Int. J. Radiation Oncology Biol. Phys., Vol. 56, No. 2, pp.

511–518, 2003
Copyright © 2003 Elsevier Inc.
Printed in the USA. All rights reserved
0360-3016/03/$–see front matter

doi:10.1016/S0360-3016(02)04611-4

CLINICAL INVESTIGATION Brain

RADIATION THERAPY FOR INTRACRANIAL GERM CELL TUMORS

DAPHNE A. HAAS-KOGAN, M.D.,*† BRIAN T. MISSETT, M.D.,* WILLIAM M. WARA, M.D.,*

SARAH S. DONALDSON, M.D.,‡ KATHLEEN R. LAMBORN, PH.D.,† MICHAEL D. PRADOS, M.D.,†
PAUL G. FISHER, M.D.,§ STEPHEN L. HUHN, M.D.,¶ BENJAMIN M. FISCH, M.D.,*
MITCHEL S. BERGER, M.D.,† AND QUYNH-THU LE, M.D.‡
Departments of *Radiation Oncology and †Brain Tumor Research Center and Neurosurgery, University of California, San
Francisco, CA; Departments of ‡Radiation Oncology, §Neurology, and ¶Neurosurgery, Stanford University Medical Center,
Stanford, CA

Purpose: To review the combined experiences of University of California, San Francisco, and Stanford University
Medical Center in the treatment of intracranial germ cell tumors (GCT) and to assess the impact of craniospinal
radiation (CSI) on patterns of relapse, progression-free survival (PFS), and overall survival (OS).
Patients and Methods: Ninety-three patients received radiation for newly diagnosed intracranial GCTs, including
49 germinomas, 16 nongerminomatous GCTs (NGGCT), and 28 with no biopsy. Median follow-up for surviving
patients was 4.5 years (range 0.25–34). Tests for variables correlating with OS and PFS were conducted using
Cox proportional hazards model.
Results: Five-year PFS and OS rates were 60% ⴞ 15% and 68% ⴞ 14% for patients with NGGCT and 88% ⴞ
5% and 93% ⴞ 4% for those with germinoma. Of 6 patients with localized NGGCT who did not receive CSI,
1 experienced an isolated spinal recurrence but was salvaged. Of 41 patients with localized germinoma, 6 who
received CSI and 35 who did not, no isolated spinal cord relapses occurred. Twenty-one patients with localized
germinoma received neither CSI nor whole brain radiation. Of these, none of 18 with ventricular radiation
relapsed. One of 3 patients with primary tumor radiation relapsed intracranially but had only received 11 Gy at
initial treatment. On multivariate analysis, germinoma histology but not CSI correlated with improved PFS and
OS.
Conclusion: CSI is not indicated in the treatment of localized germinomas. For patients with localized germi-
nomas treated with radiation alone, we recommend ventricular irradiation followed by primary tumor boost to
a total of 45–50 Gy. © 2003 Elsevier Inc.

Germinoma, Brain, Craniospinal irradiation, Pediatric.

INTRODUCTION larly craniospinal irradiation (CSI), has prompted many

Intracranial germ cell tumors (GCT) represent 3–11% of
pediatric and 1% of adult brain tumors (1). Intracranial
GCTs comprise two distinct histologic groupings: germino-
mas are the most common histologic subtype, whereas
nongerminomatous germ cell tumors (NGGCT) represent
one-third of intracranial GCTs and consist of embryonal
carcinoma, endodermal sinus (yolk sac) tumor, choriocar-
cinoma, teratoma, and GCT of mixed cellular origin (1, 2).
Historically, radiation alone, frequently encompassing a
large treatment volume, has constituted the gold standard of
treatment for intracranial GCTs (1, 3–5). However, the
morbidity of radiation therapy in children (6 –11), particu-
investigators to explore approaches that reduce the volume
and dose of radiotherapy while preserving high cure rates.
The development of strategies to reduce the morbidity of
radiotherapy has been hampered by an absence of random-
ized trials comparing treatment approaches. As pediatric
consortia and cooperative groups prepare to design a ran-
domized, phase III trial for children with intracranial GCTs,
we sought to use the combined experiences of the Univer-
sity of California, San Francisco (UCSF), and Stanford
University Medical Center (SUMC) to shed light on key
treatment issues. In particular, historical differences in the
radiation volumes prescribed in the two institutions provide
a cohort of patients that allow assessment of the utility of

Reprint requests to: Daphne A. Haas-Kogan, Department of
Radiation Oncology, University of California, San Francisco, 1600
Divisadero, San Francisco, CA 94143. Tel: (415) 353-7175; Fax:
(415) 353-9883; E-mail: hkogan@radonc17.ucsf.edu
Presented as an oral presentation at the American Society for
Therapeutic Radiology and Oncology, San Francisco, CA, Novem-
ber 4 – 8, 2001.
Supported in part by North American Brain Tumor Consortium
grant NCI U01-CA62399 (D.A.H-K.), NIH grant M01 RR01271,
Pediatric Clinical Research Center (D.A.H-K.), and CA 82103
(K.R.L. and M.S.B.).
Received Aug 19, 2002, and in revised form Dec 6, 2002.
Accepted for publication Dec 11, 2002.

