paediatrics Brussels 17

failure changed as a result of the high rate of local tumor

control and gross-total resection.

The improved rate of disease control in this study may

be attributable to factors that include the high proportion of

cases in which gross-total resection was done, systematic

targeting with three-dimensional imaging, and the rela-

tively high prescribed total dose. Gross-total resection was

performed in 84% of cases, near-total resection in 7%, and

subtotal resection in 9%; the average volume of residual

disease was only 1.2 cm

. The percentage of cases in

which gross-total resection was conducted in this study

was higher than the national average, which has ranged

from 40% to 60%.

32,33

We evaluated CNS effects in a rigorous, consistent

manner, using widely accepted tests to identify the effects of

radiation on cognitive, endocrine, and neurologic function.

The most encouraging finding from this study was the level

of function and lack of treatment-related effects in a young

and vulnerable group of children treated with high-dose

irradiation. Only a limited comparison of neurocognitive

effects can be made between patients from this study and

those treated conventionally, because prospective data from

a similarly well-characterized group of pediatric patients

with ependymoma are not available. After correcting for

other factors responsible for neurocognitive function in

pediatric patients with CNS tumors, other investigators

found that the dose and volume of irradiation seem to play

a role in altering neurocognitive status or intellectual out-

come. In a study that included 59 pediatric patients with

medulloblastoma and 37 with posterior fossa ependymoma

(including 14 patients younger than 3 years at the time

ependymoma was diagnosed), 90% of those with ependy-

moma, which was treated with irradiation to the posterior

fossa, maintained an IQ greater than 90 at 5 to 10 years after

treatment.

In the group with medulloblastoma, which

was treated with craniospinal irradiation and a boost to the

posterior fossa, only 20% of patients had an IQ greater than

90 at 5 years, and the proportion decreased to 10% at 10

years. In a separate publication, a review of multiple studies

compared the IQ of pediatric patients treated postopera-

tively with craniospinal irradiation, focal irradiation of the

primary site, or no irradiation.

Patients who received

craniospinal irradiation had significantly lower IQs than

those who did not receive such treatment; however, those

treated with focal irradiation had IQ values comparable to

those who received no irradiation. These results support

efforts to reduce the volume of irradiation. Much of the fear

instilled in those who treat young children with brain tu-

mors may be derived from reports about children with

medulloblastoma for whom a persistent and early decline in

intellectual outcome is anticipated after craniospinal irradi-

ation.

Perhaps the most direct comparison of the present

neurocognitive outcomes can be made with the results from

the study of Grill et al,

who reported a mean IQ of 85.3

(standard deviation,

13.6) for 12 long-term survivors of

ependymoma treated with conventional posterior fossa ir-

radiation at age 5 years. Spiegler et al

recently reported

on four patients with ependymoma and 30 with medullo-

blastoma in a study that was meant to show change over

time and the onset of stability for IQ measured after radia-

tion therapy. Because of the small number of serial evalua-

tions, their modeling was limited to 17 patients evaluated

within 6 months from diagnosis and followed for a median

of 3.3 years. They found that patients evaluated early in their

treatment course experienced a steep decline with eventual

leveling in the pattern of a quadratic function.

Our study is unique because it includes children

younger than 3 years at the time of irradiation. The age of 3

years has been used to define those who are at greatest risk of

the effects of irradiation and for whom trials have been

designed in an effort to delay or avoid irradiation. Age at the

time of diagnosis has also been described as an important

prognostic factor. In the present study, 13 of the 48 patients

younger than 3 years experienced disease progression. Chil-

dren in this age group in earlier studies had a worse prog-

nosis than older patients, possibly because of more

aggressive tumor biology, reluctance to give postoperative

radiation therapy, or use of lower doses of radiation.

29,39, 40

The first infant study by the Pediatric Oncology Group

attempted to delay radiation therapy by using postoperative

chemotherapy and showed a significant difference in out-

come based on age.

3,41

The 5-year PFS estimate was

12.7% 8% for the 31 patients between the ages of 0 and 23

months treated with chemotherapy for 2 years, whereas the

17 patients who were 24 to 36 months old treated with

chemotherapy for 1 year had an estimate of 54.8% 15%.

The age-related differences remained even when the analy-

sis was limited to those without metastases who had under-

gone gross-total resection: the 5-year PFS estimates were

Fig 4.

Mean Vineland Adaptive Behavior Scale scores before and after

conformal radiation therapy.

Merchant et al

3160

OURNAL OF

LINICAL

NCOLOGY