Community-Based Practice Experience of Unsuspected Breast Magnetic Resonance Imaging Abnormalities Evaluated With Second-Look Sonography

Objective. The purpose of this study was to assess the utility and efficacy of second-look breast sonography in the evaluation of abnormalities identified on presurgical breast magnetic resonance imaging (MRI) examinations. Methods. A retrospective review was performed of 152 presurgical breast MRI examinations having 196 unsuspected abnormalities to identify findings that underwent subsequent breast sonography. Eligible examinations had a suspicious abnormality identified on presurgical MRI and documentation of the location and size of the finding on MRI and subsequent second-look sonography. Fourteen examinations not meeting the criteria were excluded, with 182 abnormalities remaining. Patient medical records were reviewed. Results. Seventy percent (128 of 182) of breast MRI lesions were visible at second-look sonography; 30% (54 of 182) were sonographically occult. Ninety-five percent (121 of 128) of sonographically visible abnormalities underwent breast biopsy. Pathologic examinations of sampled sonographically visible lesions revealed 39 cancers, 9 high-risk lesions, 72 benign lesions, and I lesion not specified at surgery; 23% (9 of 128) of cancers were in the contralateral breast. Fifty-four lesions were sonographically occult. Needle biopsy was performed for 93% (50 of 54) of occult lesions, revealing 8 cancers, 1 high-risk lesion, and 41 benign lesions; 3 of the benign lesions did not have sufficient pathologic specimens but were benign at follow-up. Magnetic resonance imaging guidance was used in 86% (43 of 50) of these biopsies. One contralateral cancer was identified in the sonographically occult subset. Of the 182 lesions that underwent second-look sonography, 20% (36 of 182) had a change in management. Conclusions. Second-look sonography has value in the evaluation of abnormalities found on breast MRI. We found enhancements in 70% (128 of 182), yielding 39 cancers.