Evaluation Of The Efficacy Of EGFR-TKI-targeted Drugs In The Treatment Of Lung Adenocarcinoma By CT Radiomics

Purpose: In this study,a non-invasive radiomics method was used to extract a large amount of valuable information from patient CT images(including pre-treatment and follow-up images),and the image features were effectively combined with treatment effect evaluation,aiming to solve the problem of pulmonary glands in patients with EGFR mutation positive.The early efficacy evaluation of cancer EGFR-TKI targeted drugs provides a scientific basis for clinical drug selection and adjustment of treatment plans.Objective: To extract texture feature information from CT images(including pre-treatment and follow-up images)of patients with lung adenocarcinoma treated with EGFR-TKI targeted drugs by non-invasive radiomics method,and to evaluate the application value of CT radiomics features in the early curative effect of EGFR-TKI targeted drugs on EGFR mutation positive lung adenocarcinoma.Materials and Methods: The imaging and clinical data of 106 patients with lung adenocarcinoma treated with EGFR-TKI targeted therapy(EGFR mutation positive)in Shaoxing People’s Hospital were retrospectively analyzed,and they were divided into effective group and ineffective group according to the criteria of RECIST of curative effect of solid tumor.All cases were randomly divided into training group and verification group according to the ratio of 7: 3.Collected the CT plain scan images of patients one week before treatment and two months after treatment,used ITK-SNAP software to segment the images into regions of interest,imported the segmented images into AK software(GE Company)to extract the image omics features,used Lasso algorithm to filter the features in the training set,and conducted machine learning(logistic regression,decision tree,Ada Boost and support vector machine)to build the prediction model and carry out internal verification.Then the nomogram is constructed and the calibration degree of the model is evaluated,and the clinical benefit is evaluated by using the decision curve analysis method.Results: After dimension reduction by Lasso,three features with non-zero coefficients were extracted between groups,namely,sum entropy,Longgrunlowgreylevelemphassisangle0＿offset1,Longgrunlowgreylevelemphasis＿angl-e0＿offset4;There were no significant differences in gender and age groups(P > 0.05),but there were significant differences in long diameter before treatment and 2 months after treatment(P<0.05).Through comprehensive analysis,the logistic regression model has the best performance,with AUC of training group and verification group being 0.778 and 0.773,sensitivity being 0.83 and 0.67,specificity being 0.86 and 0.80,positive predictive value being 0.85 and 0.77,negative predictive value being 0.83,0.71 and accuracy being 0.84 and 0,respectively.Nomogram chart shows that the total score is 0～130 with the corresponding effective risk of 0.1 ～ 0.9 after three imaging omics parameters are assigned.The calibration curve shows that the predicted value is close to 45 diagonal line with the actual clinical observation value.The decision curve shows that the imaging omics model has good clinical benefits in evaluating the early curative effect of EGFR-TKI targeted drugs for lung adenocarcinoma.Conclusion: The radiomics model based on the imageomics characteristics of CT plain scan images before and after treatment has good performance in evaluating the efficacy of EGFR-TKI targeted drugs in the treatment of lung adenocarcinoma,and the Nomogram map constructed can quantitatively evaluate the efficacy,providing intuitive and reliable scientific reference for individualized treatment of patients with advanced lung adenocarcinoma treated with EGFR-TKI targeted drugs.