Contralateral Prophylactic Mastectomy Improves Survival Only Marginally in Patients with Unilateral Triple-Negative Breast Cancer: A Cohort Study Based on SEER Database

Background: The effect of contralateral prophylactic mastectomy (CPM) on the survival rate of triple-negative breast cancer (TNBC) patients is still controversial. The purpose of this study was to conrm whether unilateral TNBC patients benet from CPM. Methods: 10006 patients with unilateral TNBC in the Surveillance, Epidemiology and End Results (SEER) database were enrolled in this study, propensity score matching (PSM) was applied to balance patient assignments. After PSM,3039 pairs of patients were divided into a CPM group and no-CPM group, respectively. All the patients have undergone total mastectomy or radical mastectomy. Cox proportional hazards regression models were used to evaluate overall survival (OS) and breast cancer-specic survival (BCSS) of the two groups. Subgroup analysis was introduced to exclude the effect of confounding factors. To identify potential variables for prognosis, Kaplan–Meier survival analysis and Cox regression analysis were used and were presented by Kaplan–Meier curve and forest plot separately. Results: With a median follow ‐ up time of 34.5months (IQR 1–83 months), the estimated 5-year BCSS rates for patients in the CPM group and the no-CPM group were 81.96% and 78.71%, the 5-year OS rates were 80.10% and 75.05%, respectively. CPM improved the BCSS (hazard ratio [HR]= 0.79; 95% condence interval [CI]=0.69-0.90, p=0.001) and OS (HR= 0.74; 95% CI=0.66-0.84, p<0.001) of unilateral TNBC patients. Univariate subgroup analyses revealed that there was no signicant difference in survival time for patients in stage N3 who underwent CPM or not (p>0.05). Conclusions: CPM only limitedly improved BCSS and OS in patients with unilateral TNBC undergoing total mastectomy or radical mastectomy and was not recommended for stage N3 patients. and mastectomy was 74% and 63%, respectively. Our study showed that patients undergoing CPM have a higher 5-year overall survival rate (OS) of 80.10%, and we took a more rigorous manner to eliminate the possible effect of confounders. Reducing recurrence is one of the ways to reduce the risk of death in this special population. Several studies have shown that female TNBC patients are at a higher risk of relapse and have an associated poorer prognosis compared with those with other subtypes of breast cancer (1, 23), an early peak of recurrences at 3 years after diagnosis which was responsible for the poor prognosis of triple ‐ negative breast cancer (24). CPM could prolong the BCSS of TNBC patients by decreasing the risk of subsequent contralateral breast cancer.


Introduction
Triple-negative breast cancer (TNBC) refers to a heterogeneous collection of breast cancer that lacks the expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor Receptor 2 (HER2), accounting for 12-17% of all breast cancers (1,2). According to WHO statistics in 2018, the incidence rate of breast cancer today is nearly the same as that of lung cancer (11.6% of the total cases) in both sexes combined, accounting for almost 1 in 4 cancer cases among women (3). TNBC tends to be biologically aggressive and with a lack of commonly utilized targeting agents, which is often associated with a worse prognosis and tends to relapse earlier compared with other subtypes of breast cancer (1,(4)(5)(6). About 1-4% of patients with unilateral breast cancer developed contralateral breast cancer (CBC) within 5-12 years and the risk increases by 0.5% to 0.75% per year, in triple-negative disease cases, the risk of CBC was increased steadily over time (7)(8)(9).
The risk of death for bilateral breast cancer was approximately 2.50 times higher than unilateral breast cancer (10). Therefore, reducing the incidence of contralateral breast cancer will be helpful to reduce the mortality of breast cancer patients.
Plenty of studies have demonstrated that breast cancer patients could obtain a survival advantage and decrease cancer-related anxiety from contralateral prophylactic mastectomy (CPM) (11)(12)(13)(14)(15)(16)(17). However, CPM is usually more applicable to breast cancer patients with BRCA mutations because they have a higher risk of developing contralateral breast cancer(CBC), this risk was higher in BRCA1 compared to BRCA2 carriers(18). A retrospective study demonstrated that among patients with BRCA mutations, 57.1% have the clinical TNBC subtype (19). Despite the use of CPM has been increased in breast cancer patients these years, only a limited number of studies focused on CPM in TNBC (20,21).
In this study, we used the Surveillance, Epidemiology, and End Results (SEER) database to evaluate the impact of CPM on the clinical outcome of unilateral TNBC patients undergoing total mastectomy or radical mastectomy.

