Timely identification of vertebral compression fractures to promote early treatment interventions.
CT first, then x-ray, then MRI.
Vertebral compression fractures (spine)
Data Curating
Clinical lead: Davina Mak
Rationale
Osteoporosis is easily treatable if identified, but can lead to complex fractures at other sites if undiagnosed, especially in the hip where a > 50% mortality at 12 months over the age of 80. Currently, 70% of vertebral compression fractures are undiagnosed though they are the most common osteoporotic fracture. This project would work in conjunction with the current fracture liaison service quality improvement project.
77% of scans with vertebral compression fractures are correctly identified and 43% scans recommend referrals – both of these numbers can be improved with AI tools.
Patient pathway
Incidental findings, patients present for imaging for other reasons
Training data
Clinical datasets are available (CT, MRI, x-rays). Imaging usually performed for other indications.
Errors
A timely identification of vertebral compression fractures which indicate osteoporosis allows the clinicians to refer the patient for osteoporosis treatment. If the osteoporosis signs are missed and the patient sustains a fracture, the fracture is complex and carries a high rate of mortality in older patients
Goals
Increased rates of identification and treatment referrals after identification to reduce occurrence of complex fractures in other sites.
Success Criteria
Increased referrals and identifications of vertebral compression fractures
Alternatives
NanoxAI. Retrospective and prospective trials are planned to take place in the early 2022. The results will determine whether this clinical problem will be solved with NanoxAI’s BoneHealth solution
Status update
In 2021 we engaged with a company to test out their product but due to difficulties beyond our control the evaluation did not commence at that time.
As an alternative, we will proceed with an in-house solution which will begin development in autumn 2024. Until then, this project cannot be prioritised and will sit in the backlog.