Impact of surgery on survival in breast cancer with bone metastases only: a SEER database retrospective analysis

Background It was controversial to operate on the primary site of breast cancer (BC) with bone metastasis only. We investigated the impact of surgery on BC patients with bone metastases via a SEER database retrospective analysis. Methods A total of 2917 BC cases with bone metastasis, first diagnosed between 2010 and 2015 in the Surveillance, Epidemiology, and Results Database (SEER) of National Cancer Institute were selected. We assessed the effect of different surgical procedures on survival and prognosis. Results Compared with the non-surgical group, the primary tumor surgical group showed longer median survival time (χ2 = 146.023, P < 0.001), and the breast-conserving subgroup showed the highest median survival time of 70 months (χ2 = 157.117, P < 0.001). Compared with the non-surgery group, the median overall survival (OS) of primary surgery group was longer (HR = 0.525, 95%CI = 0.467–0.590, P < 0.001), and the breast-conserving subgroup showed the longest median operative OS (HR = 0.394, 95%CI = 0.325–0.478, P < 0.001). Conclusion This study showed that primary surgery could improve the median survival time and OS of BC patients with bone metastasis. Moreover, under the condition of low tumor burden, breast conserving surgery was a better choice.


Background
Breast cancer (BC) is the most common malignant tumor threatening women's health, with 270,000 newly diagnosed BC patients yearly and approximately 42,000 related deaths, which ranks first in morbidity and second in mortality worldwide [1]. With the improvement of people's health awareness and the development of medical technology, the screening and treatment of early BC have been significantly improved. However, about 5% of BC patients had bone metastasis at the first diagnosis, and the incidence of bone metastasis was 70% in advanced metastatic BC [2,3]. Currently, the treatment principle for BC patients with bone metastasis is to relieve pain, restore function and improve the quality of life, increase the survival time, and the main comprehensive therapy includes radiation and chemotherapy [4][5][6].
Accumulating evidence indicates that the active primary site surgery was associated with survival benefits in many retrospective studies [7][8][9][10][11][12][13][14][15], although few prospective studies produced positive results [16][17][18]. Routine screening of bone metastases in patients with localized BC was not recommended by the current guideline only if signs or symptoms occurred. The effect of operation on the primary site of BC with bone metastasis is still controversial.
In this study, we used the SEER database to investigate the effect of surgical operation on the prognosis of BC patients with bone metastases from 2010 to 2015.

Data
We obtained data from SEER Medicare database, which consists of 18 population-based cancer registries. The SEER program of the National Cancer Institute collects and publishes cancer incidence and survival data encompassing approximately 28% of the United States population. In this study, we used SEER*Stat Version 8.3.6 (http:// www. seer. cancer. gov/ seers tat) from the National Cancer Institute to survey eligible patients.

Patients
We collected BC patients with bone metastases between 2010 and 2015 based on the 7th edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual [19]. We further screened the patients through inclusion criteria and exclusion criteria, inclusion criteria: (1) Pathological diagnosis of breast cancer; (2) Age > 18 years old female; (3) Simple distant bone metastases occurred clearly at the first diagnosis; (4) Complete clinicopathological information such as tumor pathological type and histological grade; (5) Complete treatment information; (6) Prognostic information is complete. Exclusion criteria: (1) Multiple primary carcinomas; (2) Bilateral breast cancer; (3) Patients with unknown T and N stages, such as T0,TX and NX, were excluded. After relevant screening, we finally leaved 2917 patients eligible for survival analyses and related research.

Clinicopathological parameters and control variables
Patients were divided into primary surgery group and non-surgery group. The Surgery group was divided into Breast Conservion Surgery (BCS) group, subcutaneous or simple Mastectomy group (Mastectomy), modified Radical Mastectomy group and Radical Mastectomy group (Radical mastectomy). Clinical pathology staging was based on AJCC 7th edition. The following clinicopathological characteristics were used as study factors: age, race, ethnic origin, marital status, T stage, N stage, laterality, behavior, ER, PR, HER2, tumor subtype and histological grade, radiotherapy, and chemotherapy.

