Objectives The aim of this study was to identify predictive factors of axillary lymph node metastasis (ALNM) in early breast cancer (EBC).

Methods It was a retrospective study of 209 patients with T0-T1-T2 breast cancer who underwent resection of the primary tumor and axillary staging by sentinel lymph node biopsy (SLNB) and/or axillary lymph node dissection (ALND) over 2012 at our institute.The χ2 test and Fisher’s exact probability tests were used for categorical variables,and t-test for continuous variables.Predictors of ALNM were identified by univariate and multivariable logistic regression analyses using SPSS statistical software package (version 20.0).

Results Among the 209 patients, 48.8% (102 cases) had ALNM.Factors associated with ALNM in univariate analyses were tumor clinical size (23.5% in stage T0,41.9% in stage T1,55.4% in stage T2; p=0.02),multifocality (73.7% vs 43.3%,p=0.001),lymphovascular invasion (LVI) (77.1% vs 43.1%,p<0.0001),HER2 overexpression (66.7% vs 45.1%,p=0.018) and Ki67 value ≥14% (55.5% vs 41.4%,p=0.043) as well as molecular subtype (40% in luminal A,56.7% in luminal B,66.7% in HER 2 and 36.8% in triple negative subtype, p=0.049).The presence of the estrogen receptors (ER), progesterone receptors (PR) have no influence on the risk of ALNM.However,the rate of positive ER was significantly lower in patients with ALNM (72.76% ±25.76 vs 84.19 ±19.865,p=0.002). On multivariate logistic regression model, the presence of LVI (OR=4.450,CI=1.756–11.278,p=0.002), the tumor clinical size (OR=1.261,CI=1.088–1.463,p=0.002) and the rate of ER positivity (OR=0.977,CI=0.962–0.991,p=0.002) remained as independent predictors of ALNM.

Conclusions Our results suggest that LVI, tumor size and the rate of positive ER are predictive factors for ALNM in EBC.