High-grade gliomas (HGGs; WHO grades 3-4) are highly diverse, with survival times ranging from months to years. WHO 2000 grading criteria for high-grade oligodendroglial neoplasms [anaplastic oligoastrocytoma (AOA) and anaplastic oligodendroglioma (AO)] remain subjective, and the existence of grade 4 variants is controversial. Overall survival (OS) of 1,093 adult patients with a cerebral HGG newly diagnosed between 1990 and 2005 was analyzed by univariate and multivariate models for significance of the following factors: patient age, surgery type, year of diagnosis, endothelial proliferation, necrosis, oligodendroglial histology, treatment center, and chromosome 1p, 19q, 7p (EGFR), and 10q (PTEN) abnormalities by fluorescence in situ hybridization (FISH). Necrosis was a statistically significant predictor of poor OS on univariate and multivariate analyses in AOA but not in AO. Median OS for patients with necrotic AOA (22.8 months) was significantly worse than for patients with non-necrotic AOA (86.9 months; P < .0001) but was better than conventional glioblastomas (9.8 months; P < .0001). In addition to patient age, the following were significant independent prognostic factors (P .001): grade and surgery type for the entire HGG cohort; modified grade for AOA (3 v 4); and modified grade, 1p/19q codeletion status, and oligodendroglial histology for the 586 HGGs analyzed by FISH. Stratification of AOA, but not of pure AO, into grades 3 and 4 on the basis of necrosis is prognostically justified and is more powerful than the current approach. Both routine histology and genetic testing provide independent, prognostically useful information.
