Deep learning for osteoporosis screening using an anteroposterior hip radiograph image
Artit Boonrod1, Prarinthorn Piyaprapaphan1, Nut Kittipongphat1
1Department of Orthopedics, Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand.
Summary
A new deep learning model can screen for osteoporosis using hip X-rays. This tool may help identify patients needing further bone mineral density (BMD) tests.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Orthopedics
Background:
- Osteoporosis significantly increases fracture risk and mortality.
- Bone mineral density (BMD) measurement is the standard diagnostic tool but is costly and requires specialized equipment.
- Plain radiographs are inexpensive and accessible but not currently used for diagnosis.
Purpose of the Study:
- To develop and evaluate a deep learning model for osteoporosis diagnosis using anteroposterior hip radiograph images.
- To assess the diagnostic accuracy of the developed deep learning model.
Main Methods:
- Retrospective collection of 363 anteroposterior hip radiograph images from 2013-2021.
- Exclusion of images with BMD measurements older than two years.
- Training five deep learning models on 80% of images and evaluating performance on the remaining 20%, with the best model selected by AUC.
Main Results:
- The best deep learning model achieved an Area Under the Curve (AUC) of 0.91.
- The model demonstrated an accuracy of 0.82 in diagnosing osteoporosis.
- A total of 363 images were analyzed, with 150 classified as osteoporosis (T score ≤ -2.5).
Conclusions:
- Deep learning shows promise for osteoporosis screening via hip radiographs.
- The developed model could serve as a cost-effective screening tool to identify at-risk individuals for subsequent BMD testing.
