We retrospectively reviewed the clinical records in our institute from January 2012 to December 2018. The medical records of these patients were carefully reviewed to gain a comprehensive understanding of the effectiveness of breast cancer treatment and its associated factors. The inclusion criteria were as follows (1): Patients aged over 18 years. (2) Patients with HER2-0 or HER2-low (IHC 1+ or IHC 2+ and FISH-), breast cancer complete at least 4 cycles of anthracycline-taxeme-based NAC regimens.(3) Patients with a confirmed pathological diagnosis, complete clinical data, and peripheral blood indicators. (4) completed surgical treatment in our hospital with a pathological report. The exclusion criteria were as follows:(1) Patients who had distant metastasis at the time of diagnosis or bilateral breast cancer were excluded. (2) Patients who developed liver or kidney dysfunction during treatment and were unable to tolerate the treatment. (3) Patients with autoimmune diseases or other diseases that could affect peripheral blood indicators. A total of 998 patients were included in the study ( Supplementary Table 1 ).

Peripheral blood samples were collected at baseline (defined the first blood collection time for patients just admitted to the hospital) and before surgery (defined a second blood collection after at least four cycles of neoadjuvant chemotherapy is performed before the patient has surgery). The NLR was defined as the absolute neutrophil count divided by the absolute lymphocyte count. The PLR was defined as the absolute platelet count divided by the absolute lymphocyte count and the LMR was defined as the absolute lymphocyte count divided by the absolute monocyte count. The SII is defined as absolute neutrophil count multiplied by absolute neutrophil count divided by absolute lymphocyte count. Data that may affect treatment efficacy were also collected including age, ER, PR, HER-2 expression, pN, pathological type, Histological Grade, Ki67% and Molecular subtype.

Breast tumor pCR was defined as the absence of invasive carcinoma residue. Axillary lymph node pCR was defined as the absence of cancer in all dissected axillary lymph nodes.

Our study took anthracycline- (A-) based and/or taxane- (T-) based NAC regimens, repeat the cycle every three weeks for the selected regimens. All patients underwent a minimum of four cycles of NAC, antracycline and taxane regimens include: AC-T regimen: anthracyclines 100mg/m2, cyclophosphamide (C) 600mg/m2, followed by docetaxel 80-100mg/m2; TAC regimen: taxanes 75mg/m2, anthracyclines 50mg/m2, and cyclophosphamide 500mg/m2; AT regimen: taxanes 75mg/m2 and doxorubicin 60mg/m2.

Other regimens include: AC regimen: anthracyclines 90mg/m2 and cyclophosphamide 600mg/m2; TC regimen: docetaxel 80-100mg/m2 and cyclophosphamide 600mg/m2.

According to postoperative axillary lymph node pathologic response assessment, all included patients were dichotomized into two groups: pCR and non-pCR groups. The demographic,clinical and hematologic characteristics were compared between the two groups.

All statistical analyses were performed using IBM SPSS Statistics 26 software (https://www.ibm.com/cn-zh/spss). The data visualization was performed using R 4.3.0 (https://www.r-project.org/).

Categorical variables were expressed as numbers and percentages and were analyzed using the chi-square test. Normal distribution was assessed using the Shapiro-Wilk test. Normally distributed continuous variables were expressed as mean ± standard deviation (SD) and analyzed using the two-sample t-test, while non-normally distributed continuous variables were expressed as median and inter quartile range (IQR) and analyzed using the Mann-Whitney test.

A receiver operating characteristic (ROC) curve was constructed to estimate the optimal cut-off values for the baseline and preoperative NLR,PLR,MLR and SII, The cut-off values were selected based on the You-den index, which is calculated by subtracting 1 from the sum of sensitivity and specificity. Subsequently, the markers were further analyzed by dividing them into high and low values based on the optimal cut-off values. Logistic regression analysis was employed to conduct univariate analysis of the factors influencing pCR. The odds ratios (ORs) and 95% confidence intervals (CIs) were estimated for each variable. Subsequently, variables with a significance level of P<0.05 in the univariate analysis were included as potential predictors of pCR. P<0.05 is considered to indicate statistical significance for the difference.

A total of 998 patients were included in the study. The mean age at initial diagnosis was 50 years (range 23-73). Table 1 presents the baseline characteristics of the patients. The majority of patients are ER receptor positive, PR receptor positive (72.7%, 62.5%), HER2-0 (53.1%), HER2-1+ (41.4%), HER2-2+/FISH- (5.5%) and nearly 65% of patients have positive axillary lymph nodes. 78% of patients had invasive ductal carcinoma, and another 2.2% had unknown pathological type. Histology II (61.6%) and III (10.4%) were the most common grade histology, with 23.8% of histological grades unknown. A larger proportion of patients (72.3%) had a Ki67 proliferation index of ≥14%. Nearly 30% of patient’s molecular subtype belongs Luminal A, TNBC and 40% belongs Luminal B. A total of 355(35.6%) patients achieved Axillary pCR after NAT.

