Radiotherapy and Oncology, 22 (1991) 29-35
(~) 1991 Elsevier Science Publishers B.V. All rights reserved 0167-8140/91/$03.50
29
RADION 00865
A comparison of techniques for stereotactic radiotherapy by linear accelerator based on 3-dimensional dose distributions J o h n D . G r a h a m 1, A l a n E. N a h u m 2 a n d M i c h a e l B r a d a l JAcademic Unit of Radiotherapy & Oncology and 2Joint Department of Physics, Institute of Cancer Research & Royal Marsden Hospital, Sutton, Surrey. U.K.
(Received 15 January 1990, revision received 28 May 1991, accepted 29 May 1991)
Key words: Stereotactic radiotherapy; Linear accelerator; Dose-volume distribution
Summary In order to establish the appropriate beam arrangement for use in stereotactic radiotherapy using a linear accelerator, dose volume distributions were calculated for a number of spherical targets in a head phantom and assessment was made by dose sparing of normal tissue outside the target volume. Using a single isocentre, fixed beam arrangements were compared with single and multiple non-coplanar isocentric arc rotations at target sizes from 10 to 55 mm diameter on a 6 MV Philips linear accelerator. From the dose-volume histograms produced, an arrangement of 3 or 4 arcs of rotation proved most suitable, in terms of sparing of normal tissue outside the target volume to high dose irradiation, across the range of target sizes studied. There was little further benefit with increasing the number of arcs beyond this. At target sizes greater than 20 mm diameter an arrangement of 6 static non-coplanar beams achieved sparing equivalent to multiple arc rotations and may have considerable advantages in the treatment of irregular volumes where customised beam shaping could be employed.
Introduction Radiosurgery or stereotactic radiotherapy was developed in 1951 by Leksell [ 11 ] using multiple focused X-ray beams. The commercially available multi-source cobalt-60 unit (the so-called g a m m a knife) is now in operation at more than 10 centres around the world [13,22]. Other groups have used heavy-particle beams [8,14] whose physical characteristics ensure at least as good dose distributions as those achieved by multiple photon beams. However, both these methods involve high capital and running costs. In the last 10 years techniques have been developed to reproduce the treatment capabilities o f the multi-source cobalt unit using modern linear accelerators found in most radiotherapy departments [2]. Using head fixation by a stereotactic frame and multiple non-coplanar arc rotation about an
isocentre, a dose distribution similar to that from a multi-headed cobalt unit can be achieved. To date, stereotactic radiotherapy has been used mainly to treat arteriovenous malformations (AVMs) but the larger field sizes available from linear accelerator based techniques now offer opportunities to treat a wide variety o f brain tumours. Radiosurgery with the g a m m a knife has been limited to field sizes of around 20 mm diameter. Using two isocentres the volume can be increased but at the cost o f an inhomogeneous dose distribution in the target [16]. With a linear accelerator field sizes up to 60 mm diameter have been treated [ 1 ]. The rationale in using a stereotactic technique to treat malignant disease in the brain is that for a given tumour dose, the radiation burden on the surrounding normal brain can be minimised. This may allow the dose to the tumour to be increased. Stereotactic exter-
Addressforcorrespondence: Dr. M. Brada, Academic Unit of Radiotherapy & Oncology, Royal Marsden Hospital, Down's Road, Sutton, Surrey,
SM2 5PT~ U.K.
30 nal beam radiotherapy offers a non-invasive alternative to brachytherapy [ 10] and with the development of relocatable stereotactic frames [5] fractionated radiotherapy is now feasible, possibly further improving the therapeutic ratio. Various techniques have been reported using a single arc, multiple arcs (up to 11) or dynamic rotation in which the gantry and table move simultaneously [6,7,12,18]. Podgorsak et al. [ 19] assessed the different techniques by comparing the dose fall-off in various planes around the target and suggested that 4 or more arcs or dynamic radiosurgery was preferable to a single arc rotation. This comparison was limited to standard transverse, coronal and sagittal planes. Although it provides information on the dose distributions to sensitive structures lying in these planes, the planar data do not give the full picture. A volume-dose distribution is a more appropriate method of comparison and has recently been used to compare heavy-particle beams with photon beams for stereotactic radiotherapy [17]. The dose delivery using multiple non-coplanar arcs or the gamma knife is essentially spherical in shape. As the technique has been adapted to treat larger and more irregular lesions spherical volumes have been matched together using multiple isocentres [ 13]. However, before embarking on investigation of methods for treating irregular volumes, it is important to define a satisfactory method for treating simple spherical shapes in terms of arc arrangement and suitable target size. In order to address this problem we have calculated volume-dose distributions for a number of spherical target sizes in a head phantom treated by different stereotactic techniques on a 6 MV linear accelerator. This serves as a guide in the process of optimising stereotactic external beam radiotherapy even though other factors such as adjacent critical structures and the dose to the target volume which may contain some normal brain tissue have to be taken into consideration when evaluating individual treatment plans. Methods
CT scan images were taken of a Rando anthropomorphic head phantom with a slice thickness of 1.5 mm through the target volume and 5 mm through the remainder of the head. The CT information was transferred to a 3-D computer planning system developed at the Joint Center for Radiation Therapy in Boston, U.S.A. (JCPLAN), and running on a Micro Vax II computer with a Lexidata image display system. The dose at any point in the phantom was calculated from the sum of the primary and scattered components. The primary component includes inverse square law, prima-
ry in-air profiles, attenuation and penumbra functions. The scatter component is calculated using a scatter integration technique over the differential scatter-maximum ratios [4]. Circular field sizes were defined in JCPLAN using the blocking facility. Beam profiles calculated using the algorithm were compared to measured data obtained in a water phantom for the various field sizes defined by circular collimators. The calculated beam profiles (Fig. la) gave an acceptable representation of the measured profiles. Calculated depth dose curves for a 10 × 10 cm field and for circular fields of 10 and 40 mm diameter are shown in Fig. lb. Spherical target volumes were defined deep in the left parietal lobe with diameters of 10, 20, 40 and 55 mm. The first two sizes reflect typical target volumes ir-
10 x 10 c m Square Field 40 m m Circular Field 10 m m Circular Field
~-'-'Proportional Dose 1.0-
[
I
0.80.6-
1
I
•
!
I
0.4"
; /
0.2-
I I
~
I
!
I
a.
i :
1
J
0.0 • , • , • , - , • , - , - , - , • , - , - 1 0 - 8 - 6 - 4 -2 0 2 4 6 8 10 Distance
[ [ .......... I.........
Dose (%) 100
80
~
60
Off
Central
Axis
10 x 10 cm Square Field 40 m m Circular Field 10 nun Circular Field
.
"'"~.,.
