Purpose: This study examined the ability of magnetic resonance imaging (MRI) enhancement features to predict the response to neoadjuvant chemotherapy (NAC) in patients with breast cancer.
Methods: This retrospective study included 107 patients with breast cancer. All patients underwent a baseline breast MRI before NAC and follow-up MRI a mean of 3.7 months later. Breast MRI scans were evaluated using the Breast Imaging Reporting and Data System MRI lexicon. In addition, whole-breast vascularity (WBV) in the cancer-bearing breast was graded according to increased vessel number in comparison with the contralateral breast. Histopathologic tumor regression was graded semiquantitatively based on the Miller-Payne grading system. The ability of each MRI feature to predict the response was evaluated using a logistic regression analysis. Correlations between changes in MRI features and response were also evaluated using the Spearman rank correlation test.
Results: There were 73 responders (68%), including 59 partial and 14 complete responders. No significant difference in baseline MRI features was found between the responders and nonresponders, except for tumor size (P = 0.044). No dynamic enhancement feature on baseline MRI was useful for the early prediction of a response. In addition, an increased WBV did not predict a response, and the WBV change on the follow-up MRI was not correlated with the response. However, the change in the initial enhancement pattern (P = 0.007) and kinetic curve type (P = 0.003) were significantly correlated with response.
Conclusions: No baseline MRI feature described using the Breast Imaging Reporting and Data System MRI lexicon was useful for early prediction of the response to NAC.
From the Departments of *Radiology, †Surgery, and ‡Pathology, Ajou University School of Medicine, Suwon, Gyeonggi, South Korea.
Received for publication October 17, 2012; accepted December 17, 2012.
Reprints: Doo Kyoung Kang, MD, Department of Radiology, Ajou University School of Medicine, San 5, Woncheon-dong, Yongtong-gu, Suwon, Gyeonggi-do, 442-749, South Korea (e-mail: email@example.com).
The authors have no conflict of interest to declare.