Multiparameter Dual-Layer Detector Energy CT Based Radiomics In Predicting The Expression Of PD-L1 In Non-Small Cell Lung Cancer

Objective This study aimed to explore the value of radiomics for predicting the expression of PD-L1 in non-small cell lung cancer based on Dual-layer detector energy CT(DLCT)parameter Images.Methods From October 2020 to August 2022,A total of 220 patients with pathologically confirmed non-small cell lung cancer who underwent DLCT chest enhancement scanning and PD-L1 expression testing were retrospectively enrolled from Gan Su province hospital between October 2020 and August 2022.Patients were divided into a training dataset(n=176)and testing dataset(n=44).Their conventional CT visuals and Spectral CT parameters were analyzed.The volume area of interest(VOI)was outlined by the Mono-E 40 ke V image.A total of 5185 radiomics features were extracted from the conventional CT image,Mono-E 40 ke V image,iodine density(ID)map,Z-effective map,and electron density map in artethe rial phase(AP)and venous phase(VP)respectively.The Pearson correlation coefficient method was used for feature dimension reduction,Recursive feature elimination(RFE)was used for feature selection,and the logistic regression(LR)and support vector machine(SVM)algorithm was used to build the radiomics model.The area under the curve(AUC)obtained from the receiver operating curve(ROC)was used to evaluate the prediction efficiency of the model.Results Finally,the 14 and 10 most important features were selected to establish two radiomics models to predict the PD-L1 expression in the negative-positive group and the lowhigh expression groups respectively.The combined model reached the highest performance(AUC-1%,0.917;AUC-50%,0.903)better than the radiomics(AUC-1%,0.886;AUC-50%,0.892),the clinical model(AUC-1%,0.551;AUC-50%,0.712),the CT visual model(AUC-1%,0.740;AUC-50%,0.0.616)and the Spectral CT parameters model model(AUC-1%,0.733;AUC-50%,0.611)in the testing dataset.Meanwhile,the AUCs of the multiple parameters imagins were higher than any single-parameter models(the AUCs range from 0.733 to 0.868)in the testing dataset.Conclusion The radiomics model based on the multi-parameter images of DLCT can predict the expression level of PD-L1 in non-small cell lung cancer.The combination of the clinical model and radiomic model can obtain higher prediction efficiency and provide good proof for clinical immunotherapy and prognosis evaluation.