摘要
目的探讨磁共振动态增强成像(DCE-MRI)与乳腺磁共振报告数据系统(BI-RADS-MR)在乳腺肿块诊断中的价值。方法对经病理证实的75例乳腺肿块患者共计75个肿块术前行DCE-MRI检查,并测定动态增强参数。由两名高年资影像诊断医师采用盲法阅片,并根据BI-RADS-MR进行分级,构建受试者工作特征曲线(ROC)分析两名医师的判读情况,评价两名医生诊断乳腺肿块的灵敏度、特异度。结果根据BI-RADS-MR分级,38例乳腺癌患者评为Ⅰ级1例,Ⅱ级0例,Ⅲ级4例,Ⅳ级13例,Ⅴ级20例;良性Ⅰ级17例,Ⅱ级5例,Ⅲ级12例,Ⅳ级3例,Ⅴ级0例。良恶性乳腺肿块的T_(peak)、S_(max)、K_1差异有统计学意义(均P<0.05)。两名医师根据BI-RADS-MR评分方法及DCE-MRI诊断乳腺癌的ROC曲线下面积(AUC)为0.940,95%置信区间为0.887~0.994。BI-RADS-MR评分方法诊断乳腺癌的灵敏度为94.59%、特异度为92.11%。结论 DCE-MRI参数与BI-RADS-MR评分方法对于提高乳腺肿块诊断及鉴别诊断效能具有重要价值。
Objective To evaluate the diagnostic value of DCE-MRI and breast MR reporting and data system（BI-RADS-MR）in differentiation of breast lesions.Methods Seventy-five patients with 75 pathologicalconfirmed breast lesions underwent DCE-MRI before treatment and DCE-MRI quantitative analysis.Two senior doctors assessed the imaging features blindly by BI-RADS-MR.The pathology results were set as a gold standard.The receiver operating characteristic curve（ROC）interpretation of two physician′s curve analysis was used to evaluate the sensitivity,specificity of BI-RADS classification.Results According to the BI-RADS MR score,38 breast cancer patients were rated as 1 case in grade Ⅰ,0 case in grade Ⅱ,4 cases in grade Ⅲ,13 cases in grade Ⅳ,20 cases in grade Ⅴ;37 breast benign tumors were rated as 17 cases in grade Ⅰ,5 cases in grade Ⅱ,12 cases in grade Ⅲ,3 cases in grade Ⅳ and Ⅴ0 cases.The statistical difference of Tpeak,S max,K1 between benign and malignant breast tumors were significant（P〈0.05）.On ROC,the area under the curve was 0.940,95% confidence interval was from 0.887 to 0.994.The sensitivity of BI-RADS in the diagnosis of breast cancer was 94.59% and specificity was 92.11%.Conclusion DCE-MRI combined with BI-RADS-MR classification in diagnosis of benign and malignant breast tumors has an important value and can improve diagnosis and differential diagnosis level.
出处
《实用医学杂志》
CAS
北大核心
2017年第16期2780-2783,共4页
The Journal of Practical Medicine
关键词
乳腺肿瘤
乳腺疾病
磁共振成像
动态增强
breast neoplasms
breast diseases
magnetic resonance imaging
dynamic contrastenhanced
