Preventing fragility fractures: Artificial intelligence (AI)-based opportunistic osteoporosis detection in routine computed tomography (CT)
Authors' conclusion
To the best of our knowledge, this AI model is the first to classify BMD across the entire spine on both contrast and native CT scans. It demonstrates robust real-world performance and offers a promising approach for scalable opportunistic BMD screening in routine CT imaging.
Study context
- Population
- Patients with sagittal spine CT; retrospective cohort of patients who later sustained fragility fractures, CT acquired 12–24 months prior to the fractureholdout set
- Participants
- 270 patients
- Comparison
- AI-based opportunistic bone mineral density (BMD) classification on routine CT vs DXA T-scores (reference standard); vs radiologists' reports in a retrospective fragility fracture cohort
- Outcome
- Specificity and sensitivity of AI BMD classification vs DXA on the independent holdout set
Methodological detailsAI-based neural network, independent holdout set, retrospective analysis, multicentre (US and German centers), Training dataset: 1084 sagittal spine CT series from 526 patients, cervical, thoracic and lumbar regions
- AI-based neural network
- Independent holdout set
- Retrospective analysis
- Multicentre (US and German centers)
- Training dataset: 1084 sagittal spine CT series from 526 patients, cervical, thoracic and lumbar regions
Performance
As reported in the abstract
Diagnostic performance of the algorithm on the independent holdout set
Specificity of the algorithm (independent holdout set)
84%
Sensitivity of the algorithm (independent holdout set)
67%
Further findings
| Outcome | AI model |
|---|---|
| Reduced BMD flagged on CT 12-24 months before fragility fracture (retrospective study) · 12-24 months prior to fracture | |
| Osteoporotic | 87 |
| Osteopenic | 97 |
- Reduced BMD flagged on CT 12-24 months before fragility fracture (retrospective study): of 354 patients
Original publication
Preventing fragility fractures: AI-based opportunistic osteoporosis detection in routine CT . Osteoporos Int. Published 24 August 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →