How accurate are X-ray deep learning models for detecting ankle fractures?
X-ray deep learning models are very good at finding ankle fractures. A large review of 14 studies found these AI tools correctly identify about 93 out of 100 actual fractures and correctly rule out fractures in about 95 out of 100 cases without them 36. This high accuracy helps doctors diagnose trauma faster and more reliably.
What the research says
Other research on extremity fractures found an FROC curve value of 0.886 for single fractures and 0.843 for multiple fractures, showing strong detection capabilities 7. A separate study on vertebral fractures also demonstrated high accuracy for deep learning systems, though that focused on the spine rather than the ankle 5.
What to ask your doctor
- How does the AI tool used in my clinic compare to the accuracy rates found in recent studies?
- Does the AI system I am using detect both single and multiple fractures in my ankle?
- How does the AI tool handle cases where the X-ray view is limited or unclear?
- Can the AI tool help identify subtle fractures that might be missed on a standard read?
This question is drawn from common patient questions about Radiology & Imaging and answered using cited medical research. We do not provide individualized advice.