The predictive value of ductography and magnetic resonance imaging in the management of nipple discharge

Background: Standard evaluation (physical examination, mammography, sonography) often fails to identify an underlying lesion in patients with suspicious nipple discharge. The aim of this study was to determine the predictive value of ductography (DG) and magnetic resonance imaging (MRI) in this setting. Methods: Using ICD-9 codes, we retrospectively identified 376 patients who presented with suspicious nipple discharge (ND) (1995-2005); 306 patients (68%) had negative standard evaluation. Results: Among 306 patients, 186 (61%) underwent further evaluation with DG (n = 163) and/or MRI (n = 52), 35 (11%) underwent major duct excision alone (MDE), and 85 (28%) were followed clinically. Ultimately, 182/306 (59%) patients underwent surgery and/or biopsy. Overall incidence of malignant or high-risk pathology was 15% (46/306). DG was completed in 139/163 (85%) studies and detected 12 cancers and seven high-risk lesions (HRL), but failed to identify four cancers and 2 HRL (PPV 19%, NPV 63%). MRI detected seven cancers and three HRL, but failed to identify one cancer and one HRL (PPV 56%, NPV 87%). MDE alone (n = 35) detected five cancers and three HRL. Of all patients not having surgery, (142/306, 41%), one (0.01%) presented with an invasive cancer at 102 months (median follow-up, 6.3 months; range, 0-124 months). Conclusions: An underlying malignancy was identified in 30/306 (10%) patients with ND and negative standard evaluation. Ductography is a poor predictor of underlying pathology and cannot exclude malignancy. MRI's higher predictive values may allow for improved patient selection and treatment planning; however, MRI should not replace MDE as the gold standard to exclude malignancy in patients with ND and negative standard evaluation.