Pediatr Blood Cancer 2005;45:298–303
Ependymal Tumors in Childhood
Fulya Yaman Agaoglu, MD,1* Inci Ayan, MD,2 Yavuz Dizdar, MD, PhD,1 Rejin Kebudi, MD,2
Omer Gorgun, MD,2 and Emin Darendeliler, MD1
Background. Ependymal tumors are classi-
fied as ependymoma (benign or low grade)
versus anaplastic ependymoma (malignant or
high grade). Ependymomas represent 5–10% of
intracranial neoplasm in children. In this study,
demographic data and the treatment results of
pediatric patients with ependymal tumors,
treated in a single institute, is reported. Patients
and Methods. Between 1989 and 2001, 40 (22 M/
18 F) previously untreated patients with a median
age of 5.5 years (3 months–15 years), of histo-
logically proven ependymal tumors (except
ependymoblastomas) were referred to the Insti-
tute of Oncology, University of Istanbul. The
localization was supratentorial in 18, infraten-
torial in 20, both supra and infratentorial in two
patients. Histologic subgroups were 18 ependy-
momas (43.6%), and 22 anaplastic ependy-
momas (56.4%). Total tumor resection was
performed in 20 patients (50%), subtotal in 18
patients (45%), and biopsy only in 2 patients
(5%). Postoperative treatment consisted of regio-
nal (8 patients) or craniospinal (CSI) (9 patients)
radiotherapy (RT) in patients with ependymoma;
regional (7 patients) or CSI RT (14 patients) with
chemotherapy (ChT) in patients with anaplastic
ependymoma; ChT only (1 patient) in patients
less than 3 years of age. The standard technique
for posterior fossa irradiation was parallel-
opposed lateral fields and total dose was 45–
54 Gy. Between September 1989 and May 1991
patients received regimen A, which consisted of
RT followed by eight-in-one ChT, given every
4 weeks for eight courses. Patients who were
treated between June 1991 and July 1994, re-
ceived regimen B, which included two courses
of postoperative ‘‘VEC’’ (vincristine, etoposide,
Key words: adjuvant treatment; childhood brain tumors; ependymal tumors
INTRODUCTION
Currently, ependymal tumors are classified as ependy-
moma (benign or low grade) versus anaplastic ependymo-
ma (malignant or high grade). Low-grade ependymomas
are more common than high-grade ependymomas. These
tumors represent for 5–10% of all brain tumors in the
pediatric age group. It is the third most common brain
tumor in the childhood, following astrocytomas and
medulloblastomas. The tumors generally originate from
the ventricular floor where attachment is commonly noted
at the level of the obex. In approximately 75% childhood
cases, they occur below the tentorium [1–3]. Spread of
ependymoma is primarily local. The risk of meningeal
dissemination in ependymomas is about 2–30% [1].
ß 2005 Wiley-Liss, Inc.
DOI 10.1002/pbc.20212
cisplatin) ChT, administered every 3 weeks,
followed by RT applied with low dose con-
comitant cisplatin used as a radiosensitizer.
Patients with objective response to postoperative
‘‘VEC’’ continued to have ‘‘VEC’’ after comple-
tion of RT for six more courses. From August
1994 on, patients received regimen C, consist-
ing of RT and concomitant infusion of cisplatin
followed by ‘‘VCPCU’’ (vincristine, cyclopho-
sphamide, procarbazine, lomustine) adminis-
tered every 4 weeks for eight courses. Results.
A total of 40 patients were included in the
outcome and survival data. The 5-year overall
survival (OS) rate was 64.9%, and the 5-year
progression-free survival rate was 50.8% for the
whole series. Median time for progression or
relapse was 24.3 months and there were 19 pa-
tients (43.6%) with relapse or progression.
Non-metastatic patients (P ¼ 0.0008, 5-year OS
rate was 82% vs. 29%), and totally resected
patients (P ¼ 0.01, 5-year OS rate was 80% vs.
