Kaltenbach Benjamin, Brandenbusch Volker, Möbus Volker, Mall Gerhard, Falk Stephan, van den Bergh Marcus, Chevalier Frauke, Müller-Schimpfle Markus
Institute of Radiology (RZI), Klinikum Frankfurt/Main - Höchst, Academic Teaching Hospital of the University of Frankfurt/Main, Germany; Institute for Diagnostic and Interventional Radiology, Johann Wolfgang Goethe-University, Frankfurt/Main, Germany.
Diagnostic Breast Center Turmcarée, Mammography Screening, Frankfurt/Main, Germany.
Eur J Radiol. 2017 Jan;86:221-226. doi: 10.1016/j.ejrad.2016.11.022. Epub 2016 Nov 23.
The purpose of this retrospective analysis was to evaluate the likelihood of malignancy in prospectively categorized BI-RADS 4 and BI-RADS 5 calcifications.
This analysis included 849 women who underwent vacuum biopsy for BI-RADS 4 (with the subgroups 4A, 4B and 4C) or BI-RADS 5 calcifications between February 2007 and May 2015. Calcifications were classified according to the morphology and distribution descriptors of the BI-RADS lexicon (BI-RADS 4th edition lexicon). A standardized scheme (matrix) was used to combine the characteristics of the grouped calcifications with the BI-RADS assessment category.
Overall, 275/849 (32%) lesions were found to be malignant. 285/327/208/29 calcified lesions were prospectively classified as BI-RADS 4A/4B/4C/5 indicating a risk for malignancy of 16%/27%/55%/90%, respectively. The morphology descriptors predicted the risk for malignancy as follows: typically benign (n=55): 2%; indeterminate (n=676): 27%; typically malignant (n=118): 80%. The distribution descriptors correlated with a malignant histology as follows: diffuse (n=0); round or oval (n=261): 22%; regional (n=398): 33%; segmental (n=106): 42%; linear or branching (n=85): 55%. There was a significant difference between the descriptor categories (p<0.0001).
A standard scheme combining the morphology and distribution characteristics proved to be a helpful tool in diagnosis of calcifications, bridging the gap between description and classification of these lesions.
本回顾性分析的目的是评估在预先分类的乳腺影像报告和数据系统(BI-RADS)4类和5类钙化中发生恶性病变的可能性。
本分析纳入了2007年2月至2015年5月期间因BI-RADS 4类(分为4A、4B和4C亚组)或5类钙化接受真空活检的849名女性。钙化根据BI-RADS词典(第4版词典)的形态和分布描述符进行分类。采用标准化方案(矩阵)将分组钙化的特征与BI-RADS评估类别相结合。
总体而言,849例病变中有275例(32%)被发现为恶性。285/327/208/29例钙化病变被预先分类为BI-RADS 4A/4B/4C/5类,其恶性风险分别为16%/27%/55%/90%。形态描述符预测的恶性风险如下:典型良性(n = 55):2%;不确定(n = 676):27%;典型恶性(n = 118):80%。分布描述符与恶性组织学的相关性如下:弥漫性(n = 0);圆形或椭圆形(n = 261):22%;区域性(n = 398):33%;节段性(n = 106):42%;线性或分支状(n = 85):55%。描述符类别之间存在显著差异(p < 0.0001)。
结合形态和分布特征的标准方案被证明是诊断钙化的有用工具,弥合了这些病变描述与分类之间的差距。