Four AI technologies conditionally recommended that could help spot hidden spinal fractures
The technologies could help healthcare professionals spot hidden spinal fractures in patients undergoing routine CT scans.
The technologies, which have been conditionally recommended in draft guidance, can help detect vertebral fragility fractures (VFFs) on medical images taken for unrelated condition. This means they could potentially identify fractures in thousands of patients who would otherwise go undiagnosed.
By spotting these fractures opportunistically, they can be treated and future fractures avoided, which could save the NHS money and result in better outcomes for patients.
Vertebral fragility fractures (VFFs) are breaks in the spine that occur when bones are weakened, often by osteoporosis. Fractures of the spine may not always have obvious symptoms and even if symptomatic, a person may choose not to seek care. Hidden VFFs lead to complications such as a curved spine (causing the person to lean forward), height loss, immobility, pain, as well as loss of function. VFFs are also a significant predictor of future osteoporotic fractures such as hip fractures.
More than 55% of people with a hip fracture have evidence of previous VFFs. Despite their prevalence, many VFFs remain undiagnosed.
Our independent Diagnostics Advisory Committee has conditionally recommended that BriefCase-Triage, CINA-VCF Quantix, HealthVCF and IB Lab FLAMINGO can be used in the NHS while further evidence is generated.
The draft guidance emphasises that AI technologies must only be used alongside clinical judgement, not as replacements for radiologist review.
Despite ongoing efforts to raise awareness of vertebral fragility fractures, most remain undiagnosed.
Clinical evidence suggests that AI technologies can help opportunistically detect vertebral fragility fractures that would otherwise have been missed.
By identifying more people with VFFs who need treatment for the underlying cause of the fracture, we could reduce the risk of future fractures, while potentially reducing demand on other costly services such as those needed to manage hip fractures.
The incidence of VFFs increases with age. Recent data shows an incidence rate of 7.1 per 10,000 person years in adults aged over 50. Women are more commonly affected. An incidence of 12% has been reported in women aged 50 to 79 years, increasing to 20% in women over 80 years old.
The costs associated with VFFs and hip fractures are significant – the health and social care costs in the first year of post-hip fracture are over £33,000 per person.
Our committee concluded that there is a clear unmet clinical need that can be addressed by the AI technologies, noting that thousands of radiographic images are taken annually in the NHS that could be used to opportunistically detect VFFs.
There is clearly great interest in this technology. However, we need further evidence from ‘real-world’ clinical settings to establish whether investing in AI to aid opportunistic detection represents good value for money.
The four technologies have been conditionally recommended for NHS use over the next three years while further evidence is generated. Once this period has been completed, the committee will review the evidence and make recommendations on which technologies should be used going forward.
During the three-year evidence generation period, research will focus on several key areas including diagnostic accuracy compared with current NHS standard care, failure rates of the technologies, impact on referral and treatment rates, effects on healthcare professional workload, and short-term quality of life improvements.
A consultation on the draft recommendations has now begun. Healthcare professionals, commissioners, and patients are encouraged to review the full recommendations to understand how the four technologies can be used by the NHS and submit comments before Tuesday 29 July 2025.