A

RFE

tAtaoc

Int. J. Radiation Oncology Biol. Phys., Vol. 63, No. 1, pp. 47–55, 2005Copyright © 2005 American Society for Therapeutic Radiology and Oncology. Published by Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/05/$–see front matter

doi:10.1016/j.ijrobp.2005.05.024

STRO REPORT

THE AMERICAN SOCIETY FOR THERAPEUTIC RADIOLOGY ANDONCOLOGY (ASTRO) EVIDENCE-BASED REVIEW OF THE ROLE OF

RADIOSURGERY FOR MALIGNANT GLIOMA

MAY N. TSAO, M.D., MINESH P. MEHTA, M.D., TIMOTHY J. WHELAN, M.D., DAVID E. MORRIS, M.D.,JAMES A. HAYMAN, M.D., JOHN C. FLICKINGER, M.D., MICHAEL MILLS, PH.D.,

C. LELAND ROGERS, M.D., AND LUIS SOUHAMI, M.D.

The American Society for Therapeutic Radiology and Oncology, Fairfax, VA

Purpose: To systematically review the evidence for the use of stereotactic radiosurgery or stereotactic fraction-ated radiation therapy in adult patients with malignant glioma.Methods: Key clinical questions to be addressed in this evidence-based review were identified. Outcomesconsidered were overall survival, quality of life or symptom control, brain tumor control or response and toxicity.MEDLINE (1990–2004 June Week 2), CANCERLIT (1990–2003), CINAHL (1990–2004 June Week 2), EM-BASE (1990–2004 Week 25), and the Cochrane library (2004 issue 2) databases were searched using OVID. Inaddition, the Physician Data Query clinical trials database, the proceedings of the American Society of ClinicalOncology (1997–2004), ASTRO (1997–2004), and the European Society of Therapeutic Radiology and Oncology(ESTRO) (1997–2003) were searched. Data from the literature search were reviewed and tabulated. This processincluded an assessment of the level of evidence.Results: For patients with newly diagnosed malignant glioma, radiosurgery as boost therapy with conventionalexternal beam radiation was examined in one randomized trial, five prospective cohort studies, and seven retrospectiveseries. There is Level I evidence that the use of radiosurgery boost followed by external beam radiotherapy andcarmustine (BCNU) does not confer benefit with respect to overall survival, quality of life, or patterns of failure ascompared with external beam radiotherapy and BCNU. There is Level I-III evidence of toxicity associated withradiosurgery boost as compared with external beam radiotherapy alone. The results of the prospective and retro-spective studies may be influenced by selection bias. Radiosurgery used as salvage for recurrent or progressivemalignant glioma after conventional external beam radiotherapy failure was reported in zero randomized trials, threeprospective cohort studies, and five retrospective series. The available data are sparse and insufficient to make absoluterecommendations. Stereotactic fractionated radiation therapy has been reported as boost therapy with external beamradiotherapy for patients with newly diagnosed malignant glioma in only three prospective studies. As primarytherapy alone without conventional external beam radiotherapy for newly diagnosed malignant glioma patients,stereotactic fractionated radiation therapy has been reported in only one prospective study. There were only threeprospective series and two retrospective studies reported for patients with recurrent or progressive malignant glioma.Conclusions: For patients with malignant glioma, there is Level I-III evidence that the use of radiosurgery boostfollowed by external beam radiotherapy and BCNU does not confer benefit in terms of overall survival, localbrain control, or quality of life as compared with external beam radiotherapy and BCNU. The use of radiosur-gery boost is associated with increased toxicity. For patients with malignant glioma, there is insufficient evidenceregarding the benefits/harms of using radiosurgery at the time progression or recurrence. There is alsoinsufficient evidence regarding the benefits/harms in the use of stereotactic fractionated radiation therapy forpatients with newly diagnosed or progressive/recurrent malignant glioma. © 2005 American Society forTherapeutic Radiology and Oncology. Published by Elsevier Inc.

Malignant glioma, Stereotactic radiosurgery, Stereotactic fractionated radiation therapy, Evidence-based med-icine.

Reprint requests to: The American Society for Therapeuticadiology and Oncology, 12500 Fair Lakes Circle, Suite 375,airfax, VA 22033. Tel: (703) 502-1550; Fax: (703) 502-7873;-mail: research@astro.org.Supplementary data associated with this article can be found in

he online version at doi:10.1016/j.ijrobp.2005.05.024.cknowledgments—The expert panel wishes to express its grati-

ude to Drs. Jay Loeffler, Roy A. Patchell, Michael A. Vogelbaum,nd the members of the American Society for Therapeutic Radi-logy and Oncology (ASTRO) Health Services Research, Health-

tees for their thoughtful reviews of earlier versions of thisevidence-based review. ASTRO wishes to acknowledge McMasterUniversity’s Program in evidence-based Care for carrying out thesystematic review used in the development of these recommenda-tions. Further acknowledgments are extended to Alex Chambers(Research Coordinator, Program in Evidence-based Care, McMas-ter University, Hamilton, Ontario, Canada), and Jennifer Padberg,M.P.H., (Director of Research, ASTRO).

Received May 20, 2005. Accepted for publication May 20,2005.

are Policy and Economics, and Government Relations Commit-

47

AmSHesa

E

asocetscdeei

ocsjhemtt

E

Ricmia

P

g

lpo

udtpaIot

R

rdtdto

C

tpctat

padHieta

R

rmriql

A

48 I. J. Radiation Oncology ● Biology ● Physics Volume 63, Number 1, 2005

INTRODUCTION

s the use of radiosurgery becomes widespread for theanagement of certain intracranial lesions, the Americanociety for Therapeutic Radiology and Oncology (ASTRO)ealth Services Research Committee concluded that an

vidence-based review of the evidence for the use of radio-urgery or fractionated stereotactic radiation therapy indult patients with malignant glioma would be useful.

vidence-based medicineEvidence-based medicine is the conscientious, explicit,

nd judicious use of current best evidence in making deci-ions about the care of individual patients (1). The practicef evidence-based medicine means integrating individuallinical expertise with the best available external clinicalvidence from systematic research and patient choice. Sys-ematic research refers to reviews that appraise critically,ummarize, and attempt to reconcile the published evidenceoncerning a particular problem (2). External clinical evi-ence can inform, but can never replace, individual clinicalxpertise, and it is this expertise that decides whether thexternal evidence applies to the individual patient at all and,f so, how it should be integrated into a clinical decision (1).

