1589 / 1708
The median single dose was 1.6 Gy (range, 1.5–1.6 Gy).
The median boost dose was 20.0 Gy with a median single
dose of 2.0 Gy (range, 1.5–2.0 Gy). The cumulative median
dose to the posterior fossa was 55.2 Gy (range, 42–56 Gy);
10% of the children had a cumulative dose of less than 50
Gy; 2.5% of the children had a dose of more than 55.2 Gy.
Eleven children with supratentorial primary tumors un-
derwent craniospinal irradiation because the tumor had in-
filtrated the ventricular system, and 1 child underwent this
treatment because of dissemination of the disease. The
parents of 2 children refused irradiation. Both children had
localized supratentorial primary tumors and underwent
complete resection plus sandwich chemotherapy, according
to the design of the HIT 91 trial.
Table 3 shows patient characteristics according to the
volume of radiotherapy.
Survival
Follow-up ranged from 5 to 106 months (median, 38
months). The 3-year-estimated overall survival rate and
progression-free survival rate were 75.6% and 59.7%, re-
spectively (Fig. 2). The pathologic findings were not eval-
uated by the review committee in 20 patients included in the
analysis. We estimated the overall and progression-free
survival rates in the reviewed children separately and found
no difference (76% and 62%, respectively). Sixteen (29%)
children died of recurrent disease. One child died of severe
septic complications after salvage surgery. No other deaths
occurred.
Patterns of failure
At the time of the last follow-up, 30 children were free of
disease, and 25 patients showed progression. Disease re-
curred at the primary tumor site only in 20 patients (36.4%
of all patients, 80% of failures). Disease disseminated
within the CNS in 3 (5.5%) children: 1 case was meningeal
and intracranial, 1 case was intracranial, and 1 case was in
the thoracic spinal canal. Two children (3.6%) suffered
from combined distant (1, intracranial; 1, not specified) and
local failure.
Impact of clinical variables on outcome
Table 4 summarizes the correlations between the clinical
variables and both the estimated overall and progression-
free survival rates. Of the factors associated with both the
overall and progression-free survival rates, leptomeningeal
dissemination or solid metastases at the time of diagnosis
were found to be significant. All 5 children with positive
CSF cytology findings (
n
5
4) or spinal metastases (
n
5
2)
died within 2 years of surgery. In contrast, children with
localized tumors achieved a progression-free survival rate
of 65.8% at 3 years (Fig. 3). Neither age, sex, or tumor site
showed an impact on treatment outcome. The relationship
Fig. 2. Kaplan-Meier plots of the estimated overall and progres-
sion-free survival rates of 55 children.
Table 3. Patient characteristics according to treatment volume
Variable
CSI
(
n
5
40)
Local irradiation
(
n
5
13)
Median age (range) 7.2 yr (3.6–8.8 yr) 5.3 yr (3.0–16.6 yr)
Tumor site
Infratentorial
28
1
Supratentorial
12
12
Resection
Complete
20
6
Incomplete
20
7
M-Stage
M0
35
13
M1–3
5
0
CSI
5
craniospinal irradiation. yr
5
years.
Table 4. Univariate analyses of the correlation between selected
parameters and estimated progression-free survival rate in 55
children with anaplastic ependymomas
Parameter
No. of
patients
(
n
5
55)
3-yr
Progression-
free survival
rate (%)
95% CI
p
Value
Age (years)
1–6
26
66.2 46.8–85.6 0.63
.
6
29
58.4 37.6–79.2
Sex
Male
28
67.5 42.9–80.9 0.35
Female
27
57.1 36.1–78.1
Tumor site
Infratentorial
29
53.1 33.1–73.1 0.71
Supratentorial
26
72.4 52.8–92.0
Metastases
Yes
5
0
0–35.1 0.0001*
No
50
69.0 51.3–80.3
Resection
Complete
28
83.3 68.2–98.4 0.0043*
Incomplete
27
38.5 16.2–60.8
Chemotherapy
Maintenance
15
61.9 35.6–88.2 0.25
Sandwich
40
63.7 47.6–79.8
Treatment volume
CSI
40
56.3 39.8–72.8 0.44
Involved field
13
92.3 77.8–100
CSI
5
Irradiation of craniospinal axis.
* Significant.
290
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 46, Number 2, 2000