511
512 I. J. Radiation Oncology ● Biology ● Physics Volume 56, Number 2, 2003

Table 1. Patient characteristics and treatment details

No biopsy or
Germinoma NGGCT nondiagnostic tissue
Pathology (n ⫽ 49) (n ⫽ 16) (n ⫽ 28) Total (93)

Median age (range), years 15 (3–36) 11 (2–20) 17 (3–47) 15 (2–47)

Median follow-up (range),* months 64 (7–213) 40 (12–221) 44 (3–412) 57 (3–412)
Median KPS (range) 90 (60–100) 90 (60–90) 90 (40–90) 90 (40–100)
Female:male 17:32 5:11 6:22 28:65
Localized:disseminated 41:8 11:5 21:7 73:20
Extent of surgery: no surgery 3 0 22 25
Biopsy only 22 4 2 28
Subtotal resection 19 7 2 28
Gross total resection 5 5 2 12
Site (SUMC:UCSF) 14:35 8:8 9:19 31:62
PFS for relapsing patients, months 42 (.72–111) 23 (.20–64) 19 (6–312) 23 (.2–312)
Chemotherapy 9 12 3 24 (26%)
Radiation field: focal field† 21 6 14 41
Whole brain 15 2 7 24
Craniospinal irradiation 13 8 7 28
Dose: median (range) Gy Tumor dose 50 (12–55) 50 (11–54) 50 (20–60) 50 (11–60)
Whole-brain dose‡ 31 (20–49) 24 (11–36) 34 (20–51) 31 (11–51)
CSI dose 30 (24–36) 33 (31–42) 40 (26–46) 32 (24–46)

* Follow-up for surviving patients.

†
Focal field ⫽ whole ventricle or primary tumor volume plus margin.
‡
Whole-brain dose for patients who did not receive spinal irradiation.
Abbreviations: NGGCT ⫽ nongerminomatous germ cell tumor; KPS ⫽ Karnofsky performance status; PFS ⫽ progression-free survival;
SUMC ⫽ Stanford University Medical Center; UCSF ⫽ University of California, San Francisco; CSI ⫽ craniospinal irradiation.

