Yan-qi Huang, Chang-hong Liang, Lan He, Cui-shan Liang, Ze-lan Ma, and Zai-yi Liu, Guangdong General Hospital, Guangdong Academy of Medical Sciences; Yan-qi Huang, Cui-shan Liang, and Ze-lan Ma, Southern Medical University; Lan He, School of Medicine, South China University of Technology; Xin Chen, Affiliated Guangzhou First People's Hospital, Guangzhou Medical University, Guangzhou; and Jie Tian, Key Laboratory of Molecular Imaging, Chinese Academy of Sciences, Beijing, People's Republic of China.
J Clin Oncol. 2016 Jun 20;34(18):2157-64. doi: 10.1200/JCO.2015.65.9128. Epub 2016 May 2.
To develop and validate a radiomics nomogram for preoperative prediction of lymph node (LN) metastasis in patients with colorectal cancer (CRC).
The prediction model was developed in a primary cohort that consisted of 326 patients with clinicopathologically confirmed CRC, and data was gathered from January 2007 to April 2010. Radiomic features were extracted from portal venous-phase computed tomography (CT) of CRC. Lasso regression model was used for data dimension reduction, feature selection, and radiomics signature building. Multivariable logistic regression analysis was used to develop the predicting model, we incorporated the radiomics signature, CT-reported LN status, and independent clinicopathologic risk factors, and this was presented with a radiomics nomogram. The performance of the nomogram was assessed with respect to its calibration, discrimination, and clinical usefulness. Internal validation was assessed. An independent validation cohort contained 200 consecutive patients from May 2010 to December 2011.
The radiomics signature, which consisted of 24 selected features, was significantly associated with LN status (P < .001 for both primary and validation cohorts). Predictors contained in the individualized prediction nomogram included the radiomics signature, CT-reported LN status, and carcinoembryonic antigen level. Addition of histologic grade to the nomogram failed to show incremental prognostic value. The model showed good discrimination, with a C-index of 0.736 (C-index, 0.759 and 0.766 through internal validation), and good calibration. Application of the nomogram in the validation cohort still gave good discrimination (C-index, 0.778 [95% CI, 0.769 to 0.787]) and good calibration. Decision curve analysis demonstrated that the radiomics nomogram was clinically useful.
This study presents a radiomics nomogram that incorporates the radiomics signature, CT-reported LN status, and clinical risk factors, which can be conveniently used to facilitate the preoperative individualized prediction of LN metastasis in patients with CRC.
开发并验证一种基于放射组学的nomogram，用于预测结直肠癌（CRC）患者的淋巴结（LN）转移。
预测模型在一个由 326 例经临床病理证实的 CRC 患者组成的初级队列中进行开发，数据采集时间为 2007 年 1 月至 2010 年 4 月。从 CRC 的门静脉期 CT 中提取放射组学特征。使用 Lasso 回归模型进行数据降维、特征选择和放射组学特征构建。多变量逻辑回归分析用于开发预测模型，我们将放射组学特征、CT 报告的 LN 状态和独立的临床病理危险因素结合起来，构建了放射组学 nomogram。通过校准、区分度和临床实用性来评估 nomogram 的性能。进行内部验证。一个独立的验证队列包含了 2010 年 5 月至 2011 年 12 月连续的 200 例患者。
放射组学特征（由 24 个选定特征组成）与 LN 状态显著相关（两个队列的 P 值均<.001）。个体化预测 nomogram 中的预测因素包括放射组学特征、CT 报告的 LN 状态和癌胚抗原水平。将组织学分级添加到 nomogram 中并未显示出增量预后价值。该模型具有良好的区分度，内部验证的 C 指数为 0.736（C 指数为 0.759 和 0.766），校准度良好。该 nomogram 在验证队列中的应用仍具有良好的区分度（C 指数为 0.778 [95% CI，0.769 至 0.787]）和良好的校准度。决策曲线分析表明，放射组学 nomogram 具有临床实用性。
本研究提出了一种放射组学 nomogram，它结合了放射组学特征、CT 报告的 LN 状态和临床危险因素，可以方便地用于预测 CRC 患者的 LN 转移。
