Prediction Of Lymph Node Metastasis In Clinical Stage T1 Lung Adenocarcinomas Based On Radiomics

Part 1 The Predictive Value of CT Imaging Features for lymph Node Metastasis in Clinical Stage T1 Lung Adenocarcinomas ?Objective?To investigate the relationship between lymph node(LN)metastasis and conventional CT imaging features for stage T1 peripheral lung adenocarcinoma patients.?Methods and meterials?From July 2012 to March 2017,366 patients meeting the criteria were enrolled into this study.The morphology features(maximum diameter,density,spiculation,vacuole,etc.)of 242 lesions in the training set were analyzed by two thoracic radiologists,who were blinded to the clinicopathological variables.Then,the morphological characteristics of P < 0.05 are selected by using the single variable analysis algorithm,and the logistic regression method is used to establish the model.The area under the ROC curve(AUC)is used to evaluate the prediction performance of the model.Their discrimination performances were assessed using receiver operating characteristic(ROC)curve analysis and quantified by the area under the ROC curve(AUC).?Results?In this study,the total lymph node metastasis rate of the enrolled patients was 20.2%.After the feature screening by Man-whiteney U test for each feature,there were four features with statistical significance,including maximal diameter of tumor,spiculation,pleural retraction,lesion attenuation(p < 0.01).The logistic regression analysis identified the four features as independent predictors.In addition,multivariable logistic regression was used to build a model based on four morphology features.The model was tested on the validation cohort and obtained an AUC of 0.796.?Conclusion?Tumor size,spiculation,pleural traction and lesion attenuation are identified as independent risk factors for lymph node metastasis.It is important to predict the preoperative lymph node metastasis and treatment options for stage T1 lung adenocarcinoma.Part 2 The Predictive Value of Intratumoural Radiomics Signatures for Lymph Node Metastasis Status in Clinical Stage T1 Lung Adenocarcinomas?Objective?To evaluate the value of radiomics model on CT images in tumor area for preoperative predicting of lymph node(LN)metastasis in clinical stage T1 lung adenocarcinoma patients.This study aimed to develop and validate a radiomics signature of tumor area based on CT images for preoperative prediction of lymph node metastasis in clinical stage T1 lung adenocarcinoma patients.?Materials and methods?Ethical approval by the institutional review board was obtained for this retrospective analysis.From July 2012 to March 2017,a total of 366 patients with stage T1 lung adenocarcinomas were retrospectively evaluated.1946 quantitative 3D features of 366 pathologically confirmed patients were exacted from the chest CT plain image.For each patient,the 3D ROIs were defined for the tumor(GTV,gross tumor volume).The GTV radiomic features were extracted using Py Radiomics.The m RMR(maximum correlation and minimum redundancy)is used to rank the importance of features,and the Lasso regression model is used for feature selection and the establishment of radiomics features.The receiver operating characteristic(ROC)curve analysis was evaluated in terms of its calibration and validated in an independent validation cohort with 124 lung adenocarcinomas.?Results?Of the 366 clinical T1 non–small cell lung cancer,295 had pathological N0 Status and 71 had pathological lymph node metastasis.1946 3D features were extracted,and 3 most important features were selected as the most important discriminators to build the radiomics signatures.The radiomics-based nomogram demonstrated good calibration in the primary and validation cohort,and showed improved differential diagnosis performance with an AUC of 0.829 in the independent validation cohort.?Conclusion?Radiomic features of GTV have a good prediction of the risk of lymph node metastasis in patients with clinical stage T1 peripheral lung adenocarcinoma.Patients with clinical stage T1 lung adenocarcinoma will benefit from the proposed radiomics method.Part 3 Radiomics of the intratumoural and peritumoral lung parenchyma on CT: Preoperative Prediction of Lymph Node Metastasis in clinical stage T1 lung Adenocarcinomas?Objective?To evaluate the efficiency of radiomics model on CT images of intratumoral and peritumoral lung parenchyma for preoperative prediction of lymph node(LN)metastasis in clinical stage T1 peripheral lung adenocarcinoma patients.?Materials and methods?366 peripheral lung adenocarcinoma patients with clinical stage T1 were evaluated by 5 CT scanners.For each patient,two volumes of interest(VOI)on CT were defined as the gross tumor volume(GTV)and the peritumoral volume(PTV,1.5cm around the tumor)respectively.1946 radiomic features were obtained from each VOI and then refined by reproducibility and redundancy.The refined features were investigated for usefulness in building radiomic signatures by the m RMR feature-ranking method and LASSO classifier.Multivariable logistic regression analysis was used to develop a radiomic nomogram incorporating the radiomic signature and clinical parameters.The prediction performance was evaluated on the validation cohort.?Results?The radiomic signatures using the features of GTV and PTV showed good ability in predicting LN metastasis with an AUC of 0.829(95%CI,0.745-0.913),0.825(95%CI,0.733-0.918),respectively.There is no significant difference of AUC between the two groups.By incorporating the features of GTV and PTV,the AUC of radiomic signature increased to 0.843(95%CI,0.770-0.916).The AUC of radiomic nomogram was 0.869(95%CI,0.800-0.938).?Conclusions?Radiomic signatures of GTV and PTV both have a good prediction ability in the prediction of LN metastasis.The proposed nomogram can be conveniently used to facilitate the preoperative prediction of LN metastasis in T1 peripheral lung adenocarcinomas.