How AI is helping to speed up spinal fracture diagnosis

When it comes to spinal fractures, one of the biggest challenges is spotting them early on. That’s why we’re so pleased to hear that the AI-powered ADOPT study has just identified its first patients.

The study is a collaboration between the Royal Osteoporosis Society (ROS), the University of Oxford, Addenbrooke’s Hospital, Cambridge and medical imaging technology company Nanox.AI. Funding for the project has come from the Artificial Intelligence in Health and Care Award – part of the NHS AI Lab that’s led by NHSX and delivered in partnership with the Accelerated Access Collaborative (AAC) and the National Institute for Health Research (HIHR). 

It focuses on using AI to review CT scans to find undiagnosed spinal fractures across five diverse NHS test hospitals in Bradford, Cambridge, Cardiff, Nottingham and Southampton.

Currently, if a patient has an x-ray or CT scan that identifies a vertebral fracture, they should ideally be referred to a Fracture Liaison Service (FLS) if one is available in the area. But the journey isn’t always that straightforward. The radiologist may not be prompted to report back to the referring doctor, or they may use different words to describe fractures, which mean they don’t necessarily get picked up.

If it’s successful, an AI-enhanced vertebral fracture prevention pathway could not only offer potential time and cost savings for the NHS, but it could also revolutionise the patient experience, improving identification, assessment, and treatment recommendations for the future.

The patients that have now been identified will be assessed by a specialist Fracture Liaison Service – either a specialist FLS nurse or a team that specialises in osteoporosis identification, treatment, advice and more.

The importance of patient voice

The ROS is supporting the project by leading on public and patient involvement (PPI). We recruited Lead Volunteer Advocates (LVAs) for each of the sites to help drive this, along with 14 PPI representatives – some of whom have first-hand experience of spinal fractures themselves.

LVA Carol Jaensch is helping to lead the PPI at the Cambridge site. Carol said: “The PPI team members who are in place in each hospital will be able to influence the pathways for future patients, making suggestions about what was good and bad, what worked and didn’t work.

“They will sit (either virtually or in reality) together with the FLS team from their own local hospital and have proper input and influence on how things will be in the future. This will then be shared across the five test sites and hopefully across all NHS hospitals in future.”

It’s hoped that this combination of AI and real-life feedback will help move more people from initial contact to treatment more efficiently.

What is a vertebral fracture?

A vertebral fracture – also known as a compression fracture or spinal fracture - is a bone that is broken in the back. These fractures will usually be diagnosed via x-ray, CT scan or similar.

But what about those people who have scans that include the spine, but for different reasons, such as bowel or lung problems? Because the focus of the scan is another condition, fractures are either not identified or not reported. Of these cases, up to 1 in 20 scans could be showing a spinal fracture. And it’s these scans that the AI has been reviewing in the study.

“We hope the positive benefits from using AI will significantly change the lives of thousands of patients in the project and potentially hundreds of thousands of lives in the NHS,” says Associate Professor Kassim Javaid, who is leading the research at the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS), University of Oxford.

“It also represents the fruition of the combined efforts from clinicians, researchers, IT experts and the teams from NANOX and the Royal Osteoporosis Society to implement this novel care pathway into the NHS.”

The results are due in 2024, so read more about the study so far here and check back for updates over the coming months.

If you’d like to find out more about public and patient involvement, talk to our Volunteer and Public Engagement team today.