摘要
目的探讨通过MR弥散加权成像直方图分析对Luminal型乳腺癌与非Luminal型乳腺癌的鉴别诊断。方法收集我院行乳腺MR检查的女性患者共98例,经免疫组化分析获取乳腺癌分子分型。通过MATLAB软件获取肿瘤DWI图像的直方图参数,通过非参数检验以及受试者操作曲线检验直方图参数对于两种分子分型乳腺癌的鉴别诊断能力及将logistic回归分析应用于直方图参数尝试对两类分子分型乳腺癌的分类鉴别。结果在本文获取的9个直方图参数中,灰度最小值、灰度平均值、方差、均方根误差、峰度、能量值和熵值有显著性差异(P<0.05),其中灰度最小值具有最高的诊断敏感性和特异性(ROC曲线下面积0.809,敏感性0.673,特异性0.848),直方图参数Logistic回归的鉴别诊断准确率为86.7%,预测概率曲线下面积为0.852。结论弥散加权磁共振直方图分析可以实现对Luminal型乳腺癌和非Luminal型乳腺癌的鉴别诊断。
Objective To explore the differential diagnosis of Luminal breast cancer and non-Luminal breast cancer by diffusion magnetic resonance imaging histogram analysis.Methods 98 cases of breast magnetic resonance imaging(MRI)in our hospital were systematically collected from March 2017 to December 2018.All patients received MR examination.Molecular typing of breast cancer was obtained by immunohistochemical analysis,including 52 cases of Luminal type breast cancer and 46 cases of non-Luminal type breast cancer.The histogram parameters of tumor DWI images were obtained by MATLAB software,including minimum gray scale,maximum gray scale,average gray value,variance,root mean square error,deviation,kurtosis,energy value,and entropy value.Nonparametric tests and subjects'operating curves were used to test the ability of histogram parameters for differential diagnosis of two molecular types of breast cancer.Finally,this paper attempts to apply logistic regression analysis to histogram parameters to realize the classification and differentiation of two molecular types of breast cancer.Results Among the nine histogram parameters obtained in this study,there were significant differences in the minimum gray level,average gray value,variance,root mean square error,kurtosis,energy value and entropy value(P<0.05).The lowest gray scale had the highest sensitivity and specificity(area under ROC curve=0.809,sensitivity=0.673,specificity=0.848),Logistic regression),the accuracy of differential diagnosis by logistic regression was 86.7%.The area under the prediction probability curve was 0.852.Conclusion Diffusion magnetic resonance histogram analysis could be used in differential diagnosis of Luminal breast cancer and non-Luminal breast cancer.
作者
刘红霞
朱庆章
赵霞
李秉营
谢峰
王建国
LIU Hongxia;ZHU Qingzhang;ZHAO Xia;LI Bingying;XIE Feng;WANG Jianguo(Radiology Department of East Hospital District, The Affiliated Hospital of Shandong University of TCM, Jinan 250014, P.R.China;The People's Hospital of Lanling, Linyi 277700, P.R.China;The Health Service Center of Hongjialou Sub-District Community Office, Licheng District, Jinan 250100, P.R.China)
出处
《医学影像学杂志》
2020年第8期1398-1401,共4页
Journal of Medical Imaging
关键词
磁共振成像
直方图
乳腺癌
Magnetic resonance imaging
Weighted imaging,Histogram
Breast cancer
作者简介
刘红霞(1972-),女,山东济南人,本科学历,副主任医师,主要从事磁共振影像诊断工作;通讯作者:王建国,主任医师,E-mail:fskwangjianguo@126.com。
