A ground-breaking new study will test how artificial intelligence (AI) can make prostate cancer diagnosis more accurate. The study by the University of Oxford  is funded by a £1.9m grant from research charity Prostate Cancer UK.

By using AI to provide clear and detailed information about a man's cancer, the project, which uses a cutting-edge AI tool from precision medicine company Artera , will help doctors make better decisions about whether to monitor or treat it, tailoring the decision to each man's specific needs and therefore reducing the potential for both overtreatment and undertreatment. The project could also help doctors decide the best course of treatment for each man. 

Although the way we diagnose prostate cancer is safer and more accurate than ever before, it can still be challenging to determine how aggressive each man’s cancer is. 

This can have an effect on treatment decisions. In some cases, these issues lead to some men being treated with surgery or radiotherapy when they could instead be monitored, putting them at risk of life-changing side effects that could have been avoided. At the same time, other men are missing out on treatment , allowing the cancer to spread and reducing their chance of survival. 

Now researchers at Oxford University Hospitals and NHS trusts and universities in Bristol and Glasgow will investigate how the ArteraAI Prostate Biopsy Assay  tool could give patients and doctors clearer, more precise information to guide treatment choices.

Called the VANGUARD PATH study, the research is being led by Principal Investigator Professor Clare Verrill with co-leads Professor Richard Bryant and Dr Srinivasa Rao, all from the University of Oxford’s Nuffield Department of Surgical Sciences, and Professor Jonathan Aning from North Bristol NHS Trust.

The researchers will first test the AI tool on prostate biopsy samples from men who have already been diagnosed and treated for prostate cancer and have at least five years of follow-up data. This will enable the team to compare how well the AI’s predictions match to the real-world outcomes, such as which cancers were more likely to spread without treatment. 

Once they show that the AI can make accurate predictions about which cancers need treating or monitoring, the team will test the tool in real clinics, on biopsies from men as they are diagnosed – focusing on particularly difficult cases where it is harder for doctors to decide which treatment is best for each man.

Professor Verrill said: “This funding from Prostate Cancer UK will be truly transformational for men diagnosed with prostate cancer. It will pave the way for advanced AI technologies such as the Artera AI Prostate Biopsy Assay   to be rolled out across the NHS. This will enable more detailed and precise information to be provided to men who will be able to make better informed decisions with their clinical team about whether they can be safely monitored or need treatment - and, if so, help guide those decisions.” 

Dr Matthew Hobbs, Director of Research at Prostate Cancer UK, said: “AI has the potential to massively improve prostate cancer care and make sure that every man has the most accurate and best treatment plan for his specific cancer. We’ve been working with Artera for several years now and I believe that their AI tool is one of the most exciting ones to have been developed. But exciting AI tools can only make a difference if they can be properly evaluated. 

"This new project from Professor Verrill and her team is so exciting because it tests this new technology in real-world settings, meaning we can deliver the final evidence needed for it to be rolled out across the NHS. It is exactly the sort of high-impact, research the Transformational Impact Awards Scheme  supports – turning exciting innovations into proven real-life improvements for men with prostate cancer."

The AI analysis will be done alongside the usual report prepared by a pathologist and will not be used at this stage to influence the men’s treatment decisions.  However, the researchers will compare these decisions to what doctors would have decided if they had used the AI report, to see whether the tool would have helped them in those challenging cases.

As well as OUH, the project will be carried out at North Bristol NHS Trust and NHS Greater Glasgow and Clyde. Testing the tool in NHS hospitals to see how well it works in real life will also allow the researchers to understand how patients and doctors feel about using it, paving the way for smarter, more tailored prostate cancer care.

The AI leverages computer vision to analyse a digitised image of the patient's biopsy samples in ways that humans cannot, extracting detailed information about the cancer cell's appearance and structure. The AI model then generates a personalised risk score that can advise doctors on what the best treatment will be for that particular patient.

This AI model was trained on thousands of patient samples and has been validated in multiple studies to accurately predict cancer aggressiveness. One big appeal of this type of AI technology is the potential to produce personalised risk scores for prostate cancer patients quickly and cost-efficiently.