The final determinant of what treatment should be rec-mmended rests with the oncologist. It is based on his or herlinical impression of the individual case, his or her under-tanding of the relevant circumstances, and sound medicaludgment. ASTRO fully recognizes this and considers ad-erence to these recommendations voluntary. Rather, thisvidence-based review is intended to summarize the currentedical literature and identify important questions for fur-

her research. These recommendations do not apply to fur-her research in clinical trials.

METHODOLOGY

xpert panel selectionThe expert panel was assembled by the Health Services

esearch Committee of ASTRO and is composed of expertsn radiation oncology, physics, clinical research, and out-omes/health services research. A full list of expert panelembers is listed in Appendix A (online only). The primary

ntent was to provide a variety of different perspectives inn attempt to minimize bias (Table 1).

rocess overviewIn evaluating the evidence regarding the role of radiosur-

ery or fractionated stereotactic radiation therapy for ma-

Table 1. Causal pathway: A graphical repre

Patient population Interve

dult patients with malignantglioma

Stereotactic radiosurgery orradiation therapy (as booexternal beam radiotheratime of relapse after surg

radiotherapy)

ignant glioma, a systematic search of the literature waserformed. The process included an assessment of the levelf evidence (Table 2).It should be noted that Level I data are not necessarily the

ltimate requirement for a proposed treatment to be en-orsed as reasonable and necessary. The unique nature ofhe practice of oncology often precludes the opportunity toerform Level I studies. Many treatments are consideredppropriate, in some cases essential, based solely on LevelI or Level III data. As previously noted, it is the judgmentf the treating physician that ultimately determines whichherapy is indicated.

eview of the available dataPertinent information from the published literature was

etrieved and reviewed for the creation of these recommen-ations. A detailed account of the methods used to obtainhis information, including the inclusion/exclusion criteria,atabases searched, time periods covered by the search andhe search criteria used can be found in Appendix B (onlinenly).

onsensus based on the evidenceThe expert panel identified several key clinical questions

o be addressed in this review (Appendix B). The expertanel convened on numerous occasions in person and viaonference call to discuss and formulate the recommenda-ions. Drafts were developed and circulated in several iter-tions and all members of the expert panel had an oppor-unity to comment.

The expert panel did not attempt to codify establishedractice. The experts reviewed the available evidence anddded their best clinical judgment to make final recommen-ations. The content was peer-reviewed by the ASTROealth Services Research, Healthcare Policy and Econom-

cs, and Government Relations Committees as well as sev-ral external reviewers not involved in the development ofhe document. The ASTRO board of directors reviewed andpproved the document before its dissemination.

evisions/updatingEvidence-based documents are living documents, which

eflect the best practice at a particular point and include theost up-to-date evidence available. They require regular

evision to cope with changing medical practice, advancesn technology, and changes in the environment. This re-uires commitment to resources to monitor the emergingiterature so that decisions as to whether or not a document

ion of a “causal pathway” is shown below

Outcomes

onated stereotacticsurgery andas salvage at the

d external beam

1. Overall survival2. Quality of life or symptom control3. Brain tumor control or response4. Toxicity

sentat

ntion

fractist afterpy ORery an

sTvur

B

cicmtrsvtawft

rtrderoea

amCretrt

lsn

rdsTttlrs

R

pp

dsw

rnumg

B

Kgbpp

r(a

LLL

LL

49ASTRO evidence-based review of the role of radiosurgery for malignant glioma ● M.N. TSAO et al.

hould be revised, or if it has become obsolete can be made.herefore, at regular intervals, this document will be re-iewed and a determination made as to whether or not anpdate should be undertaken, or if the document should beetired.

RADIOSURGERY OR FRACTIONATEDSTEREOTACTIC RADIATION THERAPY FOR

MALIGNANT GLIOMA

ackgroundThe management of patients with malignant glioma is

hallenging. Despite best contemporary therapeutic modal-ties, most patients succumb to their disease. External beamonventional radiation therapy remains the mainstay ofanagement after surgical resection. A dose–response rela-

ionship has been reported in earlier studies of postoperativeadiation therapy. A randomized trial by the Medical Re-earch Council (4) found an improvement in median sur-ival of 9–12 months when 60 Gy of external beam radio-herapy was compared with 45 Gy. Walker et al. (5) foundn increase in median survival from 28 weeks to 42 weekshen external beam radiotherapy doses were increased

rom 50 to 60 Gy. These data were obtained retrospectivelyhrough a review of a series of sequential trials.

For malignant glioma, the majority (90% or more) of recur-ences occur within 2 cm of the enhancing edge of the originalumor (6). Furthermore, multicentric or metastatic disease isare (7). As a consequence, the possibility of escalating theose to the gross tumor volume (as boost therapy) has beenxplored through the use of radiosurgery or fractionated ste-eotactic radiation therapy. In addition, the use of radiosurgeryr fractionated stereotactic radiation therapy has been consid-red at the time of recurrent or progressive malignant gliomafter conventional primary treatment.

As the use of radiosurgery or fractionated stereotactic radi-tion therapy becomes more widespread in the management ofalignant glioma, the ASTRO Health Services Researchommittee determined that an evidence-based review of this

adiation modality was relevant and important. The aim of thisvidence-based project was to provide a clear presentation ofhe key clinical questions related to the use of stereotacticadiosurgery or fractionated stereotactic radiation therapy in

Table 2. L

Strength of the evidence: The effectiveness of radiosurgery (tumotreatment of malignant glioma was graded

evel I Evidence obtained from at least one properlevel II-1 Evidence obtained from well-designed contrevel II-2 Evidence obtained from well-designed coho

center or research groupevel II-3 Evidence obtained from comparisons betweeevel III Opinions of respected authorities, based on

Data from Woolf et al. (3).

he management of malignant glioma and to assess the pub- e

ished evidence for the effectiveness (benefits and harms) ofingle fraction or fractionated radiosurgery for patients withewly diagnosed or recurrent malignant glioma.

