Tumour growth and response to radiotherapeutic schemes is a markedly multiscale process which by no means can be reduced to only molecular or cellular events. Within this framework a new scientific area, i.e. in silico oncology has been proposed in order to address the previously mentioned hypercomplex process at essentially all levels of biocomplexity. This paper focuses on the case of imageable glioblastoma mulriforme response to radiotherapy and presents the basics of an essentially top-down modelling approach, aiming at an improved undestanding of cancer and at a patient-specific optimization of treatment.