Screening opportunistic osteoporosis through multimodal techniques of hip joint CT images: exploring 2D and 3D deep learning, radiomics, clinical data, and their integration - Report - MDSpire
Advertisement
Screening opportunistic osteoporosis through multimodal techniques of hip joint CT images: exploring 2D and 3D deep learning, radiomics, clinical data, and their integration
Clinical Report: Utilizing Multimodal Approaches for Osteoporosis Detection
Overview
This study investigates the use of hip joint CT imaging combined with deep learning and radiomics for opportunistic osteoporosis detection. The GradientBoosting model demonstrated the highest accuracy in screening.
Background
Osteoporosis is a significant public health concern, particularly among aging populations, due to its association with increased fracture risk. Current screening methods, such as dual-energy X-ray absorptiometry (DXA), are underutilized, leading to a need for alternative approaches. Opportunistic screening using hip CT could provide a cost-effective and less invasive method for early detection of osteoporosis.
Data Highlights
Model
Accuracy
AUC
GradientBoosting
0.849
0.911
densenet201 (2D)
0.817
0.884
ResNet34 (3D)
0.806
0.889
Key Findings
The study enrolled 567 patients for hip joint CT imaging analysis.
The GradientBoosting model outperformed other models in screening accuracy.
2D deep learning model densenet201 achieved an accuracy of 0.817.
3D deep learning model ResNet34 demonstrated an accuracy of 0.806.
Integration of radiomics and clinical data improved diagnostic reliability.
Clinical Implications
The findings indicate the potential for hip CT in opportunistic osteoporosis screening.
Conclusion
This study highlights the use of multimodal approaches in improving osteoporosis detection through hip CT imaging.
