Galactography: method of choice in pathologic nipple discharge?

Our objective was to assess the role of galactography in the diagnostic work-up of patients with pathological nipple discharge together with exfoliative cytology, and investigation of secretion colour. All galactographies performed in this center between 1993 and 1998 were evaluated retrospectively. In 134 cases, outcomes, defined as either a postoperative histology or a negative follow-up over 2 years after galactography, were available. In 96 of these cases, exfoliative cytology was performed, and in 65 cases the colour of the secretion was determined. Galactograms, cytological findings and colour of the secretions were compared with the outcomes. At galactography, the radiographic findings had the following individual sensitivities/specificities in detecting carcinomas: filling defect 55.6%/62.1%; ductectasia 22.2%/94%; and filling stop 5.6%/77.6%. Normal galactograms (n=33, 25%) had a sensitivity of 78% and a specificity of 93% in predicting absence of disease. Technically inadequate investigations occurred in 16 (12%) cases. Cytology showed 69.2%/66.8% for papillomatous borderline or malignant cells; 53.8%/75.3% for borderline or malignant cells; and 7.7%/100% for one case of malignant cells. For the assessment of pathological secretions, galactography is a sensitive but unspecific method for the detection of papillomas or carcinomas. Filling defects, ductectasia and ductal distortion carry the highest levels of suspicion for carcinoma. A normal galactogram is a specific yet moderately sensitive indicator of absence of localized disease. Exfoliative cytology shows low sensitivity but better specificity for carcinomas when borderline or malignant cells are found. In the present series, the colour of the secretions does not add significant information about underlying pathology.