For the purposes of this review, fractionated stereotacticadiation therapy is defined as a highly conformal technique ofelivering radiation therapy through the use of a three-dimen-ional coordinate system to locate intracranial targets precisely.he rationale of fractionation is to spare adjacent normal tissue

hrough radiobiologic concepts. Because of toxicity issues,arget sizes with a single fraction of radiosurgery are generallyimited to 4 cm or less. Theoretically, fractionated stereotacticadiation therapy may allow for larger targets to be treatedafely.

esultsThere were no studies obtained that examined the use of

rimary single fraction stereotactic radiosurgery withoutrevious external beam radiotherapy.Details regarding the studies are summarized in Appen-

ix C, Tables 3–10 (online only). The tables are divided intoections based on study design, whether patients treatedere newly diagnosed or recurrent malignant glioma.The first section reviews the evidence on the use of

adiosurgery in either newly diagnosed or recurrent malig-ant glioma. The next section reviews the evidence on these of fractionated stereotactic radiation therapy in theanagement of newly diagnosed or recurrent malignant

lioma.

RADIOSURGERY FOR MALIGNANT GLIOMA(NEWLY DIAGNOSED OR RECURRENT)

oost therapy for newly diagnosed malignant gliomaInterpretative summaryey clinical question #1: In patients treated with sur-

ery and conventional radiation, does radiosurgeryoost (either before or after conventional radiation) im-rove survival as compared to no radiosurgery boost inatients diagnosed with malignant glioma?Only one randomized trial examined the use of up-front

adiosurgery boost to the region of contrast enhancement�4 cm in diameter) followed by external beam radiother-py and Bischloroethylnitrosourea (BCNU) compared with

evidence

rol/response, survival, quality of life or symptom control) in theing to the quality of evidence obtained (3)

ned randomized controlled trialrials without randomization.se-control analytic studies, preferably from more than one

s or places with or without the intervention.l experience, descriptive studies or reports of expert committees.

evel of

r contaccord

y desigolled trt or ca

n timeclinica

xternal beam radiotherapy and BCNU in selected patients

wpbri(emepiimpebpomkbndpnbd

olfrtftreobpsRst1

ooiA1s1oipa

ff

sbrbaogt(Ratastt8stItopEsRgt

sverpscpat

tthhstbtrts

b

50 I. J. Radiation Oncology ● Biology ● Physics Volume 63, Number 1, 2005

ith glioblastoma multiforme. These results have been re-orted in abstract form and the final full report is pendingut was made available for our review in preparing thiseport (8). There was no advantage with radiosurgery boostn terms of survival. The median survival was 13.6 months95% confidence interval [CI], 11.2–15.2 months) in thexternal beam and BCNU arm as compared with 13.5onths (95% CI, 11–14.8 months) in the radiosurgery then

xternal beam and BCNU arm (p � 0.5711). A total of 203atients were randomly assigned to the two treatment armsn this trial. The study was powered to demonstrate anmprovement of 50% in median survival (12.5–18.75onths). Patients were stratified by age and Karnofsky

erformance status (KPS). A statistically significant differ-nce between treatment arms was not found in the subsetsy Radiation Therapy Oncology Group (RTOG) recursiveartitioning analysis (RPA) class III and IV nor in the subsetf patients with preoperative tumor size �40 mm. Further-ore, a comparison of radiosurgery techniques (gamma

nife or linear accelerator) in the arm with radiosurgeryoost resulted in no survival difference. However, it wasoted that 18% of radiosurgery patients had unacceptableeviations from study protocol. It is unclear whether theseatients were treated with a suboptimal radiosurgery tech-ique or were poor candidates for stereotactic radiosurgeryecause of selection criteria. The extent to which theseeviations may have affected outcome remains speculative.There have been five prospective studies (9–13) reported

n the use of radiosurgery boost for newly diagnosed ma-ignant glioma. These were relatively small series rangingrom 14 to 37 patients receiving radiosurgery boost. Agesanged from 5 to 84 years and KPS scores ranged from 20o 100. The majority of patients had glioblastoma multi-orme, although some included patients with anaplastic as-rocytoma or mixed glial tumors. Three of the five studieseported no difference in survival as compared with historicxpectation, whereas the Loeffler et al. study (10) reportedutcomes of 37 patients treated with radiosurgery boost toe better than historic expectation. Mehta et al. (11) re-orted on the survival of 31 patients who received radio-urgery boost as being superior to historic expectation byTOG RPA class. For the five studies overall, median

urvival ranged from approximately 10–26 months for pa-ients with glioblastoma multiforme and approximately9–28 months for patients with anaplastic astrocytoma.Seven retrospective studies (14–20) reported on the use

f radiosurgery boost for newly diagnosed malignant gli-ma. The largest was a combined series comprising of threenstitutions (n � 115). The smallest series had 11 patients.ges ranged from 6 to 84 years and KPS ranged from 50 to00. The majority of patients had Grade 4 glioma, althoughome anaplastic glioma patients were reported in four (14,5, 17, 18) of the six studies. One study (16) consisted ofnly glioblastoma multiforme patients, and one study (15)ncluded lower grade gliomas. Four studies (16–19) re-orted survivals as being better than historic controls. Over-

ll for the seven retrospective studies, median survivals p

rom diagnosis for glioblastoma multiforme patients rangedrom 13 to 25 months.

Although several prospective cohort and retrospectivetudies have reported survival outcomes of radiosurgeryoost after radiotherapy for malignant gliomas to be supe-ior as compared with historical controls, such analyses maye prone to selection bias. Sarkaria et al. (18) and Shrieve etl. (19), for example, compare their results with the resultsf the RTOG databases. Both studies suggest that radiosur-ery boost confers a survival advantage in comparison tohe RTOG database stratified by RPA class. Curran et al.21) analyzed 778 malignant glioma patients enrolled on theTOG 83-02 trial, a randomized Phase I/II hyperfraction-ted radiotherapy dose-escalation trial with BCNU chemo-herapy. Common stereotactic radiosurgery criteria werepplied to these patients, namely: KPS �60, well-circum-cribed tumor �4 cm, no subependymal spread, and loca-ion not adjacent to the brainstem or optic chiasm. Stereo-actic radiosurgery–eligible patients enrolled on RTOG3-02 had survival superior to that of the stereotactic radio-urgery–ineligible group. The advantage appeared to be dueo the selection of a subgroup with a high minimum KPS.rish et al. (22) also reported on a database of 101 conven-ionally treated patients with biopsy-proven malignant gli-ma with known survival times. Twenty-seven percent ofatients were deemed eligible for stereotactic radiosurgery.ligible patients had more favorable prognostic factors andignificantly longer median survival than ineligible patients.adiosurgery eligible, conventionally treated patients withlioblastoma multiforme and a group of radiosurgery-reated patients had similar survival outcomes.

