Breast Cancer: Early Prediction of Response to Neoadjuvant Chemotherapy Using Parametric Response Maps for MR Imaging

Abstract

Purpose: To prospectively compare the performance of dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging using parametric response map (PRM) analysis with that using pharmacokinetic parameters (transfer constant [K-trans], rate constant [k(ep)], and relative extravascular extracellular space [v(e)]) in the early prediction of pathologic responses to neoadjuvant chemotherapy (NAC) in breast cancer patients. Materials and Methods: The institutional review board approved this study; informed consent was obtained. Between August 2010 and December 2012, 48 women (mean age, 46.4 years; range, 29-65 years) with breast cancer were enrolled and treated with an anthracycline-taxane regimen. DCE MR imaging was performed before and after the first cycle of chemotherapy, and the pathologic response was assessed after surgery. Tumor size and volume, PRM characteristics, and pharmacokinetic parameters (K-trans, k(ep), and v(e)) on MR images were assessed and compared according to the pathologic responses by using the Fisher exact test or the independent-sample t test. Results: Six of 48 (12%) patients showed pathologic complete response (CR) (pCR) and 42 (88%) showed nonpathologic CR (npCR). Thirty-eight (79%) patients showed a good response (Miller-Payne score of 3, 4, or 5), and 10 (21%) showed a minor response (Miller-Payne score of 1 or 2). The mean proportion of voxels with increased signal intensity (PRMSI+) in the pCR or good response group was significantly lower than that in the npCR or minor response group (14.0% +/- 6.5 vs 40.7% +/- 27.2, P < .001; 34.3% +/- 26.4 vs 52.8% +/- 24.9, P =.041). Area under the receiver operating characteristic curve for PRMSI+ in the pCR group was 0.770 (95% confidence interval: 0.626, 0.879), and that for the good response group was 0.716 (95% confidence interval: 0.567, 0.837). No difference in tumor size, tumor volume, or pharmacokinetic parameters was found between groups. Conclusion: PRM analysis of DCE MR images may enable the early identification of the pathologic response to NAC after the first cycle of chemotherapy, whereas pharmacokinetic parameters (K-trans, k(ep), and v(e)) do not. (C) RSNA, 2014