CSI, because a greater proportion of patients treated at
SUMC received CSI. Furthermore, we sought to elucidate
the role of chemotherapy and the appropriate volume and
dose of radiation in the treatment of intracranial GCTs.
PATIENTS AND METHODS
We conducted an historical cohort study of 93 patients
with confirmed or suspected intracranial GCTs treated in the
departments of radiation oncology at either SUMC or
UCSF, between 1968 and 2001. Excluded patients had
received prior radiation, did not receive radiation as part of
their initial treatment regimen, or had extracranial metasta-
ses at the time of diagnosis.
Histologic confirmations of GCT were obtained in 65 of
93 cases (70%). Forty-nine patients (53%) had germinomas
confirmed by biopsy. Sixteen patients (17%) had biopsy-
proven NGGCTs, including 8 mixed GCTs, 6 teratomas, 1
embryonal carcinoma, and 1 choriocarcinoma. The remain-
ing 28 patients (30%) were presumed to have intracranial
GCTs based on clinical presentation, biochemical markers,
or neuroradiographic imaging that suggested the presence of
tumor in the pineal or suprasellar region (12, 13). Twenty-
five of these patients did not undergo biopsy either because
they were considered unacceptable surgical risks or had a
classical radiodiagnostic trial of 20 Gy in an effort to
confirm the presence of a radiation-responsive intracranial
GCT (12, 13). Six of the patients classified as having no
biopsy information available had an attempted biopsy or
resection that proved nondiagnostic.
Table 1 delineates the patient characteristics and treat-
ment factors. Treatment fields included CSI, whole brain
without spinal irradiation, or more focal fields consisting of
whole ventricular irradiation or involved-field irradiation to
the primary tumor volume plus margin. Twenty-four pa-
tients (26%) received chemotherapy as part of initial treat-
ment, most commonly: carboplatin, VP 16, bleomycin, cis-
platin, cyclophosphamide, and vincristine. For all patients
who received chemotherapy as part of their initial treatment,
chemotherapy was administered before radiation. Radiation
therapy after chemotherapy consisted of CSI in 11 patients
and focal radiation in the remaining 13 patients.
Progression-free survival (PFS) and overall survival (OS)
rates were calculated from date of diagnosis. Median fol-
low-up for surviving patients in the entire cohort was 57
months (range 3– 412 months). Thirty-six patients had fol-
low-up for longer than 5 years. Patients lost to follow-up
were defined as those with less than 5-year follow-up if they
were diagnosed before 1995, or those with no documented
follow-up after January 1, 1998, if they were diagnosed
after 1995 (January 1, 1998, represents a 2-year interval
before compilation of this database).
Life-table methods were used to estimate PFS and OS
(14). Elapsed time from diagnosis to an event or last fol-
low-up was used to compute PFS and OS probabilities.
Patients, who were followed without an event or were lost
to follow-up before disease relapse, were censored on the
date of their last imaging study. In analyses of PFS, last
follow-up dates for patients not experiencing an event were
based on imaging studies that delineated disease status. For
 Radiation for intracranial germ cell tumors
Fig. 1. Actuarial (a) progression-free survival and (b) overall survival, according to histology, of all patients with
intracranial GCT. (— —), pathologically proven germinoma; (- - -), pathologically proven nongerminomatous germ cell
tumor; (——), no biopsy.
OS, deaths, regardless of cause, were coded as events. For
PFS, an event was defined as relapse, progression, or death
during the period of active follow-up. Kaplan-Meier curves
for PFS and OS were calculated for germinoma, NGGCT,
and unbiopsied tumor subgroups. Additional Kaplan-Meier
curves were calculated to determine the PFS and OS of
patients with localized disease stratified by histologic sub-
group. A final set of Kaplan-Meier curves was calculated for
patients with localized germinoma by whether or not they
received CSI.
To assess variables influencing PFS and OS, univariate
analyses were performed to evaluate histology, total tumor
dose, treatment with chemotherapy, and administration of
CSI. To further verify the results of univariate analyses, Cox
proportional hazards multivariate analyses were conducted
adjusting for radiation dose to the primary tumor, tumor
histology, CSI or whole-brain radiation, Karnofsky Perfor-
mance Status (KPS), chemotherapy, and presence of dis-
seminated disease at diagnosis. Analyses of tumor dose
included all patients who received ⱖ30 Gy, excluding 7
patients who received a radiodiagnostic dose of 20 Gy in
lieu of histologic diagnoses. Disseminated disease was de-
fined as multifocal tumor, metastases on spinal imaging, or
CSF cytology revealing malignant cells. This study was
performed after approval by the local Human Investigations
Committees. A portion of the patients treated at UCSF and
SUMC were included in previous, separate reports (5, 15,
16).
RESULTS
PFS and OS by histology
Five-year PFS rates were 60% ⫾ 15% for patients with
biopsy-proven NGGCTs compared with 88% ⫾ 5% for
those with biopsy-proven germinomas (Fig. 1a). Five-year
OS rates were 68% ⫾ 14% and 93% ⫾ 4% for patients with
● D. A. HAAS-KOGAN et al. 513
NGGCTs and germinomas, respectively (Fig. 1b). Patients
treated without a biopsy had 5-year PFS and OS rates of
61% ⫾ 11% and 85% ⫾ 8%, respectively (Figs. 1a and 1b).
Within each histologic subtype, PFS and OS rates were not
substantially different for the subgroup of patients present-
ing with localized disease (Figs. 2a and 2b). The median
PFS for those patients suffering progression or relapse was
23 months (range 0.20 –311 months) for the entire cohort,
with 4 patients experiencing an initial relapse more than 5
years after diagnosis.
This study included 16 patients with pathologically
proven NGGCT, of whom 3 progressed, 4 relapsed, 1 died
with neither disease progression nor relapse, and 1 was lost
to follow-up. Of 49 patients with pathologically proven
germinomas, 1 progressed, 5 relapsed, 1 died with neither
disease progression nor relapse, and 11 were lost to follow-
up. For 11 germinoma patients lost to follow-up, median
follow-up was 33.1 months (range 7.2–55.5 months), and all
but 1 of these 11 were followed for at least 1 year. The
overall median follow-up for surviving germinoma patients
was 64 months (range 7–213 months).
Nongerminomatous germ cell tumors (NGGCT)
Sixteen patients with pathologically proven NGGCTs
were included in this analysis—11 with localized and 5 with
disseminated disease. Of 6 patients with localized disease
who did not receive CSI, 1 had an isolated spinal recurrence
and 1 an elevated ␤-HCG level (Table 2). Although both of
these patients were salvaged with chemotherapy, with and
without additional radiation, respectively, 2 other patients
died of disease progression evident during their course of
radiotherapy. Of 5 patients who received CSI for localized
NGGCTs, 1 died of unknown causes 101 months after
diagnosis.
Of the 5 patients presenting with disseminated NGGCTs,
3 received CSI, 1 whole brain without spinal radiation, and
514 I. J. Radiation Oncology ● Biology ● Physics Volume 56, Number 2, 2003

Fig. 2. Actuarial (a) progression-free survival and (b) overall survival, according to histology, of patients with localized
intracranial GCT. (— —), pathologically proven germinoma; (- - -), pathologically proven nongerminomatous germ cell
tumor; (——), no biopsy.

