Résumé
WHO grade II glioma, i.e. diffuse low-grade glioma, is a pre-malignant tumour, usually revealed by seizures in young patients with a normal life. This tumour has a constant growth, and will inescapably become anaplastic. Surgical resection significantly increases the overall survival by delaying the malignant transformation. Thus, the dilemma is to perform early surgery in order to optimise the extent of resection (and thus the median survival) by removing smaller tumours while preserving the quality of life. To this end, the new concept proposed in this review is to achieve surgical resection according to functional and not to oncological boundaries. In other words, the principle is to first understand the cerebral anatomo-functional organisation at the individual level (because of a major inter-individual variability), with the aim of resecting a part of the brain invaded by a diffuse chronic disease, on the condition nonetheless that this part of the brain can be functionally compensated-i.e. with no consequences on the quality of life. To this end, in addition to the preoperative functional neuroimaging and the intraoperative electrical cortical mapping in awake patients, it is also crucial to map both horizontal cortico-cortical connectivity (long-distance association fibres) as well as vertical cortico-subcortical connectivity (projection fibres), with the aim to preserve the networks underlying the minimal common core of the brain. Interestingly, this "hodotopical" workframe, based on the study of both cortical epicentres and subcortical pathways, opens the door to mechanisms of functional reshaping. These recent technical and conceptual advances in the hodotopical and plastic view of brain processing have allowed a dramatic improvement of the benefit-to-risk ratio of surgery, concerning both oncological and functional outcomes. In summary, it is time to move towards "functional neurooncology" and "preventive neurosurgery" in low-grade gliomas. Stronger interactions with fundamental neurosciences should be developed, in order (1) to build updated models of cognition and brain plasticity; (2) to elaborate biomathematical models of low-grade glioma growth and migration; (3) to study in silico the dynamic interactions between the natural course of this disease and the adaptative behaviour of its host (the brain), with the goal to adapt the best individualised therapeutic strategy.