Reprinted front Journal of Neurology, Neurosurgery, and Psychiatry 1983 ;46:797-803Copyright @ 1983 Journal of Neurology, Neurosurgery, and PsychiatryAll rights of reproduction of this reprint are reserved in all countries of the world

Stereotactic radiosurgery

LARS LEKSELL

"ENCLOSURE"

#1

British Medical Association, Tavistock Square, London WC1H 9JR

Journal of Neurology, Neurosurgery, and Psychiatry 1983;46:797-803

Occasional review

Stereotactic radiosurgeryLARS LEKSELL

From the Department of Neurosurgery, Karolinska Sjukhuset, Stockholm, Sweden

SUMMARY The development and scope of stereotactic radiosurgery is described. The technique,which combines well with the latest diagnostic methods, has already proved a safe and effectiveway of treating inaccessible cerebral lesions and in particular small arteriovenous malformations,acoustic neuroma and the solid component of craniopharyngioma, as well as playing an increas-ingly useful role in the therapy of pituitary adenoma.

Stereotactic radiosurgery is a technique for the non-invasive destruction of intracranial tissues or lesionsthat may be inaccessible or unsuitable for opensurgery. The open stereotactic method provides thebasis for radiosurgery and the first stereotacticinstrument was designed for use with probes andelectrodes.' The principle of the instrument, withthe target in the centre of a semicircular arc, made iteasily adaptable for cross-firing of the target withnarrow beams of radiation. The first attempt to sup-plant the electrodes with ionizing radiation wasmade in the early fifties, with X-rays.2

1 It was tempt-ing to try to reduce the hazards of open surgery andby the administration of a single heavy dose of radia-tion it appeared possible to destroy any deep brainstructure, without risk of bleeding or infection.

Ten years later considerable progress had beenmade, due in considerable measure to the contribu-tion of the physicists Kurt Liden and Borje Larsson,and fig 1 shows the first proton beam operation inUppsala.' 6 The heavy particle beam was an excel-lent knife blade but the synchro-cyclotron was tooclumsy. A similar technique was developed for alinear accelerator. The next step was to get to apractical, precise and simple tool which could behandled by the surgeon himself.

The first stereotactic Gamma Unit, using Cobalt60 (fig 2), was installed at the Sophiahemmet Hospi-tal in 1968 and was primarily intended for use infunctional brain surgery for the section of deep fibretracts or nuclei." The lesions were disc shaped and

Address for reprint requests: Prof. Lars Leksell, KarolinskaSjukhuset, Stockholm, Sweden.

Based on a Sir Hugh Cairns Memorial Lecture delivered in 1981 ata meeting of the Society of British Neurological Surgeons.

Received 25 March 1983 and in revised form 13 April 1983.Accepted 16 April 1983

Fig 1 Stereotactic proton beam operation. A bilateralanterior capsulotomy performed at the Gustaf WernerInstitute in Uppsala in 1960.

very sharply circumscribed 9 (fig 3). This apparatuswas also used for the irradiation of some tumoursand arteriovenous malformations.'"" The resultswere promising and a second Gamma Unit, withmore generally suitable spherical fields of radiation,was constructed and installed at the KarolinskaHospital in Stockholm in 1974. This machine wasstill, in some respects, a prototype, but it has provedvery reliable and easy to use.

The relatively slow development of radiosurgeryhas been due partly to the absolute need for a pre-cise localisation of the target. Many problems havebeen solved by the arrival of the new imaging tech-niques. However, the varying nature of the targets

797

798 Leksell

Central body

Beam sources Col limator helmet

Fig 2 Sectional diagram showing principle of cross-firingof target with narrow beams of gamma rays from multiplecobalt sources.

requires a flexible approach to the localisation and

at present several different diagnostic methods are

used.

The standard radiographic procedure in the

Fig 4 (A) stereotactic radiographic localisation in the

operating theatre. AP and lateral views are made and thelocalisation procedure can be performed in any operatingroom. (B) Stereotactic CT localisation. The coordinateframe is fixed to the table by means of a magnetic adaptor.Plastic indicator discs are temporarily mounted on the framefor the determination of the target coordinates.

operating room, when plain radiographs or air-

encephalography are used, remains very useful (fig4A). A geometric diagram

7 corrects the distortion

due to the short-distance projections and allows

rapid determination of the co-ordinates. In someinstances, particularly when angiography is used, a

stereophotogrammetric technique may be prefer-

able. The target co-ordinates on the proximal anddistal scales of the co-ordinate frame are then read

off the film and fed into a small calculator. The cor-rect co-ordinates are computed automatically and

recorded on a printer.