1 focal radiation only. Among disseminated patients treated and none of these 7 patients received CSI. Two patients
with CSI, one recurred synchronously at intracranial and failed synchronously in the brain and spine, 1 failed in the
spinal sites and one progressed while receiving radiation. spine only, 2 relapsed at the primary site, 1 progressed, and
The single patient who received whole-brain radiation ther- 1 died of an arterial thrombosis two decades after receiving
apy had an isolated primary site relapse. All 3 of these radiation therapy. Of the 7 patients with localized disease
relapsed or progressive patients died of their disease. who lacked pathologic diagnoses and relapsed, none had
received chemotherapy.
No biopsy
Twenty-eight patients had no pathologic confirmation of
their diagnoses. Twenty-one of these patients had localized Germinomas
disease and 7 had disseminated disease. Of the 21 patients Forty-nine patients with pathologically proven germino-
without a pathologic diagnosis who had localized disease, 2 mas were analyzed. The 5-year OS and PFS were 93% and
received CSI, 6 received whole-brain radiation, and 2 re- 88%, respectively. Of the 41 patients presenting with local-
ceived chemotherapy. Five of the 7 patients with no patho- ized disease, 6 received CSI, 14 whole brain without spinal
logic diagnosis and disseminated disease received CSI, 1 radiation, and 21 local radiation only. Nine patients received
received whole-brain radiation, and the remaining patient chemotherapy—5 with localized germinoma, and 4 with
received chemotherapy with focal, primary site radiation. disseminated disease.
Of the 21 patients with localized disease who lacked For patients with localized disease, none of 6 patients
pathologic confirmation, 2 received CSI and were disease- who received CSI failed (Fig. 3a). Median follow-up was
free, but these patients had only 6 months and 11 months of 90.7 months (range 14.1–139.4 months), and 5 of these 6
follow-up, respectively. Seven of these 21 patients failed, patients were followed for a minimum of 5 years. For
patients with localized disease, 5 of 35 patients who did not
receive CSI failed, but none experienced an isolated spinal
Table 2. Disease relapse in patients with localized intracranial relapse (Fig. 3a and Table 2). Of the 5 patients who failed,
germ cell tumors according to whether craniospinal irradiation 1 had progressive disease during radiotherapy, 2 had extra–
was administered central nervous system (CNS) metastases, 1 had synchro-
Localized Localized nous spinal and primary site relapses, and 1 had a primary
Variable NGGCT germinoma site failure after discontinuing treatment at 11 Gy to pursue
alternative therapy. Three of the 5 relapsed patients died
CSI Yes No Yes No with disease (Fig. 3b). Of 6 patients who received CSI for
Number of patients 5 6 6 35
Isolated spinal relapse 0 1* 0 0
localized germinoma, 1 died of anaplastic astrocytoma that
Any spinal relapse 0 1 0 1† developed 9 years after CSI (Fig. 3b), and was likely radi-
Recurrences at primary site 0 2 0 3 ation-induced given the in-field location and latency of this
Patients alive and NED 4 4 5 32 second primary tumor.
Eight germinoma patients presented with disseminated
* Patient was salvaged with chemotherapy.
†
Synchronous primary and spinal failures. disease. Seven of these patients received CSI and were
Abbreviations: NGGCT ⫽ nongerminomatous germ cell tumor; without disease at last follow-up; however, 1 of these 7
CSI ⫽ craniospinal irradiation; NED ⫽ no evidence of disease. patients failed to thrive and died of unknown etiology.
 Radiation for intracranial germ cell tumors ● D. A. HAAS-KOGAN et al. 515

Fig. 3. Actuarial (a) progression-free survival and (b) overall survival of patients with pathologically proven localized
germinoma, according to whether craniospinal irradiation was administered. (——), craniospinal irradiation; (— —), no
craniospinal irradiation.

