Diffuse low-grade gliomas are rare brain tumors of young adults. Several treatments are used by the neuro oncologist (surgery, chemotherapy, radiotherapy). Our goal is to create a decision-aid tool to ensure an individualized treatment strategy. In clinical practice, the monitoring of gliomas is based on the estimation of tumor volume, obtained from MRI. This is done either through the three diameters method, or through a manual segmentation followed by a software reconstruction ; a subjective test helped us to compare statistically the two methods. We explore also semi-automatic segmentation algorithms which seem to be a promising way. Once we studied the reliability in the calculation of the interest variable, we are interested in the modeling of the evolution of the tumor’s size, in order to help oncologists in decision making. Crucial questions include identifying subgroups of patients who could benefit from chemotherapy, determining the best time to initiate or end chemotherapy, ... Our aim is to design new predictive models dedicated to the evolution of the tumor. Preliminary but very promising results have been obtained by regression models on a database of 55 patients under neoadjuvant chemotherapy treatment. Two statistical models (linear and exponential) have been identified.