Background, rationale: Diagnosis, staging/treatment of Glioblastomas (GBM) has been commonly based on Magnetic Resonance (MR). The standard treatment includes the macroscopically complete tumour resection followed by radiotherapy treatment (RT) with concurrent temozolomide chemotherapy. Unfortunately, 90% of GBM progress within 2 years.
Hypothesis: Positron Emission Tomography (PET) based on the amino-acid radiotracer O-(2)-18F-Fluoroethyl-LTyrosine (FET) has been proposed to overcome MR limitations when differentiating local recurrence (LR) from radiogenic alterations. A personalized RT strategy based on tumour heterogeneity, defined by multimodality imaging, could allow escalating RT treatment doses to high-risk tumour subareas while sparing doses in organs at risk.
Aims: We aim to identify biologically active tumour tissue associated with LR in GBM by the best imaging modality (or combination), in order to replace the homogeneous dose distribution conventionally delivered in RT, by a dose distribution scaled based on the patient’s specific risk profile of LR.
Methods: Our project involves 410 patients 2 prospective/ 2 retrospective cohorts. 120 patients have MRs and PET before RT and 230 additionally for the follow-up. From them, 30 will be imaged by an hybrid PET/MR. Artificial intelligence will be applied for GBM segmentation, for prediction of LR time and location, for generating CT from MR and for identification of patient groups (clustering), who could benefit from a given dose escalation in RT, based on radiobiological modelling.
Expected results/potential impact: All resulted models will be joined in an open-source tool making possible the integration of results by different health institutions worldwide, in order to adapt GBM treatment based on the individual risk pattern. Our proposal represents therefore an important step in personalized medicine for GBM. An improvement in patient care and quality of life is therefore expected.