To Construct and Verify the Prediction Model of Low-dose CT Parameters on the Invasiveness of Pulmonary Mixed Glass Nodules
Objective The prediction model of lung mixed glass nodule invasiveness was established based on quantitative parameters of low-dose CT,and the prediction efficiency of the model was verified.Methods Retrospective study was performed on 102 patients with pathologically proven pulmonary mixed ground glass nodule admitted to our hospital from March 2021 to March 2023,who were divided into invasive adenocarcinoma(IA)group(n=71)and non-IA group(n=31)according to pathological results and classification criteria of lung adenocarcinoma.All patients received low-dose CT scan after admission,and quantitative parameter analysis was performed to compare the maximum diameter,volume,maximum cross-sectional area,maximum vertical length diameter,nodule mass and average CT value of the two groups.Multivariate Logistic regression model was used to analyze the factors affecting the invasiveness of pulmonary mixed glass nodules.Based on the results of regression analysis,a nomogram model was constructed,which was internally verified by receiver operating characteristic(ROC)curve and Hosmer-Lemeshow test to analyze the predictive efficiency of the model.Results Nodule maximum diameter,nodule maximum cross-sectional area,nodule maximum vertical length diameter,nodule mass and nodule volume in group IA were higher than those in non-IA group(P＜0.05).Multivariate Logistic regression analysis indicated that the increase of maximum nodule diameter(OR:2.675,95％CI:1.298-5.514)and maximum cross-sectional area(OR:2.380,95％CI:1.435-3.946)were independent risk factors for invasivity of pulmonary mixed glass nodules(P＜0.05).Based on the results of multi-factor analysis,a nomogram model is constructed.The calibration curve and decision curve indicate that the model is reliable.The ROC curve indicated that the area under the curve of the model was 0.902(standard error:0.029,P＜0.001,95％CI:0.845-0.958),sensitivity was 80.28％,and specificity was 80.65％.The Hosmer-Lemeshow test indicated a good fit(x2=2.596,P=0.581).Conclusion The maximum diameter and cross-sectional area of nodules in low-dose CT quantitative parameters are correlated with the invasiveness of pulmonary mixed glass nodules.Based on the above parameters,a model can be built to predict and evaluate the invasiveness of pulmonary mixed glass nodules.
Lung Mixed Glass NodulesInvasiveLow-dose CTQuantitative ParametersPrediction Model
万方数据
维普