55%), and
3 years of age (P ¼ 0.04, 5-year OS
rate was 75% vs. 38%) had significantly better
outcome. Conclusions. The majority of com-
plete responders were patients who had total
tumor removal. Treatment failure occurred
mainly within the first 2 years, and outcome
was dismal for patients who relapsed or had
progressive disease. The median age at diag-
nosis is 6 years in our patient group; younger
children (less than 3 years old) have less
favorable outcome. There was no significant
difference in survival or progression-free survi-
val between the two histologic subtypes. Pediatr
Blood Cancer 2005;45:298–303.
ß 2005 Wiley-Liss, Inc.
Retrospective series have shown a probability of recur-
rence after 5 years of 50%, with the majority of treatment
failures occurring at the site of the primary disease [3–8].
Treatment consideration is based on surgical resection and
——————
1
Department of Radiation Oncology, Istanbul University-Istanbul
Medical Faculty, Capa Istanbul, Turkey
2
Department of Pediatric Oncology, Istanbul University-Institute of
Oncology, Capa Istanbul, Turkey
*Correspondence to: Fulya Yaman Agaoglu, Istanbul University-
Institute of Oncology, 34390, Capa Istanbul, Turkey.
E-mail: yamanf@isbank.net.tr
Received 22 July 2004; Accepted 9 August 2004

postoperative irradiation [4]. There are no data concerning
the role of routine, adjuvant chemotherapy (ChT) [1].
ChT indications are usually for recurrent tumors. For
very young children, <3 years, immediate postoperative
radiotherapy (RT) is not accepted, and multi-agent ChT
has been given in an effort to delay or avoid RT. The
management of ependymomas remains one of the most
controversial in pediatric oncology, and may differ
between institutions from surgery alone to a combination
of surgery, RT, and ChT. The absence of a large multi-
institutional randomized study makes the optimal man-
agement of ependymoma unclear.
In this study, demographic data and the treatment
results of pediatric patients with ependymal tumors,
treated in a single institute, is reported.
PATIENTS AND METHODS
Patient and Tumor Characteristics
Between 1989 and 2001, 40 patients below 16 years of
age, with histologically verified ependymal tumor were
referred to the Istanbul University-Institute of Oncology,
for further treatment following initial surgery. There were
22 males and 18 females, aged from 3 months to 15 years
old, with a median of 5.5 years old. Eighteen tumors
were located in the supratentorial region, and 18 in the
infratentorial, and 2 in both supra and infratentorial.
There were two tumors located in the spinal axis. Based
on the findings of clinical studies (CT, MRG, or CSF
cytology), dissemination in the cerebrospinal fluid was
positive in seven cases (17%) (Table I).
All patients had a histologic diagnosis of ependymoma.
Histologic subgroups were 18 ependymomas (45%) and
22 anaplastic ependymomas (55%). Ependymoblastomas
were not included in this series.
TABLE I. Patient Characteristics
Parameter Number of patients
Sex
Male 22
Female 18
Age
<3 years 8

3 years 32
Localization
Infratentorial 18
Supratentorial 18
Supra and infratentorial 2
Spinal 2 S
Histology
Ependymoma 18
Anaplastic ependymoma 22
Neuroaxis involvement
Ependymoma 1 S þ ChT
Anaplastic ependymoma 6
Ependymal Tumors in Childhood 299
Treatment Characteristics
Total tumor resection was performed in 20 patients
(50%), subtotal in 18 patients (45%), and biopsy only in
2 patients (5%). Postoperative treatment consisted of re-
gional (n ¼ 8) or craniospinal (CSI) (n ¼ 9) RT in patients
with ependymoma; regional (n ¼ 7) or CSI RTwith ChT in
patients with anaplastic ependymoma (n ¼ 14); ChT only
(n ¼ 1) in patient less than 3 years of age. One patient with
spinal ependymoma had no adjuvant treatment after
surgery and referred to our institute with recurrent mass
after 6 years (Table II).
RT given by Co60 and LINAC treatment machines.
The standard technique for posterior fossa irradiation
was parallel-opposed lateral fields and median total dose
to the primary site was 54 Gy (range, 45–60 Gy). CSI
irradiation of the entire subarachnoid volume of the
central nervous system, both intracranially and intra-
spinally. The primary site dose is the sum of the whole
brain dose (range, 30–40 Gy, median 36 Gy) and boost to
the primary site (range, 46–60 Gy, median 54 Gy) for
those patients who received CSI irradiation. A boost to
the cribriform plate region was administered in CSI
irradiation.
