Statistical methods

DFS was calculated as the time between diagnosis and disease relapse, or death from breast cancer. To evaluate whether increased LMR, NLR, PLR, AMC, or ALC correlated with prognosis of patients with breast cancer, we chose patients with disease relapse or cancer-related death during the follow-ups and defined them as the poor prognosis group. Other patients were included in the good prognosis group. ROC curves, and AUC, were performed to further evaluate the prediction status of LMR, NLR, PLR, AMC, and ALC. Sensitivity and specificity were calculated by the optimal cutoff points on the ROC curves, which decided the maximum value (sensitivity + specificity −1) of the Youden index [22].

The Chi-square or Fisher's exact tests were used for comparing clinicopathologic features between different LMR groups. Continuous variables, reported as lymphocyte and monocyte counts, were compared by the Wilcoxon rank-sum test. Risk factors of DFS were analyzed by univariate analysis with the log-rank test and multivariate analysis with the Cox proportional hazard model. The survival curves were performed by the Kaplan–Meier method and the comparisons between groups were assessed by the log-rank test. The results were showed as hazard ratios HRs and 95% CIs. The significance of the relative differential expression of FAS was determined by the two-tailed unpaired t test. Statistical significance was decided as P-values <0.05. All statistical analyses were performed using the IBM SPSS 20.0 software (IBM, USA).