Clinicians need to study the effects of new treatments: it is sometimes possible to detect and quantify those effects by looking at evolutions in the medical images of a patient over time especially in the case of multiple sclerosis (MS) where lesions are related to clinical signs [1]. Some methods allow to compare two images to know where there are differences, typically between the last and the previous exam 2,3. However a retrospective analysis might be done on the whole set of images to find the moments when evolutions occur [4]. We propose to use the analogy between an activation in functional imaging (for instance PET, SPECT and fMRI) [5] and a signal change due to an evolving multiple sclerosis lesion. Voxels corresponding to evolving pathological areas are named ELV (Evolving Lesion Voxels) in this abstract.