Anaplastic mixed gliomas and anaplastic oligodendroglioma in children: results from the CCG 945 experience

To review interpathologist diagnosis variability and survival of children treated for either anaplastic mixed glioma (AMG) or anaplastic oligodendroglioma (AO) with surgery, irradiation and chemotherapy.

Two hundred and fifty patients with an institutional diagnosis of malignant glioma were enrolled on Children's Cancer Group CCG-945 between 1985 and 1991, and administered vincristine during involved field radiotherapy, then six cycles of prednisone, lomustine and, vincristine; or two cycles of "eight-drugs-in-one-day" (8-in-1) chemotherapy then involved-field radiotherapy followed by six cycles of 8-in-1 chemotherapy. Central review of institutional pathology was post hoc by five experienced neuropathologists.

Twenty-six children had institutional diagnoses of AMG and four had AO. Complete resection and cerebral tumor location was associated with better overall survival (OS) in patients with institutional diagnoses of AMG. However, central review established that only nine of 26 children had AMG: either mixed oligoastrocytoma (MOA) or anaplastic mixed oligoastrocytoma (AOA) and only one had AO. Central review revealed five more patients with AMG, but none with AO. Institutional and CCG central review diagnoses of AMG or AO had poor Jaccard reliabilities of 0.29 and 0.25 respectively. Five-year EFS and OS for five children with centrally confirmed MOA was 50 +/- 20%, with four centrally confirmed AOA was 37.5 +/- 17%. After central review, small samples made tests for differences in survival between regimes impossible.

Diagnosis of these tumors is challenging, with only 35% of institutional diagnoses confirmed for AMG and 25% for AO, and survival among children with these tumors is poor, despite intensive therapy. This suggests reliable diagnostic markers and new therapeutic approaches are needed.