C

Agi(tjtp

GcgbttC

ri

C1

Int. J. Radiation Oncology Biol. Phys., Vol. 65, No. 2, pp. 481–485, 2006Copyright © 2006 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/06/$–see front matter

doi:10.1016/j.ijrobp.2005.12.012

LINICAL INVESTIGATION Brain

LOW-DOSE PROPHYLACTIC CRANIOSPINAL RADIOTHERAPY FORINTRACRANIAL GERMINOMA

GORDON O. SCHOENFELD, M.D.,* ROBERT J. AMDUR, M.D.,* ILONA M. SCHMALFUSS, M.D.,†

CHRISTOPHER G. MORRIS, M.S.,* SAMEER R. KEOLE, M.D.,* WILLIAM M. MENDENHALL, M.D.,*AND ROBERT B. MARCUS, JR., M.D.‡

Departments of *Radiation Oncology and †Radiology, University of Florida College of Medicine, Gainesville, FL;and ‡Department of Radiation Oncology, Emory University, Atlanta, GA

Purpose: To report outcomes of patients with localized intracranial germinoma treated with low-dose cranio-spinal irradiation (CSI) followed by a boost to the ventricular system and primary site.Methods and Materials: Thirty-one patients had pathologically confirmed intracranial germinoma and no spinemetastases. Low-dose CSI was administered in 29 patients: usually 21 Gy of CSI, 9.0 Gy of ventricular boost, anda 19.5-Gy tumor boost, all at 1.5 Gy per fraction. Our neuroradiologist recorded three-dimensional tumor sizeon magnetic resonance images before, during, and after radiotherapy.Results: With a median follow-up of 7.0 years, 29 of 31 patients (94%) are disease free. One failure hadnongerminomatous histology; the initial diagnosis was a sampling error. Of 3 patients who did not receive CSI,1 died. No patient developed myelopathy, visual deficits, dementia, or skeletal growth problems. In locallycontrolled patients, tumor response according to magnetic resonance scan was nearly complete within 6 monthsafter radiotherapy.Conclusions: Radiotherapy alone with low-dose prophylactic CSI cures almost all patients with localizedintracranial germinoma. Complications are rare when the daily dose of radiotherapy is limited to 1.5 Gy and thetotal CSI dose to 21 Gy. Patients without a near-complete response to radiotherapy should undergo resection torule out a nongerminomatous element. © 2006 Elsevier Inc.

Germinoma, Radiotherapy.

td

t3edmfs

Tratda

s

2

INTRODUCTION

major controversy in the management of intracranialerminoma is the use of craniospinal axis irradiation (CSI)n patients with no evidence of spine metastasis at diagnosisprophylactic CSI). Proponents of prophylactic CSI arguehat the toxicity of irradiating the spine to moderate doses isustified by a cure rate of nearly 100%. Opponents arguehat the toxicity of prophylactic CSI is greater than theotential benefit.The standard dose for CSI has historically been 24–36

y at 1.8 Gy per treatment. More than 20 years ago, wehanged the radiotherapy program for germinoma, with theoal of decreasing toxicity while maintaining the potentialenefit of prophylactic CSI. In most patients, the dose perreatment is 1.5 Gy for the entire course of radiotherapy, andhe total CSI dose is 21 Gy. We use the term “low-doseSI” to refer to this basic program.The primary purpose of this article is to report our expe-

ience with low-dose prophylactic CSI in patients withntracranial germinoma. A secondary purpose is to correlate

Correspondence to: Robert J. Amdur, M.D., University of Floridaollege of Medicine, Department of Radiation Oncology, P.O. Box

00385, Gainesville, FL 32610-0385. Tel: (352) 265-0287; Fax: (352) A

481

umor response to radiotherapy-based MR scan, becauseata are scarce on this subject.

METHODS AND MATERIALS

We report the outcomes of 31 unselected patients who werereated in our department between January 1, 1982, and December1, 2002, with a tissue diagnosis of intracranial germinoma and novidence of spine metastasis. There was only 1 patient treateduring this time who is not included in this report. This patientoved out of the state immediately after treatment and was lost to

ollow-up. Outcome endpoints are disease-free survival, absoluteurvival, and treatment complications.

Table 1 summarizes the characteristics of the study group.able 2 summarizes the details of radiotherapy. Most patients

eceived 21 Gy of prophylactic CSI, a 9.0-Gy ventricular boost,nd then a 19.5-Gy tumor boost, all at 1.5 Gy per fraction, for aotal dose of 49.5 Gy (range, 40.5–51 Gy). Specifically, the CSIose was 21 Gy (21 patients), 20 Gy (4 patients), 25 Gy (2 patients),nd 24 Gy (1 patient).

