Is craniospinal irradiation required to cure children with malignant (anaplastic) intracranial ependymomas?

Between 1970 and 1989, 17 children with histologically malignant intracranial ependymomas received treatment at the University of Pennsylvania (Philadelphia, PA). Eleven were treated with prophylactic cranial or craniospinal irradiation plus a local boost (CS-XRT), five with local (L-XRT) irradiation only, and one was treated without (NRT) irradiation. With a median survival of 2 years and a median follow-up time for long-term survivors of 6.0 years, five of 11 patients who received CS-XRT are alive compared with none treated with L-XRT and none treated with NRT. Two-year actuarial survival rates are 40% (L-XRT) and 52% (CS-XRT). When examined for other factors, age and local radiation dose remain the most significant prognostic indicators of survival. The 2-year actuarial survival for children younger than 4 years at diagnosis is 20% compared with 83% for their older counterparts. Likewise, the 2-year survival for patients treated with local radiation doses over 4500 cGy was 55% compared with 0% for patients treated with lesser doses. To date there are a total of 28 recurrences. All have occurred with local components except for six (unknown) who died before the exact site(s) could be determined. There is no significant difference in the failure rates outside the original tumor bed in the three groups. These data suggest that local relapse remains the most significant component of failure. Because intrinsic and extrinsic factors such as age and radiation dose seem to be interrelated and at least as important as the use of craniospinal irradiation, the need for prophylactic treatment for children with anaplastic ependymoma could neither be substantiated nor refuted. The use of local radiation alone, however, should be restricted to carefully designed clinical trials in which meticulous pretreatment evaluation is performed, and vigilant posttreatment evaluation of the spine and brain is mandatory.