An artificial intelligence disease monitoring tool for patients with brain tumours

  • Price, Stephen J. (PI)
  • Schönlieb, Carola Bibiane (CoPI)
  • Chen, Xi (CoI)
  • Li, Chao (CoI)
  • Matys, Tomasz (CoI)
  • Das, Tilak (CoI)
  • Jenkinson, Michael (CoI)

Project: Central government, health and local authorities

Project Details


There are many people who live with a slowly-growing brain tumours – such as meningiomas or low-grade gliomas. These tumours are frequently found when they have a brain scan for unrelated problems. Patients will have regular brain scans while living with the knowledge their tumour may grow and cause them problems. Waiting for the results of these scans causes a lot of worry for patients and their families. Currently we don’t know which tumours may grow (so need careful watching) and which will not. We want to use artificial intelligence to work out which patients have tumours that are likely to grow. If we could even measure the volume of a tumour and see how it changes over time would be a great help. Even better would be to find those tumours that are high risk of growing before they do.

Our plan is to take the standard MRI images that we use, enhance them using methods we have already developed. We will then use artificial intelligence to outline the tumour and measure its volume. We plan to study scans that have already been performed on patients, so we know which tumours grew and which did not. This will let us train the computer to work out which patient is at high risk of their tumour growing.

We hope this study will be the first step of using this method to guide us treating patients. We will hold a workshop to guide our research in the future. We are especially interested asking patients and the public how they would accept decisions about their treatment that come from a computer program. We also wish to understand the best way of showing the results to patients.

By the end of this study we will have a working method to accurately measure tumour volume and predict patients at high risk of tumour growth. We hope we can use this to work out how often and individual patient is scanned – so that those at risk are carefully watched, and those at low risk have less worry.
Short title£150,000
Effective start/end date1/02/2231/01/23

Collaborative partners


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.