The introduction of computer tomography has

Fig 3 Gamma thalamotomy. Radiosurgical lesion in a produced a revolution in stereotactic localisation.

patient with intractable cancer pain. Medial thalamus, The early EMI-scanner was used for this purpose by

section 30°posterior to coronal plane. Scale in mm. Bergstr6m and Greitz'2 and now CT localisation has

Unstained formalin fixed brain, become the most common procedure (fig 4B). The

Stereotactic radiosurgery

r

%IFig 5 CT localisation of a cystic craniopharvngioma. Thefilm is superimposed on a transparent millimeter scale andthe target coordinates are read off with a ruler.

CT film is placed on a semitransparent disc withco-ordinate millimetre scales, and all three co-ordinates are easily read off (fig 5).1

When the single X-ray or proton beamn was usedthe target was cross-irradiated by adjusting the posi-tion of the patient's head. In the Gamma Unit thehead is rigidly fixed in the collimator helmet. Thissimplifies the procedure and ensures high mechani-cal accuracy. When the stereotactic localisation ofthe target and dose planning have been completedthe patient is transferred to the unit and the head isfixed so that the target structure is in the centre ofthe helmet (fig 6). The irradiation procedure is eas-ily performed by the surgeon, with one assistant.

799

General anaesthesia is not required unless thepatient is unusually apprehensive or restless, and,usually, the patient can leave the hospital directlyafter the operation if this is considered appropriate.

Clearly, the clinical results obtained must serve asthe basis for the evaluation of the method and forfuture planning long term follow-up studies arenecessary. Inevitably, before the full effects of singledose radiation were known, the early selection ofpatients was somewhat arbitrary and over the yearsa number of improvements have been made in tech-nique and dose planning. At present more than 700patients have been treated radiosurgically and mostof the applications and results have been pub-lished.' ' 1 3 24 2 6 ,7

For a long time stereotaxis was largely identifiedwith functional brain surgery. Now this field repres-ents a relatively small sector. The surgery of painwill probably remain and the early group of gamma-thalamotomies for intractable cancer pain gavemuch valuable information concerning the anatomyof the lesions and the dose of radiation required." '•Medial thalamotomy offered an attractive possibilityof obtaining pain relief without sensory loss. Theclinical results were somewhat inconsistent and usu-ally of limited duration, although more effective forthe relief of pain in the upper part of the body. Now,with a greater knowledge of the functional anatomy,this theme may well be taken up again, perhaps evenas an out-patient procedure.

The advantages of radiosurgery were obvious inthe gamma capsulotomies performed for severe anx-iety and obsessive-compulsive states, but this field issmall and psychosurgery meets much oppositionfrom ideologists in Sweden and elsewhere in theworld.

The functional visualisation obtainable by meansof the positron camera or nuclear magnetic reso-nance may lead to further changes in the whole fieldof functional stereotaxy. For example, deep epilep-tic foci could become targets for stereotactic irradia-tion.

Today stereotactic techniques have become partof the general neurosurgical armamentarium and. atthe Karolinska, about 25% of all the operative pro-cedures are stereotactic, and of these about half areradiosurgical. Table I shows the number of patientsin the various diagnostic groups. At presentstereotactic radiosurgery is most commonly used forvascular malformations, acoustic tumours andcraniopharyngiomas. The last-named form a par-ticularly interesting group because of the combina-tion of open and closed stereotactic techniques andall three groups illustrate well the way thatstereotactic treatment can complement more con-ventional surgery. For example, the successful extir-

Fig 6 Fixation of the patient's head in the collimatorhelmet for the treatment of a right-sided acoustic tumour.