Results by treatment volume yses to evaluate whether these variables are associated with
As the standard of care for localized germinoma has either improved PFS or OS, followed by multivariate anal-
shifted away from CSI (17), a point of controversy has yses to adjust for disease dissemination, histology, tumor
centered on appropriate focal radiation fields. Confusion dose, KPS, and treatment with chemotherapy, craniospinal
persists because many North American studies have not irradiation, and whole brain irradiation.
delineated whether focal radiation fields targeted the tumor Table 3 details the variables tested in multivariate anal-
volume only or included the ventricular system. To address yses and their respective associations with clinical outcome.
this issue, we analyzed recurrence patterns after treatment Analyses of histology included only biopsy-proven germi-
with focal fields including the whole ventricular system or nomas and NGGCTs. For OS, hazard ratios of 0.2 in both
focal fields targeting only the primary tumor volume. Of 41 univariate and multivariate analyses indicated improved
localized germinoma patients, 21 were treated without CSI survival of germinoma patients compared with NGGCT
or whole-brain irradiation, but with more focal radiation patients (p ⫽ 0.01 and p ⫽ 0.07 for univariate and multi-
fields that included whole ventricular radiation in 18 or the variate analyses, respectively). The superior clinical out-
primary tumor plus margin in 3. For the 18 patients treated
with an initial whole ventricular field, the median dose to
the ventricles was 32.4 Gy (range 19.8 –50.4 Gy), and in all Table 3. Multivariate analyses: overall survival and progression-
free survival
but 2 of these patients the ventricular field was followed by
a boost to the primary tumor. Only 3 of these 18 patients Hazard Hazard ratio
received chemotherapy, and none experienced disease re- Variable* ratio 95% CI p value
lapse (median follow-up 57.7 months; range 12.9 –140.0
Overall survival
months). Three patients were treated with a field consisting
Histology† 0.21 0.037–1.13 0.07
of the initial primary tumor volume plus margin. None Tumor dose‡ 0.92 0.82–1.0 0.14
received chemotherapy, and 1 patient failed. This patient’s Craniospinal irradiation 0.58 0.13–2.6 0.46
clinical course is instructive because of the pattern of failure Chemotherapy 1.5 0.33–6.7 0.61
he experienced. His initial suprasellar primary was treated Progression-free survival
Histology† 0.14 0.027–0.67 0.02
with a 2-cm margin to only 11.6 Gy, at which time the
Tumor dose‡ 0.94 0.85–1.1 0.30
patient discontinued treatment. An isolated primary site Craniospinal irradiation 0.62 0.16–2.4 0.48
relapse was treated by surgical resection and chemotherapy Chemotherapy 0.91 0.24–3.5 0.89
followed by radiation, again targeting only the primary
tumor volume plus a 2-cm margin. Six years later the patient * Each variable was tested in a Cox proportional hazards model
adjusting for other potentially confounding variables as detailed in
experienced an intracranial relapse at a site clearly distinct text.
from the original primary tumor, received CSI as salvage †
Analyses of histology included only patients with pathologi-
therapy, and is currently NED (no evidence of disease) 7 cally proven germinomas or nongerminomatous germ cell tumors.
years after completion of CSI. Germinoma histology was associated with enhanced overall sur-
vival and progression-free survival compared with nongermino-
matous germ cell tumor.
Variables predicting clinical outcome in GCT ‡
Analyses of tumor dose included patients who received ⱖ30
Several variables have been suggested as predictors of Gy, excluding those who received a radiodiagnostic dose of 20 Gy.
clinical outcome for GCTs. We performed univariate anal- Abbreviation: CI ⫽ confidence interval.
516 I. J. Radiation Oncology ● Biology ● Physics
come of germinoma histology was supported by the analy-
ses of PFS, in which univariate and multivariate analyses
revealed hazard ratios of 0.2 (p ⫽ 0.07) and 0.1 (p ⫽ 0.02),
respectively.
In examining tumor dose, identical hazard ratios of 0.9
were seen for OS and PFS in both univariate and multivar-
iate analyses. Hazard ratios approaching 1 and lack of
statistical significance suggest that higher tumor dose did
not affect tumor control and clinical outcome. In analyzing
the effect of tumor dose, we also analyzed separately the
patients with localized germinomas. Within this group, uni-
variate and multivariate analyses revealed no significant
associations between dose to the primary tumor site and
either PFS (p ⫽ 0.23 and p ⫽ 0.38, for univariate and
multivariate analyses, respectively) or OS (p ⫽ 0.08 and p
⫽ 45, for univariate and multivariate analyses, respective-
ly). Univariate and multivariate analyses indicated no asso-
ciation of OS or PFS with either treatment with chemother-
apy or administration of CSI (Table 3).
Treatment toxicity
Thirty-five patients had documented endocrinopathies be-
fore initiation of radiation (defined as diabetes insipidus,
growth retardation, hypothyroidism, adrenocorticotropic
hormone (ACTH) deficiency, or sex hormone deficiency).
Ten patients required initiation of hormone therapy after
completion of radiotherapy. Of these 10 patients, 4 had
received CSI, 1 had received whole-brain, 4 had received
whole-ventricle, and 1 had received involved field radiation.
Forty-eight patients had normal endocrine profiles and did
not require hormone replacement either before or after
irradiation.
There was one case of decreased luteinizing hormone and
