Clinical Radiology (1985) 36, 209-212 0009-9260/85/399209502.00 © 1985 Royal College of Radiologists

Case Report: Treatment of Medulloblastoma Using a Computer-controlled Tracking Cobalt Unit T E R E S A T A T E , J. A. B R A C E , T. J. D A V Y , H. M. M O R G A N , D. B. L. SKEGGS and A. J. T O O K M A N

Departments of Radiotherapy & Oncology and Medical Physics, Royal Free Hospital, London

Radiation therapy of the length of the spinal column presents various clinical and physical problems. The completed plan may be complicated to set up, be time-consuming and require daily variation to achieve reasonable dose homogeneity. A case of medulloblasto- ma is used to illustrate the steps in producing a plan for dynamic treatment using a computer-controlled track- ing cobalt unit. After definition by computed tomogra- phy, the target is considered in segments in order to develop a plan which keeps the spinal cord constantly positioned at the beam isocentre. The main computer is used to develop the patient treatment file and informa- tion is transferred to a second computer which controls and monitors the safe functioning of the cobalt unit. The cranial fields are treated separately in a conven- tional way. Good and consistent control of the dose distribution is achieved along the entire target volume. This technique is a marked improvement over all existing methods of treating the spinal axis.

A computer-control led tracking cobalt unit, a T E M MS 90 Mobal t ron, has been in use at the Royal Free Hospital for 4 years (Brace et al., 1981a). Tracking treatments are suitable for tumours which are exten- sive, irregularly shaped or awkwardly orientated in the body.

Until recently, the major obstacles to the routine use of dynamic therapy have been the difficulty in obtaining sufficient diagnostic information and the complexities of t rea tment planning. Initially, conven- tional diagnostic and orthogonal planning radiographs were used. Since 1982 a computed tomography (CT) scanner, a Siemens Somatom D R 2, has been used for planning. The t rea tment planning system is based on a Hewlet t -Packard minicomputer . This is linked to a Hewlet t -Packard microcomputer (Brace, 1982) which controls the tracking cobalt unit. The planning system does not yet produce the complete t rea tment plan, but does give all the necessary geometric projections and some dose calculations along the isocentre locus using some simplified beam profiles.

This system can be used to carry out various types of dynamic therapy and this paper describes the radiation t rea tment in a case of medulloblastoma, which demonstra tes how computer control gives a greatly improved technical and clinical approach to a standard procedure.

On examination, he was well and alert with a left facial palsy, nystagmus on lateral gaze in both eyes, full visual fields, no papilloedema and marked cerebellar ataxia.

A CT scan of the head showed a mass of slightly increased attenuation in the midline, extending to the left, with irregular cavitation which appeared to communicate with the fourth ventricle. These appearances were compatible with medulloblastoma.

At operation, approximately 90% of the tumour in the fourth ventricle and surrounding the brain stem was removed. Histology confirmed the diagnosis of medulloblastoma. Post-operatively, the child developed a left hemiparesis and brain stem signs, but remained in a stable condition. He was not taking dexamethasone at the beginning of his radiation therapy.

PLANNING TECHNIQUE

The CT scanner was used for tumour localisation, with the couch modified by a flat board insert to reproduce the patient position on the t reatment couch. Two wires run parallel down the length of the board at 15 cm separation as magnification gauges and to mark the inferior surface of the patient. The patient, lying prone in a specially constructed cast, bears one anterior midline and two horizontal lateral marks (Brace et al., 1981b) along lines defined by laser to ensure accuracy of set-up and to aid computer- controlled planning (Fig. 1). Radio-opaque fine cathe- ters are used to define these lines before localisation as they have been found to cause minimal artefact during scanning. Scans were made through the post- erior fossa and the whole spine. The scan interval can be selected according to the shape of the tumour and in this case was 30 mm. The therapist then defines the tumour volume and regions of interest, such as the kidneys, either directly onto the viewing console or, as in this case, onto magnified hard copies of each

CASE REPORT

A 4-year-old white boy presented with a 1-week history of !rritability, unsteadiness and, on the day of admission, of walking into objects.