Between September 1989 and May 1991 patients
received regimen A, which consisted of RT followed by
eight-in-one ChT (methylprednisolone, vincristine, CCNU,
procarbazine, hydroxyurea, cisplatin, Ara-C, cyclopho-
sphamide), given every 4 weeks for eight courses. Eight
monthly cycles, commencing 4 weeks after completion
of RT.
Patients who were treated between June 1991 and July
1994, received regimen B, which included two courses of
postoperative ‘‘VEC’’ (vincristine, etoposide, cisplatin)
ChT , administered every 3 weeks, followed by RT applied
with low dose concomitant cisplatin used as a radio-
sensitizer. Patients with objective response to postopera-
tive ‘‘VEC’’ continued to have ‘‘VEC’’ after completion of
RT for six more courses. From August 1994 till date,
patients received regimen C, consisting of RT and con-
comitant infusion of cisplatin followed by ‘‘VCPCU’’
(vincristine, cyclophosphamide, procarbazine, lomustine)
administered every 4 weeks for eight courses.
TABLE II. Treatment Scheme
Treatment scheme n
1
S þ regional RT 8
S þ CSI 9
S þ regional RT þ ChT 7
S þ CSI þ ChT 14
1
S, surgery; RT, radiotherapy; CSI, craniospinal RT; ChT, chemotherapy.
300 Agaoglu et al.

Statistical Analysis and Follow-Up

The overall survival (OS) time was measured from the
date of pathological diagnosis. The disease-free survival
time was measured from the pathological diagnosis to
failure. The time of failure was determined as the date at
which recurrence or progression was confirmed by
radiology or biopsy. The Kaplan–Meier method was used
to estimate survival, and the log-rank test was used for the
statistical comparison of survival and local control rates.
Multivariate analysis was performed with the Cox pro-
portional hazards model. Radiologic follow-up included
cranial and spinal CT or MRG every 3 months for the first
year after treatment, every 6 months for the second and
third year, and then every 1 year. Median follow-up time
was 38 months (range, 4–189 months).

RESULTS
The 5-year OS rate was 64.9%, and 5-year disease-free
survival rate was 50.8% for the whole series (Figs. 1 and 2).
Median time for progression or relapse was 24.3 months
and there were 13 patients (32.5%) with relapse, and
6 patients (15%) with progression. For patients receiving
CSI RT, 8 of the 23 (34%) relapsed. Seven were in field
relapse, and one was distant relapse.
For patients with local RT, 3 of 15 (20%) failed. There
were two infield relapse, and one spinal axis relapse.
The patient with spinal axis recurrent had anaplastic
ependymoma at posterior fossa. The 5-year infield failure
rate was 13% (2:15) for those treated with local RT, and
30% (7:23) for those treated with CSI RT (P ¼ 0.17). The
predominant pattern of failure was local failure. Only one
patient failed in distant organ (cervical lymph node and
occipital scalp).
Fig. 2. Five-year event-free survival for whole series.
Table III summarizes the correlations among the
clinical variables and both the estimated cumulative
probability of death and event rates. Prognostic factors
for failure and survival were determined by univariate and
multivariate analysis. Of the factors associated with both
the overall and disease-free survival rates, leptomeningeal
dissemination or solid metastases at the time of diagnosis
were found to be significant (P ¼ 0.0008, and P ¼ 0.0002)
(Fig. 3). Patients older than 3 years of age had significantly
better outcome (P ¼ 0.04, 5-year OS rate was 75% vs.
TABLE III. Univariate Analysis of Risk Factors in Children With
Ependymoma
Cumulative
probability Cumulative
of death probability of event

Parameter n Died (%) P Relapsed (%) P

Age
<3 8 62 0.04 75 0.02

3 32 25 40
Sex
Male 22 36 0.34 54 0.26
Female 18 27 38
Location
Infratentorial 20 35 0.73 50 0.84
Supratentorial 20 30 45
Total resection
Yes 20 20 0.01 45 0.09
No 20 45 50
M stage
M0 33 18 0.0008 36 0.0002
M1–3 7 71 71
Histology
Ependymoma 18 16 0.09 27 0.04
Anaplastic 22 45 63
ependymoma
Fig. 1. Five-year overall survival for whole series.

