1495 / 1708
both location (
P
¼
0.020) and age (
P
<
0.001) remained as
significant predictors of survival among pediatric ependymomas,
while radiotherapy did not retain significance. There was no
interaction between age and location.
DISCUSSION
Our study draws a large sample of 635 patients with
ependymomas from a population-based cancer registry. Both age
and location directly influence survival in children, consistent with
prior research [1–3,8–10,12].
Although the current therapy for ependymomas in children often
includes surgery followed by radiotherapy, variations in treatment
plans using or not using radiotherapy have occurred because of
concerns about neurotoxicity in young children [11]. Chemotherapy
may be provided to delay the radiation therapy until older age, and
sometimes children did not receive any irradiation for infratentorial
ependymomas [14,15]. Specifically, Grundy et al. [16] found that
among children less than 3 years without metastases from
ependymoma, chemotherapy alone following maximal surgery
provided an effective 42% 5-year cumulative incidence of freedom
from radiotherapy. Duffner et al. [15], however, recorded excessive
relapses in infants with an approach to defer radiotherapy by
chemotherapy alone longer than 12 months. Furthermore, Merchant
et al. reported that conformal irradiation in children achieves 74.7%
progression free survival at 3 years from diagnosis. Among patients
less than age 3 at time of irradiation, 3-year progression free survival
was not significantly different at 69.5%with normal neurocognitive
outcome scores [17]. In our study, radiotherapy was associated with
improved outcomes, consistent with recent work by Shu et al. [8]
that analyzed the treatment of 61 patients and found improved
survival among children with higher radiation dose. However,
following stratification, our study demonstrated a survival benefit of
Pediatr Blood Cancer
DOI 10.1002/pbc
TABLE III. Five-Year Univariate Survival Estimates by Primary Tumor Site
Primary tumor site
n
5-year survival (%)
SE (%)
95% confidence interval
Supratentorial
106
57.8
5.5
(47.0, 68.6)
Infratentorial
193
54.4
4.2
(46.1, 62.6)
Spinal
55
86.6
5.2
(76.4, 96.8)
n, number of children; SE, standard error.
Patients with ICD-0-3 histology codes 9391-
9494
n = 635
Patients Included
n = 354
Supratentorial
700, 702-714
Infratentorial
716-717
Spinal
701, 720-721
Undetermined
715,718-719, 728-729
Patients Excluded
n = 281
Primary tumor site
ICD-0-2 site codes
Fig. 1.
Patient selection for primary tumor site analysis. n, number of children.
400.00
300.00
200.00
100.00
0.00
Total survival time (months)
0.0
0.2
0.4
0.6
0.8
1.0
Cumulative survival
Supratentorial
Infratentorial
Spinal
Fig. 2.
Survival by primary tumor site. Total number of
patients
¼
354. Number of patients in each arm: supratentorial
¼
106,
infratentorial
¼
193, spinal
¼
55. Logrank Test,
P
¼
0.001.
Ependymoma Survival: A SEER Study
67