver a target dose of 50 Gy. Our patient group consisted of 18 females and 15 males with a median age of 26 years (range: 9-69). All patients were assessed by a radiosurgical team consisting of a neurosurgeon, a radiation oncologist, a neuro-radiologist, and a physicist. None of the patients were asymptomatic prior to radiosurgery. Intracranial bleeding had occurred in 22 patients and in 19 it was the presenting symptom. Seven patients presented with seizures, three with headaches without evidence of intracranial bleeding, and four with other neurological symptoms. Prior surgical intervention to resect the AVM was attempted in five patients. In another patient three embolization procedures were carried out without success. The median follow-up time postradiosurgery is 16 months (range: 7-30). The 21 patients for whom follow up DSA is reported in this paper received doses of 50 or 55 Gy prescribed at the isocenter (100%). The choice between 50 or 55 Gy essentially depended on the size and/or location of the malformation. A single fraction was given to the majority of the patients (92%) with only two patients receiving fractionated treatment of 2 fractions each. The field diameters ranged from 8 mm to 25 mm. The 21 patients fall into three categories according to the dose delivered to the edge of the AVM nidus: (a) in seven patients the nidus was covered by the 90% isodose contour (45 or 50 Gy), (b) in six it was covered by the 50% contour (25 or
711
Fig. 2. Lateral DSA images and transverse and coronal MRI scans showing a right supra sylvian region AVM fed by branch of the right middle cerebral artery (A, C, E), and the radiosurgery treatment plans (B, D, F) for such AVM. Isodose distributions calculated with the McGill stereotactic treatment planning system. The 90% isodose contour covers the spherical target volume with a diameter of 5 mm. The isodose contours displayed are 90%, 50%, and 10%.
27.5 Gy), and (c) in the remaining eight patients the periphery of the lesion was covered by lower than 50% isodose contours (dose < 25 Gy). RESULTS Up to the time of this analysis, 2 1 patients had angiographies repeated at about 12 months following radiosurgery. Table 1 lists the clinical characteristics, the treatment parameters, the complications and percent AVM obliteration obtained for these patients, whereas Tables 2 and 3 summarize the treatment results. In 38% of the patients (8/21) the AVM was completely obliterated. A further five patients (24%) had a partial obliteration, between 50% and 99%, whereas eight other patients (38%) had an obliteration between 0 and 49%. Of the 13 patients
no.
35/F 25/F 39/M 18/F 26/M 20/F 21/F 24/F 22/F 50/M 56/M 22/F 11/F 27/F 25/F 32/M 16/M 28/F 44/M 38/M 9/F
Age/sex
* As of July 1989.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
FT.
midbrain caudate nucleus splenial thalamus occipital parieto-occipital splenial temporal fronto-parietal parietal central thalamus sylvian internal capsule sylvian central occipital frontal frontal parieto-occipital splenial
Location 4.5 0.5 3.7 0.4 0.3 0.2 6.3 3.5 0.3 8.2 11.5 2.6 0.5 16.3 0.9 13.0 42.0 8.2 22.4 5.6 0.4
Volume (cm31 10 10 15 10 10 10 20 15 20 20 20 10 10 25 8 10 20 17.5 15 15 5
Field diameter (mm) 27.5 X 2 55 50 55 55 55 55 55 50 55 55 50 50 27.5 X 2 50 55 55 50 55 50 50
0360-3016/90 $3.00 + .oO Copyright 0 1990 Pergamon Press plc
19, pp. 775-782
0 Technical Innovations and Notes RADIOSURGERY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS WITH THE DYNAMIC STEREOTACTIC IRRADIATION LUIS SOUHAMI, ERVIN
B. PODGORSAK,
M.D.,* ANDRE PH.D., AND
OLIVIER,
F.C.C.P.M.,*
M.D.,
MARINA
PH.D.,
F.R.C.S.(C),+
PLA, M.Sc.,
M.C.C.P.M.*
G. BRUCE PIKE, M.E.E.+
McGill University, Montrkal, Canada From December 1986 through December 1988, 33 patients with inoperable arteriovenous malformation (AVM) were treated in our center with the dynamic stereotactic radiosurgery, which uses a standard 10 MV isocentric
linear accelerator. There were 18 females and 15 males with a median age of 26 years (range: 9-69) and a median follow-up time of 16 months (range: 7-32). The arteriovenous malformation volumes treated ranged from 0.2 to 42 cm3. The prescribed doses at the isocenter varied from 50 to 55 Gy and were given as a single fraction in the majority of the patients (31/33). Late complications consisting of intracranial bleeding and/or hemiparesis were observed in three patients. To date, 21 patients underwent repeat angiographic studies at 1 year post-treatment. A complete obliteration of the lesion was achieved in 38% of these patients. For the patients whose arteriovenous malformation nidus was covered by a minimum dose of 25 Gy, the total obliteration rate was 61.5% (g/13), whereas none of the patients who had received less than 25 Gy at the edge of the nidus obtained a total obliteration. Our preliminary analysis at 1 year post-radiosurgery reveals results comparable to those previously reported for other radiosurgical techniques for the same follow-up period. Stereotactic radiosurgery, Radiosurgery, Arteriovenous malformation, Dynamic rotation, Radiosurgery with linear accelerator. INTRODUCXION
Since the pioneering work of Leksell in 195 1 ( 15), radiosurgery has become an attractive form of therapy for intracranial arteriovenous malformations (AVM) (29, 30) and other selected intracranial lesions (6, 11, 16). Initially, radiosurgery was performed with heavy charged particle beams from cyclotrons (9, 10, 13) and later by the Gamma unit* (32). The well proven efficacy of radiosurgery has in recent years led to the development of linear accelerator (linac) based radiosurgical techniques (1, 4, 7, 8, 18, 26), making radiosurgery potentially available to most major radiation oncology centers. Several centers have developed or are developing linacbased techniques. It is estimated that by early 1990, approximately 30-40 linac-based facilities will be operational in the U.S. (12). The linac-based radiosurgical techniques
Presented at the 3 1st Annual Meeting of the American Society for Therapeutic Radiology and Oncology, San Francisco, CA, 1-6 October, 1989. * Dept. of Radiation Oncology. + Dept. of Neurosurgery. Reprint requests to: L. Souhami, M.D., Department of Radiation Oncology, Montreal General Hospital, 1650 Cedar Ave., Montreal, QuCbec, H3G IA4 Canada.
reported so far are: (a) single plane rotation (8), (b) multiple non-coplanar converging arcs ( 1, 4, 7, 1S), and (c) dynamic rotation (25, 26). A comparison study among the high energy photon beam radiosurgical techniques has been published recently (27). Several important criteria should be fulfilled before any new radiosurgical technique is used in clinical practice, including high spatial and numerical accuracy of dose delivery to the target, steep dose fall-off outside the target volume, and knowledge of dose distribution within the target volume. Recently, Podgorsak et al. (28) have studied the adequacy of linacs in performing radiosurgery. They have shown that two of the existing linac-based radiosurgical techniques (multiple non-coplanar converging arcs and dynamic rotation) meet these criteria and can be considered an adequate and less expensive option to the Gamma unit and/or proton beam radiosurgery. In
Acknowledgements-The authors wish to thank Drs. G. Bertrand and R. Leblanc for helpful discussions and referral of some of the patients and Ms. Jennifer Manal for dedicated secretarial work. Accepted for publication 29 March 1990. * Leksell Gamma unit, Electa Instrument AB, Stockholm, Sweden.
776
I. J. Radiation Oncology 0 Biology 0 Physics
this paper we report our initial results in the first 33 patients with AVM treated with the dynamic rotation at McGill University in Montreal. METHODS
AND MATERIALS
From December 1986 through December 1988,33 patients with AVM underwent stereotactic radiosurgery in our center. All patients were treated with the dynamic rotation, described in detail previously (2526). A 10 MV isocentric linac* with a remotely controlled couch rotation is used as the source of radiation. The gantry rotates from 30” to 330”, while simultaneously the couch rotates from 75” to -75” (Fig. 1). The circular beam converges at the target volume, but because of the simultaneous couch and gantry rotations, the entrance beam never coincides with an exit beam, yielding a uniform dose distribution in the target volume and a reasonably sharp dose fall-off
September 1990, Volume 19, Number 3
outside the target volume. Target localization, treatment set up, and patient immobilization during the treatment are accomplished with a stereotactic frame developed locally (2 1, 22), and compatible with computerized tomography scan (CT), magnetic resonance imaging (MRI), and digital subtraction angiography (DSA). All patients underwent MRI and DSA studies before radiosurgery to define the location and size of the AVM as well as its relationship to sensitive anatomical structures. MRI exams were performed on a 1.5 Tesla whole body system and consisted of short repetition time spin echo (T 1-weighted) acquisitions in the transverse, sagittal, and coronal orientations. The body coil was used in all studies since the stereotactic frame does not easily fit within a standard head coil. However, with slice thicknesses of 7.5 mm, good quality images were obtained with one signal measurement. We have achieved a localization accuracy off 1 mm using DSA and transverse MR images
Fig. 1. The dynamic stereotactic radiosurgical procedure starts with the couch at 75” and the gantry at 30”, as shown in Fig. IA. During treatment, the couch rotates 150” from +75’ to -75”, while the gantry simultaneously rotates 300” from 30” to 330”. Thus, each degree of couch rotation corresponds to two degrees of gantry rotation. Several sucessive positions through which the couch and gantry move during the complete radiosurgical procedure are shown, starting (A) with the gantry and couch angles of 30” and +75’, respectively, through (B) 90” and +45”, respectively, (C) 180” and O”, respectively, (D) 270” and -45”, respectively, and (E) stopping at 330” .and -75”, respectively.
* Clinac- 18, Varian Associates,
Palo Alto, CA.
Dynamic stereotactic radiosurgery 0 L. SOUHAMI efal.
(in plane); however, our sagittal and coronal MR images have been shown to be geometrically less reliable (23). DSA provides a good definition of the dynamic compartments of the AVM, namely the arterial feeders, the nidus, and the draining veins. MRI gives the picture of the bulk of the AVM without differentiating between arteries, capillaries, or veins. It delineates, however, the topographic information concerning vital cerebral structures neighboring the AVM (20). Following treatment, MRI was repeated every 3 months and DSA was obtained after 12 months or sooner if the MRI study demonstrated a major change in the lesion signal. A dedicated 3-dimensional treatment planning system developed at McGill was used for the patient dose distribution calculation. The system was verified experimentally and described in detail elsewhere (24). It was originally implemented on a large VAX computer and, more recently, a PC-based version has been used and integrated within a complete stereotactic image analysis system, capable of processing stereotactic CT, MR, and DSA images. Figure 2 shows lateral DSA as well as coronal and transverse MR diagnostic images and typical 3-dimensional isodose distributions superimposed onto these images. Specially made 10 cm thick lead collimators are used to obtain the small circular fields (5 mm to 30 mm diameter at SAD = 100 cm). The field diameter has been arbitrarily defined as the separation at 90% on the stationary beam dose profile measured at the depth of dose maximum in a water phantom. Once the patient set up is completed, it takes about 25-30 minutes to deliver a target dose of 50 Gy. Our patient group consisted of 18 females and 15 males with a median age of 26 years (range: 9-69). All patients were assessed by a radiosurgical team consisting of a neurosurgeon, a radiation oncologist, a neuro-radiologist, and a physicist. None of the patients were asymptomatic prior to radiosurgery. Intracranial bleeding had occurred in 22 patients and in 19 it was the presenting symptom. Seven patients presented with seizures, three with headaches without evidence of intracranial bleeding, and four with other neurological symptoms. Prior surgical intervention to resect the AVM was attempted in five patients. In another patient three embolization procedures were carried out without success. The median follow-up time postradiosurgery is 16 months (range: 7-30). The 21 patients for whom follow up DSA is reported in this paper received doses of 50 or 55 Gy prescribed at the isocenter (100%). The choice between 50 or 55 Gy essentially depended on the size and/or location of the malformation. A single fraction was given to the majority of the patients (92%) with only two patients receiving fractionated treatment of 2 fractions each. The field diameters ranged from 8 mm to 25 mm. The 21 patients fall into three categories according to the dose delivered to the edge of the AVM nidus: (a) in seven patients the nidus was covered by the 90% isodose contour (45 or 50 Gy), (b) in six it was covered by the 50% contour (25 or
711
Fig. 2. Lateral DSA images and transverse and coronal MRI scans showing a right supra sylvian region AVM fed by branch of the right middle cerebral artery (A, C, E), and the radiosurgery treatment plans (B, D, F) for such AVM. Isodose distributions calculated with the McGill stereotactic treatment planning system. The 90% isodose contour covers the spherical target volume with a diameter of 5 mm. The isodose contours displayed are 90%, 50%, and 10%.
27.5 Gy), and (c) in the remaining eight patients the periphery of the lesion was covered by lower than 50% isodose contours (dose < 25 Gy). RESULTS Up to the time of this analysis, 2 1 patients had angiographies repeated at about 12 months following radiosurgery. Table 1 lists the clinical characteristics, the treatment parameters, the complications and percent AVM obliteration obtained for these patients, whereas Tables 2 and 3 summarize the treatment results. In 38% of the patients (8/21) the AVM was completely obliterated. A further five patients (24%) had a partial obliteration, between 50% and 99%, whereas eight other patients (38%) had an obliteration between 0 and 49%. Of the 13 patients
no.
35/F 25/F 39/M 18/F 26/M 20/F 21/F 24/F 22/F 50/M 56/M 22/F 11/F 27/F 25/F 32/M 16/M 28/F 44/M 38/M 9/F
Age/sex
* As of July 1989.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
FT.
midbrain caudate nucleus splenial thalamus occipital parieto-occipital splenial temporal fronto-parietal parietal central thalamus sylvian internal capsule sylvian central occipital frontal frontal parieto-occipital splenial
Location 4.5 0.5 3.7 0.4 0.3 0.2 6.3 3.5 0.3 8.2 11.5 2.6 0.5 16.3 0.9 13.0 42.0 8.2 22.4 5.6 0.4
Volume (cm31 10 10 15 10 10 10 20 15 20 20 20 10 10 25 8 10 20 17.5 15 15 5
Field diameter (mm) 27.5 X 2 55 50 55 55 55 55 55 50 55 55 50 50 27.5 X 2 50 55 55 50 55 50 50