It is important to note, however, that many of the retro-pective or cohort studies that indicated an improved sur-ival benefit in patients treated with radiosurgery boostmployed radiosurgery at the completion of external beamadiotherapy rather than upfront, as per the RTOG 93-05rotocol (23). This temporal sequence could potentiallyelect out patients who improve or remain stable afterompletion of external beam radiotherapy. Whether suchatients derive a survival benefit from radiosurgery boostfter external beam radiotherapy could not be answered byhe one randomized trial reported (23).

Furthermore, optimal radiosurgery boost therapy remainso be elucidated. One possible flaw with radiosurgery boostechniques is the use of targeting the visible contrast en-ancing lesion. Magnetic resonance spectroscopy studiesave documented areas of increased metabolic activity out-ide the region of contrast enhancement (24). This leaveshe possibility that better radiosurgery boost techniques andetter patient selection may improve survival as comparedo conventionally treated patients who are treated without aadiosurgery boost. In addition, more effective systemicherapeutic agents or radiosensitizers combined with radio-urgery boost may, in theory, improve survival.

Key clinical question #2: Does the use of radiosurgeryoost improve quality of life or symptom control in

atients with newly diagnosed malignant glioma?

patemtwtaMtttm6powIyt

tadiDctpmoi

bw

rla

tmaepaea

wrgdSf

rsTtbretg

bm

tpFli4caa

ottert

bfrtt

C

eeebtob

tpsdclemf

51ASTRO evidence-based review of the role of radiosurgery for malignant glioma ● M.N. TSAO et al.

The only one randomized trial, RTOG 93-05, (23) re-orted on no improvement in quality of life (Spitzer index)t baseline compared with the end of therapy in patientsreated with radiosurgery boost. There was also no differ-nce in quality-adjusted survival or cognitive decline aseasured by the Mini-Mental Status Examination. Only

wo prospective studies commented on KPS score changeith radiosurgery boost (11, 13). Mehta et al. (11) reported

hat median KPS scores increased right after radiosurgerynd peaked again 15–18 months after radiosurgery. Inehta’s study, 14 patients were given external beam radio-

herapy followed by radiosurgery, and 15 patients werereated with radiosurgery followed by external beam radio-herapy. Shenouda et al. (12) reported on KPS scores beingaintained until 48 weeks when the median KPS dropped to

0 (range, 40–100). Of the six retrospective studies re-orted, only Shrieve et al. (19) commented on mean KPSutcomes. At diagnosis mean KPS was 91.5 (80–100),hich dropped to a mean of 87 (50–100) at last follow-up.

n Shrieve’s et al. (19) study, data on 13 patients alive 2ears after diagnosis showed that 62% were no longeraking corticosteroids.

Thus there evidence from one randomized controlled trialhat radiosurgery boost before external beam radiotherapynd BCNU does not improve quality of life and cognitiveecline. There are some prospective and retrospective stud-es that report on small patient numbers and KPS outcome.ata on these prospective and retrospective studies are

onsidered insufficient to support a quality of life or symp-om control benefit with the use of radiosurgery boost inatients with newly diagnosed malignant glioma. Further-ore, the confounding factor of steroid use and its impact

n changes in quality of life or symptom control makenterpretation of these outcomes in these studies difficult.

Key clinical question #3: Does the use of radiosurgeryoost improve brain control/tumor response in patientsith newly diagnosed malignant glioma?The duration of glioblastoma response to therapy was not

eported in the randomized trial RTOG 93-05. However,ocal failure as a component of failure was seen in 92.5% ofll patients.

In the prospective study by Gannett et al.(9), the medianime to recurrence in radiosurgery boosted patients was 7onths. In the prospective study reported by Mehta et

l.(11), the median time to failure was 22 weeks. Shenoudat al. (12) reported 25 weeks as the median time to diseaserogression. Whereas Mehta et al. (11) and Shenouda’s etl. (12) studies only included Grade 4 astrocytoma, Gannettt al. (9) included some patients with Grade 3 astrocytomand some patients with mixed glioma histology.

In the retrospective study by Buatti et al. (14), no patientith glioblastoma multiforme had a complete response to

adiosurgery boost. The median time to progression forlioblastoma multiforme patients was 3.5 months after ra-iosurgery. Sarkaria et al. (18), Shrieve et al. (19), andelch et al. (20) reported on very high local and marginal

ailures (�80%). r

Thus there remains lack of high quality evidence thatadiosurgery boost improves brain control or tumor re-ponse in patients with newly diagnosed malignant glioma.he prospective and retrospective studies suffer from selec-

ion bias and no direct comparisons are made in terms ofrain control or tumor response in patients treated withadiosurgery boost vs. not. As such, there is insufficientvidence that radiosurgery boost improves brain control/umor response in patients with newly diagnosed malignantlioma.Key clinical question #4: Does the use of radiosurgery

oost increase toxicity in patients with newly diagnosedalignant glioma?Toxicity in the RTOG 93-05 trial (23) was graded using

he morbidity scoring scheme by the RTOG and the Euro-ean Organization for Research in the Treatment of Cancer.or late radiation–related toxicities, there were four Grade 3

ate toxicities in the radiosurgery arm compared with nonen the external beam and BCNU arm. There were no Gradetoxicities related to radiotherapy. Seven patients died from

hemotherapy toxicity (5 in the external beam and BCNUrm and 2 in the radiosurgery then external beam radiother-py and BCNU arm).

For the five prospective cohort studies examining the usef radiosurgery boost in newly diagnosed malignant glioma,oxicity reports range from “no significant acute or lateoxicity” to brain necrosis in 2 of 37 patients in the Loefflert al. series (10) and 4 of 29 patients in the Mehta et al.eport (11). The retrospective studies report on a few pa-ients developing significant edema, or radiation necrosis.

