Development and internal validation of a prediction model for new vertebral compression fractures after percutaneous kyphoplasty: a retrospective study

Abstract Background This study aimed to identify risk factors for new vertebral compression fractures (NVCFs) after percutaneous kyphoplasty (PKP) for osteoporotic vertebral compression fractures (OVCF) and to develop a prediction model and nomogram for postoperative risk stratification and clinical decision-making. Methods Statistical analyses were performed using SPSS 26.0 and R 4.2.1. Continuous variables were compared using the independent-samples t test or Mann–Whitney U test according to data distribution, and categorical variables were compared using the chi-square test or Fisher’s exact test. LASSO regression was used for variable selection, and multivariable logistic regression was applied to construct the prediction model. Model performance was evaluated using ROC curves, calibration curves, and decision curve analysis. Results A total of 610 eligible patients were included in this study and randomly divided into a training cohort and a validation cohort at a ratio of 7:3. In the training cohort, 75 of 427 patients developed postoperative NVCF, whereas 27 of 183 patients developed postoperative NVCF in the validation cohort. Based on LASSO regression and multivariable logistic regression analyses, age, BMD (T-score), and cement leakage were identified as independent predictors of postoperative NVCF. The nomogram demonstrated good discriminative performance, with an area under the curve (AUC) of 0.854 in the training cohort and 0.789 in the validation cohort. Calibration curves showed good agreement between the predicted and observed probabilities, and decision curve analysis indicated favorable clinical utility. Conclusions The LASSO-logistic regression nomogram showed good predictive performance for NVCFs after PKP and may provide a useful reference for identifying high-risk patients and postoperative risk management. Further prospective validation with more comprehensive variable collection is warranted before broader clinical application.