Fig. 1 - CT scan showing target volume and markers for the coronal and sagittal reference planes, together with a 15 cm wide magnifica- tion gauge (L+40, W 300).

210 CLINICAL RADIOLOGY

/s" '~,% / '%

/ % / %

' ) I I

l i n e ~ J - --4~ . . . . . 4~ . . . . . , - -

.~, . . . . = . . . .

.-~

~, , \ %

Outline of target volume ~" I :~ ... .... Kidneys ~ /

100

K e y

Instantaneous size and position of moving beam

showing the 80% width and 50% length

- - " 4 - - - - Locus of isocentre (radiation on)

-- -<~ - - Locus of isocentre (radiation off)

i I I 1 I I I

150 200 250 300 350 400 450

Position along superior- infer ior axis (CT scan position) mm

Fig. 2 - The antero-poster ior view of the spinal target volume and the dynamic generation of a long field.

i

I

~IE=

I. I 500

scan. At present, the planning system still requires that the plan be constructed manually. A life-size, longitudinal projection of the spine was made and the entire tumour length divided into short, contiguous sections for planning purposes (Fig. 2). The length of these sections is determined by changes in the width and depth of the tumour. The shorter these lengths the more accurate will be the fit to each segment of tumour, but a compromise is dictated by the smallest possible collimator settings on the telecobalt unit and by constraints of the length of treatment time. The greater the number of individual segments, the greater will be the treatment time. In this case a segment length of 4 cm was used. The beam widths, entry angles and isocentre positions are determined for each segment in the conventional manner. In the tracking technique the target volume is effectively, straightened for treatment, with the couch vertical and lateral movements being used to bring the planned isocentre position into coincidence with the machine isocentre.

The dose to any point is proportional to the length of time this point remains in the beam and, hence, is controlled by speed and beam length, as measured in the direction of axial couch movement. The speed of longitudinal couch travel necessary to give the pre- scribed dose to each segment is calculated using the tissue-air ratio, the machine output and effective beam length (Davy, 1979). Certain geometric man- ipulations of the beam size are used to achieve equal dosage at both ends of the track (Davy, 1985).

From this information an integrated exposure time profile can be constructed and graphically displayed (Fig. 3). All of these data are entered into the minicomputer (Brace, 1979) and a patient treatment file is created containing the cobalt unit control data. To confirm that the machine is performing as ex- pected, a phantom containing both film and direct- reading dosimeters is used. A specially modified test file is created to demonstrate dose control along the track; this is especially important for keying-on fields.

Superior Direction of beam travel relative to the patient

~L B,

• ~ I

.-~I

=_ol

1 . 0

100 I I I t I

K e y

i 150

i I Radiation on

~I, 3=1.90 cm/min Radiation off

,I, Rapid couch advance

Instantaneous size (50% length) and position of moving beam Speed of axial couch movement

200

"[ Radiation on • ~" = J 3=3.80 cm/rain

R a d i a t i o n o n

, ~ . , 3=3.80 cm/min

250 300 350 400 450

Position along superior- infer ior axis (CT scan position) mm

Inferior

L

I "] Radiation on

, j' r=3.10 cm/min "1 Radiation off

I t Rapid couch advance ,,It ~ ~ "~[ Radiation on

I , J T= l .5Ocm/min

I

1500 I I

Fig. 3 - Diagram showing the pat tern of beam movement and the corresponding integrated exposure time profile.

C O M P U T E R - C O N T R O L L E D T R A C K I N G I N M E D U L L O B L A S T O M A 211

The results are plotted and should agree with the dose profile calculated from the exposure time profile. Corrections can be made easily by modifying the treatment file.

