e22204 Background: Breast cancer patients with nipple discharge (ND) may constitute a distinct clinical entity compared to those without ND, which has not been well elucidated in previous reports. We sought to analyze the demographic features and prognostic profiles of breast cancer patients presenting with nipple discharge in Chinese population. Methods: A total of 3234 patients, categorized as ND (2.47%) and non-nipple discharge (NND; 97.53%) according to different initial signs, were retrospectively analyzed. Survival curves were performed with Kaplan-Meier method and annual recurrence hazard was estimated by hazard function. Results: ND group tended to have smaller tumors and less axillary lymph node (ALN) involvement than NND group (P < 0.05). In the univariate analysis, we found significantly different recurrence-free survivals (RFS) between subgroups (P = 0.0182), with the rate of 77.78% for ND group and 72.46% for the other at the 11th year respectively. In Cox proportional hazards regression analysis, we found that tumor size (P < 0.001), ALN status (P < 0.001) were independent prognostic factors for RFS. To further evaluate whether prognostic effect of ND status remained unabated over time, the test for lack of proportionality was performed. However, it was statistically significant (global test, P = 0.039), which hinted at a demand for the employment of Cox non-proportional hazards regression in this analysis. In time dependent Cox model, ND status (P = 0.0495) as well as ERBB2 status (P = 0.017), tumor size (P < 0.001), ALN status (P < 0.001) were independent prognostic factors when ND and ERBB2 status were taken as time-varying covariates. Annual hazard rates for recurrence remained lower on ND group until approximately 4.5 years after surgery, while the situation was just the opposite thereafter. Conclusions: Our findings suggested that demographic features and prognostic profiles were quite different between breast cancers presenting with and without ND in Chinese population, which may indirectly uncover distinct biological behaviors and potentially enlighten novel therapeutic approaches for patients with different initial signs. No significant financial relationships to disclose.
Prediction score model for non-sentinel and four or more nodal metastases using a combined method of one-step nucleic acid amplification and histology in sentinel node-positive breast cancer patients