40
20 b 0
5
10 Depth
F i g . 1. ( a ) C a l c u l a t e d tom and (b) calculated
15
20
(cm)
b e a m p r o f i l e s a t 6 c m d e p t h in a w a t e r p h a n central axis depth-dose
curves for circular
f i e l d s t o t r e a t t a r g e t s o f 10 a n d 4 0 m m d i a m e t e r a n d f o r a 10 x 10 c m s q u a r e field.
31 radiated in the treatment of AVMs and fall within the range available from the gamma knife unit. The 40 and 55 mm target sizes would be used mainly in the treatment of brain tumours. The diameters of the circular collimators, chosen so as to encompass the target volume within the 90~o isodose, measured 12, 22, 45 and 60 mm in diameter respectively as defined by the 50 ~o isodose at 6 cm depth. Dose inhomogeneity within the target was less than + 7.5 ~/o. A number of treatment plans were constructed for each target volume using single arc, multiple arc, dynamic rotation and fixed beam arrangements (Table I) on a 6 MV Philips linear accelerator. However, Mazal and co-workers [ 15] have shown that the energy of the photon beam has a negligible influence on stereotactic dose distributions. TABLE
I
Details of beam arrangements. Name
A r c or b e a m
Gantry angle a
Table angle b
number Static 3 beams (coplanar)
Static 6 beams (non-coplanar)
Single arc 2 Arc
3 Arc
5 Arc
9 Arc
1
0°
0°
2 (45 ° w e d g e )
90 °
0°
3 (45 * w e d g e )
270 °
0°
1
65 °
30 °
2
115 °
30 ° 90 °
3
30 °
4
150 °
90 °
5
245 o
330 °
6
295 °
330 °
1
0°-359 °
0°
1
30°-150 °
45 °
2
210°-330 °
315 °
1
30°-150 °
18 °
2
30 ° - 150 o
90 °
3
210°-330 °
342 °
1
30°-150 °
18 °
2
30°-150 °
54 °
3
30°-150 °
90 °
4
210°-330 °
306 °
5
210°-330 °
342 °
1
30°-150 °
18 °
2
30°-150 °
36 °
3
30°-150 °
54 °
4
30°-150 °
72 °
5
30°-150 °
90 °
6
210°-330 °
288 °
7
210°-330 °
306 °
8
210°-330 °
324 °
9
210°-330 °
342 °
E q u a l w e i g h t i n g t o all fields. C i r c u l a r c o l l i m a t o r s u s e d for all b e a m s . 0 ° - g a n t r y vertical; 9 0 ° - L lateral; 2 7 0 ° - R lateral. 0 °-neutral position (transverse arc), table rotates clockwise from 0 ° t o 359 °.
J C P L A N simulates gantry rotation by calculating a number of fixed beams at 10 ° intervals. In each case the isocentre was at the centre of the spherical target volume. Three dimensional dose distributions were calculated by a method that defines a grid with a maximum of 500 points at a minimum 5 x 5 mm spacing on each transverse slice and stores the fractional volume of the chosen structure in 5~o dose bands. The doses were expressed as percentages of the isocentre dose. These were transformed into graphs of percentages of the volume of the chosen structure receiving dose above a certain level, commonly known as dose volume histograms (DVH). Comparison of different beam arrangements was made based on the volumes of normal brain tissue irradiated outside the target volume using DVH. Clearly the overall clinical tolerance of the technique may relate to the target volume dose and the amount of normal brain within this. However, our comparison of different techniques of delivery of stereotactic radiotherapy is based on the understanding that the target volume dose has been kept constant. There are particular problems with a new radiotherapy technique where there is no data relating the tolerance of normal tissue to the DVH outside the target volume. It is therefore necessary to make certain assumptions. What is clearly important is minimising the size of the volume receiving a high dose of radiation. The precise dose level has to some extent be arbitrary. Stereotactic radiotherapy as currently practised, particularly in the treatment of AVMs, is a single dose irradiation to 18-20 Gy [ 12]. The only information on tolerance of single doses as given to the whole brain is that single fractions of 1 0 G y are within clinical tolerance [3]. We have therefore chosen an isodose level of 50~o of the target dose and therefore have assumed that doses below this are of less clinical relevance. This does not exclude the possibility that other higher dose levels may be of relevance. However, in most of the DVH comparisons presented, isodoses above 50~o usually follow the same pattern as those described for
50~o. Whole brain DVHs (volume 1250 cm 3) were calculated for 3 plans at the 40 mm diameter target size. All other DVHs were based on a volume of normal brain tissue adjacent to each target. This was chosen to be a 2 cm thick shell enclosing the target volume (Fig. 2), giving shells of 50, 60, 80 and 95 mm diameter, respectively. The dose falls from maximum to below 10 ~o over a distance of 2 cm using stereotactic techniques [ 19]. The volume of normal brain tissue receiving greater than 50 ~o of the isocentre dose was tabulated. A similar technique has been used in a recent publication [ 17].
32
Spherical target Shell 2 c m s thick
•:"::.::i:!:[.......:i:~:i:i.......':.:':'::-"
Fig. 2. A transverse CT slice through the isocentre showing the position of the target and volume of interest, a shell 2 cm thick around the target. However, such calculations can only provide a theoretical b a c k g r o u n d to the choice o f treatment technique and are not a substitute for the clinical assessment o f radiation tolerance o f normal brain outside or within the target volume.
irradiated above the 20~o isodose. The D V H s calculated from the smaller volume of interest were identical to the whole brain D V H s except at low dose levels, which were not relevant to this c o m p a r i s o n (Figs. 3 and 4). D V H s based on the 2 cm thick shell of normal tissue around the target volume were calculated for a n u m b e r of b e a m arrangements at four target sizes: 10, 20, 40 and 55 m m diameter. The results for the 10 and 40 m m diameter targets are shown in Fig. 4. In order to simplify the c o m p a r i s o n between different b e a m arrangements the volumes o f normal brain tissue receiving greater than 50 ~o o f the isocentre dose were plotted as histograms (Fig. 5). The distributions using multiple arcs were superior to a traditional arrangement o f 3 static planar beams at all target sizes. A single arc rotation or a conventional arrangement of 3 fixed beams was inferior to multiple arcs at all target sizes (Fig. 5). There was no increased sparing of normal tissue above the 50~o isodose with more than 3 arcs of rotation. T w o arcs gave a distribution that was only marginally worse than 3 arcs at the 20 m m diameter target size and was slightly better at the 40 and 50 m m diameter sizes. If 6 static n o n - c o p l a n a r beams are ar100"
.... a .... .... * - ' "
% volume
Results
80
D o s e volume histograms calculated for the whole brain volume are shown in Fig. 3. T w o stereotactic techniques at the 40 m m diameter target size were c o m p a r e d to an arrangement o f 3 fixed planar beams using a circular collimator. The 3 and 5 arc rotations gave virtually identical distributions and were superior to 3 fixed beams through the sparing of normal tissue
3 beams 6 beams
"~.
m
Single arc
i~"~
"
3 arc
60 ¸
40-
20-
0 0
20
40
60
100
80
% dose 3Arc
% volume 80"
II
$ Arc
•
%
.~ -~[
volum
•
•~
----a---
't~"" ~ .~ " ~ i g
60"
"~,, ~
40
L'o.-'~.,~ , . ~
40
,
6beams Single arc
•
2
•
3arc
a r e
"
S
•
9 arc
a r c
20
-
0
,
20
-
m
.
