Accuracy of MRI in selecting patients for breast conserving therapy after neoadjuvant chemotherapy.

Abstract

#4006 Introduction: Accurate assessment of the extent of residual breast cancer is crucial for the surgical planning after neoadjuvant chemotherapy (NAC). Compared to conventional imaging, MRI shows superior ability to monitor tumor response. Nonetheless, controversy exists about the precision of MRI to visualize disease extent after NAC. The aim of this study was to determine which variables influence the accuracy of MRI.
 Methods: Retrospective analysis of 130 patients treated with NAC. Contrast-enhanced MRI studies were performed on a 1.5-T scanner with a dedicated breast coil before, during and after NAC. The outcome parameters were the accuracy of the estimation of tumor size on final MRI and the accuracy in predicting the ability to perform breast conserving therapy (BCT). Tumor size on the final MRI was therefore compared to residual tumor size on pathology. Discrepancy <10 mm between MRI and pathology was defined as correct prediction. Patients with a tumor size <30 mm on MRI were considered indicative for BCT. The MRI indication for BCT was compared with the surgical treatment feasible based on size on pathology.
 Results: In 61/130 patients (47%) the difference between size on MRI and size on pathology was >10 mm. In 50% (n=32) this discrepancy between pathology and MRI would lead to different surgical treatment. Based on the MRI indication for BCT, 19 patients would have unjustly undergone BCT (disease extent at MRI <30 mm and disease extent at pathology >30 mm). In 13 patients MRI alone would have led to mastectomy (extent >30 mm) while the actual extent at pathology was <30 mm. Thus, MRI accurately indicated selection of surgical treatment in 98 patients (75%). Inaccurate indication (25%) was caused by underestimation as well as overestimation of disease extent. Variables significantly influencing the accuracy of MRI were type of lesion on baseline MRI, pattern of tumor decline, histological and molecular subtype.
 
 Conclusions: The results show the strengths and limitations in monitoring tumor response with MRI. For preoperative assessment the final MRI should be evaluated with the knowledge of the baseline MRI and the molecular and histological subtype. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4006.