CT-based Radiomics Analysis For Prediction Of Lymph Node Metastasis In Patients With Non-small Cell Lung Cancer

Lymph node(LN)metastasis in patients with non-small cell lung cancer(NSCLC)is one of the most important clinical indicators that affect the therapeutic effect and prognosis,and also play an important role in the TNM staging of lung cancer.Accurate identification of LN involvement is crucial for prognosis and treatment strategy decision in patients with NSCLC.Imaging modalities like computed tomography(CT)are the most widely used for preoperative work-up,and are very important in detecting LN enlargement;however,they have limitations for their predictive value to differentiate benign nodes from malignant ones.Recent advances in radiomics,which extract quantitative descriptors from routinely acquired medical images noninvasively,has provided deep insights into different fields of personalized medicine in oncologic practice,including tumor detection,subtypes classification,and therapy response assessment.In view of preoperatively detecting the LN metastases in NSCLC,the author has carried out the study of radiomics analysis and its clinical applications.Based on the contrast-enhanced chest CT images,the massive "Radiomics features" were extracted.Then,combining with a number of clinical factors,the individualized prediction of LN metastasis in the patients was taken.The performance of prediction model was showed satisfactory.This study has successfully developed an radiomics-based prediction model,which could serve as an easy-to-use tool to facilitate the preoperatively individualized prediction of LN metastasis in patients with NSCLC.In the process of carrying out the relevant research work,the author of this paper as the first author(including co-first author)has published 5 articles in the journal of clinical oncology and medical imaging,and another paper in the Chinese core journals.The main contents of this paper are as follows:The first issue is to extract the radiomics features based on the contrast-enhanced CT images scan.The author extracted 591 radiomics features totally.The extracted features were studied by the author in other cancer researches.The extracted gray-level histogram features and gray-level co-occurrence matrix features have been showed satisfactory predictive performance in the diagnostic and prognostic.Secondly,in order to build and validate a method of radiomics features selection for the dimensionality reduction and feature selection of data,so as to construction of radiomics signature,which can be used to predict LN metastasis of NSCLC.A radiomics-based signature was constructed by using the least absolute shrinkage and selection operator method logistic regression analysis in our study.The radiomics signature was independently associated with LN metastasis.Finally,in order to build and validate a radiomics-based predictive model for the prediction of preoperative LN metastasis in NSCLC.The multivariable logistic regression analysis was performed,and a prediction model was derived.The performance and usefulness of built model were assessed.The predictive model can be potentially applied in the individual preoperative prediction of the LN metastasis status in NSCLC patients.The nomogram was built based on the predictive model,which showed good calibration and prediction performance.This nomogram could serve as an easy-to-use tool to facilitate the preoperatively individualized prediction of LN metastasis in patients with NSCLC.