摘要
目的:分析CT能谱成像在小肠克罗恩病疾病活动度评估中的价值。方法:回顾性分析我院2010年5月～2013年9月,经肠镜或病理证实的克罗恩病患者140例,其中活动期83例,缓解期57例。比较活动期和缓解期克罗恩病的以下几个单变量指标:①末端回肠的肠壁厚度;②肠壁动脉期和门脉期的CT值;③肠壁动脉期和门脉期的相对强化值;④动脉期和门脉期的梳状征分数;⑤肠壁的强化方式;⑥肠管的形态;⑦肠管周围的改变;⑧动脉期和门脉期的碘浓度值。应用SPSS 16.0进行分析,找出有鉴别意义的单变量指标,将有意义的单变量指标进行多变量二值Logistic回归分析,建立回归方程(数学模型),并用ROC曲线分析回归数学模型的诊断效能和确定预测诊断的诊断点,并计算曲线下面积、敏感性、特异性、准确性、阳性预测值、阴性预测值。结果:多元二值Logistic回归分析中,对鉴别活动期和缓解期克罗恩病有意义的指标是肠壁厚度、肠壁均匀一致强化以及门脉期的碘浓度值;ROC曲线下面积为0.961,敏感性、特异性、准确性、阳性预测值和阴性预测值分别为:95.8%、98.2%、96.4%、97.5%、93.3%,预测诊断界点值为0.753。结论:多元二值Logistic回归分析所建立的回归数学模型在鉴别活动期和缓解期克罗恩病的准确性为96.4%,提示用能谱CT小肠造影检查可以提高对克罗恩病活动度评估的准确性。
Purpose: To investigate value of spectral CT enterography in the evaluation of Crohn’s disease activity by logistic regression method. Methods: The spectral CT enterography data from a cohort of 140 Crohn’s disease patients from May 2010 to September 2013, who were confirmed by endoscopy and histopathology were analyzed retrospectively. Eighty-three cases were diagnosed as active and 57 cases were inactive. The parameters for differentiation between them were screened: 1. bowel thickness of the terminal ileum;2. CT value of the bowel wall in AP and PVP;3. relative contrast enhancement;4. comb sign score;5. type of enhancement pattern;6. morphology of the bowel wall;7. extra-luminal manifestations;8. iodine concentration. SPSS for Windows version 16.0 software was used to analyze the data and screen for parameters which were valuable for the differential diagnosis. Then,parameters with statistical significance were further analyzed by multivariable logistic regression. ROC analysis was performed to assess the diagnostic efficacy of regression mathematical model. The area under the ROC curve,sensitivity, specificity, accuracy, positive predictive value and negative predictive value were evaluated. Results:Parameters which were helpful in differentiating active CD from inactive CD included bowel thickness, mural homogenous enhancement and the iodine concentration in PVP by multivariable logistic regression analysis. The area under the ROC curve were 0.961, sensitivity, specificity, accuracy, positive predictive value and negative predictive were 95.8%, 98.2%, 96.4%, 97.5%, 93.3%, respectively and a predictive boundary value of 0.753 was obtained.Conclusion: The accuracy of the regression mathematical model established by using the spectral CT enterographic parameters was 96.4%, which suggests that spectral CT enterography is helpful for improving the accuracy in assessment of CD activity.
作者
赵雪松
常蕊
徐嘉旭
严嘉仪
严福华
缪飞
ZHAO Xue-song;CHANG Rui;XU Jia-xu;YAN Jia-yi;YAN Fu-hua;MIAO Fei(Department of Radiology, Ruijin Hospital Affiliate to Shanghai Jiaotong University School of Medicine,Shanghai 20025, P.R.C)
出处
《中国医学计算机成像杂志》
CSCD
北大核心
2019年第1期96-100,共5页
Chinese Computed Medical Imaging
