Résumé : BACKGROUND: Positron emission tomography (PET) with 18F-2-fluoro-2-deoxy-D-glucose (FDG) is widely applied to the study of gliomas. The histology of most gliomas is regionally heterogeneous. The relationship between histologic features and glucose metabolism evaluated by PET with FDG may therefore vary within the limits of the tumor. PET with FDG integrated in the planning of stereotactic brain biopsy allows precise comparison between local FDG uptake and histology. Using this approach, the authors investigated whether glucose metabolism of gliomas is related to anaplasia, and whether PET with FDG detects metabolic heterogeneity that parallels histologic heterogeneity of gliomas. METHODS: A total of 161 biopsy samples collected from 20 PET-guided procedures performed in patients with gliomas (8 low grade astrocytomas, 8 anaplastic astrocytomas, 1 anaplastic oligoastrocytoma, and 3 glioblastomas) were analyzed for the presence or absence of 8 histologic features. Stereotactic coordinates were used to calculate the metabolic rate of glucose (MRGlu) in the region of each biopsy sample. Gray and white matter MRGlu were used to define four metabolic grades that were compared with local histology. RESULTS: The difference in MRGlu expressed as micromoles per 100 g per minute was highly significant between anaplastic and nonanaplastic samples; the median +/- quartile deviation was 23 +/- 16 in anaplastic samples and 18 +/- 5 in nonanaplastic samples (P < 0.005). Even more significant differences were found when MRGlu was normalized to the cortex or to the white matter. Metabolic grades were different in anaplastic and nonanaplastic samples (P < 0.0001). Approximately 75% of samples metabolically graded 3 or 4 demonstrated signs of anaplasia, compared with 10% of samples graded 0 or 1. CONCLUSIONS: FDG uptake in gliomas is anatomically heterogeneous and is regionally related to the presence of anaplasia.