Predicting nipple-areolar involvement using preoperative breast MRI and primary tumor characteristics

Preoperative assessment of the nipple-areolar complex (NAC) is invaluable when considering nipple-sparing mastectomy. Our hypothesis is that breast magnetic resonance imaging (MRI) may predict involvement of the NAC with tumor.

Clinical, histopathologic, and imaging data were compiled for patients who underwent preoperative breast MRI followed by mastectomy or nipple-sparing mastectomy for malignancy between 2006 and 2009. Blinded rereview of all MRI studies was performed by a breast MRI imager and compared to initial MRI findings. Multivariate analysis identified variables predicting NAC involvement with tumor.

Of 77 breasts, 18 (23 %) had tumor involving or within 1 cm of the NAC. The sensitivity of detecting histopathologically confirmed NAC involvement was 61 % with history and/or physical examination, and 56 % with MRI. Univariate analysis identified the following variables as significant for NAC involvement: large tumors near the nipple on preoperative MRI, node-positive disease, invasive lobular carcinoma, advanced histopathologic T stage, and neoadjuvant chemotherapy. On multivariate analysis, only tumor size >2 cm and distance from tumor edge to the NAC <2 cm on MRI maintained significance. Pearson correlation coefficient for MRI size compared to histopathologic size was 0.53 (P < 0.0001).

MRI is not superior to thorough clinical evaluation for predicting tumor in or near the NAC. However, MRI-measured tumor size and distance from the NAC are correlated with increased risk of NAC involvement. The combination of preoperative history and physical examination, tumor characteristics, and breast MRI can aid the surgeon in predicting a tumor-involved nipple more than any single modality alone.