800

Table 1 Stereotactic radiosurgery 1968-194

Arterio-venous malformationArterial aneurysmAcoustic tumourCraniopharyngiomaMeningiomaPineal tumoursPituitary adenoma (non-secreting)Cushing's diseaseAcromegalyHypophysectomy

Intractable painTrigeminal neuralgiaAnxiety and obsessive-compulsive statesParkinsonismVariousTotal number of patients

Table 2 Stereotactic treatment of craniopha1966-1980 (Predominantly cystic tumours (intracystic9 0 y)

Tumour not Recpreviously aft,treated sur

Alive and well 29 16Operative mortality 0 2Recurrences 0 0Later deaths 3* 1No follow-up 2 1

34 20

*Stroke 1, hepatitis 1, leukaemia 1, (Necr

Leksell

82 15 year material19 good results were obtained in 54patients treated by stereotactic Yttrium 90 injection

205 alone (table 2). This is followed by a gradual shrink-

94 age and eventually a collapse of the cyst. Solid and22 multicystic tuinours are best treated by external2023 stereotactic irradiation2" and the sharp limitation of37 the radiation field in the Gamma Unit is a consider-9527 able advantage. The resultant shrinkage can be fol-24 lowed easily by CT (fig 7). This combined intracystic63 and precise external irradiation has proved satisfac-26 tory and the results do not depend on a skilled5 surgeon practising his technique frequently.34

762 Since 1975 95 patients with Cushing's disease andNelson's syndrome have been treated by gamma

'ryngiomas irradiation.2 Small and spatially well defined vol-

1-2 cysts), umes of tissue in the pituitary gland can be selec-tively destroyed. With the smallest collimator a kindof intrasellar "micro-radiosurgery" is possible. In a

current tumour review of 37 patients with an observation time oferpreviousr ery more than one year, complete remission was

obtained in 29 cases. This is a promising result. An

important need is to find a satisfactory method forprecise localisation of the small intrasellar tumours.

The treatment of acoustic tumours makes a fas-N = 54 cinating historical chapter. When Folke Henschen

opsy, 2 cases: wrote his thesis in 1910, 34 out of 42 patients col-lected had died from the operation. Fifty years ago

Backlund, 1980 Hugh Cairns read a paper in the Royal Society ofMedicine, which demonstrated his surgical skill andthe efforts required at that time for radical removal

res dexterity with conservation of the facial nerve."2 This opera-heir relative tion still represents a surgical challenge and in smallsurgeons to and medium sized tumours sterotactic gamma

ority of cases irradiation is a practical alternative. Hitherto 94ng cyst, and patients with acoustic tumours have been treatede injection's and a follow-up study of the first two groups oper-a review of a ated upon with the first unit in Sophiahemmet and

complete snnnkage of cysts)tCardiac infarction

pation of a craniopharyngioma requirand long surgical training. Moreover tinfrequency makes it difficult for allobtain sufficient experience. In the majthe symptoms are due to the expandiobliteration of the cystic parts by isotopsolves most of the clinical problems. In

r15 i ojorwotuctic ruousurgery jur soua crawopnaryngboma. Leir: preoperanve scan, mtaaie ana ngnt:postoperative controls 6 and 18 months after the operation. (Backlund 1980).

Sterecitactic radiosurgery 801

Table 3 Stereotactic radiosurgery in acoustic tumours (Patients treated at Sophiahemmet with gamma irradiationexclusively)

Pat Sex/Age (yr) Follow-up Tumour size Speech discrimination Cranial nerve Functional Commentsperiod (mm) score function status(Yr) initial last follow-up (8th nerve excl)

1 F/34 11 12 70% Deaf Normal Normal Bilat tumours,the left-sidedtreated in UnitII (see text)

2 F/54 8 24 64% 34% Normal Retired

due to age3 F/48 7 14 96% 42% Slight facial Normal

hypoaesthesia;transient EMGsigns

4 F/56 6 10 78% 22% Normal Normal5 M/54 7 13 58% 14% Normal Retired

due to age6 M/39 7 10 2% Deaf Normal Normal Re-irradiation

in Unit II7 M/61 5 9 40% 2% Normal Retired

due to age

with the second at the Karolinska, respectively, hasbeen published.2 3 The small group of patients withacoustic neuroma treated at the Sophiahemmet,(table 3) who have been followed for a sufficientlylong period (up to 12 years), illustrated the pos-sibilities and problems inherent with this technique.Following a sufficient dose of radiation the necrotictumour tissue usually shrinks slightly but a massremains. If the dose has been too small the tumourmay start to grow again, even after a long latency,and then has to be re-irradiated or operated on con-ventionally. Facial weakness may occur but so far ithas always been transient. Hearing is often pre-served, at least to some extent."4 The first patient inthe series is of particular interest. She had bilateraltumours, with a strong family history (her father,one brother and two children also had bilateraltumours). The technique was not yet fully developedand one segment of the right sided tumour receivedtoo low a dose. This tumour started to grow againand nearly 12 years later it was removed surgicallywith a resulting facial paralysis. Re-irradiationwould probably have been a better alternative. Inorder to evaluate fully the relative merits of openexcision or stereotactic radiation of acousticneuroma further experience is needed. However,even an experienced and skilful surgeon might havedifficulty in reproducing the results obtained withradiosurgery: that is to say no operative mortality,no permanent facial weakness and in some casespreservation of hearing. Nowadays, CT scanning hasmuch improved localisation and the treatment isalmost an out-patient procedure. Furthermore, con-ventional surgery is not precluded if required later.The policy at the Karolinska is to use stereotacticradiosurgery in the first instance, for neurinomas

