Automated osteoporosis classification and T-score prediction using hip radiographs via deep learning algorithm

“`html

Automated Osteoporosis Classification and T-Score Prediction

Practical Solutions and Value

• Using a deep learning algorithm, this study aims to classify osteoporosis and predict T-scores on the hip region through simple hip radiographs.
• A dataset of 3460 hip images from 1730 patients was used to train and test the proposed model, which showed a better performance in predicting osteoporosis compared to existing models.
• When combined with variables such as age, body mass index, and sex, the model demonstrated high consistency in T-score prediction compared to dual-energy X-ray absorptiometry (DXA).
• The proposed CNN model may provide high-accuracy identification of osteoporosis and T-score prediction, offering a cost-effective and adaptable solution for population-based osteoporosis screening.

“`