CT-Based Bone Habitat Radiomics for Predicting Risk of Vertebral Fracture in Older Adults: A Longitudinal Study.
Authors
Affiliations (3)
Affiliations (3)
- Department of Radiology, the Affiliated Hospital of Nanjing University of Chinese Medicine, Nanjing, China.
- Department of Orthopaedics, the Affiliated Hospital of Nanjing University of Chinese Medicine, Nanjing, China.
- Department of Orthopaedics, University of Rochester School of Medicine, Rochester, NY, USA.
Abstract
Study design/settingRetrospective longitudinal study.PurposeOsteoporotic vertebral fractures (OVF) are common in middle-aged and elderly populations. However, few studies have predicted the risk of OVF from the perspective of the bone heterogeneity. This study conducted a longitudinal study to predict the risk of OVF in individuals over 50Â years old based on habitat radiomics which can quantify heterogeneity of vertebral trabecular bone.MethodsIndividuals aged over 50Â years who had not experienced OVF and underwent CT scans between 2016 and 2023 were enrolled and followed up until 2024. During the follow-up period, 107 cases developed new OVF, and 270 individuals without fractures were selected as the control group. Radiomic features of each pixel within the vertebra were extracted, and the optimal segmentation of vertebral sub-regions was determined using the K-means unsupervised clustering method.ResultsThe habitat radiomics model significantly outperformed the CT value model (AUC = 0.702, DeLong test <i>P</i>-value = 0.001) and also surpassed the traditional radiomics model. The Cox proportional hazards analysis showed that the habitat radiomics risk score could serve as an independent predictor of vertebral fractures (hazards ratio = 1.092, 95% confidence interval (CI): 1.074 - 1.111, <i>P</i> < 0.001). The C-index of the habitat radiomics nomogram model was 0.803 in the training set (95% CI: 0.752 - 0.854) and 0.748 in the validation set (95% CI: 0.667 - 0.829).ConclusionThe habitat radiomics model can predict vertebral fractures based on vertebral heterogeneity, with better performance than traditional bone density prediction methods.