增强CT纹理分析在术前预测结肠癌分化程度中的应用价值  

The application of contrast enhanced CT texture analysis in preoperatively predicting the differentiation degree of colon cancer

作　　者：施婷婷[1] 李辉[1] 季长风 刘松[1] 李琳[2] 魏晓磊[1] SHI Tingting;LI Hui;JI Changfeng;LIU Song;LI Lin;WEI Xiaolei(Department of Radiology,Nanjing Drum Tower Hospital,the Affiliated Hospital of Nanjing University Medical School,Nanjing 210008,China;Department of Pathology,Nanjing Drum Tower Hospital,the Affiliated Hospital of Nanjing University Medical School,Nanjing 210008,China)

机构地区：[1]南京大学医学院附属鼓楼医院医学影像科,江苏南京210008 [2]南京大学医学院附属鼓楼医院病理科,江苏南京210008

出　　处：《实用放射学杂志》2021年第9期1480-1483,共4页Journal of Practical Radiology

摘　　要：目的利用术前增强CT(CECT)纹理分析预测结肠癌分化程度.方法回顾性分析CECT图像122例.选取横断面图像病灶最大层面,由2位放射科医师分别绘制感兴趣区(ROI)并提取静脉期纹理参数.研究采用Shapiro-Wilk检验进行正态性检验,应用Mann-Whitney U检验进行不同分化程度组间纹理参数差异性检验,通过受试者工作特征(ROC)曲线分析CT纹理参数对预测结肠癌分化程度的诊断效能.结果结肠癌低分化组与中高分化组间静脉期平均数、模式、最小值、百分位数(5/10/25)、中位数、熵、熵10~13均有显著性差异(P<0.05).当静脉期最小值≥16HU为阈值时,判断结肠癌不同分化程度的敏感度为59.49%,特异度为79.07%,曲线下面积(AUC)为0.657.结论CT纹理分析多个参数在结肠癌不同分化程度组间存在显著性差异,可作为术前有效预测结肠癌分化程度的辅助工具.Objective To predict the differentiation degree of colon cancer using contrast-enhanced CT(CECT)texture analysis preoperatively.Methods The CECT images of 122 patients were analyzed retrospectively.The maximum slice of the lesion in transverse axial images was selected.The regions of interest(ROI)were drawn by two radiologists,respectively.The texture parameters derived from venous images were extracted.The Shapiro-Wilk test was used for normality test.The difference test of texture parameters between different differentiation degree groups used the Mann-Whitney U test.The diagnostic performance of texture parameters in predicting differentiation degree of colon cancer was analyzed using receiver operating characteristic(ROC)curve analysis.Results There were significant differences in multiple texture parameters derived from venous images between poorly differentiated and moderately-highly differentiated groups,including mean,mode,minimum,percentiles(5/10/25),medium,entropy,entropy 10-13(P<0.05).When minimum ≥16 HU was the threshold,the sensitivity,specificity,and area under the curve(AUC)were 59.49%,79.07%,and 0.657,respectively.Conclusion CT texture analysis can be used as an auxiliary tool to predict the differentiation degree of colon cancer preoperatively.

关 键 词：结肠癌 计算机体层成像 纹理分析 分化程度 

分 类 号：R735.35[医药卫生—肿瘤] R814.42[医药卫生—临床医学]