Fig. 3. Five-year overall survival for metastatic and non-metastatic
patients.
38%) (Fig. 4). Neither sex, nor tumor location showed an
impact on treatment outcome. Patients with infratentorial
tumors did not have a significantly longer survival or
event-free survival compared with those with supraten-
torial tumors (P ¼ 0.73, and P ¼ 0.84). Histologically
anaplasia significantly influenced disease-free survival.
The 5-year disease-free survival rate was 37% for
anaplastic ependymomas versus 73% for ependymomas
(P ¼ 0.04). But OS was not significantly influenced by
histology (P ¼ 0.09). Anaplastic ependymomas was more
likely to occur in the supratentorial brain (P ¼ 0.03).
The disease-free survival rate by univariate analysis
was noted to be significantly improved (P ¼ 0.01) in
Fig. 4. Five-year overall survival for patients more than 3 years old
and others.
Ependymal Tumors in Childhood 301
Fig. 5. Five-year survival rate for totally resected patients and others.
patients undergoing gross-total resection at diagnosis
versus those treated with subtotal resection (Fig. 5). Six of
subtotally resected 20 patients had progression (30%).
Multivariate analysis was performed by logistic re-
gression (forward LR method). When clinical variables
were included in a multivariate analysis, non-metastatic
patients at the time of diagnosis (P ¼ 0.001), and patients
older than 3 years of age (P ¼ 0.03) were the factors asso-
ciated with the longest overall and disease-free survival.
DISCUSSION
The majority of complete responders were patients
who had total tumor removal. The completeness of the
surgical resection is the factor that has the greatest impact
on the outcome of children with ependymoma, regardless
of location [4–8]. The percentage of complete resection
differs markedly and nowadays approaches 50–60%.
In this series of patients, total resection rate was 50%.
We found that degree of surgical resection is statistically
significant factor for OS but not for event-free survival.
Five-year survival rate was 80% for totally resected
patients, and 45% for incompletely resected patients in our
series. In the retrospective series, the 5-year progression-
free survival rate for patients with negative postoperative
imaging was 51–100%, and for patients with residual
disease after surgery was 30–50% [4–8]. In the Italian
prospective AIEOP study, 5-year OS rate for totally
resected patients was 81%, and for incompletely resected
patients was 61% [9].
Currently, postoperative RT is considered to be the
standard approach of care for all children with ependy-
moma. There has never been a randomized trial comparing
surgery with or without RT although many retrospective
302 Agaoglu et al.
studies have shown a benefit for the adjuvant RT [4,7–11].
Since the use of postoperative irradiation, the survival rate
has improved from 20% to 60%. Salazar et al. reported
improved survival of intracranial ependymoma patients in
the 1970s and 1980s for those treated with CSI radiation
[10,11].The selection of treatment volume for ependymo-
mas has been an issue of considerable controversy. Several
retrospective studies have compared relapse rates with
local field, whole brain and CSI irradiation [8,12–14].
The incidence of neuroaxis dissemination is approxi-
mately 12% [12,13]. In the present series, seven patients
(17%) had positive cerebrospinal cytology. Local field RT
would be sufficient for non-disseminated (M0) ependy-
moma because the predominant site of failure was local,
and that cases of spinal relapse are usually accompanied
by relapse at the site of the primary. Paulina reported
that for non-disseminated, low-grade infratentorial epen-
dymoma, the RT volume need not include the entire
posterior fossa [14]. Metastatic relapses have been
reported in 3–15% of cases [7,8,13], and in this series
only one patient (2.5%) had distant metastasis without
local recurrence.
Dose response data for ependymoma are relatively rare
[1]. Reports confirm greater radiation efficacy with dose
levels up to 45–50 Gy [13,15], but the lack of multivariate
analysis in these series hinders any firm conclusion.