A neuroradiologist (I.M.S.) calculated three-dimensional tumorize on magnetic resonance (MR) images before and after radio-

65-7045; E-mail: amdurrj@ufl.eduReceived Sept 9, 2005, and in revised form Dec 1, 2005.

ccepted for publication Dec 9, 2005.

tvuvors

ytOTCstpg

tsm

ntstrmpc

oadoaactbtfw

swwmh

t

C

VL

%)

CPM

i

c

t

482 I. J. Radiation Oncology ● Biology ● Physics Volume 65, Number 2, 2006

herapy. A near-complete response was defined as �5% residualolume. A partial response was defined as �50% but �5% resid-al volume. Minimal response was defined as �50% residualolume. In recent years, an MR scan was done after a few weeksf CSI to plan the tumor boost. Magnetic resonance–assessedesponse after only a few weeks of radiotherapy is analyzedeparately.

RESULTS

Median follow-up was 7.0 years (range, 1.8 –20.5ears). Two patients developed recurrent tumor, meaninghat the disease-free survival rate was 94% (29 of 31).ne recurrence was in a patient who received 21 Gy CSI.his was the only recurrence in a patient who receivedSI. This patient was a male 18-year-old who in retro-

pect had nongerminoma histology (sampling error at theime of original biopsy). The pretreatment biopsy in thisatient demonstrated pure germinoma. The tumor pro-ressed locally soon after radiotherapy, and repeat resec-

Table 1. Patient characteristics

Characteristic Value (%)

Years of study 1983–2002Tissue diagnosis 31/31 (100)Median age (range) (y) 18 (10–36)Gender

Male 27 (87)Female 4 (13)

Follow-up (y)Minimum 2Median 7.0Range 1.8–20.5

CSF cytologyNegative 12/14 (86)Positive 2/14 (14)Not done 17

Spine imagingMRI/CT 23Not done 8

LocationPineal 23Suprasellar 5Pineal and suprasellar 3

Abbreviations: CSF � cerebrospinal fluid; CT � computedomography; MRI � magnetic resonance imaging.

Data are presented as n (%) unless otherwise specified.

Table 2. Tr

Radiotherapy No. of patients CSI

SI 28* 21 Gy @ 1.5 Gy/f(20–25 Gy)

entricular � local 1 —ocal only 2† —

Abbreviations: CSI � craniospinal irradiation; fx � fraction.* Adjuvant chemotherapy (carboplatin, etoposide, bleomycin) a

† Induction chemotherapy (carboplatin, etoposide, bleomycin or cispla

ion demonstrated malignant teratoma in most of thepecimen. This patient died of disseminated disease 22onths after radiotherapy.The other recurrence was in 1 of the 3 patients who did

ot receive prophylactic CSI. This patient failed only inhe spine 14 months after radiotherapy to the ventricularystem followed by a tumor boost to 50 Gy. Salvageherapy included CSI and chemotherapy, with completeesponse in the spine. A second recurrence developed 12onths after salvage CSI in the suprasellar region. The

atient died soon thereafter of complications of salvagehemotherapy.

At last follow-up, 28 patients were alive with no historyf disease recurrence. Two patients died with disease at 2nd 3 years after treatment; 1 patient died of intercurrentisease 7 years after treatment. The 5-year and 10-yearverall and relapse-free survival rates were 93% and 88%,nd 94% and 94%, respectively. The 5- and 10-year overallnd relapse-free survival rates in the 28 patients who re-eived CSI were 96% and 90%, and 96% and 96%, respec-ively. If we exclude the single recurrence in the CSI groupecause this patient had a nongerminomatous tumor at theime of relapse, the 5-year and 10-year overall and relapse-ree survival rates in the 27 patients with pure germinomaere 100% and 94%, and 100% and 100%, respectively.Table 3 shows tumor shrinkage according to the first MR

can (usually 3 months) after radiotherapy. The only patientho did not have a near-complete response was the patientho recurred locally with malignant teratoma in the speci-en. Table 4 summarizes response data for the patients who

ad an MR scan during radiotherapy. In the 12 patients who

t summary

Ventricular system Tumor boost Tumor total

9 Gy @ 1.5 Gy/fx(8–19.5 Gy)

19.5 Gy @ 1.5 Gy/fx(0–28.5 Gy)

49.5 Gy(45–51 Gy)

21 Gy 29 Gy 50 Gy— — 30, 45 Gy

tered in 1 patient.

