Magnetic resonance imaging in patients with nipple discharge: should we recommend it?

Objectives: Comparing the sensitivity of Contrast-Enhanced Magnetic Resonance Imaging (CEMRI), mammography and ultrasonography in patients with nipple discharge (ND).
Methods: We retrospectively evaluated 38 women with ND who underwent mammography, ultrasound and 1.5 T CEMRI between March 2007 and July 2009. Imaging findings, pathological diagnosis and follow-up data (mean follow-up: 20 months) were compared. Sensitivity and specificity values were reckoned. Statistical differences in sensitivity were assessed.
Results: 5/38 malignancies (13.2%; 3 invasive, 2 intraductal; 4 ipsilateral, 1 contralateral to ND), and 14/38 High-Risk Lesion (HRL--36.8%; 11 intraductal papillomas, 1 papilloma with LCIS, 1 sclerosing papilloma and 1 atypical intraductal hyperplasia, all ipsilateral) were found. CEMRI identified 5/5 cancers and 13/14 HRL (Overall Sensitivity-OSS = 94.7%; Overall Specificity-OSP = 78.9%). 3/5 cancers (1 invasive, 1 in-situ; 1 invasive contralateral) and 2/14 HRL were detected by CEMRI only. Mammography found 2/5 cancer and 3/14 HRL (OSS = 26.3%; OSP = 94.7%). Ultrasound identified 1/5 cancer and 11/14 HRL (OSS = 63.2%; OSP = 84.2). 1/14 HRL was detected by ultrasound only. Compared with mammography and ultrasound, CEMRI showed statistically significantly higher OSS values (p < 0.0001, p = 0.042 respectively).
Conclusion: In women with ND, CEMRI should be recommended when conventional imaging is negative. Unexplained ND could be considered an indication for CEMRI.