Fig 8 Stereotactic radiosurgery for arteriovenousmalformation. Preoperative angiogram and controlangiogram 2 years later.

802

measuring up to 25 mm in diameter. Several ques-tions remain to be answered, particularly the rela-tive radiosensitivity of the cochlear nerve and theoptimal dose for destruction of the tumour with pre-servation of hearing.

So far 204 patients with arteriovenous malforma-tions have been treated with stereotacticradiosurgery. In the first patient only the feedingarteries were irradiated." This attempt preceded byfar any rational knowledge concerning the effect of asingle heavy dose of gamma radiation on arteries,and it was a shot in the dark. Despite this, the selec-tive irradiation of the feeding vessels succeeded inthis first case. Some experiments on the basilarartery of the cat showed how difficult it may be toobliterate a large healthy artery by irradiationalone.-" However, from the first clinical trials a satis-factory technique for the treatment of small deepAVM's has emerged.2""7 In a series of 67 patientsfollowed for two years, where the fields of radiationcovered the whole malformation, there was totalobliteration of the pathological vessels in 83% ofcases (table 4). The long latency of at least 6 to 18months before obliteration is a disadvantage but therisk with radiosurgery in small, deep-seated, ofteninoperable AVM's is extremely small. There wereonly five cases of mild hemiparesis followingstereotactic irradiation in the whole group of 204patients.

In conclusion, it seems clear now that a firstperiod of tentative development of radiosurgery hascome to an end. The technique has been reviewed,some much needed improvements have been madeand a new improved Gamma Unit has been con-structed.

The term stereotactic radiosurgery has been useddeliberately to stress the fact that this combinationof mechanically directed instruments and modernradiation physics is still surgery, albeit using anotherphysical agent in place of the knife or radiofre-quency heat lesion. Indeed, the scalpel has neverhad a dominating place in the surgical handling ofthe brain, at least not since the introduction ofdiathermy. The Gamma Unit merely represents achange in the type of energy used. Modernneurosurgery should not rely on the open visualmethods alone but must also incorporate the newertechniques. In this way the use of narrow beams of

Table 4 Stereotactic radiosurgerv in arterio-venousmnalfornzations

Number of Observation Total Partial No changepatients period obliteration obliteration85 1 year 34 (40%) 34 (40%) 17 (20%)

67 2 yeas 56 (83.5%) 7 105%) 4 (5.9%)

Leksell

ionizing radiation has little to do with radiotherapyin its conventional meaning, but the communicationlines between the territories must remain open. Inthe same way there should be no sharp demarcationline with microsurgery; the same condition may betreated best in one patient with microsurgery and inanother by stereotaxy. Maybe the most importantlesson learnt at the Karolinska is that the simplicityof using the Gamma Unit makes this integrationpossible and that the same individual can be a com-petent microsurgeon and also a stereotacticradiosurgeon. Someone competent in both techni-ques is best fitted to decide where the boundariesbetween the two methods should lie.

A new impetus has been given to stereotaxy ingeneral by the new imaging techniques such as CT,PET, ultrasound and particularly NMR, and thesemodern, non-invasive diagnostic techniques are par-ticularly compatible with bloodless stereotacticradiosurgery. The two prototype radiosurgical unitshave functioned very well and the third generationunits should be more effective and still easier towork with. Radiosurgery has established its efficacyand safety and offers an operative system which,combined with sophisticated modern diagnosticmethods, makes the depths of the brain more access-ible.

References

Leksell, L. A stereotaxic apparatus for intracerebralsurgery. Acta Chir Scand 1949;99:229-33.

Leksell, L. The stereotaxic method and radiosurgery ofthe brain. Acta Chir Scand 1951;102:316-9.

Leksell, L, Hemer T, Liden K. Stereotaxic radiosurgeryof the brain. Report of a case . Kungl. FvsiografSdllskLund Forhandl 1955;25(17):1-10.

