The Value Of CT-Based Radiomics Nomogram In Predicting Visceral Pleural Invasion In Lung Adenocarcinoma

Objective:To evaluate the value of CT-based radiomics nomogram in preoperative prediction of visceral pleural invasion(VPI)in lung adenocarcinoma.Materials and Methods:Retrospective analysis was performed on 183 patients with lung adenocarcinoma confirmed by surgical pathology in our hospital from January 2019 to December 2020,and the patients were randomly divided into training group(n=128)and validation group(n=55)in a ratio of 7:3.Three-dimensional volume of interest(VOI)was delineated layer by layer on high resolution CT lung window images.Minimum Absolute contraction and Selection operator(LASSO)regression was used to reduce dimension of features and Support Vector Machines(SVM)was used to construct radiomics model.Univariate logistic regression was used to screen the clinical risk factors,and the imaging label was incorporated,multivariate logistic regression was used to establish clinical-imaging radiomics nomogram to predict the visceral pleural invasion of lung adenocarcinoma.Receiver operating characteristics(ROC)curve was used to analyze the efficiency of the radiomics nomogram in the training group,and it was verified again by the validation group.Calibration curves were used to assess the consistency of the predicted and observed risk of invasion in the nomogram.Decision curve analysis(DCA)was used to evaluate the clinical net benefit of the nomogram.Results:Finally,8 features were screened out,and the radiomics score was constructed according to the corresponding coefficient of the features.The radiomics nomogram was constructed by the multi-variable logistic regression of the radiomics label and conventional CT image features(including lobar sign,intratumoral necrosis and pleural traction).The area under curve(AUC)of the training group and the validation group were 0.875(95%CI:0.814-0.935)and 0.865(95%CI:0.7688-0.963)respectively.Calibration curve analysis showed that the radiomics nomogram had good consistency between prediction and actual observation.The decision curve analysis(DCA)showed that the net benefit of the combined nomogram outperformed the clinical and radiomics feature models.Conclusions:CT-based radiomics nomogram have good application value in the prediction of preoperative visceral pleural invasion in patients with lung adenocarcinoma.