Liu Yuan, Wan Lijuan, Peng Wenjing, Zou Shuangmei, Zheng Zhaoxu, Ye Feng, Jiang Jun, Ouyang Han, Zhao Xinming, Zhang Hongmei
Department of Radiology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Department of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Quant Imaging Med Surg. 2021 Jun;11(6):2586-2597. doi: 10.21037/qims-20-1049.
The aim of the present study was to investigate the potential risk factors for lymph node metastasis (LNM) in rectal cancer using magnetic resonance imaging (MRI), and to construct and validate a nomogram to predict its occurrence with node-for-node histopathological validation.
Our prediction model was developed between March 2015 and August 2016 using a prospective primary cohort (32 patients, mean age: 57.3 years) that included 324 lymph nodes (LNs) from MR images with node-for-node histopathological validation. We evaluated multiple MRI variables, and a multivariable logistic regression analysis was used to develop the predictive nomogram. The performance of the nomogram was assessed with respect to its calibration, discrimination, and clinical usefulness. The performance of the nomogram in predicting LNM was validated in an independent clinical validation cohort comprising 182 consecutive patients.
The predictors included in the individualized prediction nomogram were chemical shift effect (CSE), nodal border, short-axis diameter of nodes, and minimum distance to rectal cancer or rectal wall. The nomogram showed good discrimination (C-index: 0.947; 95% confidence interval: 0.920-0.974) and good calibration in the primary cohort. Decision curve analysis confirmed the clinical usefulness of the nomogram in predicting the status of each LN. For the prediction of LN status in the clinical validation cohort by readers 1 and 2, the areas under the curves using the nomogram were 0.890 and 0.841, and the areas under the curves of readers using their experience were 0.754 and 0.704, respectively. Diagnostic efficiency was significantly improved by using the nomogram (P<0.001).
The nomogram, which incorporates CSE, nodal location, short-axis diameter, and minimum distance to rectal cancer or rectal wall, can be conveniently applied in clinical practice to facilitate the prediction of LNM in patients with rectal cancer.
本研究旨在利用磁共振成像（MRI）探究直肠癌淋巴结转移（LNM）的潜在危险因素，并构建和验证列线图，通过逐个淋巴结的组织病理学验证来预测其发生情况。
我们的预测模型于2015年3月至2016年8月期间，使用前瞻性原发性队列（32例患者，平均年龄：57.3岁）开发而成，该队列包括来自MR图像的324个淋巴结（LNs），并进行了逐个淋巴结的组织病理学验证。我们评估了多个MRI变量，并使用多变量逻辑回归分析来开发预测列线图。通过校准、区分度和临床实用性来评估列线图的性能。在一个由182例连续患者组成的独立临床验证队列中验证了列线图在预测LNM方面的性能。
个体化预测列线图纳入的预测因素包括化学位移效应（CSE）、淋巴结边界、淋巴结短轴直径以及与直肠癌或直肠壁的最小距离。列线图在原发性队列中显示出良好的区分度（C指数：0.947；95%置信区间：0.920 - 0.974）和良好的校准。决策曲线分析证实了列线图在预测每个淋巴结状态方面的临床实用性。对于临床验证队列中读者1和读者2对淋巴结状态的预测，使用列线图时曲线下面积分别为0.890和0.841，而读者凭经验判断时曲线下面积分别为0.754和0.704。使用列线图显著提高了诊断效率（P<0.001）。
该列线图纳入了CSE、淋巴结位置、短轴直径以及与直肠癌或直肠壁的最小距离，可方便地应用于临床实践，以促进对直肠癌患者LNM的预测。
