Opportunistic Screening for Osteoporosis Using Hand Radiographs: A Preliminary Study

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Low bone mineral density (BMD) increases fracture risk but is often undiagnosed.
• Opportunistic screening for low BMD is needed in patients undergoing other imaging studies.

Purpose of the Study:

• To evaluate the efficacy of a deep learning (DL) framework using hand radiographs for opportunistic screening of osteoporosis and osteopenia.
• To correlate textural analysis of hand bones with dual-energy X-ray absorptiometry (DXA) measurements.

Main Methods:

• Retrospective analysis of 812 patients (age ≥50) with concurrent hand radiographs and DXA scans.
• Development and validation of a DL model to predict osteoporosis/osteopenia from hand radiographs.
• Random splitting of data into training/validation (n=533) and testing (n=136) sets.

Main Results:

• The DL model achieved 82.00% accuracy, 87.03% sensitivity, 61.00% specificity, and an AUC of 74.00% for detecting osteoporosis/osteopenia.
• Significant correlations were found between bone textural analysis and DXA measurements.

Conclusions:

• Hand radiographs can serve as an effective screening tool for osteoporosis and osteopenia.
• This AI-driven approach can identify patients who warrant formal DXA evaluation, potentially reducing undiagnosed cases and fracture risk.