40
i
60
•
l
80
.
- i
100
% dose
Fig. 3. Dose volume histogram of percentage volume of the whole brain receiving a percentage of the isoeentre dose with a 40 mm diameter target, comparing 3 and 5 arc rotations with 3 fixed beams.
0
b . 0
. 20
.
.
. 40
60
80
100
% dose
Fig. 4. Dose volume histograms of a 2 cm thick shell around: (a) a 10 mm diameter target and (b) a 40 mm diameter target comparing rotational and fixed beam arrangements.
33 10
1000 % of target volume 80O
volume
(cc) 8 6
6O0 4 4OO
2
2O0
o
3 beams
6
beams
1
3 arc
arc
5 arc
9 arc -
o
lo
2o
3o
Target Diameter
30
4'o (ram)
U
to
"
6;
Fig. 6. The additional sparing of normal tissue irradiated above 50% of the isocentre dose by an arrangement of 3 non-coplanar arcs over a conventional arrangement of 3 fixed beams, expressed as a p r o p o r t i o n of the target volume for targets of 10, 20, 40 and 55 m m diameter.
volume
(cc) 20
tO
0
3 beams 6 beams
1 arc
2 arc
3 arc
5 arc
9 arc
3 arc
5 arc
9 arc
120" volume
(ce) 1008060 I 40
! 20
3 beams 6 beams
1 arc
2 arc
250'
ranged so as to obtain maximal angular separation between beams, a dose distribution equivalent to the multiple arc rotations can be achieved at target sizes 20 mm and over (Fig. 5b-d), but not at 10 mm (Fig. 5a). The relative sparing of stereotactic radiotherapy using 3 arcs over a standard 3 field set up was compared at the different target sizes, The 3 arc technique spared 68 ~o of the volume of normal tissue treated by 3 beams to > 50~o of the isocentre dose for the 10 mm diameter target. This fell to 47~o at 20 mm, 18~o at 40 mm and 13~o at 55 mm target diameter although the actual volumes of normal tissue spared were 5.6cm 3, 12.1 cm 3, 19.3 cm 3 and 22.8 cm 3, respectively; showing an increase with increasing target diameter. The relative benefit from the stereotactic technique, with the sparing of normal tissue plotted as a percentage of the target volume, can be seen predominantly with target sizes less than 40 mm diameter (Fig. 6). While sparing normal brain from higher doses, the stereotactic multiple arc techniques irradiated a greater volume to doses below 25 ~o of the isocentre dose when compared with conventional or single arc technique (Figs. 3 and 4).
200"
Discussion
150" 100" 50" 0" 3 beams
6 beams
1 arc
2 arc
3 arc
Fig. 5. H i s t o g r a m s of the volume of normal brain tissue receiving greater than 5 0 3 of the isocentre dose with various b e a m and arc arrangements for target sizes: ( a ) 1 0 r a m diameter; ( b ) 2 0 m m diameter; (c) 40 m m diameter and (d) 55 m m diameter.
The comparison of dose distributions between rival plans in standard radiotherapy with coplanar beams is usually performed along the central axis in the transverse plane. Distributions in the sagittal and coronal planes may give additional information but in order to accurately estimate the volume of normal tissue being irradiated the calculation of DVHs is essential. In this comparison of stereotactic techniques it was assumed that tolerance to radiation was uniform throughout the
34 brain, although the clinical expression of late radiation damage will vary with site. Assessments were based solely on reductions in normal tissue volumes outside the target volume irradiated to > 5 0 ~ of the isocentre dose, without regard to anatomical relationships. The target volume was defined at 90 ~o isodose and with all the techniques considered the target volume dose homogeneity was + 7.5 ~o. The greatest benefit from stereotactic radiotherapy should be seen with the technique that reduces the volume of normal tissue receiving doses above tolerance. While published figures are available for the tolerance of the brain to large volume irradiation with conventional techniques and fractionation [ 21 ], there is little equivalent data for stereotactic radiotherapy [9]. Doses prescribed vary widely and have ranged from 20 to 70 Gy as a single fraction, with doses being reduced as th e target volume increases [ 1,20]. In the treatment of cerebral AVMs doses of the order of 20 Gy have been given as a single fraction [ 12]. There is little data on the tolerance of the brain to single fraction radiotherapy but an R T O G study using a single fraction of 10 Gy was within the tolerance dose of normal brain to conventional radiotherapy [3]. A comparison based on the volume of normal tissue receiving >50~o of the prescribed dose appears justified. The results presented here, based on this principle, have shown that 3 noncoplanar arcs were as effective in terms of normal tissue sparing outside the target volume as a larger number of arCS.
What is the maximum target size that can be usefully treated with stereotactic radiotherapy? Comparing 3 or more arcs of rotation with a standard 3-field plan (Fig. 6) the relative benefit, as measured by the volume of normal tissue irradiated to greater than 50~o of the prescribed dose, fell from the 20 to the 40 mm diameter target sizes but altered very little between the 40 and 55 mm diameter target sizes. Any sparing of normal tissue may be beneficial but it is at the smaller target sizes that a clinically detectable benefit is likely to be seen [23]. Most centres limit stereotactic radiotherapy with a single isocentre to a maximum target size of 30-40 mm. The arrangement of 6 static non-coplanar beams, which showed equivalent sparing to multiple-arc techniques at larger target sizes, may be of considerable advantage in the treatment of malignant disease. Problems have arisen in encompassing large irregular
volumes using the multiple-arc technique which delivers an essentially spherical high-dose volume. To date, the only practical way of treating these irregular lesions has been to use either multiple isocentres and combine the "dose spheres" with resultant dose inhomogeneity within the target, or to treat with a large sphere which as shown above has minimal benefit over conventional techniques in terms of sparing of normal brain tissue. Recent reports suggest increased late complication rates associated with multiple isocentre techniques [ 16]. By using 6 static beams it may be possible to shape the target volume using customised blocking of the beam. This would be an alternative to a conformal multiple-arc technique with dynamic beam shaping using a multi-leaf collimator. Such technology is not currently available and would entail enormous complexity in planning and delivery. It must be stressed that the theoretical calculations presented are only a guideline to the optimum arc arrangement and do not attempt to provide any information on the radiation tolerance of the technique of stereotactic radiotherapy. Considerations such as the dose and the homogeneity of dose distribution within the target volume as well as the amount of normal brain within it may also be the determinants of tolerance and will have to be evaluated in clinical studies. In summary, for spherical volumes less than 4 cm in diameter a multiple-arc technique appears to give the greatest sparing of normal tissue to high dose. There is no significant benefit from an arrangement of more than 3 arcs for the spherical target sizes studied here. Because of exit doses from a sagittal arc to the whole body and particularly the thyroid, we currently use 4 arcs for the treatment of brain tumours and occasionally up to 5 arcs in the treatment of AVMs near critical structures. For larger, irregular volumes an arrangement of a small number (usually 6) of static noncoplanar beams may be indicated as customised blocking of the beam would be practical and would achieve the best sparing of normal tissue outside the tumour volume.