Larsson B, Leksell L, Rexed B, Sourander P, MairW, Andersson B. The high-energy proton beamas a neurosurgical tool. Nature 1958;182:1222-3.

Leksell L, Larsson B, Andersson B, Rexed B, SouranderP, Mair W. Lesions in the depth of the brain producedby a beam of high-energy protons. Acta Radiol1960;54:251-64.

Leksell L, Larsson B, Rexed B. The use of high-energyprotons for cerebral surgery in man. Acta Chir Scand1963;125:1-7.

Leksell L. Stereotaxis and Radiosurgerv. An operativeSystem. Springfield, Charles Thomas, 1971.

Larsson B, Liden K, Sarby B. Irradiation of small struc-tures through the skull. Acta radiol Onicol RadiatPhys Biol 1974;13:512-34.

6 Wennerstrand J, Ungerstedt U. Cerebral radiosurgery.11. An anatomical study of Gamma radiolesions. ActaChir Scand 1970;136:133-7.

o Leksell, L. A note on the treatment of acoustic tumours.Acta Chir Scand 1971 ;137:763-5.

Steiner L, Leksell L, Greitz T, Forster DMC, BacklundE-O. Stereotaxic radiosurgery for cerebral arterio-

Stereotactic radiosurgery

venous malformations. Report of a case. Acta ChirScand 1972;138:459-64.

Bergstrom M, Greitz T. Stereotaxic computed tomogra-phy. Am J Roentgenol 1976,127:167-70.

Leksell L, Jernberg, B. Stereotaxis and tomography. Atechnical note. Acta Neurochir. 1980;52:1-7.

'" Leksell L. Cerebral radiosurgery. 1. Gammathalmotomyin two cases of intractable pain. Acta Chir Scand1968;134:585-95.

Steiner L, Forster D, Leksell L, Meyerson B, Boethius J.Gammathalamotomy in intractable pain. ActaNeurochir. 1980W52:173-84.

' Leksell L. Stereotaxic radiosurgery in trigeminal neural-gia. Acta Chir Scand 1971;137:311-4.

17 Leksell L, Backlund E-O. Stereotaxic Gammacap-sulotomy. In: Hitchcock, ER. Ballantine HT, Meyer-son BA, eds. Modern Concepts in Psychiatric Surgery.Amsterdam: Elsevier/North-Holland BiomedicalPress, 1979. 213.

" Leksell L, Liden C. A therapeutic trial with radioactiveisotopes in cystic brain tumour. In: RadioisotopeTechniques, Vol 1. Oxford: HM Stationery Office,1953.

" Backlund E-O. Stereotactic treatment ofcraniopharyngiomas-a 15-year material. ScandNeurosurg Soc 32nd ann meeting, Linkoping, 1979.(Personal communication).

803

20 Backlund E-O. Solid craniopharyngiomas treated

by stereotactic radiosurgery. In: Szikla G. ed.Stereotactic Cerebral Irradiation. InsermSymposium no. 12, Amsterdam: Elsevier/North-Holland Biomedical Press 1979, 271-281.

21 RAhn T. Stereotactic radiosurgery in Cushing's disease.(Thesis), Sundt Offset, Stockholm 1980.

22 Cairns H. On conserving the facial nerve during removalof tumours of the cerebello-pontine angle. Proc R SocMed 1931;25:1.

23 Nordn G. Stereotactic radiosurgerv in acousticneurinomas. A new therapeutic appraoch. (Thesis)Sundt Offset, Stockholm 1982.

" Hirsch A, Norfn G, Andersson H. Audiologic findingsafter stereotactic radiosurgery in nine cases of acousticneurinomas. Acta Otolaryngol 1979A88:155-60.

25 Nilsson A, Wennerstrand J, Leksell D, Backlund E-O.Stereotactic gamma irradiation of basilar artery in cat.Preliminary experiences. Acta Radiol Oncol1978;17: 150-60.

26 Steiner L, Leksell L, Forster DMC, Greitz T,Backlund E-O. Stereotaxtic radiosurgery inintracranial arterio-venous malformations. ActaNeurchir Suppl 1974,21:195-209.

27 Steiner L. Radiosurgery in arterio-venousmalformations in the brain. In: Wilson C, Stein B,eds. Current Neurosurgical Practice. In press.

Printed in Great Britain by Bourne Offset Ltd., Iver. Bucks.