Thus there is evidence that the addition of radiosurgeryoost is associated with an increased risk of toxicity rangingrom significant edema to radiation necrosis. Reoperationates varied from approximately 19–33%. A small propor-ion of patients subjected to reoperation had necrosis only inhe operative specimens retrieved.

onclusionThere is one randomized controlled trial that provides

vidence that the use of radiosurgery boost followed byxternal beam radiotherapy and BCNU as compared withxternal beam radiotherapy and BCNU does not conferenefit in terms of overall survival, quality of life, or pat-erns of failure. There appears to be a slight increased riskf late Grade 3 toxicity with the approach of radiosurgeryoost followed by external beam radiotherapy.The alternative timing strategy of external beam radio-

herapy followed by radiosurgery boost has not been re-orted in a randomized trial setting. Possible areas of re-earch include improved strategies of radiosurgery boostelivery (e.g., targeting using magnetic resonance spectros-opy), better patient selection, improved conformality al-owing for safe dose escalation, or novel approaches—forxample, combining radiosurgery with systemic agents orolecular targeted approaches. In addition, there is a need

or improved quality of life assessments after stereotactic

adiosurgery for patients with malignant glioma.

bacrTeram�iettafhbzczzcba

I

fs

sgsTstfscottw1

mtbpper

bosm

ai

t

st

nsrtf4popeTcg

sp

upgtgaNd

C

acIm

N

Ksl

52 I. J. Radiation Oncology ● Biology ● Physics Volume 63, Number 1, 2005

Historically, radiotherapy dose-escalation trials haveeen conducted with rather ineffective chemotherapy, suchs BCNU. At the 2004 American Society of Clinical On-ology meeting, Stupp et al. (25) presented preliminaryesults of The European Organization for Research andreatment of Cancer 26981, a Phase III randomized trialvaluating the role of temozolomide chemotherapy concur-ent with and also adjuvant after radiotherapy. The mediannd 2-year survival for the radiotherapy arm was 12.1onths and 10%, in comparison to 14.6 months and 26% (p0.0001) for the radiotherapy and temozolomide arm. This

s the first clear and categorical prospective randomizedvidence supporting the role of chemotherapy in glioblas-oma multiforme. Therefore, with a somewhat more effec-ive agent, interest in exploring focal boost techniques suchs radiosurgery is rejuvenated. Because radiosurgery is aocal modality that can effectively necrose tumor within theigh-dose regions, its major limitation can be speculated toe tumor regrowth from outside the high-dose regions, theone of poorly defined microscopic spread. If temozolomidean more effectively control the disease in this “peripheral”one, the impact of enhanced control within the “central”one could potentially become more meaningful. Hencelinical trials evaluating and testing this hypothesis areeing developed and are likely to represent the next gener-tion of studies in this field.

SALVAGE THERAPY FOR RECURRENT ORPROGRESSIVE MALIGNANT GLIOMA

nterpretative summaryKey clinical question #1: Does radiosurgery as salvage

or recurrent or progressive malignant glioma improveurvival?

There have been no randomized trials comparing radio-urgery for salvage in patients with recurrent malignantlioma as compared with competing interventions such asurgical debulking, chemotherapy, or best supportive care.here have been prospective (26–28) and retrospectivetudies (15, 29–31) of varying sample sizes ranging from 18o 132 patients reporting on the use of radiosurgery salvageor patients with recurrent malignant glioma; these demon-trate median survival of 6–12 months. As a point ofomparison, Wong et al. (32) reported that the medianverall survival of patients with recurrent glioblastoma mul-iforme (n � 225) or anaplastic astrocytoma (n � 150)reated on Phase II trials of agents presumed to be inactiveas 7 months (30 weeks). For glioblastoma patients, only5% were alive and progression-free at 6 months.Larson et al. (28) reported on the use of radiosurgery andarimastat in patients with recurrent Grade 3 or 4 astrocy-

oma. Median survival for these patients were reported aseing better than historical controls for Grade 3 astrocytomaatients, but worse for Grade 4 astrocytoma patients. Biases,articularly selection bias contribute to the overall survivalstimates in the published series of patients treated with

adiosurgery salvage for malignant glioma. r

Thus there is insufficient evidence to support a survivalenefit in the use of radiosurgery at the time of progressiver recurrent malignant glioma as compared with competingtrategies of management such as debulking surgery, che-otherapy, or best supportive care.Key clinical question #2: Does the use of radiosurgery

s salvage for recurrent or progressive malignant gliomamprove quality of life or symptom control?

None of the studies reported on quality of life or symp-om control outcomes.

Key clinical question #3: Does the use of radiosurgeryalvage improve brain control/tumor response in pa-ients with recurrent or progressive malignant glioma?

Chamberlain et al. (26) reported on 17 patients of whomone had complete response, 8 had partial response, and 9 hadtable disease after radiosurgery for salvage. Hall et al. (27)eported that 85% died of local or marginal failure. The medianime to progression in Larson’s et al. (28) study was 31 weeksor patients with Grade 3 astrocytoma and 15 weeks for Grade

astrocytoma. Sanghavi et al. (30) found a 4-month medianrogression free survival in 30 patients with recurrent Grade 3r 4 astrocytoma treated with salvage radiosurgery. Eightyercent of patients in Shrieve et al.’s (31) study experiencedither a local or marginal failure with salvage radiosurgery.he results on brain control/tumor response have not beenompared with competing strategies such as debulking sur-ery, chemotherapy, or best supportive care.

Key clinical question #4: Does the use of radiosurgeryalvage increase toxicity in patients with recurrent orrogressive malignant glioma?The prospective and retrospective studies examining the

se of radiosurgery salvage for patients with recurrent orrogressive malignant glioma have reported varying de-rees of toxicity. Chamberlain et al. (26) found 7/20 pa-ients to have developed early side effects from radiosur-ery. One patient had hypersomnolence syndrome 8 weeksfter radiosurgery and another patient died of herniation.ineteen of 86 patients in Shrieve’s et al. report (31)eveloped radiation necrosis.

onclusionThere is Level II-3 evidence that the use of radiosurgery

t the time of recurrence may improve survival and brainontrol in selected patients but at a cost of possible toxicity.n the absence of higher quality evidence, this interventionay be considered an option for selected patients.

