Meet our featured researcher

Using artificial intelligence to give men a safe and accurate diagnosis

We currently don’t understand enough about the factors that define a man's risk of prostate cancer. For some men, a few genetic changes, like mutations to the BRCA2 gene, greatly elevate their risk of prostate cancer. However, for most men, prostate cancer risk is polygenic, defined by mutations that have a small individual effect, but which collectively add up to significantly influence risk.

By identifying and understanding the array of individual mutations that add up to prostate cancer risk, we could pinpoint exactly who would benefit from closer monitoring. Additionally, by understanding the function of the genetic changes influencing risk, we can gain insights into the biology of prostate cancer, to identify treatment targets.

If we know we normally produce high-quality MRI images, but in this case the image is lower quality, the AI tool could help us decide what to do next. Should we run another scan right now, saving the man from going home and waiting for another appointment? Should we keep a closer eye on the man with regular PSA blood tests? Or should we go straight to biopsy?
Dr Tristan Barrett

“My research will explore a new method of delivering radiotherapy treatment to men with advanced prostate cancer to make it more effective.

Current radiotherapy treatments like radium-223 and the next generation of treatments (such as combinations of radium-224 and lead) have gained significant attention for the treatment of these men. In the UK, there is currently limited expertise in research aiming to optimise these treatments, and that’s what I want to do.

We’re also researching why men who’ve lost certain genes may respond better to radiotherapy. This understanding would allow us to deliver a really personalised treatment which hasn’t been done before for radiotherapy."

202408 Christmas DM Dr Tristan Barett Rosie Lonsdale 9
AI could help at almost every stage. Right now, for example, when we’re reviewing MRI scans, the one that comes to us first gets reported first. But there might be a scan further down the list that has more urgent findings. An AI tool could assess the scans as they come in, to make sure the ones that need reviewing first go to the top of the radiologist’s pile.
Dr Tristan Barrett
202408 Christmas DM Dr Tristan Barett Rosie Lonsdale 19

This unique approach – studying past and present treatment pathways simultaneously – will allow her to apply what what she learns from historic data almost immediately, testing any potential changes to radiotherapy with the men as they undergo their treatment.

This approach will also give me the opportunity to test some cutting-edge scans that have never been used in this way before. Working with Professor Tyler Seibert, an expert at the University of California, San Diego, she'll introduce two new scans into the radiotherapy process: restriction spectrum imaging, which maps the layout of cells in the body by studying the movement of water molecules, and oxygen-enhanced MRI, which looks for cells that are hypoxic.

Essentially, you can’t get something for nothing – if you change one factor, you compromise another. Unless you have AI in your corner. With AI post-processing, you can run that high-resolution scan in a standard amount of time, knowing you’ll get a noisier image – and afterwards, take the noise out of the image.
Dr Tristan Barrett

This research will be a vital step towards personalising prostate radiotherapy. It will show how we can make smarter use of data and technology to better inform treatment decisions before, during, and after radiotherapy.

And the results?

Crucially this appeal has already raised over £223,000 to support our lifesaving work.

We couldn't wait to share the success of the appeal with Dr Dunne.

Thank you!

Have you made a new connection, met a new collaborator, or developed a new research idea through From Ideas to Innovation? We want to hear about it!

Complete the form below to tell us how From Ideas to Innovation has impacted your research.

Note: the name and email address fields in the below form are optional, please leave blank to submit anonymously.