paediatrics Brussels 17

.0031), tumor grade (differentiated versus anaplastic;

90.3% 4.6%

43.7% 12.4%;

.0001), and history of

preirradiation chemotherapy (no chemotherapy

chemo-

therapy, 78.1% 6.0%

60.0% 14.3%;

.0446).There

was no difference in PFS estimates between patients older

than 3 years and those younger at the time of irradiation

(80.8% 7.2%

69.5% 8.6%;

.23) or between those

with infratentorial tumors and those with supratentorial

tumors (74.9% 6.3%

71.4% 13.5%;

.86). PFS

estimates were not influenced by the intervals between the

time of symptom appearance and diagnosis, the interval

between diagnosis and the start of CRT, and the number

of elapsed treatment days. High tumor grade (

.0001)

and less than gross-total resection (

.001) negatively

affected outcome and the hazard ratio for PFS in a mul-

tivariate analysis.

Neurocognitive Effects

The patients underwent a total of 316 neurocognitive

examinations to evaluate changes in intelligence quotient

(IQ), memory, academic achievement, adaptive behavior,

and visual-auditory learning. There was no statistically sig-

nificant change in the measures of these features for patients

who completed evaluation 24 months after the initiation of

CRT (more than half of the cohort; Figs 2 through 4).

However, patients younger than 3 years at the time of CRT

had a significantly lower mean IQ at the start of CRT than

did patients older than 3 years (89.7 2.8

98.7 3.1;

.034), but the IQ of those younger than 3 years im-

proved over time. There was no statistically significant dif-

ference in IQ scores for patients comparing infratentorial

and supratentorial tumor location. At the most recent

follow-up, mean scores on all neurocognitive outcomes

were within normal limits (ie, no more than 10 points

from the normative mean for the appropriate age group).

DISCUSSION

The purpose of this study was to test the hypothesis that

irradiation of a smaller-than-conventional treatment vol-

ume reduces side effects without affecting the rate of tumor

control or local pattern of failure. The results of this study

demonstrated a 3-year PFS estimate of 74.7% 5.7% for

patients with ependymoma treated with CRT using an an-

atomically confined CTV whose 10-mm margin sur-

rounded the postoperative tumor bed. The rate of failure in

the study is less than those of other studies, which have

yielded 2- to 5-year PFS estimates of only 50% to 67%.

25-30

Of the 20 patients who experienced recurrence or

progression, none had marginal failures; however, the

relatively large proportion of patients experiencing re-

lapse with disease in the neuraxis but not at the primary

site after treatment was both disappointing and informa-

tive. This proportion was higher than the expected pro-

portion, which is based on lower rates reported in some

series,

and may indicate that the overall pattern of

Fig 1.

Event-free survival estimates for patients treated with postoperative

conformal radiation therapy.

Fig 2.

Estimated mean intelligence quotient (IQ) before and after conformal

radiation therapy.

Fig 3.

Mean Wechsler Individual Achievement Test (WIAT) scores before

and after conformal radiation therapy. CVLT-C, California Verbal Learning

Test: Child; VAL, visual-auditory learning test.

Conformal RT for Pediatric Ependymoma

www.jco.org

3159