STEREOTACTIC FRACTIONATED RADIATIONTHERAPY FOR MALIGNANT GLIOMA

ewly diagnosed malignant gliomaInterpretative summaryey clinical question #1: Does the use of fractionated

tereotactic radiation therapy for newly diagnosed ma-ignant glioma improve survival?

There has only been one study of fractionated stereotactic

adiation therapy used in the primary treatment for malig-

nptc

tnpRatvicdcw

rrtit

tpsg

rsto

slc

t

slt

tapdcrrb

sl

n1

C

sfiailcbanaeto

S

Kss

rtgGgva(srsra

sst

swhitsn

sst

53ASTRO evidence-based review of the role of radiosurgery for malignant glioma ● M.N. TSAO et al.

ant glioma (Grade 3 or 4 astrocytoma) (33). This smallrospective study (n � 19 patients) found no difference inime to progression or median survival from historicalontrols.

One additional prospective, multi-institutional Phase IIrial of fractionated stereotactic boost combined with exter-al beam radiation, limited only to glioblastoma multiformeatients, has been completed by the RTOG. This trial,TOG BR-0023, is a Phase II trial of accelerated radiother-py using weekly stereotactic conformal boosts for supra-entorial glioblastoma. This trial examined the use of con-entional external beam radiotherapy to a dose of 5000 cGyn 25 daily fractions with stereotactic radiosurgery boostonsisting of four treatments of 5 or 7 Gy, once per weekuring Weeks 3–6. BCNU was given within 1 month afterompleting radiotherapy for six cycles. Target sample sizeas 76. Accrual has been completed. Results are pending.The other four studies (34–37) combined external beam

adiotherapy with a fractionated regimen of stereotacticadiation given as boost. All of these studies included as-rocytoma Grades 3 and 4; the study by Regine et al. (36)ncluded some low-grade glioma patients in addition tohose with malignant glioma.

Baumert et al. (34) reported 1- and 2-year survival rateso be 77% and 42%, respectively. These data were com-ared to a historical group and were reported as beinguperior for the fractionated stereotactic radiation therapyroup of patients.Because of heterogeneity, small sample size, and lack of

andomized data, it is not possible to discern whether aurvival benefit exists with the use of fractionated stereo-actic radiation therapy for newly diagnosed malignant gli-ma patients.Key clinical question #2: Does the use of fractionated

tereotactic radiation therapy for newly diagnosed ma-ignant glioma improve quality of life or symptomontrol?

None of the studies reported on quality of life or symp-om control outcomes.

Key clinical question #3: Does the use of fractionatedtereotactic radiation therapy for newly diagnosed ma-ignant glioma improve local brain control or brainumor response?

Baumert et al. (34) reported on a mean time of 11 monthso progression after fractionated stereotactic radiation ther-py. Nieder et al. (33) reported that the median time torogression was 5 months in his study with no significantifference as compared with historical controls treated withonventional external beam radiotherapy. Regine et al. (36)eported that the median time to response (at least 20%eduction in the multiple of the cross sectional diameterased on magnetic resonance imaging) was 4 months.Key clinical question #4: Does the use of fractionated

tereotactic radiation therapy for newly diagnosed ma-ignant glioma increase toxicity?

Reported toxicities included reoperation from radiation

ecrosis in 1 of 17 patients in Baumert’s study (34) to 4 of2 patients having radionecrosis in Cardinale’s report (35).

onclusionBecause of the heterogeneity of studies, small sample

izes, and lack of randomized Phase III data, there is insuf-cient evidence that fractionated stereotactic radiation ther-py for newly diagnosed malignant glioma confers benefitn terms of survival, quality of life, symptom control, orocal brain control/brain tumor response as compared withonventional management consisting of surgery followedy external beam radiotherapy with or without chemother-py. As such, fractionated stereotactic radiation therapy forewly diagnosed malignant glioma given as primary ther-py after surgery or as boost therapy after conventionalxternal beam radiotherapy remains experimental. Of note,he RTOG has completed a prospective Phase II trial, resultsf which are currently unavailable.

ALVAGE AT THE TIME OF RECURRENCE ORPROGRESSION

Interpretative summaryey clinical question #1: Does the use of fractionated

tereotactic radiation therapy for recurrent or progres-ive malignant glioma improve survival?

There have been three prospective (38–40) and twoetrospective studies (41, 42) that reported the use of frac-ionated stereotactic radiation therapy for recurrent or pro-ressive malignant glioma. The majority of patients hadrade 4 astrocytoma and a few had progressive Grade 3lioma or transformed low-grade glioma. Survivals wereariable ranging from a median survival of 6.7 to 12 monthsfter stereotactic fractionated radiation therapy. Cho et al.41) compared the results of 46 patients treated with radio-urgery and 25 patients treated with fractionated stereotacticadiation therapy for progressive malignant glioma. Medianurvival was not statistically different in those patientseceiving radiosurgery salvage as compared with fraction-ted stereotactic radiation therapy for salvage.

Key clinical question #2: Does the use of fractionatedtereotactic radiation therapy for recurrent or progres-ive malignant glioma improve quality of life or symp-om control?

Laing et al. (39) was the only study that reported onymptom control. Seventy-seven percent of patients treatedith fractionated stereotactic radiation therapy at recurrencead improved Barthel scores at 3 months, whereas 67% hadmproved Barthel indices at 6 months. The ability for pa-ients to taper or discontinue steroid use with fractionatedtereotactic radiation therapy at the time of recurrence hasot been adequately reported.Key clinical question #3: Does the use of fractionated

tereotactic radiation therapy for recurrent or progres-ive malignant glioma improve brain control or brainumor response?

Lederman et al. (38) reported that 7.1% of patients

tootpl

ss

n2t

C

sfiabowmcastsm

s

mt

amttsogsc

o(ftdpatmastR

54 I. J. Radiation Oncology ● Biology ● Physics Volume 63, Number 1, 2005

reated with fractionated stereotactic radiation therapy dem-nstrated stable disease. No patient demonstrated completer partial response. Laing et al. (39) reported decreasedumor size in 35% of patients in his series. The majority ofatients in the Laing et al. (39) and Cho et al. (41) series hadocal or marginal failure.