follicle-stimulating hormone levels thought to be secondary
to chemotherapy as a result of timing of the declining
hormone levels, and one case of panhypopituitarism attrib-
uted to the radiation treatment, although the time of onset of
the endocrinopathy was not delineated.
Three patients died after treatment with no evidence of
disease. One patient with disseminated germinoma was
diagnosed with failure to thrive after treatment with CSI and
chemotherapy and died 43 months after diagnosis, 1 patient
died of an intracranial arterial thrombosis 23 years after
whole-brain irradiation, and 1 patient died of a cerebellar
anaplastic astrocytoma 9 years after CSI. Two patients died
of unknown causes after loss to follow-up 9 and 11 years
after treatment, respectively.
DISCUSSION
Controversy remains in the treatment of intracranial
GCTs, due largely to the complexity and rarity of these
tumors, coupled with a dearth of prospective randomized
trials. Our data, in conjunction with the current literature,
support distinct therapeutic approaches for germinomas and
NGGCTs. The only factor predictive of superior clinical
outcome in this study was the diagnosis of germinoma, as
Volume 56, Number 2, 2003
compared with NGGCT. Although many published stud-
ies group germinoma with NGGCT, their distinct prog-
nosis points to the need to approach these clinical entities
individually.
Published reports corroborate poorer prognosis for pa-
tients with NGGCT compared with those with germinoma
(2, 18). For NGGCTs, radiation alone produces 5-year sur-
vival rates of only 30 – 40% (15). Similarly, we found a
5-year OS rate of only 68% for NGGCT compared with
93% for germinoma. More recent studies that demonstrate
excellent response rates to chemotherapy have shifted the
standard treatment of NGGCT to combined modality ther-
apy consisting of chemotherapy and radiation.
The choice of irradiation field in the treatment of local-
ized NGGCT remains unclear. The lack of an association in
this study between CSI and either PFS or OS may result
from the small sample size of NGGCT patients. There is
scarce literature addressing the appropriate radiation field
for localized NGGCT and only small patient numbers. In a
study by the French Society of Pediatric Oncology, chemo-
therapy and focal radiation resulted in five relapses among
24 patients with localized NGGCT (19). A similar regimen
in a separate study resulted in 3 of 18 patients experiencing
disease recurrence, 2 with isolated spinal relapses (20).
In 5 of the 7 NGGCT patients with disease progression or
relapse, the primary site was a component of failure, repre-
senting the dominant form of disease relapse. A high inci-
dence of primary site failure in NGGCT has been reported
by others and highlights the need for more intensive che-
motherapy or higher radiation doses to the primary site for
patients with poor responses to chemotherapy or radiation
(18).
Of the various histologic subtypes of intracranial GCTs,
germinoma represents the most common and prognostically
favorable subgroup. Historically, intracranial germinomas
have been treated with radiation alone, producing excellent
cure rates (1, 3, 5). The significant long-term toxicities of
radiation, particularly in children (6 –11), have prompted
investigations into alternative treatments that minimize the
dose and volume of irradiation. However, these alternative
approaches must preserve the high cure rates achieved by
radiation alone.
Several series have reported excellent clinical outcome
with preirradiation chemotherapy followed by focal irradi-
ation (21–27). The appeal of this approach lies in lower
radiation doses and smaller radiation fields that presumably
translate into reduced long-term toxicity. In our retrospec-
tive study the addition of chemotherapy, as assessed by
univariate and multivariate analyses, did not influence either
PFS or OS. Of 49 germinomas, only 9 received chemother-
apy, yet survival rates were excellent. Of note, only 3 of 41
patients with localized germinoma received focal irradiation
directed at the primary tumor only, although only 1 of these
patients received chemotherapy, as would be consistent
with the current paradigm of chemoradiation. One of these
patients failed, but had been treated suboptimally with only
11 Gy at the time of initial treatment. The remaining 38
 Radiation for intracranial germ cell tumors
patients with localized germinoma received whole-ventri-
cle, whole-brain, or CSI fields. Therefore, our data cannot
address the possible benefit afforded by adding chemother-
apy to focal radiation. Radiation alone continues to be the
gold standard for the treatment of intracranial germinomas.
Promising approaches such as combined chemotherapy and
focal, reduced-dose irradiation must be compared with ra-
diation alone in a prospective, randomized trial.
In addition to the role of chemotherapy, a point of con-
troversy in the treatment of localized germinomas has been
the appropriate field of radiation (17). Previous studies have
reported prospective data suggesting the effectiveness of
CSI in preventing locoregional recurrence. In analyzing the
combined experiences of SUMC and UCSF, we sought to
capitalize on historical differences in the fields of radiation
prescribed for patients with localized germinoma. Such
patients commonly received CSI at SUMC but generally
received more limited radiation fields at UCSF. CSI did not
influence PFS or OS for those with localized germinomas.
However, we must emphasize that only 6 patients with
localized germinoma received CSI, significantly limiting
the power of statistical analyses examining CSI. In addition,
