Using Artificial Intelligence to Diagnose Osteoporotic Vertebral Fractures on Plain Radiographs
Authors' conclusion
Therefore, the AI_OVF_SH system is an efficient tool to assist radiologists and clinicians to improve the diagnosing of vertebral fractures.
Study context
- Population
- Plain radiographs from six community centers in Shanghai (training/internal validation) and outpatients of Shanghai Sixth People's Hospital (external validation); reference standard radiologists' Genant semiquantitative grading11,397 radiographs
- Outcome
- Accuracy, sensitivity, specificity, PPV and NPV for vertebral fracture diagnosis
Methodological detailsdeep-learning model development and validation, internal and external validation
- Deep-learning model development and validation
- Internal and external validation
Performance
As reported in the abstract
All fractures, internal validation set
| Test or model | Accuracy | Sensitivity | Specificity |
|---|---|---|---|
| AI_osteoporotic vertebral fracture (OVF)_SH | 97.41% | 84.08% | 97.25% |
AI_osteoporotic vertebral fracture (OVF)_SH
Accuracy97.41%
Sensitivity84.08%
Specificity97.25%
Further findings
Moderate fractures, internal validation set
- Sensitivity
- 88.55%
- Specificity
- 99.74%
Severe fractures, internal validation set
- Sensitivity
- 92.30%
- Specificity
- 99.92%
All fractures, external validation set
- Accuracy
- 96.85%
- Sensitivity
- 83.35%
- Specificity
- 94.70%
Moderate fractures, external validation set
- Sensitivity
- 85.61%
- Specificity
- 99.85%
Severe fractures, external validation set
- Sensitivity
- 93.46%
- Specificity
- 99.92%
Original publication
Using Artificial Intelligence to Diagnose Osteoporotic Vertebral Fractures on Plain Radiographs. J Bone Miner Res. Published 2 August 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →