NICE extends the evidence period for AI fracture detection on X‑ray

On 23 October 2026 NICE updated its early value assessment of AI tools that help detect fractures on X-rays in urgent care (HTG739). The only change is to the timetable: the evidence generation period has been extended by one month and now runs until about February 2027.

What's new

HTG739 was first published on 14 January 2025. It is an early value assessment, which means NICE allows some technologies to be used in the NHS while more evidence is collected, rather than recommending them for routine use. NICE's page now says the evidence generation period has been extended by one month to give more time to fill the evidence gaps, and that the date may change again.

The recommendations themselves are unchanged. Four technologies can be used during the evidence period as options to help healthcare professionals detect fractures on X-rays in urgent care: Rayvolve and TechCare Alert for people of any age, and BoneView and RBfracture for people aged 2 and over. This is conditional: the evidence set out in NICE's evidence generation plan must be collected, and each product needs the right regulatory approval, including NHS England's Digital Technology Assessment Criteria (DTAC). A fifth technology, qMSK, should only be used in research.

When the period ends, the companies submit their evidence and NICE decides whether the tools can be adopted routinely. NICE may withdraw the guidance if the conditions are not met. The evidence NICE has asked for includes how accurately clinicians read X-rays with and without AI, the costs and consequences of missed fractures, fracture clinic referral rates, changes to treatment decisions, software failure rates, whether the tools pick up or miss findings other than fractures, and how well they perform in different groups of patients.

What it means for learners

  • Nothing changes in practice yet. This is a short delay to a review, not a new recommendation. Whether an AI fracture tool is in use where you work is a local decision.
  • AI is an aid, not the reader. NICE describes these tools as helping healthcare professionals, and one of its open questions is how accurate clinicians are with and without them. Build your own systematic approach first: see how to read a skeletal X-ray and how to describe a fracture.
  • Read the image before the AI marks. The RCR suggests trainees should be able to view images without AI output first; see our item on RCR guidance on AI in reporting workflows.
  • Fractures are not the only finding. NICE explicitly wants evidence on non-fracture findings. A tool trained to flag fractures will not necessarily mention a dislocation, effusion or bone lesion, so complete the whole film yourself; commonly missed injuries are covered in scaphoid fractures and the paediatric elbow.

Source

  • National Institute for Health and Care Excellence. Artificial intelligence (AI) technologies to help detect fractures on X-rays in urgent care: early value assessment. Health technology guidance HTG739. Published 14 January 2025; last updated 23 October 2026 (evidence generation period extended). nice.org.uk/guidance/htg739; recommendations at chapter 1.

Also in What's new: RCR guidance on AI in reporting workflows · ChatGPT-5 and distal radius fractures.

For education only. Not a substitute for a formal report, senior review or your local policy. Never use this page to make decisions about a real patient.