Acknowledgement This work was supported by the Professor Bloom Research Fund, The Cancer Research Campaign & The Royal Marsden Hospital.
References 1 Betti, O. O., Munari, C. and Rosier, R. Stereotactic radiosurgery with the linear accelerator: treatment of arteriovenous malformations. Neurosurgery 24: 311-321, 1989.
2 Betti, O.O. and Derechinsky, V. Multiple beam stereotaxic irradiation. Neurochirurgie 29: 295-298, 1983. 3 Borgelt, B., Gelber, R., Larson, M., Hendrickson, F., Griffin, T.
35
4
5
6
7
8
9
10
11 12
13
and Roth, R. Ultra-rapid high dose irradiation schedules for the palliation of brain metastases: final results of the first two studies by the Radiation Therapy Oncology Group. Int. J. Radiat. Oncol. Biol. Phys. 7: 1633-1638, 1981. Chin, L. M., Cheng, C. W., Siddon, R. L., Rice, R. K., Mijnheer, B. J. and Harris, J. R. Three-dimensionalphoton dose distributions with and without lung corrections for tangential breast intact treatments. Int. J. Radiat. Oncol. Biol. Phys. 17: 1327-1335, 1989. Gill, S. S., Warrington, A. P., Thomas, D. G. T. and Brada, M. Relocatable frame for stereotactic external beam radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 20: 599-603, 1991. Hartmann, G. H., Schlegel, W., Sturm, V., Kober, B., Pastyr, O. and Lorenz, W.J. Cerebral radiation surgery using moving field irradiation at a linear accelerator facility. Int. J. Radiat. Oncol. Biol. Phys. 11: 1185-1192, 1985. Houdek, P. V., Fayos, J. V., Van-Buren, J. M. and Ginsberg, M.S. Stereotaxic radiotherapy technique for small intracranial lesions. Med. Phys. 12: 469-472, 1985. Kjellberg, R.N. Stereotactic Bragg peak proton beam radiosurgery for cerebral arteriovenous malformations. Ann. Clin. Res. 18 (Suppl 47): 17-19, 1986. Kjellberg, R. N., Hanamura, T., Davis, K. R., Lyons, S. L. and Adams, R.D. Bragg-peak proton-beam therapy for arteriovenous malformations of the brain. N. Engl. J. Med. 309: 269-274, 1983. Leibel, S. A., Gutin, P. H., Wara, W. M., Silver, P. S., Larson, D. A., Edwards, M. S. B., Lamb, S.A., Ham, B., Weaver, K. A., Barnett, C. and Phillips, T. Survival and quality of life after interstitial implantation of removable high activity iodine-125 sources for the treatment of patients with recurrent malignant gliomas. Int. J. Radiat. Oncol. Biol. Phys. 17:1129-1139, 1989. Leksell, L. The stereotaxic method and radiosurgery of the brain. Acta Chir. Scand. 102: 316-319, 1951. Loeflter, J. S., Alexander, E., Siddon, R. L., Saunders, W. M., Coleman, C. N. and Winston, K.R. Stereotactic radiosurgery for intracranial arteriovenous malformations using a standard linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 17: 673-677, 1989. Lunsford, L. D., Flickinger, J., Linder, G. and Maitz, A. Stereotactic radiosurgery of the brain using the first United States 201
cobalt-60 source gamma knife. Neurosurgery 24: 151-159, 1989. 14 Lyman, J.T., Kanstein, L., Yeater, F., Fabrikant, J.I. and Frankel, K.A. A helium ion beam for stereotactic radiosurgery of central nervous system disorders. Med. Phys. 13: 695-699, 1986. 15 Mazal, D.A., Rosenwald, J-C., Gaboriaud, G. and Porcher, J, B. Implication of basic physical phenomena on the dose distribution in stereotactic x-ray external irradiation (Abstr.). 17th International Congress of Radiology, p. 217, Paris, July 1-8, 1989. 16 Nedzi, L.A., Kooy, H., Alexander, E., Gelman, R.S. and Loeffler, J.S. Variables associated with the development of complications from radiosurgery of intracranial tumours. Proc. Am. Soc. Ther. Radiat. Oneol., October 15-19th, 1990. Int. J. Radiat. Oncol. Biol. Phys. 19 (Suppl 1): 149, 1990. 17 Phillips, M. H., Frankel, K. A., Lyman, J. T., Fabrikant, J. I. and Levy, R.P. Comparison of different radiation types and irradiation geometries in stereotactic radiosurgery. Int. J. Radiat. Oncol. Biol. Phys. 18: 211-220, 1990. 18 Podgorsak, E. B., Olivier, A., Pla, M., Lefebvre, P. Y. and Hazel, J. Dynamic stereotactic radiosurgery. Int. J. Radiat. Oncol. Biol. Phys. 14: 115-126, 1988. 19 Podgorsak, E. B., Pike, G. B., Olivier, A., Pla, M. and Souhami, L. Radiosurgery with high energy photon beams: a comparison among techniques. Int. J. Radiat. Oncol. Biol. Phys. 16: 857-865, 1989. 20 Sturm, V., Kober, B., Hover, K. H., Schlegel, W., Boesecke, R., Pastyr, O., Hartmann, G. H., Schabbert, S., zum-Winkel, K., Kunze, S. and Lorenz, W.J. Stereotactic percutaneous single dose irradiation of brain metastases with a linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 13: 279-282, 1987. 21 Tiver, K. Treatment ofCNS tumours with conventional radiotherapy: the importance of dose and volume factors in tumour control and CNS radiation tolerance. Australas. Radiol. 33" 15-22, 1989. 22 Walton, L., Bomford, C.K. and Ramsden, H.N.C. The Sheffield stereotactic radiosurgery unit: physical characteristics and principles of operation. Br. J. Radiol. 60: 897-906, 1987. 23 Withers, H.R., Taylor, J. M. G. and Maciejewski, B. Treatment volume and tissue tolerance. Int. J. Radiat. Oncol. Biol. Phys. 14: 751-759, 1988.
