paediatrics Brussels 17

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

291

Fig. 5. Relationship between the extent of resection and the esti- mated progression-free survival rate. compl. resec, complete re- section; incompl. resec., incomplete resection.

Fig. 3. Relationship between initial dissemination and the esti- mated progression-free survival rate.

between the tumor site and the progression-free survival rate is shown in Fig. 4.

distribution is shown in Fig. 6. Because of very uniformly administered radiotherapy, it is difficult to draw conclu- sions; however, we did not find an impact of fraction size or total dose on the survival rate. DISCUSSION Ependymomas account for 3% to 4% of childhood can- cers (1). There is little information on the outcome of different treatments for ependymomas and still no consen- sus on the optimal therapy. Most studies have investigated low- and high-grade ependymomas, despite several reports about a worse outcome in patients with anaplastic ependy- momas (12–15). Only WHO grade III ependymomas were included in the two German brain tumor trials described here. The outcome in patients with ependymomas remains suboptimal, although the survival rates have increased from 24% (16) to 60% and 70% (12, 17). Regardless of the therapy administered in the patients in the present study, those with unfavorable factors, such as incomplete resection and tumor dissemination, had a poor outcome, with progres- sion-free survival at 3 years of 38% and 0%, respectively. Disease recurrence at the primary site is still the main obstacle to cure, occurring in 88% of all cases of progres- sion in our study. Similar rates have been observed in other

Impact of treatment variables on outcome The treatment-related variables associated with progres- sion-free survival are also summarized in Table 4. The patients with macroscopically complete resection ( n 5 28) fared significantly better, with an estimated overall survival rate of 91.5% at 3 years, than those who underwent incom- plete resection ( n 5 27), with an estimated overall survival rate of 56.1% ( p 5 0.046). The estimated progression-free survival was also significantly better for children with com- pletely resected tumors (Fig. 5). The maintenance chemotherapy or sandwich chemother- apy did not alter the prognosis. Specifically, children who were treated for disease in the neuraxis with an additional boost to the tumor site showed no difference in outcome compared with the children who were treated with irradia- tion at the tumor site only. Of the children who did not receive any radiotherapy, 1 is alive after 5 years (her tumor specimen was not reviewed) and 1 died of local and distant disease progression after 1.5 years. The distribution of risk factors in patients with supraten- torial tumors given radiotherapy to the craniospinal axis or the tumor region is shown in Table 5, and the survival rate

Table 5. Characteristics of 24 supratentorial ependymomas according to treatment volume*

Variable

CSI

Local field

Total number* of tumors

12

12

Resection

Complete Incomplete

4 8

6 6

M stage M0

11

12

M1–3

1

0

* Of the 26 children with supratentorial ependymomas, 2 were not irradiated.

Fig. 4. Relationship between tumor site and estimated progression- free survival rate.