Application Of CT-based Radiomics In Predicting EGFR Mutation Status And Subtypes In Non-small Cell Lung Cancer

Part One Application of CT-based Radiomics in Predicting EGFR Mutation Status in Non-small Cell Lung Cancer Objective: To investigate the application value of CT radiomics in the prediction of epidermal growth factor receptor(EGFR)mutant status in non-small cell lung cancer(NSCLC).Methods: A total of 779 patients with non-small cell lung cancer were included in this study(519 in the training set and 260 in the validation set).All patients underwent CT examination and EGFR gene test.Region of interest(ROI)was manually segmented in plain CT images and radiomics features were extracted.The least absolute contraction and selection operator(LASSO),random forest(RF)and sequential feature selection(SFS)are used to select key features.The support vector machine(SVM)model was constructed.By calculating the area under the receiver operating characteristic curve(AUC),the predictive efficacy of the clinical model,the radiomics model,and the combined model was compared.A Nomogram was established based on the radiomics scores and clinical variables,and the consistency and practicability of the model were evaluated by calibration curve and clinical decision curve.Results: In the training set,the AUC values of clinical model 1,radiomics model 2and the combined model 3 were: 0.75(95%CI,0.70-0.80)VS.0.73(95%CI,0.71-0.75)VS.0.85(95%CI,0.84-0.87);In the validation set,the AUC values of clinical model 1,radiomics model 2 and the combined model 3 were: 0.70(95% CI,0.68-0.71)VS.0.68(95% CI,0.67-0.70)VS.0.81(95% CI,0.79-0.83).Compared with clinical model and radiomics model,the combined model has relatively high predictive efficacy.Conclusion: CT radiomics have good predictive value for EGFR mutation status in non-small cell lung cancer patients,and can provide guidance for individualized molecular targeted therapy.Part Two Predicting EGFR Mutation Subtypes in Patients with Non-small Cell lung Cancerby CT RadiomicsObjective: To investigate the application value of CT radiomics in the prediction of epidermal growth factor receptor(EGFR)mutant subtypes in non-small cell lung cancer(NSCLC).Methods: A total of 287 patients with non-small cell lung cancer(201 in the training set and 86 in the validation set)were enrolled in this study,including 149 exon 19deletion(19 Del)mutations and 138 exon 21 L858 R missense(L858R)mutations.All patients underwent CT examination and EGFR gene test.Region of interest(ROI)was manually segmented in plain CT images and radiomics features were extracted.The least absolute contraction and selection operator(LASSO),random forest(RF)and sequential feature selection(SFS)are used to select key features.The support vector machine(SVM)model was constructed.By calculating the area under the receiver operating characteristic curve(AUC),the predictive efficacy of the clinical model,the radiomics model,and the combined model was compared.A Nomogram was established based on the radiomics scores and clinical variables,and the consistency and practicability of the model were evaluated by calibration curve and clinical decision curve.Results: In the training set,the AUC values of clinical model 1,radiomics model 2and the combined model 3 were: 0.70(95% CI,0.67-0.73)VS.0.69(95% CI,0.66-0.71)VS.0.76(95% CI,0.74-0.78);In the validation set,the AUC values of clinical model 1,radiomics model 2 and the combined model 3 were: 0.60(95% CI,0.88-0.62)VS.0.63(95% CI,0.61-0.65)VS.0.70(95% CI,0.68-0.72).Compared with the clinical model and the radiomics model,the predictive efficacy of the combined model is relatively high.Conclusion: CT radiomics have good predictive value for EGFR mutant subtypes in non-small cell lung cancer patients,and can provide guidance for individualized molecular targeted therapy.