(~) 1991 Elsevier Science Publishers B.V. All rights reserved 0167-8140/91/$03.50
29
RADION 00865
A comparison of techniques for stereotactic radiotherapy by linear accelerator based on 3-dimensional dose distributions J o h n D . G r a h a m 1, A l a n E. N a h u m 2 a n d M i c h a e l B r a d a l JAcademic Unit of Radiotherapy & Oncology and 2Joint Department of Physics, Institute of Cancer Research & Royal Marsden Hospital, Sutton, Surrey. U.K.
(Received 15 January 1990, revision received 28 May 1991, accepted 29 May 1991)
Key words: Stereotactic radiotherapy; Linear accelerator; Dose-volume distribution
Summary In order to establish the appropriate beam arrangement for use in stereotactic radiotherapy using a linear accelerator, dose volume distributions were calculated for a number of spherical targets in a head phantom and assessment was made by dose sparing of normal tissue outside the target volume. Using a single isocentre, fixed beam arrangements were compared with single and multiple non-coplanar isocentric arc rotations at target sizes from 10 to 55 mm diameter on a 6 MV Philips linear accelerator. From the dose-volume histograms produced, an arrangement of 3 or 4 arcs of rotation proved most suitable, in terms of sparing of normal tissue outside the target volume to high dose irradiation, across the range of target sizes studied. There was little further benefit with increasing the number of arcs beyond this. At target sizes greater than 20 mm diameter an arrangement of 6 static non-coplanar beams achieved sparing equivalent to multiple arc rotations and may have considerable advantages in the treatment of irregular volumes where customised beam shaping could be employed.
Introduction Radiosurgery or stereotactic radiotherapy was developed in 1951 by Leksell [ 11 ] using multiple focused X-ray beams. The commercially available multi-source cobalt-60 unit (the so-called g a m m a knife) is now in operation at more than 10 centres around the world [13,22]. Other groups have used heavy-particle beams [8,14] whose physical characteristics ensure at least as good dose distributions as those achieved by multiple photon beams. However, both these methods involve high capital and running costs. In the last 10 years techniques have been developed to reproduce the treatment capabilities o f the multi-source cobalt unit using modern linear accelerators found in most radiotherapy departments [2]. Using head fixation by a stereotactic frame and multiple non-coplanar arc rotation about an
isocentre, a dose distribution similar to that from a multi-headed cobalt unit can be achieved. To date, stereotactic radiotherapy has been used mainly to treat arteriovenous malformations (AVMs) but the larger field sizes available from linear accelerator based techniques now offer opportunities to treat a wide variety o f brain tumours. Radiosurgery with the g a m m a knife has been limited to field sizes of around 20 mm diameter. Using two isocentres the volume can be increased but at the cost o f an inhomogeneous dose distribution in the target [16]. With a linear accelerator field sizes up to 60 mm diameter have been treated [ 1 ]. The rationale in using a stereotactic technique to treat malignant disease in the brain is that for a given tumour dose, the radiation burden on the surrounding normal brain can be minimised. This may allow the dose to the tumour to be increased. Stereotactic exter-
Addressforcorrespondence: Dr. M. Brada, Academic Unit of Radiotherapy & Oncology, Royal Marsden Hospital, Down's Road, Sutton, Surrey,
SM2 5PT~ U.K.
30 nal beam radiotherapy offers a non-invasive alternative to brachytherapy [ 10] and with the development of relocatable stereotactic frames [5] fractionated radiotherapy is now feasible, possibly further improving the therapeutic ratio. Various techniques have been reported using a single arc, multiple arcs (up to 11) or dynamic rotation in which the gantry and table move simultaneously [6,7,12,18]. Podgorsak et al. [ 19] assessed the different techniques by comparing the dose fall-off in various planes around the target and suggested that 4 or more arcs or dynamic radiosurgery was preferable to a single arc rotation. This comparison was limited to standard transverse, coronal and sagittal planes. Although it provides information on the dose distributions to sensitive structures lying in these planes, the planar data do not give the full picture. A volume-dose distribution is a more appropriate method of comparison and has recently been used to compare heavy-particle beams with photon beams for stereotactic radiotherapy [17]. The dose delivery using multiple non-coplanar arcs or the gamma knife is essentially spherical in shape. As the technique has been adapted to treat larger and more irregular lesions spherical volumes have been matched together using multiple isocentres [ 13]. However, before embarking on investigation of methods for treating irregular volumes, it is important to define a satisfactory method for treating simple spherical shapes in terms of arc arrangement and suitable target size. In order to address this problem we have calculated volume-dose distributions for a number of spherical target sizes in a head phantom treated by different stereotactic techniques on a 6 MV linear accelerator. This serves as a guide in the process of optimising stereotactic external beam radiotherapy even though other factors such as adjacent critical structures and the dose to the target volume which may contain some normal brain tissue have to be taken into consideration when evaluating individual treatment plans. Methods
CT scan images were taken of a Rando anthropomorphic head phantom with a slice thickness of 1.5 mm through the target volume and 5 mm through the remainder of the head. The CT information was transferred to a 3-D computer planning system developed at the Joint Center for Radiation Therapy in Boston, U.S.A. (JCPLAN), and running on a Micro Vax II computer with a Lexidata image display system. The dose at any point in the phantom was calculated from the sum of the primary and scattered components. The primary component includes inverse square law, prima-
ry in-air profiles, attenuation and penumbra functions. The scatter component is calculated using a scatter integration technique over the differential scatter-maximum ratios [4]. Circular field sizes were defined in JCPLAN using the blocking facility. Beam profiles calculated using the algorithm were compared to measured data obtained in a water phantom for the various field sizes defined by circular collimators. The calculated beam profiles (Fig. la) gave an acceptable representation of the measured profiles. Calculated depth dose curves for a 10 × 10 cm field and for circular fields of 10 and 40 mm diameter are shown in Fig. lb. Spherical target volumes were defined deep in the left parietal lobe with diameters of 10, 20, 40 and 55 mm. The first two sizes reflect typical target volumes ir-
10 x 10 c m Square Field 40 m m Circular Field 10 m m Circular Field
~-'-'Proportional Dose 1.0-
[
I
0.80.6-
1
I
•
!
I
0.4"
; /
0.2-
I I
~
I
!
I
a.
i :
1
J
0.0 • , • , • , - , • , - , - , - , • , - , - 1 0 - 8 - 6 - 4 -2 0 2 4 6 8 10 Distance
[ [ .......... I.........
Dose (%) 100
80
~
60
Off
Central
Axis
10 x 10 cm Square Field 40 m m Circular Field 10 nun Circular Field
.
"'"~.,.
