Diagnostic Values of Breast Imaging Reporting and Data System and ultrasound Elastography in Benign and Malignant Breast Tumor

Objective: To evaluate the diagnostic values of Breast Imaging Reporting and Data System (BI-RADS), ultrasound elastography (UE) and the combination in differentiating benign and malignant breast tumor. Methods: The BI-RADS and UE image features of 248 breast cancer patients (a total of 260 lesions) proved by surgery and pathology from February 2013 to March 2015 were retrospectively analyzed. With the pathologic results as the gold standard, the sensitivity, specificity, positive and negative predictive values, and accuracy were calculated for BI-RADS, UE and the combination. On the basis of the sensitivity and specificity, they were analyzed by receiver operating characteristic (ROC) curve. Results: In all 260 lesions, 71 lesions were benign and 189 were malignant according to UE diagnosis; 50 lesions were benign and 210 were malignant proved by BI-RADS; 55 lesions were benign and 205 were malignant diagnosed by the combination. The sensitivity (86.09%), specificity (61.64%), positive predictive value (85.19%), negative predictive value (63.38%), and accuracy (79.23%) of ultrasound elastography were all less than that of BI-RADS (98.39%, 64.38%, 88.85%, 87.62%, 94.00%) and the combination (99.47%, 73.97%, 92.31%, 90.73%, 98.18%). The areas under the ROC curve for UE, BI-RADS and the combination were respectively 0.746[95%CI(0.673–0.818)], 0.814[95%CI(0.744–0.884)] and 0.867[95%CI(0.805–0.929)]. Conclusion: Ultrasonic BI-RADS can be the first choice for diagnosing breast cancer, with UE as the auxiliary method. The combined application can further improve the diagnosis rate of benign and malignant breast tumor.