Statistical analysis
Patient and tumor characteristics were summarized using descriptive statistics and compared using a two-sided χ 2 test for categorical variables and Student's t test for continuous variables. In the survival analysis, Kaplan-Meier method was used to calculate the median survival time, and log-rank test was used for univariate analysis. Statistically significant factors in univariate analysis results were incorporated into COX proportional hazard regression model for multivariate analysis. All the above statistical analyses were calculated by SPSS 22.0 statistical software. Two-sided P values < 0.05 were considered statistically significant.

Patient characteristics
A total of 2917 BC patients with bone metastasis only were included in this study, and clinical characteristics were listed in Table 1. The primary surgery group included 1245 cases (42.7%, 1235/2917), and the non-surgery group consisted of 1672 cases (57.3%, 1672/2917). The clinicopathological parameters showed significant differences between two groups included marital status, age, histological grade, T stage, N stage, ER, molecular subtype, radiotherapy and chemotherapy. Compared with the non-operative patients, the proportion of young patients undergoing surgery was higher (χ 2 = 21.613, P < 0.001). Compared with unmarried patients, married patients had a higher surgical rate (46.8% vs 38.8%). Patients in advanced stage had more ratio of radiotherapy and chemotherapy. Compared with ER positive, ER + / HER2− and ER + /HER2 + patients, ER negative, TNBC, and HER2-enriched BC patients had higher surgical rate (P < 0.001 and P = 0.009). There were no significant differences in race, Ethnic origin, behavior, laterality, HER2 and PR expression between the two groups.

Univariate analysis and multivariate analysis on the prognosis of BC with bone metastasis
Univariate analysis was performed for the 2917 cases (Table 2), and the clinicopathological parameters influencing the prognosis of patients were race, histologic grade, marital status, age, histology, T stage, tumor radiotherapy and chemotherapy, ER, PR, and HER2, subtypes, primary tumors surgery and surgical procedure (P< 0.05). Ethnic origin, laterality, and N staging showed no significant difference (P > 0.05). According to the results of univariate analysis, primary site surgical group significantly affected the prognosis of patients ( Fig. 1) compared to the non-surgical group (χ 2 = 146.023, P<0.001); Among the surgery group, the survival time of the breast-conserving group (Fig. 2) benefitted most for 70 months (χ 2 = 157.117, P < 0.001). Advanced stage patients showed decreased median survival time (P < 0.001). Compared with the black, the white and others got a longer survival time (χ 2 = 35.071, P < 0.001); The expression of ER, PR, and HER2, histology of invasive ductal carcinoma, married patients, younger age, radiotherapy, and

Comparison of baseline data among 1245 patients with different surgical modes
To further analyze the difference in survival analysis among different surgical methods, we used χ 2 test to further analyze the baseline data of 1245 patients with different surgical methods (Table 4). T staging, N staging, radiotherapy and chemotherapy showed significant differences (P < 0.05) in baseline data of 1245 patients, while no significant differences in other variables were found. In the BCS group, the proportion of T and N stages, radiotherapy and chemotherapy was higher: the clinical stages were lower, the tumor burden was lower, and the treatment was more active.