Patient selection
All data used in the current study were obtained from Surveillance, Epidemiology, and End Results (SEER) 18 registries Custom Data (with additional treatment elds) of the National Cancer Institute in the United States. We used SEER*Stat version 8.3.6 to download data les directly. No informed consent or institutional review board approval was required in the present study due to the publicavailable data of SEER. The inclusion criteria were:(a) Female patients diagnosed with breast cancer between 2010 and 2016.
(b)Unilateral single primary stage M0 triple-negative breast cancer (TNBC) subtype. (c)Patients undergoing total mastectomy or radical mastectomy. Patients missing any of the following information were excluded: marital status, race recode, age at diagnosis, AJCC TNM stage, tumor size, the status of regional nodes, radiation recode, vital status recodes, and follow-up. Patients before 2010 or after 2016 were excluded because these patients were not recorded in SEER 18 registries Custom Data. People with postoperative survival time of less than one month were also excluded. The selection process was shown in Fig 1.

Data collection
The following variables were collected from all patients: age at diagnosis, marital status, race recode, laterality, grade, AJCC TNM stage, tumor size, regional nodes status, surgery type, radiation recode, chemotherapy recode, survival months, and vital status recode. In the study, breast cancer-speci c survival (BCSS) was used as the primary study outcome and was calculated from the date of diagnosis to the date of death caused by breast cancer. The secondary endpoint was overall survival (OS) (time from diagnosis to death from any cause).

Statistical analysis
The main demographic data include age at diagnosis, marital status, race recode. Age was divided into three subgroups (<40,40-59 and ≥ 60 years). Marital status comprised married and not married (divorced, widowed, single, and separated). Race categories included white, black, and other (American Indian/AK Native, Asian/Paci c Islander). Tumor features included TNM stage, tumor size, and regional nodes status. Treatment methods were classi ed into surgery (radical mastectomy and total mastectomy), chemotherapy, and radiotherapy. Propensity score matching (PSM) was applied to balance patient assignments to properly assess the effects of CPM. The Chi-square test was used to evaluate the differences between the variables. OS and BCSS were estimated by the Kaplan-Meier method, and differences between groups were assessed using the log-rank test. Univariate and multivariate Cox proportional hazards regression models were performed to identify the independent factors with a signi cant impact on patient survival. Hazard ratios (HRs) and 95% con dence intervals (CIs) were calculated for each prognostic factor. All the statistical analyses were performed using SPSS statistical software, version 26.0 (SPSS, Chicago, IL, USA). A two-tailed P < 0.05 was considered statistically signi cant.

Patient characteristics
Based on the inclusion and exclusion criteria, 10006 patients were included in our study, of which 3791(37.9%) patients underwent CPM while 6215(62.1%) patients did not undergo contralateral prophylactic mastectomy(no-CPM). Most variables showed a signi cant difference between the two groups(p<0.05), except for laterality and radiotherapy(p>0.05). Compared with those patients who did not receive CPM, patients in the CPM group were more <60 years old (90.2% vs 62.0%), more white race (80.2% vs 67.9%), more married (65.2% vs 52.4%), and more patients underwent total mastectomy (65.0% vs 49.4%). There was little difference in proportion distribution among other variables. After PSM, a total of 3039 pairs of cases were matched. We analyzed the baseline data of each subgroup and found that there was no signi cant statistical difference between these two groups besides the N stage. Table 1 summarizes the demographic and clinical characteristics of the patients in both groups before and after PSM.

