paediatrics Brussels 17

B. Timmermann et al. / Radiotherapy and Oncology 77 (2005) 278–285

283

[4,31,40,41,48] . The irradiation volume encompassed the whole neuraxis in most of the children, which led to significant late sequelae [16,17,19,22,23] . Later trials attempted to delay or omit radiation therapy in infants. The Baby-POG I trial treated 48 children with intracranial ependymoma under the age of 3 years. First analysis resulted in promising 3-year overall survival rates of 61.8% [9] . These findings did not persist. The children developed late recurrences and new analysis revealed the benefit of not delaying radiotherapy for any longer than 1 year, with 5-year overall survival rates of 63.3%. When radiotherapy was delayed for 2 years, 5-year overall survival rate was only 25.7% [8] . They concluded that delay of radiotherapy of more than 1 year adversely affected survival. Several trials tried to delay or avoid radiotherapy: the CCG trial by administering postoperatively the ‘eight-in-one’ regimen, the M.D. Anderson Cancer Center study MOPP chemother- apy, and the Australia–New Zealand study by giving Vincristine, etoposide and intensive cyclophosphamide. But all of these studies included only 5–15 children and could not lead to strong conclusions regarding radiotherapy [1,12,49] . In our trial, the 3-year overall survival rate of 21 children who were irradiated was 66.7%, compared to only 38.5% for those receiving no radiotherapy. We failed to answer the question if radiotherapy can be delayed until recurrence. However, only two of nine children receiving radiotherapy after developing progression survived without any further progression. Nowadays, the irradiation of intracranial ependymoma without dissemination has changed signifi- cantly. Most of the brain tumor groups recommend encompassing the tumor bed only, because as with our findings, no benefit could be detected of irradiating the whole CNS of patients with non-disseminated ependymoma [2,40] . This reduction of irradiated central nervous tissue will lead to reduction of the risk of late sequelae. The efficacy of chemotherapy is still unclear and the predomi- nance of local failures indicates the need for local treatment approaches. Therefore, we feel that the avoidance or long delay of local radiotherapy even in young children with ependymoma is not yet justified. This is in accordance with the current Children’s Oncology Group trial (COG ACNS0121) implementing RT even for very young children with localized ependymoma [30] . Regarding the optimal doses for radio- therapy, we cannot draw any conclusion from our analysis because of the small group and inhomogeneous treatments. Retrospective series indicate that total doses greater than 45 Gy must be delivered to the tumor site [13,48] . Merchant et al. studied anaplastic ependymoma and found a positive influence of increasing the local dose [29] . There is consensus that local relapse is the major cause of failure in ependymoma [3,18,29,33,34,36,44,47] . In accord- ance in our analysis, all children failed locally and, in 76% of the cases, the site of failure was solely local. Age at diagnosis was found to be an important prognostic factor for ependymomas. In previous studies, children below 3–6 years had a lower survival rate than older patients [14,33,38] . There is little data concerning the impact of age in the subgroup of infants and babies. Duffner et al. revealed a significant difference between the two age groups (A: 0–24 months vs. B: 24–36 months) with 5-year survival rates of 25.7% compared to 63.3%, but the younger children had a

radiotherapy [12] . Six of the total 15 children had anaplastic histology and no further specific clinical parameters were listed separately for ependymoma patients. In the series of White et al., 5/14 children younger than 4 years of age with ependymomas survived after receiving a VETOPEC-based early chemotherapy [49] . All patients had M0 stage though, and information about the proportions of anaplastic tumors and irradiated patients is missing in the report. Radiotherapy was reserved for relapse in the report of Ater et al., who administered MOPP-Chemotherapy to infants less than 3 years of age [1] . Five ependymomas were included, two survived; one of them received salvage radiotherapy. Local control is the most important aspect of treatment in ependymomas; most treatment failures occur locally. Several trials have shown that complete surgery is a strong prognostic factor in these tumors [3,18,29,33,34,36,44,47] . In Children’s Cancer Group (CCG) Protocol 921, patients with gross total resection had a 5-year progression free survival of 66%, compared to 11% for those with residual disease [39] . In the analysis of the German HIT trials for children with anaplastic ependymoma above 3 years of age, a 3-year PFS of 83.3% could be achieved after complete resection, compared to only 38.5% after incomplete surgery [44] . In our analysis, babies and infants receiving complete surgery also show an advantage, with 3-year PFS of 41.2%, compared to only 12.5% after incomplete resection. Still, the importance of the operative procedure is very clear and there have been groups reporting successful treatment of intracranial ependymoma with surgery alone [21] . Palma et al. reported that six out of 12 children survived without any adjuvant therapy; only one child experienced late recur- rence [35] . In young children omission of adjuvant, therapy could significantly reduce the risk of late morbidity, thus, potentially favoring an attempt to remove remaining tumor at second surgery. On the other hand, the risks of aggressive surgery are high in young children [5] and the role of surgery without adjuvant radiation is still uncertain. Ependymomas have not been supposed to be very chemo-responsive in the past. In a randomized CCG trial for children over 3 years of age, vincristine and lomustine were found to be of no benefit for ependymoma [10] . In the CCG trial for infants, including five children with measurable postoperative disease, no child with intracranial ependymoma achieved complete or even partial response after receiving chemotherapy [12] . Response rates reported in other retrospective studies range from 0 to 48% [9,15] . In the Australia–New Zealand trial, seven children were evaluable for response to chemother- apy, and six achieved either complete or partial response, but only five of 14 children with ependymoma survived, none of those with initial dissemination [49] . Duffner et al. concluded from her re-analysis of the Baby-POG I trial that ependymoma might be chemo-sensitive, but not chemo- curable, because long delay of radiotherapy reduced survival rates despite intensive interposed chemotherapy [8] . In our series, only three out of 13 children treated with chemotherapy alone survived. Historically, surgery alone resulted in 5-year survival rates of less than 30%, comprising all ranges of histological grade and degree of surgery [27] . After employing adjuvant irradiation routinely, survival could be improved signifi- cantly, with survival rates of up to 60% in older children