Detection of malignant and benign breast lesions with an automated US system: Results in 120 cases

PURPOSE: To evaluate clinically an automated ultrasound (US) system for detecting benign and malignant breast lesions. MATERIALS AND METHODS: A prototype automated US system was used to examine 119 patients: 38 patients with 39 proved malignant breast lesions (7-50 mm), 41 patients with 41 proved benign breast lesions (8-40 mm), and 40 patients without breast lesions. The device yields a three-dimensional set of B-mode scans and reconstructed US images comparable to mammograms. All patients had undergone mammography. Four radiologists who had not performed the examinations independently assessed the mammograms and US images to detect benign and malignant breast lesions. RESULTS: Each of the four readers did not recognize one to three detectable malignant lesions on mammograms, one to two detectable malignant lesions on US images, two to four detectable benign lesions on mammograms, and five to seven detectable benign lesions on US images. All readers identified the 39 cancers with at least one of the modalities. The 40 cases without lesions were diagnosed correctly more frequently on the US images by three readers and on the mammograms by one reader. CONCLUSION: Depiction of breast lesions at automated US is reproducible. Automated US is complementary to mammography.