Key clinical question #4: Does the use of fractionatedtereotactic radiation therapy for recurrent or progres-ive malignant glioma result in toxicity?

Laing et al. (39) reported on 4 of 22 patients developingeurologic deterioration due to radiation toxicity. Three of5 patients in Cho’s et al. (41) series underwent reopera-ions for radiation necrosis or progression.

onclusionBecause of the heterogeneity of studies, small sample

izes, and lack of randomized Phase III data, there is insuf-cient evidence that fractionated stereotactic radiation ther-py for recurrent or progressive malignant glioma confersenefit in terms of survival, quality of life, symptom control,r local brain control/brain tumor response as comparedith competing strategies such as debulking surgery, che-otherapy, or best supportive care. Also, there is insuffi-

ient evidence that fractionated stereotactic radiation ther-py provides superior or equivalent benefit in terms ofurvival, quality of life, symptom control, local brain con-rol/brain tumor response, or toxicity as compared with aingle fraction of radiosurgery for recurrent or progressivealignant glioma.Overall, there is insufficient evidence that fractionated

tereotactic radiation therapy for recurrent or progressive o

REFEREN

an adjunct to surgery and external beam radiotherapy in the

1

1

1

1

1

1

1

1

alignant glioma confers benefit over competing interven-ions and should be regarded as experimental.

ONGOING TRIALS

The RTOG is conducting a Phase II trial (BR-0023) ofccelerated radiotherapy using weekly stereotactic confor-al boosts for supratentorial glioblastoma multiforme. This

rial examines the use of conventional external beam radio-herapy to a dose of 5000 cGy in 25 daily fractions withtereotactic radiotherapy boost consisting of four treatmentsf 5 or 7 Gy, once per week during Weeks 3–6. BCNU isiven within 1 month after completion of radiotherapy forix cycles. Target sample size is 76. Accrual has beenompleted and results are pending.

The European Organization for Research and Treatmentf Cancer (22972) and the Medical Research CouncilBR10) are conducting a Phase III randomized trial of focalractionated conformal stereotactic boost following conven-ional radiotherapy of high-grade gliomas. This study ran-omizes patients with high-grade gliomas after surgery toartial brain irradiation (54–60 Gy) followed by a fraction-ted stereotactic boost (4 � 5 Gy daily) or no furtherreatment. Eligibility criteria include small high-grade glio-as (World Health Organization III/IV) �4 cm in diameter,

ge �65 years, World Health Organization performancetatus 0–1. The trial was planned requiring 550–600 pa-ients to detect a 10% difference in 2-year survival (34).esults are pending.Additionally, several single-institution trials are ongoing

r planned.

CES

1. Sackett DL, Richardson WS, Rosenberg W, et al. Evidence-based medicine: How to practice and teach EBM. New York:Churchill Livingstone; 1997.

2. Jadad AR, Cook DJ, Browman GP. A guide to interpretingdiscordant systematic reviews. CMAJ 1997;156:1411–1416.

3. Woolf SH, Battista RN, Anderson GM, et al. Assessing theclinical effectiveness of preventive maneuvers: Analytic prin-ciples and systematic methods in reviewing evidence anddeveloping clinical practice recommendations. A report by theCanadian Task Force on the Periodic Health Examination.J Clin Epidemiol 1990;43:891–905.

4. Bleehen NM, Stenning SP. A Medical Research Council trialof two radiotherapy doses in the treatment of grades 3 and 4astrocytoma. The Medical Research Council Brain TumourWorking Party. Br J Cancer 1991;64:769–774.

5. Walker MD, Strike TA, Sheline GE. An analysis of dose-effect relationship in the radiotherapy of malignant gliomas.Int J Radiat Oncol Biol Phys 1979;5:1725–1731.

6. Hochberg FH, Pruitt A. Assumptions in the radiotherapy ofglioblastoma. Neurology 1980;30:907–911.

7. Erlich SS, Davis RL. Spinal subarachnoid metastasis fromprimary intracranial glioblastoma multiforme. Cancer 1978;42:2854–2864.

8. Souhami L. Personal communication; August 2004.9. Gannett D, Stea B, Lulu B, et al. Stereotactic radiosurgery as

treatment of patients with malignant gliomas. Int J RadiatOncol Biol Phys 1995;33:461–468.

0. Loeffler JS, Alexander E, 3rd, Shea WM et al. Radiosurgeryas part of the initial management of patients with malignantgliomas. J Clin Oncol 1992;10:1379–1385.

1. Mehta MP, Masciopinto J, Rozental J, et al. Stereotacticradiosurgery for glioblastoma multiforme: Report of a pro-spective study evaluating prognostic factors and analyzinglong-term survival advantage. Int J Radiat Oncol Biol Phys1994;30:541–549.

2. Shenouda G, Souhami L, Podgorsak EB, et al. Radiosurgeryand accelerated radiotherapy for patients with glioblastoma.Can J Neurol Sci 1997;24:110–115.

3. Stea B, Rossman K, Kittelson J, et al. A comparison ofsurvival between radiosurgery and stereotactic implants formalignant astrocytomas. Acta Neurochir Suppl 1994;62:47–54.

4. Buatti JM, Friedman WA, Bova FJ, et al. Linac radiosurgeryfor high-grade gliomas: The University of Florida experience.Int J Radiat Oncol Biol Phys 1995;32:205–210.

5. Larson DA, Gutin PH, McDermott M, et al. Gamma knife forglioma: Selection factors and survival. Int J Radiat Oncol BiolPhys 1996;36:1045–1053.

6. Nwokedi EC, DiBiase SJ, Jabbour S, et al. Gamma knifestereotactic radiosurgery for patients with glioblastoma mul-tiforme. Neurosurgery 2002;50:41–47.

7. Prisco FE, Weltman E, de Hanriot RM, et al. Radiosurgical

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

55ASTRO evidence-based review of the role of radiosurgery for malignant glioma ● M.N. TSAO et al.

boost for primary high-grade gliomas. J Neuro-Oncol2002;57:151–160.

8. Sarkaria JN, Mehta MP, Loeffler JS, et al. Radiosurgery in theinitial management of malignant gliomas: survival compari-son with the RTOG recursive partitioning analysis. RadiationTherapy Oncology Group. Int J Radiat Oncol Biol Phys 1995;32:931–941.