Table 3. Post-RT MRI response

Response�6 mo

after RT (%) Last MRI (

omplete response (0–5%) 17 (81) 19 (90)artial response (5–50% residual) 3 (10)* 1 (5)inimal response (�50%) 1 (3)† 1 (5)†

Abbreviations: RT � radiotherapy; MRI � magnetic resonancemaging.

Data are presented as n (%).* 7%, 10%, and 23% residual abnormality of 0.15 cm3, 0.81

m3, and 0.09 cm3.† In retrospect, this patient had nongerminoma histology. See

ext.

eatmen

x

dminis

tin, etoposide) administered in 1 patient.

hgcc

prdvtfp

rmrs

dottiur

C

ttpwovbotTg

R

t1otCt

vabfwf

L

tnGdt

stCCpop

PitpcT

1tf5

EVVNSSDM

CPM

o

t

483Low-dose CSI for intracranial germinoma ● G. O. SCHOENFELD et al.

ad data available regarding �-subunit human chorioniconadotropin levels, none was elevated, and therefore noomparisons could be made with tumor response or diseaseontrol.

Table 5 summarizes complications data. The majority ofatients with endocrinopathy had hormone dysfunction beforeadiotherapy. Twenty patients had visual problems before ra-iotherapy, most commonly Parinaud’s related syndrome orisual field deficit. In 11 of 20 patients (55%), visual problemshat were present before treatment did not improve at lastollow-up examination. No patient without visual findings atresentation had new visual problems after radiotherapy.

DISCUSSION

To date there are more than 20 published series with dataelated to the use of CSI in patients with intracranial ger-inoma. We include these citations with this article for

eaders who want a comprehensive reference list on thisubject (1–21).

Some children do not develop normally after standard-ose CSI (24–36 Gy at 1.8 Gy per treatment); however, it isften unclear whether growth problems are the result ofumor or a side effect of radiotherapy (6, 12). There arehree main ways to decrease the toxicity of radiotherapy forntracranial germinoma: use chemotherapy in place of CSI,se radiotherapy alone but do not treat the spine, or useadiotherapy alone with low-dose CSI.

hemotherapy instead of CSIThis strategy attempts to treat subclinical disease

hroughout the craniospinal axis with chemotherapy. Mul-iple studies report a low relapse rate in the spine withlatinum-containing regimens and radiotherapy to thehole brain or ventricular system (1, 3, 5, 8–10, 21). Mostf these series report relapses outside the local radiotherapyolume, but the salvage rate is high. In our opinion, it isetter to prevent relapses with low-dose CSI than it is tomit CSI and plan to salvage recurrences. Another issue ishe lack of long-term follow-up in the chemotherapy series.umor recurrence might occur years after treatment of

Table 4. Tumor response by MR scan during RT

Response

MRI responseduring RT

(%)*

omplete response (0–5%) 5 (25)artial response (5–50% residual) 14 (70)inimal response (�50%) 1 (5)†

Abbreviations as in Table 3.Data presented as n (%).* Magnetic resonance scan during treatment was not done in 11

f the 31 patients in this study. See text.† In retrospect, this patient had nongerminoma histology. See

ext.

erminoma (19).

adiotherapy alone without CSIMultiple series describe patients who do well with radio-

herapy alone to the whole brain or ventricular system (11,3, 16, 18, 19). When an MR scan of the spine and a samplef craniospinal fluid from the spine are negative for tumor,he chance of failure in the spine without chemotherapy orSI is �5%. The level of human chorionic gonadotropin in

he serum is not prognostically important (14).We see no weak points in the combined Stanford–Uni-

ersity of California, San Francisco, series (18). Haddock etl. (11) and Wolden et al. (19) report good results at 5 years,ut survival drops to approximately 80% after 10 years ofollow-up. All of the recurrences in our series presentedithin 2 years of the completion of CSI, with a median

ollow-up of 7 years.

ow-dose CSIOur strategy has been to lower the dose of CSI in a way

hat will eliminate subclinical disease with a low rate oformal tissue complications. The results with 21 Gy at 1.5y per fraction could not be better: 100% cure rate and noocumented complications. There are five main publica-ions reporting results after low-dose CSI, detailed below.

University of Pennsylvania (2004). In this retrospectiveeries (17), the dose of prophylactic CSI was decreased overime. Six patients were treated with radiotherapy alone withSI at 23.4–27 Gy and 3 patients with radiotherapy alone withSI at 18–19.8 Gy. The dose per treatment was 1.8 Gy. Noatient experienced tumor recurrence with a median follow-upf 7.2 and 5.5 years. It is not clear from this report whetheratients experienced complications from low-dose CSI.

