Development and internal validation of a clinical-radiological nomogram for osteoporosis screening: A cohort retrospective study

Retrospective cohort study. The goal of this study was to identify the clinical and radiological characteristics of patients with osteoporosis and to develop a practical clinical prediction model for patients for accurately predicting the risk of osteoporosis. This study included 954 patients from September 2020 to September 2024 at our hospital. Independent risk factors were selected by the least absolute shrinkage and selection operator method (LASSO) regression. Then, a prediction model (nomogram) was established. Randomly split internal validation cohorts were used to test the nomogram model's calibration, discrimination, and clinical utility. Six independent prediction factors, age, female, glucocorticoid use, chronic obstructive pulmonary disease (COPD), cut-off values for Hounsfield unit (HU) and vertebral quality (VBQ) scores, were identified, and based on this a nomogram model was developed for predicting patient prognosis. The C-index of the prediction nomogram was 0.86 in training set. The area under the receiver operating characteristic curve (AUC) was 0.87 in both the training and validation sets. The model has good practicability for clinics according to the decision curve analysis (DCA) and clinical impact curve (CIC). The nomogram model has good predictive performance and clinical practicability, which could provide a certain basis for simplifying osteoporosis diagnosis.