longer follow-up may be required to identify all failures.
These limitations notwithstanding, the finding that not a
single isolated spinal relapse occurred among 35 patients
with localized germinomas treated without CSI, further
corroborates the lack of benefit of CSI. In a single patient,
the spine was a component of widespread CNS relapse;
however, no spinal failures were observed in the remaining
4 who recurred. Current evidence substantiates omitting
CSI from the treatment of localized germinoma. Spinal
failure rates of ⬍10% in the absence of CSI, reported in
most contemporary series, do not justify routine incorpora-
tion of CSI into treatment strategies for localized germi-
noma (4, 16, 18, 28 –30).
Nevertheless, the precise irradiation field appropriate for
localized germinoma remains unclear. In an attempt to
clarify this issue, we reviewed the irradiation fields of
patients with localized germinoma. Of 21 patients who had
neither CSI nor whole-brain irradiation, 3 received fields
encompassing only the primary plus margin and 18 received
whole-ventricular irradiation. Although no recurrences took
place after whole-ventricular fields, 1 intracranial relapse
occurred among 3 patients treated with radiation to the
initial tumor volume plus margin, although this patient
received a suboptimal initial dose of only 11 Gy. Because
the number of patients in our cohort who received irradia-
tion confined to the primary tumor only is small, we cannot
confirm the adequacy of such a focal field. However, several
REFERENCES
1. Hoffman HJ, Otsubo H, Hendrick EB, et al. Intracranial
germ-cell tumors in children. J Neurosurg 1991;74:545–551.
2. Kretschmar CS. Germ cell tumors of the brain in children: A
● D. A. HAAS-KOGAN et al. 517
studies report higher recurrence rates associated with radi-
ation targeting the tumor volume only (4, 18, 31–33).
We recommend whole-ventricular irradiation followed
by a boost to the primary tumor for localized germinoma
when using radiation alone. Some question the rationale of
whole-ventricular irradiation given the continuity of CSF
space throughout the entire CNS. Several published studies
indicate that germinomas may be characterized by multifo-
cality and intracranial relapses at foci separate from the
primary tumor with lesser propensity for diffuse CNS in-
volvement. These features distinguish germinomas from
other CNS malignancies, such as medulloblastoma, that
require CSI for cure (34). Whether more generous local
radiation fields sterilize occult multifocal disease or target
direct ventricular invasion, there appears a role for whole-
ventricular irradiation in the treatment of localized germi-
nomas.
Univariate and multivariate analyses failed to document
an association between total dose administered to the pri-
mary tumor and either PFS or OS. Although our data fail to
show a dose–response, the majority of our patients (72%)
received 50 –54 Gy. Therefore we cannot firmly assert that
lower doses of radiation will not compromise clinical out-
come. In fact, most contemporary and older studies report
superior tumor control rates after radiation doses greater
than 40 – 45 Gy (29, 35, 36). We believe the literature
supports a dose of ⱖ45 Gy to the primary tumor for ger-
minomas treated with radiation alone.
The high cure rate of intracranial GCTs allows assess-
ment of long-term toxicities of treatment. Forty-five patients
required hormone replacement for various endocrinopa-
thies, although the majority presented with endocrinopa-
thies before initiation of any treatment. These findings are
consistent with a report of germinomas, in which long-term
endocrine deficiencies after irradiation were associated ex-
clusively with similar deficiencies before treatment (37).
Although curative radiation doses may clearly lead to pitu-
itary dysfunction (38 – 40), the true contribution of irradia-
tion to long-term endocrine deficiencies seen in patients
with intracranial GCTs remains uncertain. The only death in
a germinoma patient treated with CSI resulted from an
anaplastic astrocytoma that developed 9 years after irradia-
tion. Similar reports of fatal CNS second primary tumors
highlight the need to explore alternative approaches that
minimize both the volume and dose of radiation (37). The
rarity and complexity of intracranial GCTs underscore the
need to address such issues prospectively in a collaborative
group setting.
review of current literature and new advances in therapy.
Cancer Invest 1997;15:187–198.
3. Bamberg M, Kortmann RD, Calaminus G, et al. Radiation
518 I. J. Radiation Oncology ● Biology ● Physics
therapy for intracranial germinoma: Results of the German
cooperative prospective trials MAKEI 83/86/89. J Clin Oncol
1999;17:2585–2592.
4. Aoyama H, Shirato H, Kakuto Y, et al. Pathologically-proven
intracranial germinoma treated with radiation therapy. Ra-
diother Oncol 1998;47:201–205.
5. Wolden SL, Wara WM, Larson DA, et al. Radiation therapy
for primary intracranial germ-cell tumors. Int J Radiat Oncol
Biol Phys 1995;32:943–949.
6. Radcliffe J, Packer RJ, Atkins TE, et al. Three- and four-year
cognitive outcome in children with noncortical brain tumors
treated with whole-brain radiotherapy. Ann Neurol 1992;32:
551–554.
7. Ellenberg L, McComb JG, Siegel SE, et al. Factors affecting
intellectual outcome in pediatric brain tumor patients. Neuro-
surgery 1987;21:638–644.
8. Packer RJ, Sutton LN, Atkins TE, et al. A prospective study
of cognitive function in children receiving whole-brain radio-
therapy and chemotherapy: 2-year results. J Neurosurg 1989;
70:707–713.
