Nipple aspirate fluid (NAF) cytology and pathologic parameters predict residual breast cancer and nodal spread after excisional biopsy

We reported that atypical/malignant NAF cytology predicted residual breast cancer (RC) and tumor size after excisional biopsy (EB), although normal NAF cytology did not exclude RC. Tumor size correlates with the risk of lymph node (LN) metastases. We hypothesized that pathologic factors known after EB (tumor distance from the biopsy margins, multifocal and multicentric disease, subtype of DCIS or invasive cancer (IC), grade of DCIS or IC, tumor and specimen size, tumor and biopsy cavity location, presence or absence of an extensive intraductal component, and biopsy scar distance from the nipple), combined with NAF cytology, would predict the presence of RC and LN spread. Methods: After IRB approval and informed consent, 72 NAF specimens were collected from 70 subjects aged 30-79 (median 52) years who required reexcision or mastectomy due to close or positive margins on diagnostic EB. Nipple fluid was aspirated by a trained physician or nurse clinician using a modified breast pump. Cytologic review was performed using a standardized classification scheme. Pathologic factors were obtained from patient records and pathology reports. Stepwise logistic regression was performed based on bivariate analyses of the association between NAF cytology and pathologic parameters with RC and LN spread. We analyzed the data two ways: 1) subjects of all cytology classes, and 2) subjects with normal cytology, for in our prior study normal NAF cytology did not exclude RC. Results: All Subjects Normal Cytology Sensitivily(%) Specificity(%) Sensitivity(%) Specificity(%) RC 94 77 92 80 LN Spread 77 81 71 81 Our model was highly sensitive in predicting RC and, to a lesser extent, tumor spread to one or more LN. The model was more sensitive than NAF cytology (36%) or pathology parameters (83%) alone. When only subjects with normal NAF cytology were evaluated by LR, the combined model was again more sensitive (92 vs 83%) than information on pathologic parameters alone. Conclusion: NAF cytology combined with pathologic parameters are highly sensitive in predicting the presence of RC in women with close or positive margins after EB.