Peripheral blood samples were collected within 7 days before the first treatment and within 7 days before surgery Table 2 , higher baseline neutrophil count (P = 0.017), higher baseline PLR count (P = 0.019), higher baseline and preoperative NLR (P = 0.005 and 0.028, respectively) and higher baseline and preoperative SII (P = 0.002 and 0.024, respectively) were observed in the non-pCR group than in the pCR group. However, no significant differences were observed between the two groups with respect to other hematological parameters.

Based on the outcome variable pCR, ROC analysis was conducted for hematological markers to select the optimal cutoff values.

Figure 1 shows the ROC curves for baseline hematological parameters. Figure 1A represents the ROC curve for baseline lymphocyte, with an AUC of 0.513 and an optimal cutoff of 1.895. Figure 1B represents the ROC curve for baseline neutrophil, with an AUC of 0.544 and an optimal cutoff of 2.875. Figure 1C represents the ROC curve for baseline monocyte, with an AUC of 0.491 and an optimal cutoff of 0.325. Figure 1D represents the ROC curve for baseline platelet, with an AUC of 0.536 and an optimal cutoff of 225.5. Figure 1E represents the ROC curve for baseline NLR, with an AUC of 0.546 and an optimal cutoff of 1.30. Figure 1F represents the ROC curve for baseline PLR, with an AUC of 0.535 and an optimal cutoff of 116.97. Figure 1G represents the ROC curve for baseline LMR, with an AUC of 0.529 and an optimal cutoff of 6.66. Figure 1H represents the ROC curve for baseline SII, with an AUC of 0.549 and an optimal cutoff of 468.09.

When the AUC value of baseline hematological parameters was greater than 0.05, indicating that these indicators were more closely related to axillary PCR before neoadjuvant therapy. Figure 1I .

Figure 2 shows the ROC curves for preoperative hematological parameters. Figure 2A represents the ROC curve for preoperative lymphocyte, with an AUC of 0.51 and an optimal cutoff of 1.635. Figure 2B represents the ROC curve for preoperative neutrophil, with an AUC of 0.534 and an optimal cutoff of 4.035. Figure 2C represents the ROC curve for preoperative monocyte, with an AUC of 0.522 and an optimal cutoff of 0.655. Figure 2D represents the ROC curve for preoperative platelet, with an AUC of 0.499 and an optimal cutoff of 349.5. Figure 2E preoperative the ROC curve for baseline NLR, with an AUC of 0.531 and an optimal cutoff of 4.00. Figure 2F represents the ROC curve for preoperative PLR, with an AUC of 0.508 and an optimal cutoff of 117.07. Figure 2G represents the ROC curve for preoperative LMR, with an AUC of 0.532 and an optimal cutoff of 2.07. Figure 2H represents the ROC curve for preoperative SII, with an AUC of 0.527 and an optimal cutoff of 835.04.

When the AUC value of preoperative hematological parameters was greater than 0.05, indicating that these indicators were more closely related to axillary PCR after neoadjuvant therapy. Figure 2I .

The proportions of patients achieving pCR based on different clinical and tumor characteristics are presented in Table 3 . In univariate logistic regression, patients with ER≥1%, PR≥1%, Ki67>14% were more likely to achieve pCR compared to those with ER<1%(P <0.001), PR<1% (P <0.001), Ki67 ≤ 14% (P = 0.006). pCR is more readily available to patients with invasive ductal carcinoma (P = 0.000), patients with higher values of baseline neutrophil, baseline platelet, baseline NLR, baseline PLR, preoperative PLR and preoperative LMR were more likely to achieve pCR compared to patients with lower values (P = 0.002; P = 0.01; P = 0.004; P = 0.036; P = 0.012; P = 0.03 respectively). On the other hand, patients with lower values of baseline LMR, baseline SII, preoperative neutrophil, preoperative NLR, preoperative SII were more likely to achieve pCR compared to patients with higher values (P = 0.031; P = 0.009; P = 0.009; P = 0.012; P = 0.022 respectively).

In the multivariate analysis, after including variables with P values<0.05 from the univariate analysis, two clinical and tumor characteristics and two hematologic markers still maintained their independent predictive roles. These factors include: ER at biopsy (OR=2.18; 95% CI 1.43–3.32; P <0.001), pathological type at biopsy (OR=0.51; 95% CI 0.40–0.65; P <0.001), baseline platelet (OR=1.45; 95% CI 1.02–2.05; P =0.037), preoperative PLR (OR=1.63; 95% CI 1.0–2.64; P =0.046). Incorporate the above factors as shown in Figure 3 .

The proportion of ER status patients with different clinical and hematological parameters is shown in Table 4 . In univariate logistic regression, among ER-negative patients, 93% had PR < 1% (P <0.001),71% had zero HER2 status (P <0.001),64.3% had invasive ductal carcinoma (P <0.001) and 78% had Ki67 > 14% (P <0.001). The proportion of LuminalA 279patients reached 94% (P <0.001) and nearly 53% achieved pCR (P <0.001). However, there was no statistical difference between ER status in terms of age, baseline and preoperative hematologic parameters (P >0.05).