When a patient’s cancer is curable, a big part of the treatment is determined by estimating the risk that the cancer will come back. Accurate risk estimation is important, because patients at higher risk of recurrence require a more intensive, and ideally more tailored treatment. More information is required to understand what makes a particular cancer higher risk, and if there are better treatments for these higher risk cancers. Artificial intelligence techniques will be used to generate digital footprints of the tumour and to teach computers to understand microscope pictures. In other words this will allow computers to 'see' things which a human pathologist could miss and thereby improve risk estimates and provide patients with the best care for their situation.