Discussion
De novo stage IV BC refers to a tumor that has metastasized to other organs of the body with poor prognosis. The therapeutic goal is to improve the OS and the quality of life of patients, and systemic treatment is the first choice. The intervention guideline of local surgery for BC is to relieve symptoms, remove tumor rupture, bleeding, fungal infection and cancer pain without affecting the life of the patient [6]. In our study, the pathology of BC patients with bone metastasis was mostly LuminalA type (ER+/HER2−) (72.5%), which had good prognosis due to the stable endocrine therapy and low proliferative index. Primary site surgical group significantly affected the prognosis of patients (Fig. 1) compared to the non-surgical group (χ 2 = 146.023, P < 0.001), benefitting most for 70 months in BCS subgroup (Fig. 2) (χ 2 = 157.117, P < 0.001). The expression of ER, PR, and HER2, histology of invasive ductal carcinoma, married patients, younger age, radiotherapy, and chemotherapy were all protective factors for BC with bone metastasis (P < 0.001). Although the treatment of BC is surgery-based comprehensive treatment, the survival benefit from primary surgery may be related to the following conditions: (1) Surgery could remove the primary tumor site, get rid of the primary tumor cell and tumor stem cells, and reduce the possibility of peripheral release and spread of circulating tumor cells [20][21][22]; (2) Primary surgery could play an important role at local control of patients: tumor ulcer, infection and other aspects, which improved patients' physical and psychological quality of life; (3) Primary surgery could reduce the burden of tumor and improve the curative effect of tumor chemotherapy [23,24]; (4) Resection of the primary tumor could reduce the immune suppression of the tumor on the body, activate CD4 and CD8 T lymphocytes, and stimulate the immune response of the body to tumor cells [12,25,26].
In the past related studies about advanced invasive carcinoma [23,27,28], a phenomenon had been observed in gastric cancer, ovarian cancer, colon cancer that the reduction in tumor burden and an increase in OS were associated, but it was controversial that surgery did not take a survival benefit in de novo stage IV BC [5,[16][17][18]. However, in recent years, some retrospective studies [7][8][9][10][11][12][13][14][15] had shown that resection of the primary site of de novo stage IV BC could bring survival benefits, which were most obvious in young patients with positive estrogen receptor, low tumor burden, negative human epidermal growth factor receptor, and bone metastasis only.
We analyzed previous prospective clinical studies. Firstly, the Translational Breast Cancer Research Consortium 013 (TBCRC-013 study) was a prospective multi-institutional registry trial which aimed to evaluate the role of surgery in stage IV breast cancer. Patients diagnosed with stage IV BC at presentation (group A, n = 112) or stage IV within 3 months of diagnosis (group    Early results [29] from this study showed that surgery was associated with improved survival on multivariate analysis (HR 0.28, 95% CI 0.10-0.74, P = 0.01); In addition, 3-year overall survival results showed no difference among patients who responded to first-line therapy [16]. Secondly, the prospective clinical trial was initiated at Tata Memorial Centre in India enrolling 350 patients to receive locoregional treatment (n = 173) or no locoregional treatment (n = 177). The result indicated the surgery could not take survival benefit because of unreasonable systemic therapy. In their study, most patients (92%) with HER2-positive BC did not receive trastuzumab therapy; In addition, they had more metastases ratio (75% vs 25%) while less bone metastases ratio (29%) [17]. At last, the MF07-01 trial conducted by the Turkish Federation seemed to produce positive result. OS was improved for the surgery group at 41.6% as compared to 24.4% in the no surgery group at 5 years. (46 versus 37 months, P = 0.005). Subgroup analysis showed that the survival benefit was associated with ER positive and HER2/neu-negative disease, age under 55, and bone metastases only [18]. Different primary tumor surgery methods took different survival benefit, which might be associated with baseline of patients in terms of T stage, N stage, chemotherapy, radiation therapy. There was lower tumor load, T stage, N stage levels and higher proportion of chemotherapy, radiation therapy in the BCS group, consistent with the fact that the prognosis was better in patients with bone metastasis only from BC with a lower tumor burden. Expanding the scope of surgery and lymph node dissection showed that there was no significant survival benefit [7,30,31], consistent with the results of multifactor analysis in our study, but surgical margin status was correlated with patient prognosis [7]. In addition, Her-2 overexpression was statistically significant in univariate and multivariate analyses (P < 0.05), and Her-2 overexpression might be a protective factor affecting BC bone metastasis, which might be related to anti-Her-2 targeted therapy [32,33].
Some limitations still exist in this study. Firstly, we are lack of the comprehensive information about systemic treatment, such as endocrine therapy, HER2-targeted therapy, or chemotherapy, which might lead to some bias in the survival analysis. Also, the short of data in the SEER database on events associated with bone metastasis as well as related systematic treatment [34][35][36] might lead to bias in the conclusions. Another potential issue is   PR the possibility of incomplete or inaccurate claim entry. The tumor burden of patients selected for surgery was relatively low, which was also a part of the surgical bias and might affect the results [37].

Conclusion
In conclusion, this study shows that primary surgery could improve the prognosis and OS of De novo stage IV BC patients with bone metastasis only. Under the premise of low tumor burden and comprehensive treatment, BCS is a better choice.