CT征象联合纹理特征在局灶性自身免疫性胰腺炎与胰腺癌鉴别诊断中的价值分析  

CT signs combined with texture features in the differential diagnosis between focal autoimmune pancreatitis and pancreatic cancer

作　　者：李莎莎 武志峰[2] 鄂林宁[2] LI Sha-sha;WU Zhi-feng;E Lin-ning(School of Medical Imaging,Shanxi Medical University,Taiyuan 030001,China;Department of Radiology,Shanxi Academy of Medical Sciences,Shanxi Bethune Hospital,Taiyuan 030032,China)

机构地区：[1]山西医科大学医学影像学院,山西太原030001 [2]山西医学科学院山西白求恩医院放射科,山西太原030032

出　　处：《中国临床医学影像杂志》2021年第5期347-350,共4页Journal of China Clinic Medical Imaging

基　　金：山西省自然科学基金资助项目(项目编号201801D121200)。

摘　　要：目的:探讨CT征象联合纹理特征用于局灶性自身免疫性胰腺炎(Focal autoimmune pancreatitis,f-AIP)与胰腺癌(Pancreatic cancer,PC)鉴别诊断的价值。方法:回顾性收集经临床或活检病理证实的50例f-AIP患者和60例PC患者的CT图像,通过评价这两个病变的CT形态征象,和使用MaZda软件从动脉期厚层CT图像提取纹理特征参数,比较f-AIP组与PC组间差异,并将获得的有统计学意义的指标进行二元logistic回归分析,建立CT征象、纹理特征及CT征象联合纹理特征诊断模型,最后运用ROC曲线对各模型进行评价。结果:所建立的CT征象模型、纹理特征模型及联合模型的AUC值分别为0.850、0.881、0.937,且CT征象联合纹理特征模型的敏感度及特异度均高于前两者(敏感度82%;特异度93.3%)。结论:采用CT征象联合纹理特征建立模型,诊断效能明显优于单一模型。Objective:To analyze the value of CT signs combined with texture features in the differential diagnosis between focal autoimmune pancreatitis(f-AIP)and pancreatic cancer(PC).Methods:Retrospectively collecting CT images of 50 f-AIP patients and 60 PC patients confirmed by clinical or biopsy pathology.CT morphological signs of these two lesions were analyzed,and MaZda software was used to extract the arterial phase CT images texture feature parameters.The differences between f-AIP group and PC group were compared,and the statistically significant indexes were analyzed by binary logistic regression.The diagnostic models of CT signs,texture features and CT signs combined with texture features were established.Finally,the ROC curve was used to evaluate the models.Results:The AUC values of the CT sign model,texture feature model and joint model were 0.850,0.881,0.937,respectively.The sensitivity and specificity of the CT sign joint texture feature model were higher than the former two(sensitivity:82%;specificity:93.3%).Conclusion:The diagnostic efficiency of the combined model based on CT signs and texture features is much better than that of single model.

关 键 词：胰腺炎 胰腺肿瘤 诊断 鉴别 体层摄影术 螺旋计算机 

分 类 号：R576[医药卫生—消化系统] R735.9[医药卫生—内科学]