MRI平扫结合扩散加权成像对大鼠乳腺良恶性病变的诊断  

Diagnosis of benign and malignant rat breast lesions through combination of MRI plain scanning and diffusion weighted imaging

作　　者：张松[1] 邹利光[1] 逄鑫[1] 廖翠薇[1] 梁开运[1] 

机构地区：[1]第三军医大学新桥医院放射科,重庆400037

出　　处：《第三军医大学学报》2011年第1期69-73,共5页Journal of Third Military Medical University

基　　金：国家自然科学基金(30770609;81071197)~~

摘　　要：目的研究磁共振成像(magnetic resonance imaging,MRI)平扫结合扩散加权成像(diffusion weighted imaging,DWI)在大鼠诱发性乳腺良恶性病变中的诊断价值。方法 Wistar雌性大鼠50只,7,12-二甲基苯蒽(DMBA)灌胃诱发大鼠乳腺肿块,MRI常规T1WI、T2WI和压脂序列扫描,以及DWI成像,测量乳腺肿块的T1WI信号强度(SIT1)、T2WI信号强度(SIT2)和最大横截面积(Amax),DWI图像后处理得到乳腺肿块的表观扩散系数(apparent diffusion coefficient,ADC)和指数化表观扩散系数(exponential apparent diffusion coefficient,eADC)。取感兴趣区组织行HE染色,观察乳腺肿块的病理表现,将其分为乳腺良性增生和乳腺癌2组。采用判别分析法,得出MRI各参数值联合应用的线性判别公式。应用受试者工作特性(ROC)曲线确定各指标鉴别乳腺良恶性病变的最佳切点值。结果 43只大鼠诱发出乳腺肿块,其中乳腺良性增生18只,乳腺癌25只。除SIT1外(P>0.05),乳腺良性病变与乳腺癌的SIT2、Amax、DWI ADC值和eADC值均有显著性差异(P<0.01)。MR平扫结合DWI成像对乳腺良恶性病变诊断的线性判别得分值(Score)=-0.007×SIT1+0.008×SIT2+0.302×Amax-2.042×ADC-1.639。Score值的ROC曲线下面积为0.951,大于任一单个参数的ROC曲线下面积,其联合诊断作用大于单个参数的诊断作用。结论大鼠诱发性乳腺肿瘤是MR成像研究的适合模型,MRI平扫结合DWI成像有助于提高对乳腺良恶性病变的诊断价值。Objective To study the differential diagnosis value of magnetic resonance imaging（MRI） plain scanning combined with diffusion weighted imaging（DWI） in induced benign and malignant rat breast lesions.Methods Fifty female Wistar rats were subjected to intragastric administration of 7,12-dimethyl-1,2-benzanthracene（DMBA） to induce breast tumors.The rats were examined by MRI T1WI,T2WI,fat-suppressed T2WI and DWI.After post-processing of the images,T1WI signal intensity（SIT1）,T2WI signal intensity（SIT2）,the maximum cross-sectional area（Amax）,apparent diffusion coefficient（ADC） value and exponential apparent diffusion coefficient（eADC） value of breast tumors were evaluated.The breast tumor tissues within the regions of interest were subjected to HE staining to observe pathological manifestations,and then divided into two groups of benign breast hyperplasia and breast cancer.The linear discriminant equation of combined MRI parameters was obtained by discriminant analysis.Receiver operator characteristic（ROC） curve was applied to determine the best cut-off value for differentiating benign and malignant breast lesions.Results Forty-three rats developed breast tumors,18 with benign breast hyperplasia and 25 with breast cancer.Except for SIT1（P0.05）,the indexes of SIT2,Amax,ADC value and eADC value showed significant differences between benign breast hyperplasia and breast cancer（P0.01）.The linear discriminant equation calculated by combining MRI plain scan and DWI parameters was Score=-0.007×SIT1 ＋0.008×SIT2 ＋0.302×Amax-2.042×ADC-1.639.The area under the ROC curve of Score value was 0.951.Both ROC area and diagnostic value of combined parameters were larger than those of any individual parameter.Conclusion Rats with induced breast cancer are suitable models for MRI study.Combining MRI plain scan and DWI together can help improving the differential diagnosis of malignant and benign breast lesions.

关 键 词：磁共振成像 扩散加权成像 乳腺癌 乳腺良性增生 

分 类 号：R655.804[医药卫生—外科学] R445.2[医药卫生—临床医学]