Optimizing osteoporosis pre-screening (OOPS) through AI-driven models and validation in the Asian population.

BACKGROUND

Osteoporosis is a major concern in public health, leading to an increased number of fractures, particularly in older adults. Early screening is essential for timely intervention and subsequent management. This study introduces Optimizing Osteoporosis Pre-Screening (OOPS), an AI-based Osteoporosis pre-screening tool, designed to provide healthcare professionals with a decision support system (DSS) that can help in the early detection of Osteoporosis. OOPS was validated against Dual-energy X-ray Absorptiometry (DXA) scan, the gold standard for Osteoporosis diagnosis, worldwide.

METHODOLOGY: This cross-sectional study included 1100 participants (550 males and 550 females) after informed consent. Demographic data were noted on a questionnaire-based study proforma, followed by a DXA scan of the left femur to diagnose Osteoporosis. Machine learning (ML) techniques were applied to select the most relevant features from the obtained data to make a predictive AI model, OOPS. Three feature selection techniques/ methods were used to simplify the model, followed by the implementation of five ML models, which were evaluated for performance based on seven metrics.

RESULTS

The CatBoost (CB) model, combined with Random Forest (RF) feature selection, established the highest performance, achieving an Area Under the Curve (ROC/AUC) of 0.98, sensitivity of 0.95, specificity of 0.85, and accuracy of 0.93. Additional metrics, including F1-Score, Positive Predictive Value (PPV), and Negative Predictive Value (NPV), were 0.97, 0.95, and 0.96, respectively.

CONCLUSION

The Optimizing Osteoporosis Pre-Screening (OOPS) tool, integrated with its Decision Support System (DSS), is an effective and reliable tool for early Osteoporosis screening. Its strong performance and simple interface make it particularly beneficial in settings with limited access to advanced diagnostic technologies such as DXA. This study highlights the potential of AI-driven tools to enhance public health outcomes by enabling early identification and management of Osteoporosis. This may help in reducing the fracture-related morbidity in aging populations.