40
20 b 0
5
10 Depth
F i g . 1. ( a ) C a l c u l a t e d tom and (b) calculated
15
20
(cm)
b e a m p r o f i l e s a t 6 c m d e p t h in a w a t e r p h a n central axis depth-dose
curves for circular
f i e l d s t o t r e a t t a r g e t s o f 10 a n d 4 0 m m d i a m e t e r a n d f o r a 10 x 10 c m s q u a r e field.
31 radiated in the treatment of AVMs and fall within the range available from the gamma knife unit. The 40 and 55 mm target sizes would be used mainly in the treatment of brain tumours. The diameters of the circular collimators, chosen so as to encompass the target volume within the 90~o isodose, measured 12, 22, 45 and 60 mm in diameter respectively as defined by the 50 ~o isodose at 6 cm depth. Dose inhomogeneity within the target was less than + 7.5 ~/o. A number of treatment plans were constructed for each target volume using single arc, multiple arc, dynamic rotation and fixed beam arrangements (Table I) on a 6 MV Philips linear accelerator. However, Mazal and co-workers [ 15] have shown that the energy of the photon beam has a negligible influence on stereotactic dose distributions. TABLE
I
Details of beam arrangements. Name
A r c or b e a m
Gantry angle a
Table angle b
number Static 3 beams (coplanar)
Static 6 beams (non-coplanar)
Single arc 2 Arc
3 Arc
5 Arc
9 Arc
1
0°
0°
2 (45 ° w e d g e )
90 °
0°
3 (45 * w e d g e )
270 °
0°
1
65 °
30 °
2
115 °
30 ° 90 °
3
30 °
4
150 °
90 °
5
245 o
330 °
6
295 °
330 °
1
0°-359 °
0°
1
30°-150 °
45 °
2
210°-330 °
315 °
1
30°-150 °
18 °
2
30 ° - 150 o
90 °
3
210°-330 °
342 °
1
30°-150 °
18 °
2
30°-150 °
54 °
3
30°-150 °
90 °
4
210°-330 °
306 °
5
210°-330 °
342 °
1
30°-150 °
18 °
2
30°-150 °
36 °
3
30°-150 °
54 °
4
30°-150 °
72 °
5
30°-150 °
90 °
6
210°-330 °
288 °
7
210°-330 °
306 °
8
210°-330 °
324 °
9
210°-330 °
342 °
E q u a l w e i g h t i n g t o all fields. C i r c u l a r c o l l i m a t o r s u s e d for all b e a m s . 0 ° - g a n t r y vertical; 9 0 ° - L lateral; 2 7 0 ° - R lateral. 0 °-neutral position (transverse arc), table rotates clockwise from 0 ° t o 359 °.
J C P L A N simulates gantry rotation by calculating a number of fixed beams at 10 ° intervals. In each case the isocentre was at the centre of the spherical target volume. Three dimensional dose distributions were calculated by a method that defines a grid with a maximum of 500 points at a minimum 5 x 5 mm spacing on each transverse slice and stores the fractional volume of the chosen structure in 5~o dose bands. The doses were expressed as percentages of the isocentre dose. These were transformed into graphs of percentages of the volume of the chosen structure receiving dose above a certain level, commonly known as dose volume histograms (DVH). Comparison of different beam arrangements was made based on the volumes of normal brain tissue irradiated outside the target volume using DVH. Clearly the overall clinical tolerance of the technique may relate to the target volume dose and the amount of normal brain within this. However, our comparison of different techniques of delivery of stereotactic radiotherapy is based on the understanding that the target volume dose has been kept constant. There are particular problems with a new radiotherapy technique where there is no data relating the tolerance of normal tissue to the DVH outside the target volume. It is therefore necessary to make certain assumptions. What is clearly important is minimising the size of the volume receiving a high dose of radiation. The precise dose level has to some extent be arbitrary. Stereotactic radiotherapy as currently practised, particularly in the treatment of AVMs, is a single dose irradiation to 18-20 Gy [ 12]. The only information on tolerance of single doses as given to the whole brain is that single fractions of 1 0 G y are within clinical tolerance [3]. We have therefore chosen an isodose level of 50~o of the target dose and therefore have assumed that doses below this are of less clinical relevance. This does not exclude the possibility that other higher dose levels may be of relevance. However, in most of the DVH comparisons presented, isodoses above 50~o usually follow the same pattern as those described for
50~o. Whole brain DVHs (volume 1250 cm 3) were calculated for 3 plans at the 40 mm diameter target size. All other DVHs were based on a volume of normal brain tissue adjacent to each target. This was chosen to be a 2 cm thick shell enclosing the target volume (Fig. 2), giving shells of 50, 60, 80 and 95 mm diameter, respectively. The dose falls from maximum to below 10 ~o over a distance of 2 cm using stereotactic techniques [ 19]. The volume of normal brain tissue receiving greater than 50 ~o of the isocentre dose was tabulated. A similar technique has been used in a recent publication [ 17].
32
Spherical target Shell 2 c m s thick
•:"::.::i:!:[.......:i:~:i:i.......':.:':'::-"
Fig. 2. A transverse CT slice through the isocentre showing the position of the target and volume of interest, a shell 2 cm thick around the target. However, such calculations can only provide a theoretical b a c k g r o u n d to the choice o f treatment technique and are not a substitute for the clinical assessment o f radiation tolerance o f normal brain outside or within the target volume.
irradiated above the 20~o isodose. The D V H s calculated from the smaller volume of interest were identical to the whole brain D V H s except at low dose levels, which were not relevant to this c o m p a r i s o n (Figs. 3 and 4). D V H s based on the 2 cm thick shell of normal tissue around the target volume were calculated for a n u m b e r of b e a m arrangements at four target sizes: 10, 20, 40 and 55 m m diameter. The results for the 10 and 40 m m diameter targets are shown in Fig. 4. In order to simplify the c o m p a r i s o n between different b e a m arrangements the volumes o f normal brain tissue receiving greater than 50 ~o o f the isocentre dose were plotted as histograms (Fig. 5). The distributions using multiple arcs were superior to a traditional arrangement o f 3 static planar beams at all target sizes. A single arc rotation or a conventional arrangement of 3 fixed beams was inferior to multiple arcs at all target sizes (Fig. 5). There was no increased sparing of normal tissue above the 50~o isodose with more than 3 arcs of rotation. T w o arcs gave a distribution that was only marginally worse than 3 arcs at the 20 m m diameter target size and was slightly better at the 40 and 50 m m diameter sizes. If 6 static n o n - c o p l a n a r beams are ar100"
.... a .... .... * - ' "
% volume
Results
80
D o s e volume histograms calculated for the whole brain volume are shown in Fig. 3. T w o stereotactic techniques at the 40 m m diameter target size were c o m p a r e d to an arrangement o f 3 fixed planar beams using a circular collimator. The 3 and 5 arc rotations gave virtually identical distributions and were superior to 3 fixed beams through the sparing of normal tissue
3 beams 6 beams
"~.
m
Single arc
i~"~
"
3 arc
60 ¸
40-
20-
0 0
20
40
60
100
80
% dose 3Arc
% volume 80"
II
$ Arc
•
%
.~ -~[
volum
•
•~
----a---
't~"" ~ .~ " ~ i g
60"
"~,, ~
40
L'o.-'~.,~ , . ~
40
,
6beams Single arc
•
2
•
3arc
a r e
"
S
•
9 arc
a r c
20
-
0
,
20
-
m
.
