A retrospective study was performed on 2156 female breast cancer patients who underwent surgery and were pathologically diagnosed as having invasive ductal carcinoma after surgery from January 2014 to January 2017 in Harbin Medical University Cancer Hospital. Data collected included basic demographic characteristics of patients such as age, body mass index (BMI) and menopausal status, family history of tumors, date of diagnosis of breast cancer, and pathological features. All patients received standard treatment.

Inclusion criteria:

Exclusion criteria:

The survival status and treatment information of all patients were obtained through telephone follow-up and outpatient review in Harbin Medical University Cancer Hospital. The primary endpoint of this study was recurrence free survival (RFS), which was defined as the period from the date of surgery to the date of disease recurrence and metastasis or the end of follow-up. The status of recurrence and metastasis (including local recurrence, regional recurrence, isolated recurrence and distant metastasis) of patients was mainly judged by imaging examination (breast ultrasound, molybdenum target, chest CT, bone scan, etc.) or pathological results of tissue biopsy. The secondary endpoint was overall survival (OS), defined as the time from the date of surgery to any date of death or the date of the last follow-up. The follow-up period was until December 30, 2021.

Statistical analysis was performed using SPSS 26.0 software. The optimal cut-off point of the CONUT score for predicting recurrence, metastasis and death was determined by the receiver operating characteristics curve (ROC curve) and maximum Youden index. CONUT score and clinicopathological parameters of breast cancer patients were evaluated by Pearson’s χ² test or Fisher exact test method. Kaplan-Meier method was used to draw a survival curve. The Cox regression model was used to conduct univariate and multivariate analysis on the relationship between clinicopathological features and patients’ OS and RFS. P<0.05 for a statistically significant difference.

A total of 1216 women with breast cancer after surgery was finally eligible, of which 697 (57.3%) were confirmed to have HER2-low expression and were included in this study. The mean age of the patients was (52.0 ± 9.5) years The median age was 52 years (range: 19-79 years). The age of enrolled patients was divided into the <52 years old group and the ≥52 years old group. Total mastectomy was performed in 93.7%(653/697)patients. 104(19.4%)patients had no history of breastfeeding; 53.8%(375/697)were menopausal at the time of diagnosis; about 39% were overweight, of whom 6% were severely obese. There were 428(61.4%) patients in stage T1, 240(34.4%)patients in stage T2, and 29(4.2%)patients in stage T3; the positive rate of lymph nodes was 45.7%(319/697); The postoperative TNM stage was 269 (38.6%) for stage I patients, 303 (43.5%) for stage II patients and 125 (17.9%) for stage III patients; 546 (78.3%) for histological grade I-II and 152 (21.7%) for grade III; 577 (82.8%) for ER positive patients and 545 (78.2%) for PR positive patients. There were 118 cases of triple-negative breast cancer, accounting for 16.9% of all patients, 282 cases of Luminal A breast cancer (40.5%), and 297 cases of Luminal B breast cancer (42.6%). The baseline characteristics of the patients are shown in Table 2.

The optimal cut-off point for the CONUT score to predict postoperative recurrent metastasis in breast cancer patients was 2.5 with an AUC value of 0.716 (95% CI=0.660~0.771), as shown by the ROC curve. Based on this cut-off point, all patients were divided into a high CONUT group (≥3 points) and a low CONUT group (<3 points) (Figure 2) The results showed that CEA (χ² = 8.863, P=0.003), CA153 (χ² = 5.087, P=0.024), NLR(χ² = 7.579, P=0.006), T-stage (χ² = 9.773, P=0.008), lymph node infiltration (χ² = 17.128, P=0.001), TNM stage (χ² = 21.737, P<0.01), radiotherapy (χ² = 14.737, P<0.01), and possibly CONUT high subgroup breast cancer were influential factors (Table 2).

The median follow-up time for the whole group was 70 months (range: 9-96 months),107 (15.4%) patients developed recurrent metastases during the follow-up period, with RFS rates of 93.2%, 86.8% and 85.9% at 3, 5 and 7 years respectively. Kaplan-Meier curves demonstrated a significantly higher rate of recurrence and metastasis in the high CONUT subgroup compared to the low CONUT subgroup, with 5-year RFS rates of 78.9% and 89%, respectively, with statistically significant differences (χ² = 38.384, P<0.01) (Figure 3A). Subgroup analysis further analyzed the prognosis of breast cancer patients in the CONUT score subgroup at different stages and showed that the TNM stage was a poor prognostic factor for breast cancer patients in the high CONUT subgroup. For stage I patients, the 5-year RFS rates in the high CONUT group and low CONUT group were 87.5% and 95.3% respectively, with a statistically significant difference (χ² = 4.882, P=0.027)(Figure 3B); For stage II patients, the 5-year RFS rates in the two groups were 75.4% and 91.3% respectively, with a statistically significant difference (χ² = 14.465, P<0.01) (Figure 3C); For stage III patients, the 5-year RFS rates for the two groups were 52.8% and 73.9%, respectively, with statistically significant differences (χ² = 4.495, P=0.034) (Figure 3D).

