A single institution review of new breast malignancies identified solely by sonography

The role of sonography as a sole identifier of breast malignancy remains undetermined. Currently, the American College of Radiology Imaging Network (ACRIN) trial, protocol 6666, is accruing high risk patients, with dense breast tissue on mammography, to evaluate the performance of screening sonography.

We conducted a retrospective single institution review to identify the detection rate of nonpalpable, mammographically occult breast malignancies identified solely by sonography.

A total of 34,694 breast sonograms were performed at our outpatient radiology affiliate between April 1998 and April 2006. This number includes unilateral and bilateral examinations, and, 6-month followup examinations. Computerized databases and individual charts were retrospectively reviewed. Sonographic and mammographic findings were compared to determine concordance of identified lesions. Four thousand ninety-one ultrasound guided biopsies, (fine needle aspiration biopsy [FNA] and core needle biopsy [CNB]) were performed over the 8-year period, and 452 malignancies were identified. Discarded fluid contents of pure cyst aspirations were not included in this number. Seven percent (32) of the biopsy-proved malignancies were mammographically occult, of which 59% (19) were palpable, and 31% (10) were not palpable. Of all cancers diagnosed, 2.2% were identified solely by sonography. The number of cancers identified solely by sonography relative to the total number of sonograms done was 0.03%.

Given these results, we advocate the selective use of sonography in appropriate populations, namely, in those with palpable findings, mammographic abnormalities, and in women with dense tissue who have personal or family history of breast cancer.