Glioblastoma multiforme still remains, at present, the most frequent and deadly primary malignant glioma in adult. Despite safer and larger neurosurgical resections, patients almost always relapse very close or inside the tumor bed. Since more than 20 years, radiation therapy (RT) continue delivering the same dose of 60 Gy in 6 weeks, more precisely guided with CT-scanner and magnetic resonance imaging (MRI) in the treatment position. If morbidity has decreased with "non whole-brain" volumes, RT is nearly always failing locally, as surgery. Until now, all the series evaluating escalating doses (up to 80-90 Gy) in limited volumes have failed. One can really question : is the good dose delivered in the adequate volume? Main goal of new imaging techniques is to better visualize microscopic extension of malignant glioma cells. As based on metabolic principles, areas of abnormalities visualized with functional imaging have a different meaning, often complementary from conventional data. The four evaluated techniques are : magnetic resonance spectroscoy (MRS), functional MRI (fMRI), 18FDG or methionine PET, IMT (123iodine-alpha-methyl-thyrosine) SPECT. Each technique has potential interests and limits, MRS and fMRI appearing the most promising : they have both acceptable spatial resolution and can be executed just after conventional MRI acquisition. Areas of functional abnormalities are only partially including areas of hyperintensity in T1, T2 weighted MRI. It is therefore highly possible that, using it complementary to conventional CT and MRI for RT treatment planning, they add some precious informations; consequently, the very limited efficacy/toxicity ratio could be increased. This hypothesis will only be confirmed by prospective studies registering in parallel both functional and morphological abnormalities, linking them with sites of local recurrence. Once "targeted" the real microscopically invaded areas, one can speculate on new escalating dose studies, delivering RT in "adequate" volumes, combining it with new "targeted" drugs, as already recently demonstrated in head and neck cancers.