Xiaomei Huang1, Zixuan Cheng2, Yanqi Huang3, Cuishan Liang4, Lan He2, Zelan Ma3, Xin Chen5, Xiaomei Wu3, Yexing Li3, Changhong Liang6, Zaiyi Liu7. 1. The Second School of Clinical Medicine, Southern Medical University, Guangzhou, Guangdong, China. 2. Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, 106 Zhongshan Er Road, Guangzhou, 510080, China; School of Medicine, South China University of Technology, Guangzhou, Guangdong, China. 3. Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, 106 Zhongshan Er Road, Guangzhou, 510080, China. 4. The Second School of Clinical Medicine, Southern Medical University, Guangzhou, Guangdong, China; Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, 106 Zhongshan Er Road, Guangzhou, 510080, China. 5. Department of Radiology, Guangzhou First People's Hospital, Guangzhou Medical University, Guangzhou, China. 6. The Second School of Clinical Medicine, Southern Medical University, Guangzhou, Guangdong, China; Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, 106 Zhongshan Er Road, Guangzhou, 510080, China. Electronic address: cjr.lchh@vip.163.com. 7. The Second School of Clinical Medicine, Southern Medical University, Guangzhou, Guangdong, China; Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, 106 Zhongshan Er Road, Guangzhou, 510080, China. Electronic address: zyliu@163.com.
Abstract
RATIONALE AND OBJECTIVES: To develop and validate a computed tomography-based radiomics signature for preoperatively discriminating high-grade from low-grade colorectal adenocarcinoma (CRAC). MATERIALS AND METHODS: This retrospective study was approved by our institutional review board, and the informed consent requirement was waived. This study enrolled 366 patients with CRAC (training dataset: n = 222, validation dataset: n = 144) from January 2008 to August 2015. A radiomics signature was developed with the least absolute shrinkage and selection operator method in training dataset. Mann-Whitney U test was applied to explore the correlation between radiomics signature and histologic grade. The discriminative power of radiomics signature was investigated with the receiver operating characteristics curve. An independent validation dataset was used to confirm the predictive performance. We further performed a stratified analysis to validate the predictive performance of radiomics signature in colon adenocarcinoma and rectal adenocarcinoma. RESULTS: The radiomics signature demonstrated discriminative performance for high-grade and low-grade CRAC, with an area under the curve of 0.812 (95% confidence interval [CI]: 0.749-0.874) in training dataset and 0.735 (95%CI: 0.644-0.826) in validation dataset. Stratified analysis demonstrated that radiomics signature also showed distinguishing ability for histologic grade in both colon adenocarcinoma and rectal adenocarcinoma, with area under the curve of 0.725 (95%CI: 0.653-0.797) and 0.895 (95%CI: 0.838-0.952), respectively. CONCLUSIONS: We developed and validated a radiomics signature as a complementary tool to differentiate high-grade from low-grade CRAC preoperatively, which may make contribution to personalized treatment.
RATIONALE AND OBJECTIVES: To develop and validate a computed tomography-based radiomics signature for preoperatively discriminating high-grade from low-grade colorectal adenocarcinoma (CRAC). MATERIALS AND METHODS: This retrospective study was approved by our institutional review board, and the informed consent requirement was waived. This study enrolled 366 patients with CRAC (training dataset: n = 222, validation dataset: n = 144) from January 2008 to August 2015. A radiomics signature was developed with the least absolute shrinkage and selection operator method in training dataset. Mann-Whitney U test was applied to explore the correlation between radiomics signature and histologic grade. The discriminative power of radiomics signature was investigated with the receiver operating characteristics curve. An independent validation dataset was used to confirm the predictive performance. We further performed a stratified analysis to validate the predictive performance of radiomics signature in colon adenocarcinoma and rectal adenocarcinoma. RESULTS: The radiomics signature demonstrated discriminative performance for high-grade and low-grade CRAC, with an area under the curve of 0.812 (95% confidence interval [CI]: 0.749-0.874) in training dataset and 0.735 (95%CI: 0.644-0.826) in validation dataset. Stratified analysis demonstrated that radiomics signature also showed distinguishing ability for histologic grade in both colon adenocarcinoma and rectal adenocarcinoma, with area under the curve of 0.725 (95%CI: 0.653-0.797) and 0.895 (95%CI: 0.838-0.952), respectively. CONCLUSIONS: We developed and validated a radiomics signature as a complementary tool to differentiate high-grade from low-grade CRAC preoperatively, which may make contribution to personalized treatment.