Development of a prediction model for ascending aortic dilatation in patients with normally functioning bicuspid aortic valves using LASSO-Logistic regression
10.3760/cma.j.cn131148-20250710-00366
- VernacularTitle:基于LASSO-Logistic回归构建瓣膜功能正常的二叶式主动脉瓣患者升主动脉扩张预测模型
- Author:
Sijing HE
1
;
Ning YAN
;
Xue YANG
;
Lili WANG
;
Xingyue YANG
;
Lisha NA
Author Information
1. 宁夏医科大学第一临床医学院，银川 750004
- Publication Type:Journal Article
- Keywords:
Echocardiography;
Bicuspid aortic valve;
Ascending aortic dilatation;
Matrix- metalloproteinase;
LASSO-Logistic regression
- From:
Chinese Journal of Ultrasonography
2025;34(11):967-975
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To identify the risk factors for ascending aortic dilatation and develop a prediction model using LASSO-Logistic regression in patients with normally functioning bicuspid aortic valves（BAV）.Methods:Eight-four adult patients with BAV diagnosed as having normal valve function by transthoracic echocardiography who attended to the General Hospital of Ningxia Medical University between June 2024 and April 2025 were prospectively selected，there were 42 patients with ascending aortic dilatation and 42 patients without dilatation.The patients were divided into a training set（60 cases）and a test set（24 cases）using stratified random sampling at a ratio of 7 to 3 via the R caret package. All subjects underwent transthoracic echocardiography，lipid and plasma matrix metalloproteinases（MMP）and tissue inhibitors（TIMP）levels detection. The inverse probability of treatment weighting（IPTW）method was employed to control for potential confounding factors. LASSO and multifactorial binary Logistic regression was applied to screen the independent risk factors for ascending aortic dilatation of BAV and development a nomogram prediction model. The accuracy，consistency and clinical applicability of the prediction model were evaluated by applying the ROC curve，calibration curve，and decision curve analysis（DCA），respectively.Results:①After adjustment for IPTW，LASSO-Logistic regression analysis identified left ventricular global longitudinal strain（LVGLS）and plasma MMP-2 levels as independent risk factors for ascending aortic dilatation in BAV patients with normal valve function.②The nomogram prediction model constructed based on the above screening features，and the area under the curve（AUC）of ROC for the training set and test set were 0.917 and 0.903，with specificities of 0.867 and 0.917，and sensitivities of 0.933 and 0.916，respectively. ③Calibration curves demonstrated satisfactory alignment，with C-indices of 0.908（95% CI=0.879～0.937）for the training set and 0.903（95% CI=0.874 - 0.932）for the test set. Hosmer-Lemeshow goodness-of-fit tests indicated strong consistency between predicted and observed outcomes，with P-values of 0.138 and 0.750 for the training and test sets，respectively.The DCA curve demonstrated that within a threshold probability range of 0.04 - 0.90 in the test set，the clinical decision-making model provided a higher net benefit rate for patients with BAV. Conclusions:LVGLS absolute value reduction and elevated plasma MMP-2 levels are independent risk factors for predicting ascending aortic dilatation in BAV patients with normal valve function. The prediction model based on LASSO-Logistic regression has good predictive value，providing a scientific basis for clinical decision-making in patients with BAV and aortic diseases.
