To determine the predictive value of preoperative breast MRI in selecting patients appropriate for nipple-sparing mastectomy.

Abstract

e11579 Background: The presence of the nipple areolar complex (NAC) is fundamental to the identity and esthetic of the breast. The objective of the nipple-sparing mastectomy is to preserve this in the appropriate setting. However, accurate patient selection is critical prior to performing a nipple-sparing mastectomy (NSM). Currently, there is a paucity of evidence to assist surgeons in developing an oncologically safe selection criteria. The aim of this study is to determine the utility of preoperative breast MRI to more accurately select patients for this procedure. Methods: All patients undergoing mastectomy with a preoperative MRI (within 3 months) were reviewed over a two-year period at a single institution. Preoperative breast MRI results were compared with the final pathological mastectomy specimen report. Patients were excluded if they had a lumpectomy prior to MRI. Patients were also excluded if Paget’s disease of the nipple-areolar complex was present and if inflammatory breast cancer was suspected. Patients receiving neo-adjuvant chemotherapy or who had a nipple-sparing mastectomy were also excluded. Results: 128 patients (187 mastectomies specimens) were identified who had a preoperative MRI prior to mastectomy. Indications for mastectomy were 67 invasive ductal cancers, 62 prophylactic/benign mastectomies, 41 had DCIS and 17 with invasive lobular cancers. The overall sensitivity of the preoperative MRI was 57% and specificity was 89.75%. The positive and negative predictive values were 41.4% and 94.3% respectively. Further analysis revealed that 66.6% of the false negatives were patients with DCIS identified in the lactiferous ducts of the NAC. On excluding these patients, the sensitivity and specificity was 75% and 91% respectively and the positive and negative predictive values were 42.9% and 97.6% respectively. Conclusions: Preoperative MRI is beneficial in selecting patients suitable for NSM with invasive cancer or in the prophylactic setting. Interestingly, in the presence of DCIS, the preoperative MRI may be misleading if used to select patients. Therefore, intra-operative retro-areolar frozen section is recommended in this subset.