paediatrics Brussels 17

Anaplastic ependymomas in childhood ● B. T IMMERMANN et al.

289

Table 1. Characteristics of 55 children with anaplastic ependymomas treated in Germany and Austria, 1988–1997

Hospital, University of Wu¨rzburg, Wu¨rzburg, Germany. Additional data about radiotherapy were collected and mon- itored by the Department of Radiooncology, University of Tu¨bingen, Tu¨bingen, Germany. The follow-up period was defined as extending from the date of surgery to the date of last patient contact or last event. The length of survival was calculated from the date of surgery. Terminal events were defined as the date of death from any cause (overall surviv- al), the date of progression (progression-free survival), or the date of diagnosis, progression, or death (event-free survival). For all patients alive without events, the length of survival for the statistical analysis was considered the last date of the documented contact with the patient. Data for patients who died without evidence of progression were censored. The Kaplan-Meier method was used to estimate survival, and the log-rank test was used for the statistical comparison of survival estimates. We planned to perform a multivariate analysis (Cox regression) only if more than one of the potential prognostic factors showed a remarkable influence in the univariate analysis ( p , 0.1) and a minimum of 10% of the patients had one of these risk factors. All statistical analyses in this study were done for descriptive purposes. Data are presented with nominal two-tailed p values (unad- justed for multiple comparisons) and 95% confidence inter- vals. All analyses were carried out with the SAS system for Windows 6.1 software (SAS Institute, Cary, NC). Patient population Seventy-one children were treated for anaplastic ependymoma. Pathologic findings were reviewed in 51 (71.8%) children. Sixteen children were excluded from evaluation because the pathologic results of the review committee revealed low-grade ependymoma ( n 5 1), ependymoblastoma ( n 5 5), primitive neuroectodermal tumor ( n 5 2), astrocytoma ( n 5 1), medulloblastoma ( n 5 5), or glioblastoma ( n 5 2). Of 55 patients eligible for the study (27 females and 28 males; age range 3.0 – 16.6 years; median, 6.2 years), pathologic review was performed in 35 children. In 26 patients, the tumor site was supratentorial (47.3%); in 29 children, the tumor site was infratentorial (52.7%) (Table 1). In 20 patients, leptomeningeal dissemination was evalu- ated by CSF cytologic studies at presentation. Four patients had positive CSF cytology findings; in 35 children, no CSF findings were available, but the children underwent cranio- spinal MRI. Metastases to the central nervous system (CNS) were found in 2 children: 1 child had a spinal tumor in the lumbosacral region, and the other child had cerebral dis- semination as well as positive CSF findings. No metastases outside the CNS were found (Table 2). Treatment Surgery. All children underwent surgery. The extent of resection was assessed by postoperative CT/MRI and was RESULTS

No. of patients

Characteristic

Percentage

Median age (range): 6.2 years (3–16) Sex Male

28 27 29 26

50.9 49.1 52.7 47.3

Female

Site

Infratentorial Supratentorial

Metastases M0

50

91.0

M1

3 2

5.4 3.6

M2/3

Surgery

Incomplete resection Complete resection

27 28 18 37 15 40 40 13

49.1 50.9 32.7 67.3 27.3 72.7 72.7 23.6

HIT

88/89

91

Chemotherapy Maintenance

Sandwich

Radiotherapy* CSI

Local field

CSI 5 craniospinal irradiation. * 2 children received no irradiation.

considered macroscopically complete in 28 children and incomplete in 27. Chemotherapy. Eighteen children in HIT 88/89 and 37 children in HIT 91 were treated. In the HIT 91 trial, 81% of the patients were randomized; the parents of 7 children refused randomization. All patients received adjuvant che- motherapy; maintenance chemotherapy was performed in 15 children and sandwich chemotherapy in 40 children. Radiotherapy. Thirteen children were irradiated at the primary tumor site only. The median total dose was 54.0 Gy (1 patient received a dose of 55.8 Gy). The median dose per fraction was 2.0 Gy (range, 1.8–2.0 Gy). Forty children received craniospinal irradiation with an additional boost to the posterior fossa. The median total dose to the neuraxis was 35.2 Gy (range, 24.0–39.6 Gy); 15.0% of the patients received a dose of less than 30.0 Gy.

Table 2. Pattern of dissemination at time of diagnosis in 5 children

Age (years), sex

Primary tumor site Infratentorial Supratentorial Infratentorial

CSF positive

Distant metastases

M stage

3.4, F 5.3, F 4.0, F

Yes Yes Yes Yes

No No No

M1 M1 M1 M2

7.9, M Infratentorial 8.7, M Infratentorial

Cerebral

Unknown Lumbosacral M3

CSF 5 cytologic analysis of cerebrospinal fluid.