Affiliation:
1. Department of Radiology, Xiaogan Hospital Affiliated to Wuhan University of Science and Technology, Xiaogan, PR China
2. Department of Orthopedics, Xiaogan Hospital Affiliated to Wuhan University of Science and Technology, Xiaogan, PR China
Abstract
Background We aimed to establish a novel model using a radiomics analysis of magnetic resonance (MR) images for predicting osteoporosis. Purpose To investigate the effectiveness of a radiomics approach utilizing magnetic resonance imaging (MRI) of the lumbar spine in identifying osteoporosis. Material and Methods In this retrospective study, a total of 291 patients who underwent MRI were analyzed. Radiomics features were extracted from the MRI scans of all 1455 lumbar vertebrae, and build the radiomics model based on T2-weighted (T2W), T1-weighted (T1W), and T2W + T1W imaging. The performance of the combined model was assessed using metrics such as the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. The AUCs of these models were compared using the DeLong test. Their clinical usefulness was assessed using a decision curve analysis. Results T2W, T1W, and T1W + T2W imaging retained 27, 27, and 17 non-zero coefficients, respectively. The AUCS about radiomics scores based on T2W, T1W, and T1W + T2W imaging were 0.894, 0.934, and 0.945, respectively, which all performed better than the clinical model significantly. The rad-signatures based on T1W + T2W imaging, which exhibited a stronger predictive power, were included in the creation of the nomogram for osteoporosis diagnosis, and the AUC was 0.965 (95% confidence interval (CI)=0.944–0.986) in the training cohort and 0.917 (95% CI=0.738–1.000) in the test cohort. The calibration curve indicated that the radiomics nomogram had considerable clinical usefulness in prediction, observation, and decision curve analysis. Conclusion A reliable and powerful tool for identifying osteoporosis can be provided by the nomogram that combines the T1W and T2W imaging radiomics score with clinical risk factors.