9. Waber DP, Tarbell NJ, Fairclough D, et al. Cognitive sequelae
of treatment in childhood acute lymphoblastic leukemia: Cra-
nial radiation requires an accomplice. J Clin Oncol 1995;13:
2490–2496.
10. Chapman CA, Waber DP, Bernstein JH, et al. Neurobehav-
ioral and neurologic outcome in long-term survivors of pos-
terior fossa brain tumors: Role of age and perioperative fac-
tors. J Child Neurol 1995;10:209–212.
11. Moore BD, Ater JL, Copeland DR. Improved neuropsycho-
logical outcome in children with brain tumors diagnosed dur-
ing infancy and treated without cranial irradiation. J Child
Neurol 1992;7:281–290.
12. Spiegel AM, Di Chiro G, Gorden P, et al. Diagnosis of
radiosensitive hypothalamic tumors without craniotomy: En-
docrine and neuroradiologic studies of intracranial atypical
teratomas. Ann Intern Med 1976;85:290–293.
13. Inoue Y, Takeuchi T, Tamaki M, et al. Sequential CT obser-
vations of irradiated intracranial germinomas. AJR Am J
Roentgenol 1979;132:361–365.
14. Kaplan EL, Meier P. Nonparametric estimation from incom-
plete observations. J Am Stat Assoc 1958;53:457–481.
15. Fuller BG, Kapp DS, Cox R. Radiation therapy of pineal
region tumors: 25 new cases and a review of 208 previously
reported cases. Int J Radiat Oncol Biol Phys 1994;28:229–
245.
16. Linstadt D, Wara WM, Edwards MS, et al. Radiotherapy of
primary intracranial germinomas: The case against routine
craniospinal irradiation. Int J Radiat Oncol Biol Phys 1988;
15:291–297.
17. Paulino AC, Wen BC, Mohideen MN. Controversies in the
management of intracranial germinomas. Oncology 1999;13:
513–521.
18. Matsutani M, Sano K, Takakura K, et al. Primary intracranial
germ cell tumors: A clinical analysis of 153 histologically
verified cases. J Neurosurg 1997;86:446–455.
19. Baranzelli MC, Patte C, Bouffet E, et al. Carboplatin-based
chemotherapy (CT), and focal irradiation (RT) in primary
cerebral germ cell tumors (GCT). J Clin Oncol 1999;18:140a.
20. Robertson PL, DaRosso RC, Allen JC. Improved prognosis of
intracranial non-germinoma germ cell tumors with multimo-
dality therapy. J Neurooncol 1997;32:71–80.
21. Allen JC, DaRosso RC, Donahue B, et al. A phase II trial of
preirradiation carboplatin in newly diagnosed germinoma of
the central nervous system. Cancer 1994;74:940–944.
Volume 56, Number 2, 2003
22. Kitamura K, Shirato H, Sawamura Y, et al. Preirradiation
evaluation and technical assessment of involved-field radio-
therapy using computed tomographic (CT) simulation and
neoadjuvant chemotherapy for intracranial germinoma. Int J
Radiat Oncol Biol Phys 1999;43:783–788.
23. Fouladi M, Grant R, Baruchel S, et al. Comparison of survival
outcomes in patients with intracranial germinomas treated
with radiation alone versus reduced-dose radiation and che-
motherapy. Childs Nerv Syst 1998;14:596–601.
24. Sawamura Y, Shirato H, Ikeda J, et al. Induction chemother-
apy followed by reduced-volume radiation therapy for newly
diagnosed central nervous system germinoma. J Neurosurg
1998;88:66–72.
25. Baranzelli MC, Patte C, Bouffet E, et al. Nonmetastatic intra-
cranial germinoma: The experience of the French Society of
Pediatric Oncology. Cancer 1997;80:1792–1797.
26. Bouffet E, Baranzelli MC, Patte C, et al. Combined treatment
modality for intracranial germinomas: Results of a multicentre
SFOP experience. Société Française d’Oncologie Pédiatrique.
Br J Cancer 1999;79:1199–1204.
27. Buckner JC, Peethambaram PP, Smithson WA, et al. Phase II
trial of primary chemotherapy followed by reduced-dose ra-
diation for CNS germ cell tumors. J Clin Oncol 1999;17:933–
940.
28. Dattoli MJ, Newall J. Radiation therapy for intracranial ger-
minoma: The case for limited volume treatment. Int J Radiat
Oncol Biol Phys 1990;19:429–433.
29. Sugiyama K, Uozumi T, Arita K, et al. Clinical evaluation of
33 patients with histologically verified germinoma. Surg Neu-
rol 1994;42:200–210.
30. Haddock MG, Schild SE, Scheithauer BW, et al. Radiation
therapy for histologically confirmed primary central nervous
system germinoma. Int J Radiat Oncol Biol Phys 1997;38:
915–923.
31. Shibamoto Y, Abe M, Yamashita J, et al. Treatment results of
intracranial germinoma as a function of the irradiated volume.
Int J Radiat Oncol Biol Phys 1988;15:285–290.
32. Shirato H, Nishio M, Sawamura Y, et al. Analysis of long-
term treatment of intracranial germinoma. Int J Radiat Oncol
Biol Phys 1997;37:511–515.
33. Aoyama H, Shirato H, Ikeda J, et al. Induction chemotherapy
followed by low-dose involved-field radiotherapy for intracra-
nial germ cell tumors. J Clin Oncol 2002;20:857–865.
34. Paulino AC. Radiotherapeutic management of medulloblas-
toma. Oncology 1997;11:813–823.
35. Abay EO 2nd, Laws ER Jr, Grado GL, et al. Pineal tumors in
children and adolescents. Treatment by CSF shunting and
radiotherapy. J Neurosurg 1981;55:889–895.
36. Kersh CR, Constable WC, Eisert DR, et al. Primary central
nervous system germ cell tumors. Effect of histologic confir-
mation on radiotherapy. Cancer 1988;61:2148–2152.
37. Hardenbergh PH, Golden J, Billet A, et al. Intracranial ger-
minoma: The case for lower dose radiation therapy. Int J
Radiat Oncol Biol Phys 1997;39:419–426.
38. Probert JC, Parker BR, Kaplan HS. Growth retardation in
children after megavoltage irradiation of the spine. Cancer
1973;32:634–639.
39. Schriock EA, Schell MJ, Carter M, et al. Abnormal growth
patterns and adult short stature in 115 long-term survivors of
childhood leukemia. J Clin Oncol 1991;9:400–405.
40. Clayton PE, Shalet SM. Dose dependency of time of onset of
radiation-induced growth hormone deficiency. J Pediatr 1991;
118:226–228.