Garrett et al. [16], and Stuben et al. [17] reported a similar
dose-response; patients that received doses greater than
45 Gy had significantly improved local control and
survival rates. In this study the median total dose to the
primary site was 54 Gy (range, 45–60 Gy) and there was
no great dose inhomogenity between patients. We could
not perform dose-response analysis.
The histologic grade distribution is very heterogeneous
in the literature [7,8,15], and this series is containing
high percentage of anaplastic ependymoma (55%). The
influence of histologic grade on outcome is among the
most controversial of prognostic factors of childhood
ependymomas. The most recent reports, using the WHO
or the modified WHO classification, failed to establish a
link between anaplasia and poorer outcome. Two factors
may explain this failure: the lack of consensus in clas-
sifying ependymomas, and perhaps the positive impact
of treatment for patients with ependymomas [15]. Manage-
ment often differs according to tumor grade, such that
anaplastic tumors receive CSI RT while local fields treat
benign ependymomas. In the review by Vanuytsel and
Brada [18], tumor grade and location were found to be
factors that may influence the risk of spinal seeding, but
there was no evidence that failure in the neuraxis was
prevented by CSI irradiation. Our study shows that low-
grade tumors had a better disease-free survival but not OS.
There is currently no evidence that more aggressive RT
(i.e., CSI RT) benefits children with anaplastic tumors
[13,19–21].
The role of adjuvant ChT has not been defined. The
German Prospective Trial HIT 88/89 and HIT 91evaluated
combined postoperative irradiation and ChT for anaplastic
ependymomas and there was not effect on survival [22].
The Children’s Cancer Group (CCG) study for adjuvant
ChT of childhood posterior fossa ependymoma did not
show any benefit of adjuvant ChT [23]. However, Needle
et al. reported in 1997 a survival benefit for adjuvant ChT
with hyperfractionated RT [24]. In our institute, adjuvant
ChT is not a routine therapy, but patients with metastatic
disease and children less than 3 years old, and patients with
recurrent or progressive disease are given ChT protocol.
The prognostic influence of sex has been analyzed in
our study, but we did not find any significant difference.
Some authors have reported a worse prognosis in male
patients [4,23], in contrast Zorlu et al. [25] and Timmer-
man et al. [22] did not show any significant difference, as
confirmed by our results.
Ependymomas are most common in young children,
and half of the cases occur before age 5 [1]. The median
age at diagnosis was 5.5 years in our patient group;
younger children (less than 3 years old) have less favorable
outcome. In our study, patients below 3 years were 20%,
and those below 5 years were 40%. Most series have
reported that younger children had a lower survival than
older patients [5,8,15,17]; however, some authors have
reported no age-related effect on survival [11,26]. Treat-
ment failure occurred mainly within the first 2 years, and
outcome was dismal for patients who relapsed or had
progressive disease. The predominant site of failure is
tumor bed, so surgery and diagnostic work up is very
important for ependymal tumors. Dose intensification by
three-dimensional conformal RT or hyperfractionated
RT with radiosensitizers might be better local control
strategies for residual tumor.
REFERENCES
1. Halperin EC, Constine LS, Tarbell NJ, et al. Pediatric radiation
oncology. 3rd edition. Lippincott: Williams & Wilkins. 1999.
2. Sutton LN, Goldwein J, Perigolongo G, et al. Prognostic fac-
tors in childhood ependymomas. Pediatr Neurosurg 1990;16:
57–65.
3. Carrie C, Mottolese C, Bouffet E, et al. Non-metastatic childhood
ependymomas. Radiother Oncol 1995;36:101–106.
4. Vanuytsel LJ, Bessell EM, Ashley SE, et al. Intracranial ependy-
moma: Long term results of a policy of surgery and radiotherapy.
Int J Radiat Oncol Biol Phys 1992;23:313–319.
5. Nazar GB, Hoffman HJ, Becker LE, et al. Infratentorial
ependymomas in childhood: Prognostic factors and treatment.
J Neurosurg 1990;72:408–417.
6. Healey E, Barnes P, Kupsky W, et al. The prognostic significance of
postoperative residual tumor in ependymoma. Neurosurgery
1991;28:666–672.
7. Perilongo G, Massimino M, Sotti G, et al. Analyses of prognostic
factors in a retrospective review of 92 children with ependymoma:
Italian Pediatric Neuro-Oncology Group. Med Pediatr Oncol
1997;29:79–85.

