Validated Multidetector-Row Computed Tomography-Based Radiomic Analysis for Preoperative Survival Prediction in Gastric Cancer: A Multicenter Study
| Autor: | Haidong Xiang, Yi Wang, Di Dong, Yingjiang Ye, Mengjie Fang, Jie Tian, Xiaoxuan Jia, Siwen Wang, Jing Zhou, Zaiyi Liu, Hailin Li, Caizhen Feng |
|---|---|
| Rok vydání: | 2019 |
| Předmět: | |
| Zdroj: | SSRN Electronic Journal. |
| ISSN: | 1556-5068 8177-1924 |
| Popis: | Background: Preoperative noninvasive evaluation of clinical outcome remains challenging for gastric cancer. Herein, we identified a multidetector-row computed tomography (MDCT)-based radiomic signature as a novel prognostic factor in survival prediction for gastric cancer patients. Methods: A retrospective dataset of 326 gastric cancer patients were enrolled from two centers and allocated to three cohorts: training cohort (n=166) and internal validation cohort (n=83) from one center, and external validation cohort (n=77) from another. Quantitative feature extraction was derived from portal venous phase MDCT images. The least absolute shrinkage and selection operator penalized Cox regression was adopted to construct a radiomic signature. A radiomic nomogram integrating radiomic signature and significant clinical risk factors was developed. A preoperative tumor-node-metastasis staging model was also built for comparison. All models were evaluated in terms of risk stratification, discrimination, calibration, and clinical use. Findings: A four-feature radiomic signature was established and showed robust prognosis power in both the internal and external validation cohorts. Subgroup analysis indicated radiomic signature could well risk stratify gastric cancer patients with advanced clinical stage. The radiomic nomogram combining radiomic signature, extramural vessel invasion, and alpha fetoprotein surpassed the other models in disease-free survival prediction with a concordance index of 0.768 in training cohort and 0.721 in internal validation cohort. Meanwhile, the area under two-year time-dependent receiver operating characteristic curve achieved 0.785 and 0.809 in both cohorts, respectively. Interpretation: The proposed MDCT-based radiomic signature was verified as a prognostic factor of gastric cancers, holding potential in promoting preoperative clinical prognosis and treatment strategies. Funding: This work was supported by the National Key R&D Program of China (2017YFC1308700, 2017YFA0205200, 2017YFC1309100), National Natural Science Foundation of China (81771924, 81501616, 81227901, 81671851, 81527805), the Beijing Natural Science Foundation (L182061), the Bureau of International Cooperation of Chinese Academy of Sciences (173211KYSB20160053), the Instrument Developing Project of the Chinese Academy of Sciences (YZ201502), and the Youth Innovation Promotion Association CAS (2017175). Declaration of Interest: The authors declare no conflicts of interest. Ethical Approval: This retrospective analysis was ethically granted by the Institutional Review Board of Peking University People's Hospital (PKUPH) and Guangdong General Hospital (GDGH) in compliance with the Health Insurance Portability and Accountability. Informed consent was not required from patients. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|