40
i
60
•
l
80
.
- i
100
% dose
Fig. 3. Dose volume histogram of percentage volume of the whole brain receiving a percentage of the isoeentre dose with a 40 mm diameter target, comparing 3 and 5 arc rotations with 3 fixed beams.
0
b . 0
. 20
.
.
. 40
60
80
100
% dose
Fig. 4. Dose volume histograms of a 2 cm thick shell around: (a) a 10 mm diameter target and (b) a 40 mm diameter target comparing rotational and fixed beam arrangements.
33 10
1000 % of target volume 80O
volume
(cc) 8 6
6O0 4 4OO
2
2O0
o
3 beams
6
beams
1
3 arc
arc
5 arc
9 arc -
o
lo
2o
3o
Target Diameter
30
4'o (ram)
U
to
"
6;
Fig. 6. The additional sparing of normal tissue irradiated above 50% of the isocentre dose by an arrangement of 3 non-coplanar arcs over a conventional arrangement of 3 fixed beams, expressed as a p r o p o r t i o n of the target volume for targets of 10, 20, 40 and 55 m m diameter.
volume
(cc) 20
tO
0
3 beams 6 beams
1 arc
2 arc
3 arc
5 arc
9 arc
3 arc
5 arc
9 arc
120" volume
(ce) 1008060 I 40
! 20
3 beams 6 beams
1 arc
2 arc
250'
ranged so as to obtain maximal angular separation between beams, a dose distribution equivalent to the multiple arc rotations can be achieved at target sizes 20 mm and over (Fig. 5b-d), but not at 10 mm (Fig. 5a). The relative sparing of stereotactic radiotherapy using 3 arcs over a standard 3 field set up was compared at the different target sizes, The 3 arc technique spared 68 ~o of the volume of normal tissue treated by 3 beams to > 50~o of the isocentre dose for the 10 mm diameter target. This fell to 47~o at 20 mm, 18~o at 40 mm and 13~o at 55 mm target diameter although the actual volumes of normal tissue spared were 5.6cm 3, 12.1 cm 3, 19.3 cm 3 and 22.8 cm 3, respectively; showing an increase with increasing target diameter. The relative benefit from the stereotactic technique, with the sparing of normal tissue plotted as a percentage of the target volume, can be seen predominantly with target sizes less than 40 mm diameter (Fig. 6). While sparing normal brain from higher doses, the stereotactic multiple arc techniques irradiated a greater volume to doses below 25 ~o of the isocentre dose when compared with conventional or single arc technique (Figs. 3 and 4).
200"
Discussion
150" 100" 50" 0" 3 beams
6 beams
1 arc
2 arc
3 arc
Fig. 5. H i s t o g r a m s of the volume of normal brain tissue receiving greater than 5 0 3 of the isocentre dose with various b e a m and arc arrangements for target sizes: ( a ) 1 0 r a m diameter; ( b ) 2 0 m m diameter; (c) 40 m m diameter and (d) 55 m m diameter.
The comparison of dose distributions between rival plans in standard radiotherapy with coplanar beams is usually performed along the central axis in the transverse plane. Distributions in the sagittal and coronal planes may give additional information but in order to accurately estimate the volume of normal tissue being irradiated the calculation of DVHs is essential. In this comparison of stereotactic techniques it was assumed that tolerance to radiation was uniform throughout the
34 brain, although the clinical expression of late radiation damage will vary with site. Assessments were based solely on reductions in normal tissue volumes outside the target volume irradiated to > 5 0 ~ of the isocentre dose, without regard to anatomical relationships. The target volume was defined at 90 ~o isodose and with all the techniques considered the target volume dose homogeneity was + 7.5 ~o. The greatest benefit from stereotactic radiotherapy should be seen with the technique that reduces the volume of normal tissue receiving doses above tolerance. While published figures are available for the tolerance of the brain to large volume irradiation with conventional techniques and fractionation [ 21 ], there is little equivalent data for stereotactic radiotherapy [9]. Doses prescribed vary widely and have ranged from 20 to 70 Gy as a single fraction, with doses being reduced as th e target volume increases [ 1,20]. In the treatment of cerebral AVMs doses of the order of 20 Gy have been given as a single fraction [ 12]. There is little data on the tolerance of the brain to single fraction radiotherapy but an R T O G study using a single fraction of 10 Gy was within the tolerance dose of normal brain to conventional radiotherapy [3]. A comparison based on the volume of normal tissue receiving >50~o of the prescribed dose appears justified. The results presented here, based on this principle, have shown that 3 noncoplanar arcs were as effective in terms of normal tissue sparing outside the target volume as a larger number of arCS.
What is the maximum target size that can be usefully treated with stereotactic radiotherapy? Comparing 3 or more arcs of rotation with a standard 3-field plan (Fig. 6) the relative benefit, as measured by the volume of normal tissue irradiated to greater than 50~o of the prescribed dose, fell from the 20 to the 40 mm diameter target sizes but altered very little between the 40 and 55 mm diameter target sizes. Any sparing of normal tissue may be beneficial but it is at the smaller target sizes that a clinically detectable benefit is likely to be seen [23]. Most centres limit stereotactic radiotherapy with a single isocentre to a maximum target size of 30-40 mm. The arrangement of 6 static non-coplanar beams, which showed equivalent sparing to multiple-arc techniques at larger target sizes, may be of considerable advantage in the treatment of malignant disease. Problems have arisen in encompassing large irregular
volumes using the multiple-arc technique which delivers an essentially spherical high-dose volume. To date, the only practical way of treating these irregular lesions has been to use either multiple isocentres and combine the "dose spheres" with resultant dose inhomogeneity within the target, or to treat with a large sphere which as shown above has minimal benefit over conventional techniques in terms of sparing of normal brain tissue. Recent reports suggest increased late complication rates associated with multiple isocentre techniques [ 16]. By using 6 static beams it may be possible to shape the target volume using customised blocking of the beam. This would be an alternative to a conformal multiple-arc technique with dynamic beam shaping using a multi-leaf collimator. Such technology is not currently available and would entail enormous complexity in planning and delivery. It must be stressed that the theoretical calculations presented are only a guideline to the optimum arc arrangement and do not attempt to provide any information on the radiation tolerance of the technique of stereotactic radiotherapy. Considerations such as the dose and the homogeneity of dose distribution within the target volume as well as the amount of normal brain within it may also be the determinants of tolerance and will have to be evaluated in clinical studies. In summary, for spherical volumes less than 4 cm in diameter a multiple-arc technique appears to give the greatest sparing of normal tissue to high dose. There is no significant benefit from an arrangement of more than 3 arcs for the spherical target sizes studied here. Because of exit doses from a sagittal arc to the whole body and particularly the thyroid, we currently use 4 arcs for the treatment of brain tumours and occasionally up to 5 arcs in the treatment of AVMs near critical structures. For larger, irregular volumes an arrangement of a small number (usually 6) of static noncoplanar beams may be indicated as customised blocking of the beam would be practical and would achieve the best sparing of normal tissue outside the tumour volume.
