Does power Doppler ultrasonography improve the BI-RADS category assessment and diagnostic accuracy of solid breast lesions?

Background: Power Doppler ultrasonography (US) is used in addition to B-mode US as a tool for the differential diagnosis between benign and malignant lesions. However, there are few data showing how useful power Doppler US is for BI-RADS category assessment. Purpose: To assess how much power Doppler US contributed to increasing the category diagnosis from Breast Imaging Reporting and Data System (ACR BI-RADS(R)-US) category 3 to category 4. Material and Methods: A total of 2426 patients (age range 16-91 years, mean 52 years) who underwent B-mode and power Doppler breast US using a linear 10-MHz transducer were analyzed. We devised integrated US category classification (BI-RADS-Integral) based on BI-RADS descriptors and vascularity. Masses with a circumscribed margin, parallel orientation, and negative vascularity on power Doppler US were defined as category 3. Irregular-shaped masses with a spiculated margin or echogenic halo were defined as category 5. Masses with one of the four suspicious findings (no circumscribed margin, no parallel orientation, microcalcifications, and positive vascularity) were defined as category 4a, and masses with two or more suspicious findings were defined as category 4b. Results: There were 98 breast cancer cases, and the average sizes of invasive and non-invasive cancers were 15 mm and 18 mm, respectively. The frequency of breast cancer according to category was: 0.4% (3/714) of the category 3 cases, 2.7% (7/256) in category 4a, 64% (47/73) in category 4b, and 100% (41/41) in category 5. Among the cases with a category 4a mass lesion, there were 183 lesions whose category diagnosis had been raised from category 3 to category 4a based on the Doppler blood flow findings alone, and 3 (1.6%) of those lesions were breast cancer. All three of them were ductal carcinoma in situ (DCIS) cases. When the cut-off point was placed between category 3 and category 4, BI-RADS-Integral and BI-RADS-US without power Doppler had a sensitivity of 96.9% and 93.9%, respectively, specificity of 72.1% and 90.4%, respectively, and accuracy of 74.4% and 90.7%, respectively. Conclusion: The presence of Doppler blood flow increases the malignancy pick-up rate, but at the expense of a significant decrease in specificity and diagnostic accuracy and an increase in biopsy rate.