数字乳腺断层结合合成二维图像在乳腺病变中的诊断价值探讨  

To Investigate the Diagnostic Value of DBT Combined with S2D Images in Breast Lesions

作　　者：张鹏 周会 孙小会[1] 刘红艳[1] 黄奕妆 邓树芳 唐美群 吴文英 Zhang Peng;Zhou Hui;Sun Xiaohui;Liu Hongyan;Huang Yizhuang;Deng Shufang;Tang Meiqun;Wu Wenying(Department of Radiology,The Sixth Affiliated Hospital of Guangzhou Medical University(Qingyuan People’s Hospital),Qingyuan,Guangdong 511518)

机构地区：[1]广州医科大学附属第六医院(清远市人民医院)放射科,广东清远511518

出　　处：《现代医用影像学》2022年第9期1629-1633,共5页Modern Medical Imageology

基　　金：清远市科技计划项目(编号:清科[2019]24号)。

摘　　要：目的:探讨DBT合成二维图像(S2D)能否取代FFDM图像。方法:收集我院2020年11月至2021年10月术前均同时行FFDM和DBT检查,具有术后病理结果的100例乳腺疾病患者(120个病灶)。一次检查获得三种图像:FFDM图像、DBT薄层图像、S2D图像(重组DBT薄层数据得到类似FFDM图像),并将其分成四组图像:FFDM、S2D、DBT+FFDM、DBT+S2D;对每一个病灶进行BI-RADS分类,根据分类结果判断良恶性,0-4A类判为良性,4B类及以上判为恶性。记录FFDM和DBT的平均腺体剂量。采用卡方检验比较四组图像在BI-RADS分类中的差异;采用ROC曲线评价四组图像鉴别病灶良、恶性的能力,以病理结果作为判定病灶良恶性的标准;采用独立样本t检验比较FFDM和DBT平均腺体剂量的差异。结果:78个良性病灶、42个恶性病灶。FFDM与S2D、DBT+FFDM与DBT+S2D的BI-RADS分类差异无统计学意义(P>0.05);DBT+S2D与S2D的BI-RADS分类差异有统计学意义(P<0.05)。DBT+FFDM、DBT+S2D鉴别病灶良、恶性的ROC曲线下面积均为0.96,敏感度、特异度均为88.1%、94.9%,优于FFDM和S2D的ROC曲线下面积(分别为0.817和0.821)。FFDM和DBT的单个体位平均腺体剂量分别为(0.92±0.17)mGy、(1.50±0.27mGy),差异有统计学意义(t值为1.97,P<0.05)。结论:S2D图像可代替FFDM图像,DBT+S2D的检查模式能够在确保安全的辐射剂量下,提高乳腺病变的诊断效能,值得推广应用。Objective:To investigate whether DBT combined with S2 D images in breast lesions can replace FFDM images.Methods:100 patients(120 lesions)with breast lesions with postoperative pathological results who underwent FFDM and DBT from November 2020 to October 2021 were collected retrospectively.Three sets of images are obtained in one time inspection:includes FFDM images,DBT thin-layer images and S2 D images(similar to FFDM images was obtained by reorganizing DBT thin-layer data),and then divided them into four groups:FFDM,S2 D,DBT+FFDM and DBT+S2 D;Each lesion was classified by BI-RADS.0-4 A were classified as benign,and 4 B and above were classified as malignant.The mean glandular doses of FFDM and DBT were recorded.The chi-square test was used to compare the differences of the four groups of images in the BI-RADS classification;According to the pathological results,ROC was used to evaluate the efficacy of different types of images in distinguished benign and malignant breast lesions;The difference of average glandular dose between FFDM and DBT was compared by independent sample t-test.Results:There were 78 benign lesions and 42 malignant lesions.There was no significant difference in BI-RADS classification between FFDM and S2 D images and DBT combined with FFDM and DBT combined with S2 D images(P>0.05);There was significant difference in BI-RADS classification between DBT combined with S2 D images and S2 D images(P<0.05).The area under ROC of DBT combined with FFDM and DBT combined with S2 D images in the differential diagnosis of benign and malignant breast lesions was 0.96,and the sensitivity and specificity were 88.1%and 94.9%,which was better than the ROC area of FFDM and S2 D images(0.817 and 0.821).The average glandular doses of FFDM and DBT were 0.92±0.17 mGy and 1.50±0.27 mGy.There was significant difference in the average glandular dose of FFDM and DBT(t value was 1.97,P<0.05).Conclusion:S2 D images in breast lesions can replace conventional FFDM images.DBT and S2 D images can improve the diagnostic effi

关 键 词：乳腺病变 全屏数字化乳腺X线成像 合成二维图像 

分 类 号：R655.8[医药卫生—外科学] R816[医药卫生—临床医学]