Glioblastoma is the most common, aggressive and lethal malignant brain tumour, with an average survival time of 14 months. This poor prognosis is because the disease does not respond to the most usual treatments of chemotherapy and radiotherapy, tends to reappear after surgery and is very heterogeneous. This heterogeneity means that every part of the tumour mass has highly varied characteristics and therapeutic responses. However, clinical practice often employs the same standard procedure to treat it.
It was necessary to advance towards personalised medicine for these patients. This is what we have tried to do, based on image. Starting out with the image sequences obtained in the clinical routine, in the ONCOhabitats project we have developed a system based on artificial intelligence that enables us to characterise the internal heterogeneity of the tumour, delimitating those zones that may behave differently and analysing which characteristics were related to prognosis and survival. We have conducted one validation of the tool in an international, multicentre trial, participants in which included Hospital Vall d’Hebron, Hospital Clínic de Barcelona and Oslo University Hospital, among others.