INTRIGUE - INvestigating Treatment Response In Glioblastoma Using radiomic Evaluation (360G-Wellcome-225513_Z_22_Z)

Background to the research problem The most aggressive type of brain tumour is called glioblastoma (GBM) and most patients have an MRI scan to plan treatment. After surgery, radiotherapy, and chemotherapy, tumour will often come back nearby, due to tumour cells that have spread into adjacent brain tissue, which current MRI techniques struggle to detect. It may be possible to train computers to detect this, however. Radiomics, a type of artificial intelligence, which uses sophisticated software and maths so that the computer can look for patterns which may detect tumour spread. Our approach We will use computer software to identify a pattern of change on MRI scans of patients with GBM, including the region surrounding the main tumour site, thought to include areas of tumour cell invasion. Using the patterns identified and seeing where tumour recurs after treatment, we are aiming to determine the relationship between the radiomics and high-risk areas of recurrence. Expected impact of our work Being able to use MRI scans to predict areas of tumour recurrence may help to personalise patient treatment through better targeting of radiotherapy to high-risk areas and avoid giving too much treatment to normal areas.