Yue Wang1, Wei Liu1, Yang Yu2, Wei Han3, Jing-Juan Liu1, Hua-Dan Xue1, Jing Lei1, Zheng-Yu Jin4, Jian-Chun Yu5. 1. Department of Radiology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, No. 1, Shuaifuyuan, Dongcheng District, Bejing, 100730, People's Republic of China. 2. CT Collaboration, Siemens Healthineers Ltd, 59# Beizhan Road, Shenyang, 110013, People's Republic of China. 3. Department of Epidemiology and Biostatistics, Institute of Basic Medical Sciences, School of Basic Medicine, Chinese Academy of Medical Sciences, Peking Union Medical College, No. 5, Dongdansantiao Street, Beijing, 100005, People's Republic of China. 4. Department of Radiology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, No. 1, Shuaifuyuan, Dongcheng District, Bejing, 100730, People's Republic of China. jinzy@pumch.cn. 5. Department of General Surgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, No. 1, Shuaifuyuan, Dongcheng District, Beijing, 100730, People's Republic of China. yu-jch@163.com.
Abstract
OBJECTIVE: The purpose of the study was to investigate the role of CT radiomics for the preoperative distinction of intestinal-type gastric adenocarcinomas. MATERIALS AND METHODS: A total of 187 consecutive patients with preoperative contrast CT examination and pathologically proven gastric adenocarcinoma were retrospectively collected. Patients were divided into a training set (n = 150) and a test set (n = 37). Arterial phase (AP), portal phase (PP), and delay phase (DP) images were retrieved for analysis. A dedicated postprocessing software was used to segment the lesions and extract radiomics features. Random forest (RF) algorithm was applied to construct the classifier models. A nomogram was developed by incorporating multiphase radiomics scores. Receiver operating characteristic (ROC) curves were used to evaluate the performance of the radiomics model and nomogram in both sets. RESULTS: The radiomics model showed a favorable capability in the distinction of intestinal-type gastric adenocarcinomas. The areas under curves (AUCs) of the AP, PP, and DP radiomics models were 0.754 (95% CI: 0.676, 0.820), 0.815 (95% CI: 0.744, 0.874), and 0.764 (95% CI: 0.688, 0.829) in the training set, respectively, which were confirmed in the test set with AUCs of 0.742 (95% CI: 0.572, 0.872), 0.775 (95% CI: 0.608, 0.895), and 0.857 (95% CI: 0.703, 0.950), respectively. The nomogram yielded excellent performance for distinguishing intestinal-type adenocarcinomas in both sets, with AUCs of 0.928 (95%: 0.875, 0.964) and 0.904 (95% CI: 0.761, 0.976). CONCLUSIONS: The multiphase CT radiomics nomogram holds promise for the individual preoperative discrimination of intestinal-type gastric adenocarcinoma. KEY POINTS: • CT radiomics has a potential role in the distinction of intestinal-type gastric adenocarcinomas. • Single-phase enhanced CT-based radiomics showed favorable capability in distinguishing intestinal-type tumors. • The nomogram which incorporates the multiphase radiomics scores could facilitate the individual prediction of intestinal-type lesions.
OBJECTIVE: The purpose of the study was to investigate the role of CT radiomics for the preoperative distinction of intestinal-type gastric adenocarcinomas. MATERIALS AND METHODS: A total of 187 consecutive patients with preoperative contrast CT examination and pathologically proven gastric adenocarcinoma were retrospectively collected. Patients were divided into a training set (n = 150) and a test set (n = 37). Arterial phase (AP), portal phase (PP), and delay phase (DP) images were retrieved for analysis. A dedicated postprocessing software was used to segment the lesions and extract radiomics features. Random forest (RF) algorithm was applied to construct the classifier models. A nomogram was developed by incorporating multiphase radiomics scores. Receiver operating characteristic (ROC) curves were used to evaluate the performance of the radiomics model and nomogram in both sets. RESULTS: The radiomics model showed a favorable capability in the distinction of intestinal-type gastric adenocarcinomas. The areas under curves (AUCs) of the AP, PP, and DP radiomics models were 0.754 (95% CI: 0.676, 0.820), 0.815 (95% CI: 0.744, 0.874), and 0.764 (95% CI: 0.688, 0.829) in the training set, respectively, which were confirmed in the test set with AUCs of 0.742 (95% CI: 0.572, 0.872), 0.775 (95% CI: 0.608, 0.895), and 0.857 (95% CI: 0.703, 0.950), respectively. The nomogram yielded excellent performance for distinguishing intestinal-type adenocarcinomas in both sets, with AUCs of 0.928 (95%: 0.875, 0.964) and 0.904 (95% CI: 0.761, 0.976). CONCLUSIONS: The multiphase CT radiomics nomogram holds promise for the individual preoperative discrimination of intestinal-type gastric adenocarcinoma. KEY POINTS: • CT radiomics has a potential role in the distinction of intestinal-type gastric adenocarcinomas. • Single-phase enhanced CT-based radiomics showed favorable capability in distinguishing intestinal-type tumors. • The nomogram which incorporates the multiphase radiomics scores could facilitate the individual prediction of intestinal-type lesions.