Comparison of lesion detection in the transverse and coronal views on automated breast sonography.

Abstract

The purpose of our study was to compare the accuracy of detection of lesions in the coronal and transverse views on automated breast sonography. Three breast radiologists independently interpreted automated breast sonograms from 113 women: 14 with negative findings and 99 with known breast lesions (99 benign and 53 malignant findings). The readers were asked to detect the presence or absence of the abnormalities using the transverse and coronal views in different reading sessions. If a lesion was detected, we evaluated its location, sonographic features, including shape, orientation, margin, boundary type, and echo pattern, and final American College of Radiology Breast Imaging Reporting and Data System assessment category. The reading time was also evaluated. Of 456 potential detections (152 lesions with 3 readers), 80.5% of the detections were made with the transverse view, and 67.3% were made with the coronal view. The detection rates for both malignant and benign lesions were significantly higher for the transverse view than for the coronal view (P < .001). The reading time for the coronal view was less than that for the transverse view (mean ± SD, 3.83 ± 1.71 versus 5.57 ± 2.21 minutes). The intraclass correlation coefficient values for lesion location, distance from the nipple, and lesion size for both views all exceeded 0.7. Although the coronal view on automated breast sonography showed comparable reliability and a shorter reading time, the detection rate for the coronal view was significantly lower than that for the transverse view in both benign and malignant lesions. It is preferable to reserve the coronal view as an additional method to the transverse view.