Yang Hong, Zhang Huimin, Liu Wei, Tan Bei, Guo Tao, Gao Xiang, Feng Rui, Wu Kaichun, Cao Qian, Ran Zhihua, Liu Zhanju, Hu Naizhong, Zhu Liangru, Lai Yamin, Wang Congling, Han Wei, Qian Jiaming
Department of Gastroenterology, Peking Union Medical College Hospital, Chinese Academy Medical Sciences and Peking Union Medical College, Beijing, China.
Department of Radiology, Peking Union Medical College Hospital, Chinese Academy Medical Sciences and Peking Union Medical College, Beijing, China.
Front Oncol. 2022 May 2;12:856345. doi: 10.3389/fonc.2022.856345. eCollection 2022.
Differential diagnosis of Crohn's disease (CD) and ulcerative primary intestinal lymphoma (UPIL) is a tough problem in clinical practice.
Our study identified key differences between CD and UPIL patients and aimed to further establish a scoring model for differential diagnosis.
A total of 91 CD and 50 UPIL patients from 9 tertiary inflammatory bowel disease centers were included. Univariate and multivariate analyses were used to determine significant markers for differentiating CD and UPIL. A differential scoring model was established by logistic regression analysis.
The differential model was based on clinical symptoms, endoscopic and imaging features that were assigned different scores: intestinal bleeding (-2 points), extraintestinal manifestation (2 points), segmental lesions (1 point), cobblestone sign (2 points), homogeneous enhancement (-1 point), mild enhancement (-1 point), engorged vasa recta (1 point). A total score of ≥1 point indicates CD, otherwise UPIL was indicated. This model produced an accuracy of 83.66% and an area under the ROC curve of 0.947. The area under the ROC curve for validation using the 10-fold validation method was 0.901.
This study provided a convenient and useful model to differentiate CD from UPIL.
克罗恩病(CD)与原发性肠道溃疡淋巴瘤(UPIL)的鉴别诊断是临床实践中的一个难题。
我们的研究确定了CD患者与UPIL患者之间的关键差异,并旨在进一步建立一种鉴别诊断的评分模型。
纳入了来自9个三级炎症性肠病中心的91例CD患者和50例UPIL患者。采用单因素和多因素分析来确定区分CD和UPIL的显著标志物。通过逻辑回归分析建立鉴别评分模型。
鉴别模型基于临床症状、内镜和影像学特征,这些特征被赋予不同的分数:肠道出血(-2分)、肠外表现(2分)、节段性病变(1分)、鹅卵石征(2分)、均匀强化(-1分)、轻度强化(-1分)、直小血管充血(1分)。总分≥1分提示为CD,否则提示为UPIL。该模型的准确率为83.66%,ROC曲线下面积为0.947。采用10倍交叉验证法进行验证时,ROC曲线下面积为0.901。
本研究提供了一种方便且有用的模型来区分CD和UPIL。