Hypoxia is recognized to be an independent predictor of clinical outcome in oncology. PET using [18F]-FMISO has been described to be useful for the non invasive assessment of hypoxia in cancer. The use of this radiotracer for brain tumours is very limited and there is no standard to acquire and quantify [18F]-FMISO uptake. So there is a need for a methodological evaluation of this PET tracer The purpose of this research is to define optimal parameters for acquisition and data exploitation to quantify [18F]-FMISO uptake and so predict clinical outcome in glioblastomas.

Low sensitivity to radiation of glioblastoma is partly caused by hypoxia. Hypoxia in tumours is not predicted by tumour size. Detecting and monitoring tissue oxygenation are of great interest to modify therapeutic strategies, including local dose escalation for radiotherapy or select chemotherapeutic agents with better impact in glioblastomas.

PET with appropriate radiotracers, especially [18F]-FMISO, enables non-invasive assessment of hypoxia. [18F]-FMISO only accumulates in viable hypoxic cells. So, it has been demonstrated that PET using 18F-FMISO is suitable to localize and quantify hypoxia. But there isn't any optimal acquisition protocol or standardized images quantification treatment. Thus, the interpretation of [18F]-FMISO PET images and the predictive value of [18F]-FMISO SUV (Standardized Uptake Value) remain unclear explaining the need of methodological approaches.

Hypoxia is one of the worst prognostic factors of clinical outcome in glioblastomas. Today, it is well admitted that hypoxia is heterogeneous, variable within different tumour types and varied spatially and temporally. Hypoxia induced proteomic and gene expression changes that lead to increase angiogenesis, invasion and metastasis. So the hypoxic fraction in solid tumours reduces their sensitivity to conventional treatment modalities, modulating therapeutic response to ionizing radiation or certain chemotherapeutic agents. This is particularly important in glioblastomas. Hypoxic cells in solid tumours could influence local failure following radiotherapy and has been associated with malignant progression, loco regional spread and distant metastases and represents an increasing probability of recurrence.

Thus, the non-invasive determination and monitoring of the oxygenation status of tumours is of importance to classify patients' outcome and modify therapeutic strategies in those tumours. Actually the oxygenation status of individual tumours is not assessed routinely. Numerous different approaches have been used to identify hypoxia in tumours. Eppendorf oxygen probe measurements (pO2 histography) may be considered as a 'gold standard' for hypoxia in human malignancies. However, it is an invasive method being confined to superficial, well accessible tumours and requires many measures. PET using [18F]Fluoro-deoxyglucose ([18F]-FDG), allows non-invasive imaging of glucose metabolism and takes a growing place in cancer staging, but [18F]-FDG can't assess correctly the oxygenation status of tumours and is not suitable for brain tumor. PET with appropriate radiotracers enables non-invasive assessment of presence and distribution of hypoxia in tumours. Nitroimidazoles are a class of electron affinic molecules that were shown to accumulate in hypoxic cells in cultures and in vivo. [18F]-FMISO is the most frequently employed tracer; its intracellular retention is dependent on oxygen concentration. Consequently [18F]-FMISO has been used as a non-invasive technique for detection of hypoxia in human. Different authors have demonstrated that it is suitable to localize and quantify hypoxia. Thus, [18F]-FMISO PET has been studied to evaluate prognosis and predict treatment response. However, some investigators report an unclear correlation between Eppendorf measurements and standardized uptake values (SUV). This observation may be explained by the structural complexity of hypoxic tumour tissues. Nevertheless, there is a need of standardized procedures to acquire and quantify [18F]-FMISO uptake. Actually the use of this tracer is very limited in clinic and the academic studies have included small populations of patients and suffer of the heterogeneity of technical procedures.

The aim of this study is to determine the optimal acquisition protocol and treatment parameters enable to describe [18F]-FMISO uptake in glioblastomas known to be hardly influenced by hypoxia. Then, validate [18F]-FMISO-PET as a prognostic maker of recurrence.

We will introduce a pretherapy [18F]-FMISO PET-CT in the treatment planning of patients suffering of different newly diagnosed glioblastoma and eligible to a radical treatment with curative intent, consisting of conformational radiotherapy and chemotherapy. [18F]-FMISO PET-CT results will not be take into account for the patient management. We will test different acquisition protocols and use a wild panel of quantification parameters issued from published studies and original ones developed by our team enable to describe [18F]-FMISO uptake. Patients will be followed clinically and para-clinically during one year after the end of the treatment according to the edited recommendations of each tumour type and grade to analyze outcome (failure is define as persistent disease in the primary site, progression of disease, locoregional relapse after complete response or distant metastasis). Thus we will be able to measure failure free survival and determine overall survival.

Inclusion Criteria:

• Patients over 18
• Patients with a malignant tumour glioblastomas proposed for a radical treatment consisting in conformational radiotherapy and/or chemotherapy
• Signed informed consent

Exclusion Criteria:

• Patients who can't undergo radiotherapy or chemotherapy
• Patients with distant metastases known before inclusion except renal cancer where patients with metastases can be included
• Patients suffering of a second cancer or treated before by radiotherapy in the tumour site.
• Pregnant and breast feeding women, women in age to procreate without contraception