Diagnostic model for distinguishing fresh or old osteoporotic vertebral compression fractures based on modified computed tomography window: a retrospective cohort study.

Objective: To develop and validate a Computed Tomography (CT) -based nomogram for distinguishing fresh from old osteoporotic vertebral compression fractures (OVCFs), with magnetic resonance imaging (MRI) as the reference standard.

Methods: In this retrospective study, OVCF patients from The First Affiliated Hospital of Chongqing Medical University (August 2022-December 2023) were analyzed. Modified CT window parameters (width: 400; level: 200) were applied to quantify vertebral features, including CT values, height reduction, endplate integrity, trabecular sparsity, Schmorl's nodes, and high-density shadows. Predictive variables were selected via univariate and multivariate logistic regression, followed by nomogram construction. Model performance was assessed using receiver operating characteristic (ROC) curves (area under the curve, AUC), calibration plots, Hosmer-Lemeshow testing, and decision curve analysis (DCA). 10-fold cross-validation was used to evaluate the generalization performance of the model.

Results: The nomogram incorporated seven imaging biomarkers, achieving AUCs of 0.941 (training cohort) and 0.974 (validation cohort). The calibration accuracy was assessed using the Hosmer-Lemeshow test (χ²=3.30, P = 0.95). DCA demonstrating significant clinical utility across probability thresholds. In cross-validation, the mean AUC on the validation sets was 0.911 ± 0.080 (mean ± standard deviation).

Conclusions: The proposed CT-based nomogram achieved adequate performance in identifying fresh and old OVCFs.