THE COMPLETE PLAN

The overall plan consisted of three components. (1) Parallel opposed isocentric 6°Co 18 c m x l 8 cm

portals to treat the whole cranium to the second cervical vertebra with individually constructed lead blocks to the eyes and nasopharynx. A midplane dose of 4320 cGy in 27 daily fractions over 38 days to the whole brain was prescribed to give a biologically equivalent dose, calculated from TDF tables, of 4000 cGy in 20 fractions.

(2) A cone-down to the posterior fossa using 8cmx6.5 cm parallel opposed portals to give a mid- plane dose of 1000 cGy in five fractions over- 9 days.

(3) The spinal field was tracked from the lower border of the second sacral vertebra to key-on to the 50% border of the cranial fields as calculated at the midplane. The spine dose was 3840 cGy target absorbed dose (International Commission on Radia- tion Units and Measurements, 1978) in 24 daily fractions over 33 days, the TDF equivalent of 3500 cGy in 17 fractions.

The beam configurations and dose distributions from the spinal track and the cranial fields in the region of the join are shown in Fig. 4. A homogeneous dose distribution is demonstrated and this was con- tinued down the entire length of the track. The centre of the head fields was expressed as an exact distance

/ / /

/

//

/

u3

z ~- O 50 -J z

O x [:3 <

Latera l

I, 4, I . . . . . . . . . . . . .

l

)" . . . . . . . . . . . . + . . . . t Spinal tracked field

~ L a t e r a [ f i e l d i s o c e n t r e

( 9 0 c m S A D )

14,

I .~ co I o c

i o

I , c c

g 5 ,

~: t~0 z <~

J I Q_

~Z o z

----- 100

---~Mid-line

Tota l Dose ( b ra in*sp ine f ie tds) : ~ .

Dose from spinal t rack

J \ i Dose f rom b ra in f ie lds

0 t i , i *

0 50 100 150 200 250

POSITION ALONG S U P E R I O R - INFERIOR AXIS ( CT SCAN POSITION ) mm

Fig. 4 - Beam configurations and dose distributions along the spinal axis in the region of the join of the spinal and cranial fields (cast lead shielding blocks not shown).

212 CLINICAL RADIOLOGY

from the couch position at the end of the spinal track. At each treatment session the radiographers recorded the couch finishing position and made the necessary displacement to centre the cranial fields. This position was confirmed by a centre marked on the head shell. The running time of the spinal track program was 11 rain.

SET-UP AND MACHINE O P E R A T I O N

As well as the t reatment parameters , the pat ient 's file contains identification data, a record of the prescription and progress through t rea tment and set- up instructions for the radiographers. The computer will not allow treatment to start until all the set-up details are confirmed. Various safety mechanisms are built into the program and the radiographer can halt the t reatment manually at any t ime (Brace, 1982).

COURSE DURING T R E A T M E N T

For the first 3 days thermoluminescent dosimeters were used to measure eye doses. These indicated a total dose of 400 cGy to the right eye and 500 cGy to the left during the first phase of the t reatment . On review of the lead positioning it was felt that no increase in eye shielding could be made without compromising the t rea tment of the frontal lobe dura.

After 5 days of t rea tment the pat ient 's condition deteriorated and a new CT scan demonstra ted hyd- rocephalus as well as residual tumour around the brain stem. A ventriculo-peritoneal shunt was inserted and a course of dexamethasone 2 mg three times a day was commenced. This resulted in clinical improvement and radiation was continued without incident. In particu- lar, no symptoms of nausea, vomiting or headache developed and the blood count remained within nor- mal limits throughout. Oral sedation was used for most of the t reatment course.

The child was considered to have .a poor prognosis and was started on weekly injections of vincristine 1 mg during irradiation. At the complet ion of treat- ment he was generally well but with persistent brain s tem signs and a left hemiparesis.

