Osteoporosis prediction using lumbar CT Hounsfield units: comparative performance and clinical implications of seven machine learning models

Key Points

• This research aims to evaluate the performance of various machine learning models for predicting osteoporosis using lumbar CT Hounsfield units.
• Evaluated seven machine learning models including KNN, Logistic Regression, and SVM.
• Used L1-L4 Hounsfield units from CT scans for analysis.
• Assessed diagnostic performance comparing accuracy and classification methods.
• KNN demonstrated the most robust and balanced diagnostic performance.
• Logistic Regression and SVM provided stable classification thresholds.
• Machine learning models show potential in osteoporosis screening where DXA is unavailable.