Using artificial intelligence to enhance radiotherapy
Just like the human body, each cancer has biological differences. These differences are a major reason why cancer treatment may be successful in one person, but not in another – even if they have the same type of cancer. Radiotherapy is an important part of treatment for many cancer patients, however, in current practice it offers little capacity for personalisation. Treatment regimes are designed based on the cancer type and size of the tumour only.
New technology that combines radiotherapy with detailed imaging (magnetic resonance imaging or MRI) offers a new degree of precision. Called MRI-Linacs, this new generation of radiotherapy is able to target a tumour more accurately, meaning fewer healthy cells are damaged through treatment. Professor Price has identified an opportunity to further enhance treatment with MRI-Linacs by considering the biological characteristics of an individual’s tumour.
The research
In this project, Professor Price and his multidisciplinary team will focus on four cancers that often have poorer outcomes from standard radiotherapy. These are pancreatic, oesophageal, liver, and rectal cancers. The team will use ultra-high strength MRI scanners to produce ‘microscopic’ resolution images of tumour samples obtained from a biobank. These highly detailed images will allow the team to characterise the biological differences between tumours. In a world-first, the team will then use a specialised form of artificial intelligence, called ‘deep learning’, to transfer this knowledge into clinical MRI scanners to enhance the resolution of imagery in MRI-Linacs. This will allow clinicians to predict the effectiveness of treatment and enable personalisation.
The impact
The enhanced capacity of MRI-Linacs will have several significant benefits for patients with these four cancers that are otherwise difficult to treat successfully. For example, in pancreatic cancer, this new approach will enable clinicians to target areas of the patient’s tumour that are likely to be resistant to radiotherapy (for example, areas with lower levels of oxygen) with a higher dose. In rectal cancer, the high-resolution MRI will enable clinicians to see exactly how successful treatment by radiotherapy has been. If a patient has had a complete response to treatment, then surgery and its associated quality of life negative impacts can be avoided. If the imagery shows areas of treatment resistance, then these can be targeted with a higher dose of radiotherapy before surgery is considered.
Successful implementation of this new enhanced imaging technology along with the precision of MRI-Linacs has the potential to greatly improve treatment outcomes and survival. While the team is focused on four particularly problematic cancers in this study, the new approach could eventually be used to enhance radiotherapy effectiveness for any solid cancers.