DISCUSSION

This case report describes a computer-control led t reatment technique which has many advantages over conventional static field methods for whole neuro-axis therapy. The whole spine is t reated with a continuous moving field of megavoltage radiation, at standard source-axis distance (SAD), thus eliminating the need for gaps and the risk of over- or underdosage. The beam divergence of the end of the fields is kept at a minimum by the use of small fields at both ends. The tracking technique will allow the generation of a homogeneous dose distribution along the whole length of the spinal cord, rather than the dose being pre- scribed to an average tumour depth (Chang et al., 1969). At the same time the need for compensators is eliminated.

The method gives an easy set-up and the t reatment t ime of 11 min compares very favourably with that of 15-20 min using two single fields with lead shielding (Landberg et al., 1980) or 20-60 min using a single orthovoltage field at extended source-skin distance (Bottrill et al., 1965). This decrease in both the actual t reatment t ime and the number of set-ups ensures bet ter co-operat ion of the patient.

The production of three-dimensional plans is very t ime-consuming and involves much detailed work, but in this instance the development of a relatively simple plan with a single track took less t ime than a conven- tional plan. Work to produce a computerised system of planning is nearing complet ion and should allow machine control data to be available within 1 day.

Because the skin gap between spinal and cranial fields is controlled by the collimator setting and the couch position and is not dependent upon marks, there is no difficulty with overdosage to the cervical cord if steroid oedema necessitates the construction of a new head shell during treatments .

The Royal Free automated t reatment system has proved extremely reliable and does not encroach into t reatment t ime for maintenance any more than a standard unit.

Acknowledgement. The authors wish to thank Dr L. F. N. Senanayake, Consultant Radiotherapist, for permission to report on his patient, and also the Imperial Cancer Research Fund for financial support. T.T. and J.A.B. are funded by the Cancer Research Campaign.

REFERENCES

Bottrill, D. O., Rogers, R. T. & Hope Stone, H. F. (1965). A composite filter technique and special patient jig for the treat- ment of the whole brain and spinal cord. British Journal of Radiology, 38, 122-130.

Brace, J. A. (1982). A computer controlled telecobalt unit. Interna- tional Journal of Radiation, Oncology, Biology, Physics, 8, 2011-2013.

Brace, J. A. (1979). A computer system for the dynamic control of a telecobalt unit. In Fundamentals in Technical Progress, pp. II.12.1-18. Proceedings of a Symposium, Liege, Belgium, May 1979, Presses Universitaires de Liege.

Brace, J. A., Davy, T. J. & Skeggs, D. B. L. (1981a). Computer controlled cobalt unit for radiotherapy. Medical and Biological Engineering and Computing, 19, 612-616.

Brace, J. A., Davy, T. J., Skeggs, D. B. L. & Williams, H. S. (1981b). Conformation therapy at the Royal Free Hospital. A progress report on the tracking cobalt unit. British Journal of Radiology, 54, 1068-1074.

Chang, C. H., Housepian, E. M. & Herbert, C., Jr (1969). The operative staging system and a megavoltage radiotherapeutic technique for cerebellar medulloblastomas. Radiology, 93, 1351- 1359.

Davy, T. J. (1979). Dynamic treatment using a computer controlled telecobalt-60 unit. In Fundamentals' in Technical Progress, pp. II.11.1-17. Proceedings of a Symposium, Liege, Belgium, May 1979. Presses Universitaires de Liege.

Davy, T, J. (1985). Physical aspects of conformation therapy using computer controlled tracking units. In Progress in Medical Radiation Physics, ed. Orton, C., Vol. 2. Plenum, New York. In press.

International Commission on Radiation Units and Measurements (1978). Dose specification for reporting external beam therapy with photons and electrons. ICRU Report 29, Section 2,2,2.

Landberg, T. G., Lindgren, M. L., Cavallin-Stahl, E. K., Svahn- Tapper, G. O., Sundbarg, G., Garwicz, S. et al. (1980). Improvements in the radiotherapy of medulloblastoma 1946- 1975. Cancer, 45, 670--678.