8. Rousseau P, Habrand JL, Sarrazin D, et al. Treatment of intracranial
ependymomas of children: Review of a 15 year experience. Int J
Radiat Oncol Biol Phys 1993;23:313–319.
9. Massimino M, Gandola L, Giangaspero F, et al. Hyperfractionated
radiotherapy and chemotherapy for childhood ependymoma: Final
results of the first prospective AEIOP (Associazone Italiana di
Ematologia-Oncologia Pediatrica) Study. Int J Radiat Oncol Biol
Phys 2004;58:1336–1345.
10. Salazar OM, Rubin P, Bassano D, et al. Improved survival of
patients with intracranial ependymomas by irradiation: Dose
selection and field extension. Cancer 1975;35:1563–1573.
11. Salazar OM, Castro-Vita H, Van Houtte P, et al. Improved survival
in cases of intracranial ependymoma after radiation therapy:
Late report and recommendations. J Neurosurg 1983;59:652–
659.
12. Rezai AR, Woo HH, Lee M, et al. Disseminated ependymomas of
the central nervous system. J Neurosurg 1996;85:618–624.
13. Goldwein JW, Corn BW, Finlay JL, et al. Is craniospinal irradiation
required to cure children with malignant (anaplastic) intracranial
ependymomas? Cancer 1991;67:2766–2771.
14. Paulino AC. The local field in infratentorial ependymoma: Does
the entire posterior fossa need to be treated? Int J Radiat Oncol
Biol Phys 2001;49:757–761.
15. Bouffet E, Perilongo G, Canete A, et al. Intracranial ependymomas
in children: A critical review of prognostic factors and a plea for
cooperation. Med Pediatr Oncol 1998;30:319–331.
16. Garrett PG, Simpson WJ. Ependymomas: Results of radiation
treatment. Int J Radiat Oncol Biol Phys 1983;9:1121–1124.
17. Stuben G, Stuschke M, Kroll M, et al. Postoperative radiotherapy
of spinal and intracranial ependymomas: Analysis of prognostic
factors. Radiother Oncol 1997;45:3–10.
Ependymal Tumors in Childhood 303
18. Vanuytsel L, Brada M. The role of prophylactic spinal irradiation in
localized intracranial ependymoma. Int J Radiat Oncol Biol Phys
1991;21:825–830.
19. Merchant TE, Haida T, Wang MH, et al. Anaplastic ependymoma:
Treatment of pediatric patients with or without craniospinal
radiotherapy. J Neurosurg 1997;86:943–949.
20. Gerszten PC, Pollack IF, Martinez AJ, et al. Intracranial ependy-
moma of childhood. Lack of correlation of histopathology and
clinical outcome. Pathol Res Pract 1996;192:515–522.
21. Ross GW, Rubinstein LJ. Lack of histopathological correlation of
malignant ependymomas with postoperative survival. J Neurosurg
1989;70:31–36.
22. Timmermann B, Kortmann RD, Kühl J, et al. Combined post-
operative irradiation and chemotherapy for anaplastic ependy-
momas in childhood: Results of the German Prospective Trials
HIT 88/89 and HIT 91. Int J Radiat Oncol Biol Phys 2000;46:287–
295.
23. Evans AE, Anderson JR, Bodreaux HBL, et al. Adjuvant chemo-
therapy of childhood posterior fossa ependymoma: Cranio-spinal
irradiation with or without adjuvant CCNU, vincristine, and
prednisone: A Children’s Cancer Group Study. Med Ped Oncol
1996;27:8–14.
24. Needle MN, Goldwein JW, Grass J, et al. Adjuvant chemotherapy
for the treatment of intracranial ependymoma of childhood. Cancer
1997;80:341–347.
25. Zorlu AF, Atahan IL, Akyol FH, et al. Intracranial ependymomas:
Treatment results and prognostic factors. Radiat Med 1994;12:
269–272.
26. Foreman NK, Love S, Thorne R. Intracranial ependymomas:
Analysis of prognostic factors in a population-based series. Pediatr
Neurosurg 1996;24:119–125.