Acknowledgement This work was supported by the Professor Bloom Research Fund, The Cancer Research Campaign & The Royal Marsden Hospital.
References 1 Betti, O. O., Munari, C. and Rosier, R. Stereotactic radiosurgery with the linear accelerator: treatment of arteriovenous malformations. Neurosurgery 24: 311-321, 1989.
2 Betti, O.O. and Derechinsky, V. Multiple beam stereotaxic irradiation. Neurochirurgie 29: 295-298, 1983. 3 Borgelt, B., Gelber, R., Larson, M., Hendrickson, F., Griffin, T.
35
4
5
6
7
8
9
10
11 12
13
and Roth, R. Ultra-rapid high dose irradiation schedules for the palliation of brain metastases: final results of the first two studies by the Radiation Therapy Oncology Group. Int. J. Radiat. Oncol. Biol. Phys. 7: 1633-1638, 1981. Chin, L. M., Cheng, C. W., Siddon, R. L., Rice, R. K., Mijnheer, B. J. and Harris, J. R. Three-dimensionalphoton dose distributions with and without lung corrections for tangential breast intact treatments. Int. J. Radiat. Oncol. Biol. Phys. 17: 1327-1335, 1989. Gill, S. S., Warrington, A. P., Thomas, D. G. T. and Brada, M. Relocatable frame for stereotactic external beam radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 20: 599-603, 1991. Hartmann, G. H., Schlegel, W., Sturm, V., Kober, B., Pastyr, O. and Lorenz, W.J. Cerebral radiation surgery using moving field irradiation at a linear accelerator facility. Int. J. Radiat. Oncol. Biol. Phys. 11: 1185-1192, 1985. Houdek, P. V., Fayos, J. V., Van-Buren, J. M. and Ginsberg, M.S. Stereotaxic radiotherapy technique for small intracranial lesions. Med. Phys. 12: 469-472, 1985. Kjellberg, R.N. Stereotactic Bragg peak proton beam radiosurgery for cerebral arteriovenous malformations. Ann. Clin. Res. 18 (Suppl 47): 17-19, 1986. Kjellberg, R. N., Hanamura, T., Davis, K. R., Lyons, S. L. and Adams, R.D. Bragg-peak proton-beam therapy for arteriovenous malformations of the brain. N. Engl. J. Med. 309: 269-274, 1983. Leibel, S. A., Gutin, P. H., Wara, W. M., Silver, P. S., Larson, D. A., Edwards, M. S. B., Lamb, S.A., Ham, B., Weaver, K. A., Barnett, C. and Phillips, T. Survival and quality of life after interstitial implantation of removable high activity iodine-125 sources for the treatment of patients with recurrent malignant gliomas. Int. J. Radiat. Oncol. Biol. Phys. 17:1129-1139, 1989. Leksell, L. The stereotaxic method and radiosurgery of the brain. Acta Chir. Scand. 102: 316-319, 1951. Loeflter, J. S., Alexander, E., Siddon, R. L., Saunders, W. M., Coleman, C. N. and Winston, K.R. Stereotactic radiosurgery for intracranial arteriovenous malformations using a standard linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 17: 673-677, 1989. Lunsford, L. D., Flickinger, J., Linder, G. and Maitz, A. Stereotactic radiosurgery of the brain using the first United States 201
cobalt-60 source gamma knife. Neurosurgery 24: 151-159, 1989. 14 Lyman, J.T., Kanstein, L., Yeater, F., Fabrikant, J.I. and Frankel, K.A. A helium ion beam for stereotactic radiosurgery of central nervous system disorders. Med. Phys. 13: 695-699, 1986. 15 Mazal, D.A., Rosenwald, J-C., Gaboriaud, G. and Porcher, J, B. Implication of basic physical phenomena on the dose distribution in stereotactic x-ray external irradiation (Abstr.). 17th International Congress of Radiology, p. 217, Paris, July 1-8, 1989. 16 Nedzi, L.A., Kooy, H., Alexander, E., Gelman, R.S. and Loeffler, J.S. Variables associated with the development of complications from radiosurgery of intracranial tumours. Proc. Am. Soc. Ther. Radiat. Oneol., October 15-19th, 1990. Int. J. Radiat. Oncol. Biol. Phys. 19 (Suppl 1): 149, 1990. 17 Phillips, M. H., Frankel, K. A., Lyman, J. T., Fabrikant, J. I. and Levy, R.P. Comparison of different radiation types and irradiation geometries in stereotactic radiosurgery. Int. J. Radiat. Oncol. Biol. Phys. 18: 211-220, 1990. 18 Podgorsak, E. B., Olivier, A., Pla, M., Lefebvre, P. Y. and Hazel, J. Dynamic stereotactic radiosurgery. Int. J. Radiat. Oncol. Biol. Phys. 14: 115-126, 1988. 19 Podgorsak, E. B., Pike, G. B., Olivier, A., Pla, M. and Souhami, L. Radiosurgery with high energy photon beams: a comparison among techniques. Int. J. Radiat. Oncol. Biol. Phys. 16: 857-865, 1989. 20 Sturm, V., Kober, B., Hover, K. H., Schlegel, W., Boesecke, R., Pastyr, O., Hartmann, G. H., Schabbert, S., zum-Winkel, K., Kunze, S. and Lorenz, W.J. Stereotactic percutaneous single dose irradiation of brain metastases with a linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 13: 279-282, 1987. 21 Tiver, K. Treatment ofCNS tumours with conventional radiotherapy: the importance of dose and volume factors in tumour control and CNS radiation tolerance. Australas. Radiol. 33" 15-22, 1989. 22 Walton, L., Bomford, C.K. and Ramsden, H.N.C. The Sheffield stereotactic radiosurgery unit: physical characteristics and principles of operation. Br. J. Radiol. 60: 897-906, 1987. 23 Withers, H.R., Taylor, J. M. G. and Maciejewski, B. Treatment volume and tissue tolerance. Int. J. Radiat. Oncol. Biol. Phys. 14: 751-759, 1988.
