Sonographic features of dense breast imaging reporting and data system 4 (BI-RADS-US4) for non-palpable breast lesions

The aim of this study was to investigate the diagnostic value of various features for differentiating between malignant and benign lesions by using the sonography for dense breast imaging reporting and data system 4 (BI-RADS-US4) for non-palpable breast lesions (dBI-RADS-US4-nPBL). The sonographic features and pathological results of a total of 102 dBI-RADS-US4-nPBLs were retrospectively analyzed so as to evaluate the diagnostic values of sonography for such patients. Among the 102 lesions included, 47 cases were benign and 55 cases were malignant, and the overall positive predictive value was 53.9%; the positive predictive values of the subcategories BI-RADS-US4A, 4B, and 4C were 18.2%, 55.0%, and 93.1%, respectively. The diagnostic performance for BI-RADSUS classification was determined using ROC curve analysis. For the cut-off point of 4B (sensitivity, 57.4%; specificity, 89.1%), the area under the receiver operating characteristic curve was 81.8%. The diagnostic sensitivity for lesions with an irregular shape was the highest (89.1%), and features such as burr, sublobe, microcalcification, rear echo attenuation, and mass walking pattern had higher diagnostic specificities for this category of lesions (95.7%, 87.2%, 89.4%, 80.9%, and 87.2%, respectively). Mass walking pattern (odds ratio [OR]=5.861, 95% confidence interval [1.862, 20.846], p=0.006) and margin (OR=4.462, [0.932, 4.462], p=0.000) were the two most significant predictors of malignant lesions in multivariate analysis. For BI-RADS-US4-nPBLs, the feature of irregular shape had the highest diagnostic sensitivity, and features such as burr, sublobe, microcalcification, rear echo attenuation, and mass walking pattern had higher diagnostic specificities. patients with BI-RADS-US4B and 4C-npBLs should be recommended to undergo biopsy, but those with BI-RADS-US4A-nPBLs should be appropriately cautious about the need for a biopsy.