Evaluation of contrast-enhanced ultrasonography breast predictive model in optimizing breast lesions BI-RADS classification

Abstract

Objective To explore the value of contrast-enhanced ultrasonography (CEUS) breast predictive model in the optimization of BI-RADS classification of breast lesions. Methods A total of 1 049 BI-RADS 4, 5 breast lesions were obtained from 1 039 patients in 8 centers. CEUS examination was performed prior to biopsy or surgery. According to the classification of the model, class 3, 4A, 4B and 4C were selected as biopsy thresholds, and the ROC curve was drawn. The diagnostic sensitivity, specificity, accuracy, positive predictive value, negative predictive value and Jordanian index were calculated for the biopsy threshold. The biopsy rate of breast lesions before and after angiography, cancer detection rate, follow-up cases of malignant risk were compared. Results There were benign lesions 586 (55.9%), malignant lesions 463 (44.1%) in the 1 049 breast lesions. The area of ROC with thresholds of 3, 4A, 4B and 4C were 0.695, 0.838, 0.847 and 0.757, respectively (all P<0.01). Ultrasonography had a certain diagnostic effect on benign and malignant breast lesions. The diagnostic sensitivity, specificity, accuracy, positive predictive value and negative predictive value with class 4A after CEUS set as the biopsy threshold were 93.32%, 75.65%, 82.75%, 75.57% and 93.35%, respectively, and the Jordanian index was 0.690. When chass 3 after CEUS was set as the biopsy threshold, the biopsy rate was reduced from 100% to 76.74%, the detection rate was increased from 44.23% to 56.77%, and the risk of cancer was only 0.67% in the follow-up cases. When class 4A was set as the biopsy threshold, the biopsy rate was reduced from 100% to 55.58% after CEUS. The detection rate of cancer increased from 44.23% to 74.44%. The risk of cancer was 2.96%. Conclusions The biopsy rate of breast lesions in category 4 and 5 would be reduced and cancer detection rate of them would be increased after CEUS, however, the risk of malignancy in the follow-up cases would be controlled as low as category 3 and 4A in previous BI-RADS. Thus, CEUS has a good prospect of in optimizing BI-RADS and reducing biopsy rate in unnecessary lesions. Key words: Contrast-enhanced ultrasound; Breast neoplasms; Breast imaging reports and data systems; Biopsy, needles