Maligue Keimzelltümoren and International Society ofaediatric Oncology German trials (2000). After the Mal-

gue Keimzelltümoren trial (8) with CSI of 30 Gy at 1.5 Gy perreatment, the dose was lowered in a subsequent group of 13atients to CSI 24 Gy followed by a tumor boost of 16 Gy (nohemotherapy). The dose per treatment was not specified.here were no relapses with a median follow-up of 4.4 years.St. Jude (2000). This retrospective analysis (12) included

2 children with histologically confirmed germinoma, allreated with CSI of 25.6 Gy (23.4–32 Gy) at 1.5–1.8 Gy perraction, with a boost to the primary site to a total dose of0.8 Gy (45–54 Gy). There were no failures with a median

Table 5. Complications data

Complication n (%)

ndocrinopathy 11/31 (35)isual problems before treatment 20/31 (65)isual problems after treatment 11/31 (35)ew visual problems after radiotherapy —econd malignancy 1/31 (3)*keletal growth problems —ementia —yelopathy —

* Developed AML 7 years after chemotherapy/XRT.

fiNMdrwC

wrbfrptr

tpctoe

T

s

wg�Aagtrm

ntsotaaptdaei

484 I. J. Radiation Oncology ● Biology ● Physics Volume 65, Number 2, 2006

ollow-up of 69 months. Detailed neuropsychological test-ng was performed in 8 patients before and after treatment.o differences in intelligence quotient scores were noted.ost children (7 of 12) had endocrinopathy at time of

iagnosis, and only 1 child developed endocrinopathy afteradiation therapy. The importance of this study is that thereere no reported complications from low–moderate-doseSI in a patient population at highest risk of late effects.Maligue Keimzelltümoren German trials (1999). This

as a prospective study (2) of 49 patients treated withadiotherapy alone with CSI of 30 Gy followed by a tumoroost of 15 Gy, both at 1.5 Gy per treatment. Medianollow-up was 56 months. The 5-year relapse-free survivalate was 89%. There was only one spine relapse. Thisatient relapsed in the cervical spine at the junction betweenhe cranial and spine fields. Complications were reportedlyare, but details were not given.

Joint Center for Radiation Therapy (1997). In a subset ofhis retrospective series (6), the investigators report on 9atients who were treated with radiotherapy alone that in-luded CSI of �22 Gy (fractionation not specified). Ofhese patients, 2 had positive results on cytologic evaluationf cerebrospinal fluid. All 9 patients where cured, with novidence of progression or growth retardation.

umor response by MR scanOurs is the only series that formally analyzes tumor

hrinkage during and soon after radiotherapy in patients p

REFEREN

the treatment of pure intracranial germinoma: Impact of treat-

1

1

1

1

1

1

ith intracranial germinoma. Our data confirm that pureerminoma is highly radiosensitive. Most patients have50% tumor shrinkage after a few weeks of radiotherapy.n MR scan 3–6 months after radiotherapy should show

lmost no residual mass. This information should be used touide clinical management decisions. A patient with �75%umor shrinkage 3 months after completing �45 Gy ofadiotherapy should be considered to have a nongermino-atous element until proven otherwise.

CONCLUSIONS

Before treatment, all patients with intracranial germi-oma should have a lumbar puncture specimen analyzed forumor cells and a contrast-enhanced MR scan of the cranio-pinal axis. In patients with no evidence of spine metastasis,ur preference is to use radiotherapy alone to avoid theoxicity of chemotherapy. The toxicity of CSI seems to becceptable with the dose schedule that we describe in thisrticle. We continue to recommend 21 Gy of CSI in allatients, followed by 9 Gy to the ventricular system andhen 15 Gy to the primary site. It is important to keep theose per treatment as low as possible, which is why we usefraction size of 1.5 Gy. Modifications that we are consid-

ring to decrease the potential for toxicity include decreas-ng the final dose to 45 Gy and going directly from CSI to

rimary site boost.

CES

1. Aoyama H, Shirato H, Ikeda J, et al. Induction chemother-apy followed by low-dose involved-field radiotherapy forintracranial germ cell tumors. J Clin Oncol 2002;20:857–865.

2. Bamberg M, Kortmann RD, Calaminus G, et al. Radiationtherapy for intracranial germinoma: Results of the Germancooperative prospective trials MAKEI 83/86/89. J Clin Oncol1999;17:2585–2592.

