X-Ray Occult Carpal Detection

CSC Lead: Dika

MRI imaging is superior in identification of occult carpal fracture, but is not always accessible. Imaging from X-rays can give suboptimal views, and the presentation of arthritis can make small fractures difficult to see. An AI tool to aide clinical diagnosis of occult carpal fractures using x-rays would increase diagnostic sensitivity in areas and situations where MRI is not available. A computer aided diagnosis tool which would automatically run when either a scaphoid fracture is suspected or if a patient is referred for a hand/wrist x-ray from A&E would increase sensitivity and confidence of diagnosis. Carpal fractures can be difficult to identify and patients with high clinical suspicion are put in a splint and referred to the fracture clinic even if a fracture isn’t seen on the x-ray by the clinician. Subtle lucency of an un-displaced fracture and the significance of a small bone fragment is currently easily missed. A successful tool would therefore increase diagnostic confidence and accuracy and reduce repeated x-rays and needless fracture clinic referrals

Clinical lead: Davina Mak

Project Plan

3. Dataset curation (retrospective).

4. Model training

5. Model testing

6. Implementation

7. Audit

References:
TOHETI trial results