乳腺影像学报告及数据系统:超声良恶性病变鉴别诊断初探

1992年美国放射学会制定了乳腺影像学报告及数据系统(breast imaging reporting and data system,BI-RADS),2003年第四版修订时包括了超声诊断(BI-RADS-US)。本文旨在初步探讨BI-RADS-US对于乳腺良恶性病变鉴别诊断的能力。我们收集了2006年1月至2007年3月期间、获得病理诊断的乳腺占位性病变共69例72灶。均为女性,年龄14-83岁,平均41.6岁。首先根据一般的非BI-RADS标准,做出良恶性鉴别。然后用BI-RADS-US标准进行鉴别和分级(0-6级),其中2级和3级为良性和良性可能性大,4级为性质待定,5级为恶性可能性大。计算并比较上述两种方法的准确性、敏感性、特异性。结果在72病灶中,根据非BI-RADS标准,诊断为良性者40灶,符合病理诊断者38灶(95.0%),诊断为恶性者23灶,符合病理者18灶(78.3%);诊断为性质待定者9灶。根据BI-RADS-US标准和分级,2级和3级共有37灶,与病理结果符合者为36灶(97.2%);5级有31灶,与病理符合者为22灶(71.0%);4级有4灶。非BI-RADS和BI-RADS-US的准确性各为83.3%和81.9%(P=1.000),敏感性各为91.7%和95.8%(P=1.000),特异性各为79.2%和75.0%(P=0.754)。总之,第一版BI-RADS-US在乳腺良恶性病变鉴别诊断方面的初步应用表现出较高的敏感性,而与非BI-RADS方法比较则尚无明显差别。

BI-RADS Ultrasonography in Differentiation of Benign from Malignant Lesions of the Breast: A Preliminary Result

Breast imaging reporting and data system (BI-RADS) was established by the American College of Radiology in 1992 and contained ultrasonography (BI-RADS-US) when BI-RADS was revised in 2003. The aim of this article was to assess the capacity of BI-RADS-US in differentiation between benign and malignant lesions of the breast.Between January 2006 and March 2007, 69 female patients with breast disease having 72 masses or nodules were examined using ultrasound.The age of them range 14-83 years (mean= 41.6). The differentiation of benign from malignant lesions was relied on traditional non-BI-RADS method firstly and then undertaken based on BI-RADS-US criterion in which the BI-RADS-US lexicon and category were included.The categories 2 and 3 were benign and probably benign findings, the category 4 was indeterminate in nature, and the category 5 was highly suggestive malignancy.The sensitivity, specificity, and accuracy of both evaluations were calculated and compared.Among 72 foci, 40 were diagnosed as the benign and 23 were considered as the malignant according to the non-BI-RADS, with accordance to pathologic diagnoses in 38 (95.0%) and 18 (78.3%), and 9 were indeterminate.Based on the BI-RADS-US, 37 lesions were classified as the categories 2 or 3(benign) and 31 as the category 5 (malignant), which having 36 (97.2%) and 22(71.0%) compatible to pathology, and there were 4 of the category 4 (indeterminate in nature). There were no statistically significant differences between the evaluations of non-BI-RADS and BI-RADS-US in aspects of accuracy (83.3% and 81.9%,P=1.000), sensitivity (91.7% and 95.8%,P=1.000), and specificity (79.2% and 75.0%, P=0.754).In our preliminary study, the first edition of BI-RADS-US has a higher sensitivity in deferential diagnosis of breast benign from malignant lesions, but no apparent differences when compared with non-BI-RADS.