Osteoporosis is a significant public health concern, particularly among the elderly, as it leads to decreased bone density and a substantially increased risk of fractures [1]. However, identifying osteopenia—an intermediate stage before progression to osteoporosis—at an early stage is crucial as it predisposes individuals to fractures if left undiagnosed and untreated [2]. Early detection of osteopenia allows for timely intervention and effective fracture prevention. Despite its importance, the gold standard for diagnosing osteoporosis and osteopenia, dual-energy X-ray absorptiometry (DXA), has limited accessibility due to high costs and restricted availability in many healthcare settings. As a result, a substantial portion of at-risk individuals remain undiagnosed, leading to preventable fractures and associated morbidity.

To improve early detection of osteopenia, we developed a deep learning model that utilizes readily available posteroanterior (PA) chest X-ray (CXR) images to categorize bone health into three classes: normal, osteopenia, and osteoporosis.. CXRs are one of the most commonly performed radiographic examinations, offering a more accessible and cost-effective alternative for opportunistic bone health assessment. Our approach leverages a three-class classification framework trained on CXR images labeled with corresponding DXA-derived bone mineral density (BMD) categories. By integrating advanced deep learning techniques, our model aims to accurately differentiate between normal bone density, osteopenia, and osteoporosis, enabling early screening and risk stratification.