Table 3 further analyzes the univariate and multivariate analysis affecting RFS, and the results show that the CONUT score is predictive of recurrent metastasis in breast cancer patients. Multivariate Cox regression analysis showed that lymph node metastasis (HR=3.105, 95%CI=1.998-4.826, P<0.001), tumor size (HR=0.543 95%CI=0.349-0.845, P=0.007), Ki67 status (HR=2.058, 95%CI=1.277-3.317, P=0.003) and CONUT score (HR=2.591, 95%CI=1.736-3.809, P<0.001) were independent predictors of recurrent metastasis in patients with early to mid-stage HER2 low expression breast cancer receiving regular treatment. The results showed that the CONUT score had a predictive effect on recurrence and metastasis in breast cancer patients (Table 3).

A total of 68 (9.8%) deaths occurred during the follow-up period, with overall patient survival rates of 99.2%, 96.5%, and 90.5% at 3, 5, and 7 years respectively. Kaplan-Meier curves demonstrated significantly higher mortality in the high CONUT group compared to the low CONUT group, with 5-year OS rates of 87.9% and 92.5%, respectively, with statistically significant differences (χ² = 26.3215, P<0.01) (Figure 4A); Subgroup analysis further analyzed the prognosis of breast cancer patients in the CONUT score subgroup at different stages and showed that the TNM stage was a poor prognostic factor for breast cancer patients in the high CONUT subgroup.

For stage I patients, the 5-year OS rates for the high CONUT group and low CONUT group were 96.4% and 98.5%, respectively, with a statistically significant difference (χ² = 9.200, P=0.002)(Figure 4B); For stage II patients, the 5-year OS rates for the two groups were 93.1% and 96.0%, respectively, with statistically significant differences (χ² = 5.088, P=0.024) (Figure 4C); For stage III patients, the 5-year OS rates for the two groups were 80.6% and 90.1%, respectively, with a statistically significant difference (χ² = 4.845, P=0.028) (Figure 4D).

Table 4 incorporates relevant clinicopathological variables for further univariate and multivariate analyses affecting OS, and the results indicate that the CONUT score is also predictive of overall survival in breast cancer patients. Multivariate Cox regression analysis showed that radiotherapy (HR=2.111, 95%CI=1.243~3.587, P=0.006), lymph node infiltration (HR=3.064, 95%CI=1.578-5.950, P=0.001), Ki67 status (HR=1.896, 95%CI=1.041-3.453, P=0.036) and CONUT score (HR=2.520, 95%CI=1.549-4.099, P<0.001) were independent factors affecting the survival prognosis of patients with HER2 low expression breast cancer. The above findings further illustrate the feasibility of the CONUT score in predicting survival and recurrence in breast cancer patients.

We analyzed the relationship between strong ER positivity and CONUT scores. Using 10% as the cut-off point for strong ER positivity, 513 of 697 patients were strongly ER-positive, of which 129 were in the high CONUT group and 384 were in the low CONUT group. Strongly positive ER(χ² = 1.884, P=0.170) by chi-squared test, not statistically significantly different from the CONUT group.

However, we analyzed the relationship between ER strongly positive patients in different CONUT score groups and the use of endocrine drugs by Kaplan-Meier survival curves. 114 patients in the high CONUT subgroup used endocrine drugs, and RFS (χ² = 13.606, P<0.001) and OS (χ² = 16.151, P<0.001) were statistically different. Thirty-five patients in the low CONUT group were not on endocrine medication and although there was no statistically significant difference in RFS(χ² = 3.841, P=0.050) and OS(χ² = 2.060, P=0.151), the curves changed significantly. It was concluded from the Kaplan-Meier survival curves that patients using endocrine drugs had longer RFS and OS than those not using endocrine drugs, both in the high CONUT subgroup and in the low CONUT subgroup (Figure 5).

Kaplan-Meier survival curves to statistically analyze the duration of endocrine drug use in patients in different CONUT score groups. Of the 114 patients using endocrine drugs in the CONUT high subgroup, 63 patients had been on endocrine drugs for at least 5 years, with better overall survival for patients on drugs compared to those on drugs for less than 5 years. There was no statistical difference between RFS(χ² = 1.572, P=0.210) and OS(χ² = 2.725, P=0.099). Of the 349 patients using endocrine medication in the low CONUT subgroup, 113 used medication for less than 5 years and 236 patients used endocrine medication for no less than 5 years. The analysis concluded that patients with longer duration of drug use had better survival compared to those with shorter duration of use. No statistical difference between RFS(χ² = 0.307, P=0.579) and OS(χ² = 3.573, P=0.059). The above analysis concluded that patients who had been on endocrine drugs for a longer period had relatively better overall and relapse-free survival in both the high CONUT and low CONUT subgroups.

Kaplan-Meier survival curves to statistically analyze the duration of endocrine drug use in patients in different CONUT score groups. Of the 114 patients using endocrine drugs in the CONUT high subgroup, 63 had been on endocrine drugs for at least 5 years. The analysis yielded no statistical difference in the effect of the duration of drug use in the CONUT high group on patients’ RFS(χ² = 1.572, P=0.210) and OS(χ² = 2.725, P=0.099). Of the 349 patients using endocrine drugs in the low CONUT subgroup, 113 had been using the drugs for less than 5 years and 236 patients had been using endocrine drugs for no less than 5 years. There was also no statistical difference in RFS (χ² = 0.307, P=0.579) and OS(χ² = 3.573, P=0.059) for patients by the length of drug use in the lower subgroups. However, the curves show that patients with longer duration of endocrine drug use had relatively better overall survival and relapse-free survival in both the CONUT high and CONUT low subgroups (Figure 6).