| Objective To establish a novel prediction model for osteoporotic vertebral compression fracture (OVCF) in postmenopausal women by combining the efficient assessment tools for bone, bone microstructure and paravertebral muscle. Methods The retrospective study collected 210 postmenopausal female patients who attended the Department of Spine Surgery of Shenzhen People's Hospital (the Second Clinical Medical College of Jinan University and the First Affiliated Hospital of Southern University of Science and Technology) from November 2020 to November 2022. The patients were randomly divided into model set and validator set, and the set was divided into non-OVCF group and OVCF group according to whether OVCF occurred. The statistically different parameters were screened out from the Modeling population through univariate statistical analysis for binary Logistic regression analysis and ROC curve comparison. According to the statistical results, the prediction model was established by incorporating sensitive parameters and the Nomogram plot was drawn. Finally, the C-index, calibration curve and decision curve analyses of the model construction set and the validator set were compared to verify the predictive ability of the model. Results Univariate statistical analysis showed that bone mineral density (BMD), quantitative computed tomography (QCT), vertebral body quality (VBQ), cross section area of psoas muscle (CSAPS), cross section area of erector spinae muscle and multifidus muscle (CSAES+MF), age, bone-specific alkaline phosphatase (BAP), body mass index (BMI) were associated with OVCF in postmenopausal women (P＜0.05). Binary Logistic regression analysis showed that BMD (OR=0.266, P=0.003), QCT (OR=0.965, P=0.008), VBQ (OR=4.346, P=0.044), CSAES+MF (OR=0.806, P=0.028) and CSAPS (OR=0.588, P=0.025) were independent risk factors for OVCF in postmenopausal women. According to ROC curve analysis, BMD (AUC=0.931, P＜0.001), QCT (AUC=0.890, P＜0.001), VBQ (AUC=0.784, P＜0.001), CSAES+MF (AUC=0.697, P＜0.001) and CSAPS (AUC=0.830, P＜0.001) all had a certain predictive effect on the occurrence of OVCF. The AUC of BMD, QCT and CSAPS was found to be higher (P＜0.05) through ROC curve comparison. Finally, the C-index of the Nomogram plot was 0.964 (0.935-0.993), and the C-index of the Validator group was 0.872 (0.802-0.941), and the calibration curve and decision-making curve results showed that the fit and clinical utility of the model were good. Conclusion In this study, we found that BMD, QCT, VBQ, CSAES+MF and CSAPS were independent risk factors for OVCF in postmenopausal women, and a Nomogram plot was drawn through the above parameters, and a new OVCF prediction model with good clinical utility and high prediction efficiency was established for postmenopausal women in China. |