3. Bouffet E, Baranzelli MC, Patte C, et al. Combined treatmentmodality for intracranial germinomas: Results of a multicentreSFOP experience. Societe Francaise d’Oncologie Pediatrique.Br J Cancer 1999;79:1199–1204.

4. Balmaceda C, Heller G, Rosenblum M, et al. Chemotherapywithout irradiation—a novel approach for newly diagnosedCNS germ cell tumors: Results of an international cooper-ative trial. The First International Central Nervous SystemGerm Cell Tumor Study. J Clin Oncol 1996;14:2908 –2915.

5. Gandola L, Navarria P, Massimino M, et al. Combined treat-ment of nonmetastatic intracranial germinoma in childhood:Impact of radiotherapy volume extent [Abstract]. Int J RadiatOncol Biol Phys 2002;54:203.

6. Hardenbergh PH, Golden J, Billet A, et al. Intracranial ger-minoma: The case for lower-dose radiation therapy. Int JRadiat Oncol Biol Phys 1997;39:419–426.

7. Zissiadis Y, Dutton S, Kieran M, et al. Stereotactic radiother-apy for pediatric intracranial germ cell tumors. Int J RadiatOncol Biol Phys 2001;51:108–112.

8. Kortmann RD, Calaminus G, Becker G, et al. Radiotherapy in

ment volume and dose prescription on outcome [Abstract]. IntJ Radiat Oncol Biol Phys 2000;48:204.

9. Matsutani M. Combined chemotherapy and radiation therapyfor CNS germ cell tumors—the Japanese experience. J Neu-rooncol 2001;54:311–316.

0. Buckner JC, Peethambaram PP, Smithson WA, et al. Phase IItrial of primary chemotherapy followed by reduced-dose ra-diation for CNS germ cell tumors. J Clin Oncol 1999;17:933–940.

1. Haddock MG, Schild SE, Scheithauer BW, et al. Radiationtherapy for histologically confirmed primary central nervoussystem germinoma. Int J Radiat Oncol Biol Phys 1997;38:915–923.

2. Merchant TE, Sherwood SH, Mulhern RK, et al. CNS germi-noma: Disease control and long-term functional outcome for12 children treated with craniospinal irradiation. Int J RadiatOncol Biol Phys 2000;46:1171–1176.

3. Ogawa K, Shikama N, Toita T, et al. Long-term results ofradiotherapy for intracranial germinoma: A multi-institutionalretrospective review of 126 patients. Int J Radiat Oncol BiolPhys 2004;58:705–713.

4. Ogino H, Shibamoto Y, Takanaka T, et al. CNS germinomawith elevated serum human chorionic gonadotropin level:Clinical characteristics and treatment outcome. Int J RadiatOncol Biol Phys 2005;62:803–808.

5. Shikama N, Toita T, Nakamura K, et al. Intracranial ger-minoma multicenter analyses: Should spinal irradiation beused for all patients? [Abstract]. Int J Radiat Oncol Biol

Phys 2003;57:S370.

1

1

1

1

2

2

485Low-dose CSI for intracranial germinoma ● G. O. SCHOENFELD et al.

6. Shikama N, Ogawa K, Tanaka S, et al. Lack of benefit of spinalirradiation in the primary treatment of intracranial germinoma: Amultiinstitutional, retrospective review of 180 patients. Cancer2005;104:126–134.

7. Maity A, Shu HK, Janss A, et al. Craniospinal radiation in thetreatment of biopsy-proven intracranial germinomas: Twenty-five years’ experience in a single center. Int J Radiat OncolBiol Phys 2004;58:1165–1170.

8. Haas-Kogan DA, Missett BT, Wara WM, et al. Radiationtherapy for intracranial germ cell tumors. Int J Radiat Oncol

Biol Phys 2003;56:511–518.

9. Wolden SL, Wara WM, Larson DA, et al. Radiation therapyfor primary intracranial germ-cell tumors. Int J Radiat OncolBiol Phys 1995;32:943–949.

0. Shibamoto Y, Oda Y, Yamashita J, et al. The role of cerebro-spinal fluid cytology in radiotherapy planning for intracranialgerminoma. Int J Radiat Oncol Biol Phys 1994;29:1089–1094.

1. Shirato H, Aoyama H, Ikeda J, et al. Impact of margin fortarget volume in low-dose involved field radiotherapy afterinduction chemotherapy for intracranial germinoma. Int J Ra-

diat Oncol Biol Phys 2004;60:214–217.