Reaction–diffusion models are commonly employed to describe the growth and infiltration of low-grade gliomas, often assuming that tumor diffusion aligns with white-matter architecture inferred from diffusion tensor imaging (DTI). However, clinical observations in [1] indicate that glioma progression does not necessarily follow DTI-defined pathways. In this talk, we investigate alternative image-informed formulations for modelling low-grade glioma growth that do not rely on DTI-derived diffusion. We perform a patient-specific comparison of three models: a reaction–diffusion model in which diffusion is defined through image-informed tissue deformation extracted from longitudinal imaging, and a reaction–diffusion–advection model where advection is driven by a velocity field obtained from the same image-informed deformation. All models are calibrated using longitudinal tumor segmentations, and their ability to reproduce observed spatio-temporal growth patterns is evaluated quantitatively.