paediatrics Brussels 17

volume at risk of recurrence while sparing normal tissues.

These techniques incorporate three-dimensional imaging

(computed tomography and magnetic resonance imaging)

into the planning process and use sophisticated software to

delineate and display the treatment volume and important

normal tissue structures for selective targeting and optimi-

zation of dosimetry. Initially developed for the treatment of

adults with prostate and head and neck cancer, CRT has

been successful in reducing side effects and improving tu-

mor control

6,7

and holds the promise of reducing radiation-

related treatment effects in children with brain tumors,

although no long-term clinical trials have yet been reported.

The successful application of CRT to ependymoma in chil-

dren may improve outcomes by reducing side effects and

thereby permit the reintroduction of radiation therapy as a

treatment option for very young children. Nevertheless,

guidelines for the use of CRT will be needed to ensure that

the appropriate volume receives the prescription dose and

that disease control is not compromised.

We designed and conducted a phase II trial to test the

hypothesis that irradiation of a smaller-than-conventional

treatment volume reduces side effects without affecting the

rate of tumor control or local pattern of failure. We selected

an anatomically confined margin of 10 mm around the

tumor, tumor bed, or both as the clinical target volume for

a prospective phase II trial. These guidelines were used

prospectively to treat 88 pediatric patients with ependy-

moma, the largest trial to date for such patients. The neu-

rocognitive function of these patients was evaluated before

and after CRT in a comprehensivemanner that has not been

previously reported in the literature.

PATIENTS AND METHODS

Patients

From July 1997 through January 2003, 88 pediatric patients

with intracranial ependymoma were enrolled onto a phase II trial

approved by the institutional review board. Criteria for enroll-

ment included age between 1 and 21 years at the time of irradia-

tion; histologic confirmation of intracranial ependymoma; no

evidence of dissemination; no prior irradiation; no ongoing che-

motherapy; adequate performance status; and written informed

consent. The protocol was amended to allow enrollment of pa-

tients as old as 25 years; only one patient older than 21 years

entered the study. Sixteen patients received chemotherapy be-

fore irradiation; most received multiagent chemotherapy in-

cluding cyclophosphamide, cisplatin or carboplatin, etoposide,

and vincristine.

Extent of Resection Definitions

All patients underwent resection before radiation therapy.

Gross-total resection was defined as resection after which the only

tumor cells that remained were visible with the use of the operat-

ing microscope; patients for whom this type of resection was

achieved had no evidence of disease on postoperative neuroimag-

ing. Near-total resection was defined as resection after which only

residual tumor 5-mm thick was visible on postoperative neuro-

imaging. Subtotal resection was defined as resection that left

behind residual tumor 5-mm thick on postoperative neuroim-

aging. Further surgery was systematically applied to maximize the

extent of resection before irradiation.

CRT

Target volume definitions and planning and treatment pa-

rameters have been previously described.

The gross tumor vol-

ume (GTV) contained the tumor bed, residual tumor, or both.

The clinical target volume (CTV) contained the GTV with an

added margin of 10 mm, which was included so that subclinical

microscopic disease beyond the GTV could be treated. The CTV

was anatomically confined; that is, it was limited by normal tissue

structures through which tumor extension was unlikely. The plan-

ning target volume included the CTV surrounded by an additional

margin of 3 to 5 mm expanded in three dimensions to account for

uncertainty in patient positioning and image registration. Con-

ventional fractionation (1.8 Gy per day) was used to treat all

patients, and the prescribed dose was 59.4 Gy. Exceptions included

children younger than 18 months and three children older than 18

months who received 54.0 Gy after gross-total resection.

Neurocognitive Testing

Neurocognitive testing was performed before (baseline) and

6, 12, 24, 36, 48, and 60 months after the start of CRT. Baseline

testing was delayed slightly if the institution of CRT was given

logistic priority. Age-appropriate tests included those for cognitive

development (Bayley Scale of Infant Development–II,

Wechsler

Preschool and Primary Scales of Intelligence–Revised,

Wechsler

Intelligence Scale for Children–III,

and Wechsler Adult Intelli-

gence Scale III

), verbal memory and recall (California Verbal

Learning Test: Child and Adult versions

13,14

), academic achieve-

ment (Wechsler Individual Achievement Test,

) activities of daily

living (Vineland Adaptive Behavior Scale Survey

), and visual-

auditory paired associate learning (Visual-Auditory Learning

Test

). All tests have well-documented reliability and validity and

result in age-corrected standard scores. The testing regimen was

based on patient age. When more than one instrument was age-

appropriate for a patient, the selection of instrument reflected the

desire to maintain consistency in the use of consecutive instru-

ments and to conduct at least two evaluations by using the same

instrument. Clinical judgment was used to select the instru-

ment with which the child was expected to achieve the most

valid performance.

Statistical Methods

Progression-free survival (PFS) was measured from the initi-

ation of radiation therapy to the neuroimaging documented time

of tumor recurrence, where tumor recurrence included local-only

failure, distant-only failure, or simultaneous local and distant

failure. One patient who died of reasons unrelated to radiation

therapy was censored at the date of death in the estimate of PFS.

The rate of local failure was monitored by using group sequential

boundaries obtained according to the sequential conditional

probability ratio tests.

Patients were followed every 3 months for

2 years and every 6 months thereafter. PFS was estimated by using

the Kaplan-Meier method.

Estimates based on categoric clinical

variables were compared by using the log-rank test,

and esti-

mates based on continuous clinical variables were compared by

using Cox regression analysis.

Variables for which

.10 in the

univariate analysis were included in a Cox regression model.

Local control was estimated using cumulative incidence meth-

ods,

with distant tumor recurrence and death (one patient

Conformal RT for Pediatric Ependymoma

www.jco.org

3157