Screening of opportunistic osteoporosis in patients with hip joint injury by radiomics analysis of chest CT scanning

The objective of our study is to create and validate radiological features based on chest CT scans, which can be used to identify osteoporosis in these patients. This study enrolled a total of 449 patients with hip joint fractures, who were randomly allocated into a training set and a testing set. The patients' baseline information was utilized to develop clinical information models. Additionally, radiomics technology was employed to extract radiomics features from chest CT images for building radiomics models. Ultimately, the two sets of models were merged to create the Nomogram model, which was utilized for the diagnosis of osteoporosis. The efficacy of all three models in diagnosing osteoporosis was assessed and compared. A total of 17 radiomics features were selected for the development of radiomics models, while gender, age, and HU values were utilized for constructing clinical information models. The resulting Nomogram model, formed by merging the two sets of models, exhibited excellent diagnostic performance for osteoporosis. Using radiological features extracted from chest CT scans of the thoracic spine, hip fracture patients with or without osteoporosis can be differentiated before undergoing surgery.