9. Shrieve DC, Alexander E, 3rd, Black PM et al. Treatment ofpatients with primary glioblastoma multiforme with standardpostoperative radiotherapy and radiosurgical boost: prognosticfactors and long-term outcome. J Neurosurg 1999;90:72–77.

0. Selch MT, DeSalles AAF, Goetsch SJ, et al. Single-fractionradiosurgery for primary and recurrent malignant gliomas. JRadiosurg 1998;1:155–168.

1. Curran WJ, Jr., Scott CB, Weinstein AS, et al. Survivalcomparison of radiosurgery-eligible and -ineligible malignantglioma patients treated with hyperfractionated radiation ther-apy and carmustine: A report of Radiation Therapy OncologyGroup 83-02. J Clin Oncol 1993;11:857–862.

2. Irish WD, Macdonald DR, Cairncross JG. Measuring bias inuncontrolled brain tumor trials—to randomize or not to ran-domize? Can J Neurolog Sci 1997;24:307–312.

3. Souhami L, Scott C, Brachman D, et al. Randomized prospec-tive comparison of stereotactic radiosurgery (SRS) followedby conventional radiotherapy (RT) with BCNU to RT withBCNU alone for selected patients with supratentorial glioblas-toma multiforme (GBM): Report of RTOG 93-05 protocol. IntJ Radiat Oncol Biol Phys 2002;54:94–95.

4. Grosu AL, Weber W, Feldmann HJ, et al. First experiencewith I-123-alpha-methyl-tyrosine SPECT in the 3-D radiationtreatment planning of brain gliomas. Int J Radiat Oncol BiolPhys 2000;47:517–526.

5. Stupp R, Mason WP, Van Den Bent J, et al. Concomitant andadjuvant temozolomide and radiotherapy (RT) for newly di-agnosed glioblastoma multiforme (GBM). Conclusive resultsof a randomized phase III trial by the EORTC Brain and RTgroups and the NCIC Clinical Trials Group [Abstract]. Pro-ceedings from the 40th Annual Meeting of the AmericanSociety of Clinical Oncology, June 5–8, 2004; New Orleans,LA. Alexandria, VA: American Society of Clinical Oncology,2004. p. 1s.

6. Chamberlain MC, Barba D, Kormanik P, et al. Stereotacticradiosurgery for recurrent gliomas. Cancer 1994;74:1342–1347.

7. Hall WA, Djalilian HR, Sperduto PW, et al. Stereotacticradiosurgery for recurrent malignant gliomas. J Clin Oncol1995;13:1642–1648.

8. Larson DA, Prados M, Lamborn KR, et al. Phase II study ofhigh central dose gamma knife radiosurgery and marimastat inpatients with recurrent malignant glioma. Int J Radiat OncolBiol Phys 2002;54:1397–1404.

9. Sarkar A, Pollock BE, Brown PD, et al. Evaluation of

gamma knife radiosurgery in the treatment of oligodendro-

gliomas and mixed oligodendroastrocytomas. J Neurosurg2002;97:653– 656.

0. Sanghavi S, Skrupky R, Badie B, et al. Recurrent malignantgliomas treated with radiosurgery. J Radiosurg 1999;2:119–125.

1. Shrieve DC, Alexander E 3rd, Wen PY, et al. Comparison ofstereotactic radiosurgery and brachytherapy in the treatment ofrecurrent glioblastoma multiforme. Neurosurgery 1995;36:275–282; discussion 282–274.

2. Wong ET, Hess KR, Gleason MJ, et al. Outcomes and prog-nostic factors in recurrent glioma patients enrolled onto phaseII clinical trials [see comment]. J Clin Oncol 1999;17:2572.

3. Nieder C, Nestle U, Walter K, et al. Hypofractionated stereo-tactic radiotherapy for malignant glioma: A phase I/II Study.J Radiosurg 1999;2:107–111.

4. Baumert BG, Lutterbach J, Bernays R, et al. Fractionatedstereotactic radiotherapy boost after post-operative radiother-apy in patients with high-grade gliomas. Radiother Oncol2003;67:183–190.

5. Cardinale RM, Schmidt-Ullrich RK, Benedict SH, et al. Ac-celerated radiotherapy regimen for malignant gliomas usingstereotactic concomitant boosts for dose escalation. RadiatOncol Invest 1998;6:175–181.

6. Regine WF, Patchell RA, Strottmann JM, et al. Combinedstereotactic split-course fractionated gamma knife radiosur-gery and conventional radiation therapy for unfavorable glio-mas: A phase I study. J Neurosurg 2000;93(Suppl. 3):37–41.

7. Cho KH, Hall WA, Lo SS, et al. Stereotactic radiosurgeryversus fractionated stereotactic radiotherapy boost for patientswith glioblastoma multiforme. Technol Cancer Res Treatment2004;3:41–49.

8. Lederman G, Wronski M, Arbit E, et al. Treatment of recur-rent glioblastoma multiforme using fractionated stereotacticradiosurgery and concurrent paclitaxel. Am J Clin Oncol 2000;23:155–159.

9. Laing RW, Warrington AP, Graham J, et al. Efficacy andtoxicity of fractionated stereotactic radiotherapy in the treat-ment of recurrent gliomas (phase I/II study). Radiother Oncol1993;27:22–29.

0. Nieder C, Wiedenmann N, Stark S, et al. Phase II trial offunctional imaging-optimized stereotactic fractionated radio-therapy plus temozolomide for recurrent high-grade glioma.Proceedings from the 40th Annual Meeting of the AmericanSociety of Clinical Oncology, June 5–8, 2004; New Orleans,LA. Alexandria, VA: American Society of Clinical Oncology;2004. p. 120S.

1. Cho KH, Hall WA, Gerbi BJ, et al. Single dose versusfractionated stereotactic radiotherapy for recurrent high-gradegliomas. Int J Radiat Oncol Biol Phys 1999;45:1133–1141.

2. Selch MT, DeSalles AAF, Solberg TD, et al. Hypofraction-ated stereotactic radiotherapy for recurrent malignant gliomas.

J Radiosurg 2000;3:3–12.