Objective:To analyze the mammography and MRI findings in patients with mammographically detected microcalcifications with BI-RADS category3-5, to compare mammography and contrast-enhanced MRI in evaluating suspicious microcalcifications.Materials and Methods:The mammography and MRI findings of112mammographically detected microcalcifications with BI-RADS category3-5were restrospectively analyzed, according to the lexicon of Breast Imaging Reporting and Data System (BI-RADS) designed by American College of Radiology (ACR), including morphology and distribution of microcalcification on mammography, and morphology, internal enhancement characteristics, kinetic enhancement curve on MRI. By using pathological diagnosis and clinical follow-up as reference standards, characterize the correlation mammography and MRI with histopathology. SPSS17.0software was used to make the following statistic analysis:(1) the positive predictive values (PPVs) of the morphological descriptors and distribution of microcalcifications were estimated, and fisher exact test was used to calculate the odd ratios (OR) to help stratify the risk of malignancy.(2)The positive predictive values (PPVs) of each imaging characteristics on MRI were estimated.(3) The sensitivity (SEN), specificity (SPE), positive predictive value (PPV), negative predictive value (NPV), and accuracy (ACC) were calculated in the diagnostic evaluation of microcalcifications on mammography and MRI, χ2test was used for testing the difference of each group, and further analyzing the mammography and MRI characteristics of the false-positive and false-negative lesions.Results:Of the112lesions, the pathologic diagnosis of microcalcifications was made in94patients:74malignant lesions, including invasive ductal carcinoma (IDC) in46(62.1%), ductal carcinoma in situ (DCIS) in22(29.7%), invasive micropapillary carcinoma in4(5.4%), mixed mucinous carcinoma in one (1.4%), and pleomorphic carcinoma in one (1.4%);20benign lesions, including fibroadenoma in6, adenosis in6, papilloma in4, atypical ductal hyperplasia2, and chronic inflammation in2. The remaining18patients underwent mammographic or clinical follow-up for more than six months to ensure benignity.Of the112mammographically detected microcalcifications with BI-RADS category3-5,19cases (17%) of BI-RADS category3were benign, the most feature for its’s morphology were round (73.7%); for distribution, diffuse (31.6%) and clustered (42.1%) microcalcification were the most characteristics.19cases (17%) of BI-RADS category4a were5for malignancy and14for benign, the lesions were more common in the round (42.1%) and amorphous (42.1%) microcalcifications with clustered (42.1%) and regional (36.8%) distribution.18cases (16.1%) of BI-RADS category4b were14for malignancy and4for benign, the lesions were more common in the amorphous (61.1%) microcalcifications with clustered (72.2%) distribution.22cases (19.6%) of BI-RADS category4c were21for malignancy and1for benign, the most feature for its’s morphology were fine pleomorphic (54.5%); for distribution, clustered (45.5%) microcalcification were the most characteristics.34cases (30.3%) of BI-RADS category5were malignant, the lesions were more common in the fine pleomorphic (35.3%) and fine linear/branching (61.8%) microcalcifications with segmental (41.2%) distribution. Results of the Fisher exact test revealed that the fine pleomorphic and fine linear/branching microcalcifications indicate significantly increased risks of malignancy (95%CI for odds ratios excludes1) compared with that of round, amorphous, and coarse heterogeneous descriptors. For distribution descriptors, the odds ratios of malignancy were0.683(95%CI:0.515-0.792;P=0.005) for segmental versus clustered microcalcifications,1.611(95%CI:1.212-2.141; P=0.004) for segmental versus regional microcalcifications, which indicate that segmental distribution was a increased risk of malignancy.MRI findings:of the BI-RADS category3, the characteristics on MRI were no enhancement (36.8%,7/19) and regional progressive enhancement (36.8%,7/19); the features of the BI-RADS category4a were regional progressive enhancement (26.3%,5/19) and non-mass-like enhancement (42.1%,8/19); the characteristics of the BI-RADS category4b lesions showed mass (44.4%,8/18) and non-mass-like enhancement (50.0%,9/18); the characteristics of the BI-RADS category4c lesions showed mass (54.5%,12/22) and non-mass-like enhancement (54.5%,12/22); BI-RADS category5lesions showed mass (32.4%,11/34)and non-mass-like enhancement (67%,23/34). On MRI, ten (8.9%) lesions did not show any enhancement after contrast agent injection.13(11.6%) lesions showed regional progressive enhancement,35(31.3%) lesions were mass enhancement,54(48.2%) lesions were non-mass-like enhancement. On mammography, the microcalcifications with no enhancement on MRI, were round (50%) for morphology, clustered (40%) and regional (40%) for distribution. The regional progressive enhancement showed round (61.5%) for morphology, and clustered (53.8%) for distribution. The mammographic findings for mass enhancement on MRI showed amorphous (31.4%), fine pleomorphic (31.4%), and clustered (62.9%) microcalcifications. The non-mass-like enhancement on MRI were more common in the fine pleomorphic (27.8%) and fine linear/branching (31.5%) findings, regional (33.3%), segmental (24.1%), and clustered (25.9%) distributions.The SEN, SPE, PPV, NPV, and ACC were93.2%,86.8%,91.1%,93.2%,86.8%versus97.3%,81.6%,91.9%,91.1%,93.9%, respectively, in the diagnostic evaluation of microcalcifications on mammography and MRI.Conclusion:The BI-RADS category3and4a lesions showed no enhancement and regional progressive enhancement on MRI; the features of the BI-RADS category4b,4c and5showed mass and non-mass-like enhancement. DCE-MRI can not indicate microcalcifications intuitively,but DCE-MRI is valued to differentiat malignant from benign breast lesions associated with microcalcification. |