
Construction and Verification of a Predictive Model for the Progression of Aortic Valve Calcification
Abstract
Background: The primary objective of this study is to develop and validate a predictive model assessing the likelihood of disease progression in individuals with aortic valve calcification (AVC).
Methods: For the second and third visits, 2,533 patients were followed up. They were randomly assigned to a train set and a validation set at a ratio of 7:3. After employing the Least Absolute Shrinkage and Selection Operator (LASSO) and multiple Cox regression to filter predictors, the selected variables were input into the Cox proportional risk model for model construction. Calibration curve, Consistency Index (C-index), Receiver Operating Characteristic (ROC) curve, and Decision Curve Analysis (DCA) were employed to validate the model. Patients were categorized into low- and high-risk groups based on the model’s predicted risk score, and survival analysis was conducted using Kaplan-Meier (K-M) plots. An online platform was used to enhance the clinical utility.
Results: The incidence of AVC progression was 9.63%. LASSO-Cox regression analysis identified seven variables significantly correlated with AVC progression. In both the training and validation sets, the Area Under the Curve (AUC) and C-index of the prediction model exceeded 0.8. The calibration curve aligned closely with the diagonal line. Decision Curve Analysis (DCA) underscored the clinical application value of the model. Survival analysis demonstrated a significantly higher progression rate in the high-risk group compared to the low-risk group. The online platform visualized the probability of progression.
Conclusion: The developed predictive model has proven reliability and accuracy in forecasting the 2-, 3-, and 4-year progression rates of patients with AVC. It offers a dependable framework for estimating progression and facilitating individualized comprehensive prevention strategies for individuals with AVC.
© 2025 Zhen Guo, Zhenyu Xiong, Chaoguang Xu, Jingjing He, Shaozhao Zhang, Rihua Huang, Menghui Liu, Jiaying Li, Xinxue Liao, Xiaodong Zhuang, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.