Sonographic Evaluation of Early-Stage Breast Cancers That Undergo Neoadjuvant Chemotherapy

Abstract
We prospectively evaluated low-stage breast cancers treated with neoadjuvant chemotherapy using whole-volume sonography and color Doppler imaging. Thirty-four women with breast cancer (mean maximum size, 2.4 cm) received neoadjuvant chemotherapy with doxorubicin and docetaxel. Targeted whole-volume sonography of tumor sites was performed before and after chemotherapy to assess mass size, color pixel speed-weighted density, and American College of Radiology Breast Imaging Reporting and Data System sonographic characteristics. After chemotherapy, tumor sites were excised by lumpectomy or mastectomy. Three (11.3%) of 34 patients had a complete histologic response. After chemotherapy, correlation was r = 0.716 between final histologic and sonographic sizes. Compared with histologic residual tumors, sonography had 4 false-negative results, 3 false-positive results, and 27 true-positive results (sensitivity, 87%), with no false-negative results among a subgroup of tumors of 7 mm and larger (sensitivity, 100%). The 3 cases with false-positive results were histologic fibrosis or biopsy changes. Mean speed-weighted density was 0.015 before and 0.0082 after chemotherapy (P = .03). After chemotherapy, vascularity was less common within (P = .06) or adjacent to (P = .009) masses or in tumor sites (P = .05). Prechemotherapy variables of gray scale characteristics and vascularity were compared with final histologic size, and all had P > .20. Postchemotherapy sensitivity of sonography was high for residual tumors of 7 mm or larger. Correlation was moderate between histologic and sonographic final tumor sizes. False-positive results were caused by fibrosis or biopsy-related changes. False-negative results occurred with residual tumor size of 6 mm or smaller. After chemotherapy, vascularity usually decreased, and this was not specific for complete response. Before chemotherapy, no vascular or gray scale feature at initial imaging predicted complete responders.
Funding Information
  • National Institutes of Health (RO1 CA 91713)