A Hierarchical Opportunistic Screening Model for Osteoporosis Using Deep Learning

Osteoporosis has become the third most chronic disease worldwide. More than 200 million patients worldwide suffer from osteoporosis, with an average of one fragility fracture occurring every 3 seconds. In view of the above, we propose a framework for osteoporosis screening, which first uses deep learning segmentation network to segment computed tomography (CT) images, and then uses CT images, laboratory indicators, and demographic characteristics to classify osteoporosis based on deep learning multimodal fusion classification network, which can effectively realize the auxiliary screening of osteoporosis.