Survival analysis of all population
The median follow-up time was 34.5months (IQR 1-83 months). 860(14.1%) patients died from breast cancer speci cally, 378 (12.4%) in the CPM group and 482(15.8%) in the no-CPM group. The estimated 5-year breast cancer-speci c survival (BCSS) rates for patients in the CPM group and no-CPM group were 81.96% and 78.71%, the 5-year overall survival (OS) rates were 80.10% and 75.05%, respectively. The Kaplan Meier analysis revealed that patients who underwent CPM had better survival than the patients in the no-CPM group in both BCSS and OS (both log-rank p < 0.005, Fig.2). Univariate and multivariate Cox proportional hazard models were estimated to investigate the effect of risk factors on BCSS and OS.
In the univariate analysis of BCSS, all factors were associated with BCSS in TNBC patients compared with each reference group, with the exclusion of laterality(P>0.05). All relevant factors were integrated into the multivariate Cox regression analysis of BCSS. The multivariate analysis demonstrated that patients of grade IV, more advanced T stage(T3-T4), stage N3, tumor size>50mm and more regional nodes positive resulted in an adverse survival outcome of TNBC (HR>1, P<0.05). CPM, other races, chemotherapy, and midlife (40-59 years old) were protective factors for women with TNBC (HR<1, P<0.05). The results of the multivariate analysis were shown in Table 2.
The univariate Cox regression analysis of overall survival (OS) indicated laterality and radiotherapy were not signi cantly associated with OS. According to previous clinical experience, radiotherapy was one of the conventional treatment protocols for TNBC, so radiotherapy was also included in the multivariate Cox proportional hazards model. Results showed that unmarried, grade IV, more advanced T stage(T3-T4), stage N3, larger tumors, and more regional nodes positive were proved to be positively associated with breast cancer-speci c mortality (HR>1, P<0.05). Patients in CPM, chemotherapy, midlife (40-59 years old), and the other race (American Indian/AK Native, Asian/Paci c Islander) group experience better survival (HR<1, P<0.05). CPM was shown to be the factor associated with improved cancer outcomes in both univariate and multivariate analysis of OS (univariate analysis: HR=0.78, CI= 0.69-0.88, P<0.001, and multivariate analysis: HR=0.74, CI= 0.66-0.84, P<0.001), detailed in Table 3.
In conclusion, these analyses suggested that TNBC patients undergoing radical mastectomy or total mastectomy could bene t from CPM and chemotherapy. K-M survival curve for BCSS and OS was shown in Fig. 2.

Subgroup analysis
The baseline of stage N remained unbalanced after propensity score matching (Table 1), therefore stage N may be a potential confounder and was included in the subgroup analysis. According to the results of BCSS and OS multivariate Cox survival analysis, the N stage was divided into N0-N2 group and N3 group because there was no signi cant difference between N1 -N2 patients and N0 patients (log-rank P>0.05). K-M survival curves ( Fig.3 and Fig.4) showed that stage N0-N2 patients in the CPM group had superior BCSS and OS than those in the no-CPM group (log-rank P<0.05), while stage N3 patients in both groups did not derive survival bene t from CPM (log-rank P>0.05).
The subgroup analyses were illustrated by forest plots of HRs, which revealed that factors positively associated with the prognosis of TNBC patients included CPM, chemotherapy, and middle age (40-59 years old). These two forest plots also more visualize that higher tumor stage (stage T3-T4 and N3), larger tumor tissue (tumor size >50mm), and more regional positive lymph nodes were risk factors for TNBC patients. (Fig.5 and Fig.6 ).

