Radiomics approach for preoperative identification of stages Ⅰ-Ⅱand Ⅲ-Ⅳ of esophageal cancer

作     者：Lei WU Cong Wang Xianzheng Tan Zixuan Cheng Ke Zhao Lifen Yan Yanli Liang Zaiyi Liu Changhong Liang 

作者机构：School of Medicine South China University of Technology Guangzhou 510006 China Department of Radiology Guangdong General HospitalGuangdong Academy of Medical Sciences Guangzhou 510080 China School of Automation Science and Engineering South China University of Technology Guangzhou 510641 China 

出 版 物：《Chinese Journal of Cancer Research》 (中国癌症研究（英文版）)

年 卷 期：2018年第30卷第4期

页      面：396-405页

核心收录：

学科分类：1002[医学-临床医学] 1001[医学-基础医学(可授医学、理学学位)] 100214[医学-肿瘤学] 100106[医学-放射医学] 10[医学] 

基　　金：supported by the National Key R&D Program of China (No. 2017YFC1309100) National Natural Scientific Foundation of China (No. 81771912) Science and Technology Planning Project of Guangdong Province (No. 2017B020227012) 

主　　题：Esophageal cancer tumor staging diagnostic imaging tumor volume 

摘      要：Objective： To predict preoperative staging using a radiomics approach based on computed tomography （CT）images of patients with esophageal squamous cell carcinoma （ESCC）.Methods： This retrospective study included 154 patients （primary cohort： n： t 14; validation cohort： n：40） withpathologically confirmed ESCC. All patients underwent a preoperative CT scan from the neck to abdomen. Highthroughput and quantitative radiomics features were extracted from the CT images for each patient. A radiomicssignature was constructed using the least absolute shrinkage and selection operator （Lasso）. Associations betweenradiomics signature, tumor volume and ESCC staging were explored. Diagnostic performance of radiomicsapproach and tumor volume for discriminating between stages Ⅰ-Ⅱand Ⅲ-Ⅳ was evaluated and compared usingthe receiver operating characteristics （ROC） curves and net reclassification improvement （NRI）.Results= A total of 9,790 radiomics features were extracted. Ten features were selected to build a radiomicssignature after feature dimension reduction. The radiomics signature was significantly associated with ESCCstaging （P〈0.001）, and yielded a better performance for discrimination of early and advanced stage ESCC comparedto tumor volume in both the primary [area under the receiver operating characteristic curve （AUC）： 0.795 vs. 0.694,P=0.003; NRI=0.424）] and validation cohorts （AUC： 0.762 vs. 0.624, P=0.035; NRI=0.834）.Conclusions： The quantitative approach has the potential to identify stage Ⅰ-Ⅱand Ⅲ-Ⅳ ESCC beforetreatment.