Discussion
It is still controversial that whether patients with TNBC can get survival bene ts from the contralateral prophylactic mastectomy. In this study, the clinical characteristics of patients with TNBC and the effect of CPM on the prognosis of TNBC patients were analyzed retrospectively, which was based on a large population from the SEER database. Our results suggest that patients with unilateral TNBC undergoing total mastectomy or radical mastectomy may obtain survival bene ts from CPM.
In a previous cohort study of 1325 TNBC patients, Adkins et al. (2011) investigated the effect of surgical approaches on the prognosis in TNBC patients (22). This study reported the 5-year overall survival rate (OS) of patients undergoing breast-conserving therapy (BCT) and mastectomy was 74% and 63%, respectively. Our study showed that patients undergoing CPM have a higher 5-year overall survival rate (OS) of 80.10%, and we took a more rigorous manner to eliminate the possible effect of confounders.
Reducing recurrence is one of the ways to reduce the risk of death in this special population. Several studies have shown that female TNBC patients are at a higher risk of relapse and have an associated poorer prognosis compared with those with other subtypes of breast cancer (1,23), an early peak of recurrences at 3 years after diagnosis which was responsible for the poor prognosis of triplenegative breast cancer (24). CPM could prolong the BCSS of TNBC patients by decreasing the risk of subsequent contralateral breast cancer.
In addition, TNBC complicated with occult cancer may be another reason for poor prognosis. The overall rate of occult cancer in breast cancer was about 7.1%, simple total mastectomy or radical mastectomy cannot completely remove occult cancer that may coexist (25). Therefore, recurrence or progression will occur. King, TA, et al. (2015) reported a part of occult breast cancer was detected in contralateral prophylactic mastectomy specimens (26). Prophylactic resection of the contralateral breast could minimize the risk of this occult cancer. Although our study could not directly prove that CPM could increase the resection rate of occult cancer, it also indirectly showed that the expanded resection could reduce the risk of death, and we speculated that this risk might be caused by occult cancer, so further histopathological examination to verify the existence of occult cancer in the contralateral breast is necessary.
Although studies by Bedrosian, et al. (2010) (14)and Brewster et al. (27)Suggested that ER-negative or hormone receptor-negative breast cancer patients were more likely to receive bene t from CPM, their studies were not speci c to the triple-negative subtype and therefore may underestimate the effect of CPM on receptor-positive patients.
In our observation, we found that not all unilateral TNBC patients could obtain a better prognosis from CPM. Unlike N0-N2 patients, there was no signi cant difference in survival of stage N3 patients with or without CPM ( Fig.3 and Fig.4). Identifying subgroups more likely to achieve better treatment outcomes from CPM may be of great signi cance in choosing a more personalized approach for patients.
There are several limitations to this study, including those inherent to a retrospective observational study design. First, the study may be affected by selection bias. Second, the demographic information provided by the SEER database does not include complications and other detailed factors. The contribution of these factors to survival bene ts cannot be evaluated. Third, the interval between surgery and radiotherapy and chemotherapy and the data of chemotherapy regimens may also have an impact on survival outcomes and provide important implications for clinical practice. Fourth, some studies have shown that BRCA1/2 mutation carriers with breast cancer have a high risk of contralateral disease, carriers of the BRCA1 gene mutation are more likely to develop triple-negative breast cancer, and these women usually choose to undergo contralateral mastectomy, to reduce the likelihood of recurrence (28, 29). However, the SEER database does not provide data on the status of BRCA1/2 mutation. Lack of availability of this information in the database limits us to evaluate the in uence of these factors. Despite certain limitations, our study is a population-based study that includes a large number of unilateral TNBC patients, and the results are convincing.

Conclusion
In conclusion, our study demonstrates that CPM does improve survival in patients with unilateral triple-negative breast cancer, excluding those with stage N3, but this improvement is very limited. Therefore, clinicians should be extra cautious when choosing whether to apply CPM. Further prospective clinical trials are still needed to validate our ndings. Declarations Ethics approval and consent to participate No ethics approval, consent to participate or consent for publication was required for this study.

Consent for publication All authors are informed and consent for publication.
Funding The authors received no nancial support for the research and authorship of this article.

Con icts of interest The authors declared no con icts of interest.
Availability of data and material All the data and material were true and availability.
Code availability Not applicable.
Authors' contributions QH, PY conceptualized the paper. PY, QL and, TL analyzed the data. The data was collected by PY, JS, TN, TL, and QL. PY and YC wrote the initial draft with all authors providing critical feedback and edits to subsequent revisions. All authors approved the nal draft of the manuscript. QH is the guarantor. The corresponding authors attest that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Variables
Before PSM p After PSM p  Figure 1