Additional radiotherapy to breast‐conserving surgery is an optional treatment for de novo stage IV breast cancer: A population‐based analysis

Abstract Background We aim to assess the value of locoregional treatment (LRT) including breast‐conserving surgery (BCS), mastectomy (MAST), and radiotherapy (RT) in patients with de novo stage IV breast cancer. Methods Patients with de novo stage IV breast cancer were retrospectively identified from the Surveillance, Epidemiology, and End Results database between 2004 and 2014. Kaplan‐Meier analysis, log‐rank tests, propensity score matching (PSM), and the multivariate Cox proportional model were used for statistical analysis. Results A total of 5798 patients were identified including 849 (14.6%), 763 (13.2%), 2338 (40.3%), and 1848 (31.9%) who received BCS alone, BCS+RT, MAST alone, and MAST+RT, respectively. The proportions of receiving BCS decreased from 35.9% in 2004 to 26.2% in 2014 (p = 0.002), and the probability of patients receiving MAST increased from 64.1% in 2004 to 74.8% in 2014 (p = 0.002). Before PSM, there was a significant difference in breast cancer‐specific survival (BCSS) among the treatment arms. Patients who received RT had better BCSS, the 5‐year BCSS was 40.5%, 52.3%, 41.5%, and 47.7% in patients treated with BCS alone, BCS+RT, MAST alone, and MAST+RT, respectively (p < 0.001). In the PSM cohort, patients treated with BCS alone had lower 5‐year BCSS compared to those treated with BCS+RT (43.9% and 52.1%, p = 0.002). However, there were comparable 5‐year BCSS between BCS+RT and MAST alone groups (51.3% and 50.1%, p = 0.872), and BCS+RT and MAST+RT cohorts (51.5% and 55.7%, p = 0.333). Similar results were confirmed in multivariate analysis. Conclusions Postoperative RT improves BCSS in patients with de novo stage IV breast cancer, and BCS+RT shows a non‐inferior outcome compared to MAST+RT. BCS+RT may be the optimal local management of de novo stage IV breast cancer.

BCS+RT and MAST+RT cohorts (51.5% and 55.7%, p = 0.333). Similar results were confirmed in multivariate analysis. Conclusions: Postoperative RT improves BCSS in patients with de novo stage IV breast cancer, and BCS+RT shows a non-inferior outcome compared to MAST+RT.
BCS+RT may be the optimal local management of de novo stage IV breast cancer.

K E Y W O R D S
breast-conserving surgery, mastectomy, metastatic breast cancer, radiotherapy, survival

| INTRODUCTION
Approximately 2 million women were newly diagnosed with breast cancer in 2018 worldwide, accounting for 11.6% of all cancer types, and breast cancer is the leading cause of cancer death in females. 1 De novo stage IV breast cancer, which is defined by metastasis from the breast and axilla to distant sites, was approximately 3.6%-6% of all newly diagnosed breast cancer. 2,3 Bone (66.3%), lung (30.3%), liver (26.1%), and brain (7.3%) were the most common sites of distant metastasis. 4 The clinical outcome in patients with de novo stage IV breast cancer was significantly poor with median survival ranging from 16 to 29 months. [5][6][7] Systemic therapy was a mainstay of treatment in patients with de novo stage IV breast cancer. 5,[8][9][10][11] Locoregional treatments (LRTs), including surgery or radiotherapy (RT), are recommended just for relieving symptoms, and whether LRT should be administrated to primary tumor remains controversial. Two large prospective studies of local surgery have yielded conflicting results 7,12 ; however, most of the current retrospective studies showed a survival advantage with local surgery to the primary tumor. [13][14][15][16] RT is still deemed to be one of the significant ways of LRT, whereas related studies are rare and inconclusive compared with surgery. A retrospective study from France regarded the effect of RT in de novo stage IV breast cancer, and they found that RT alone provided comparable metastasis progression-free survival and overall survival (OS) rates compared to the combination of RT and surgery. 17 However, another population-based study including 2207 patients with de novo stage IV breast cancer showed that post-surgery RT increased breast cancer-specific survival (BCSS) rates compared to surgery alone. 18 Given the controversy, several prospective randomized trials are in progress to sort out this concept. [19][20][21] However, these trials all focused on the value of surgery in de novo stage IV breast cancer. To the best of our knowledge, the survival benefit of RT has been rarely elucidated in the prospective study; therefore, it is worth exploring the value of postoperative RT in this disease. In addition, the optimal procedure of surgical approaches and whether additional RT after surgery has survival benefit remain unclear. In this population-based retrospective study, we aim to compare the differences in survival between breast-conserving surgery (BCS) and mastectomy (MAST) in de novo stage IV breast cancer. In addition, we further analyzed the effect of postoperative RT on survival outcomes in this patient subset.

| Database source and patient's collection
The Surveillance, Epidemiology, and End Results (SEER) database was used to identify patients. The SEER database is a population-based, open-access resource collecting information on cancer prevalence, management, and survival for approximately 28% of the United States population. 22 We included patients who met the following criteria: (a) female patients with newly diagnosed metastatic breast cancer from 2004 to 2014; (b) receiving systemic chemotherapy and local therapy including BCS and MAST; and (c) availability of data on age, race/ethnicity, tumor (T) stage, nodal (N) stage, estrogen receptor (ER), progesterone receptor (PR), surgical approaches, and RT records. Patients without positive pathological diagnosis and receiving non-beam irradiation were excluded. Non-beam irradiation was defined as RT technologies including radioactive implants and radioisotopes. We obtained the access to the SEER database, and using the data was exempt from the approval of the Institutional Review Board for not involving any private information.

| Variables
The following baseline variables were identified: age at diagnosis (<65 years, ≥65 years), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and Other), histology (invasive ductal carcinoma [IDC], invasive lobular carcinoma [ILC], and other), grade (welldifferentiated, moderately differentiated, and poorly differentiated/undifferentiated), T stage (T1, T2, T3, and T4), N stage (N0, N1, N2, and N3), ER status (ER positive and ER negative), PR status (PR positive and PR negative), treatment procedures (BCS alone, BCS+RT, MAST alone, and MAST+RT). The TNM stage was defined using the 7th edition of the American Joint Committee on Cancer staging system. 23 The primary objective of this study was BCSS, defined as the time interval from diagnosis to death from breast cancer.

| Statistical analysis
Chi-square test was used to analyze the differences in demographic and clinicopathological characteristics among the treatment groups. Survival curves were delineated with the Kaplan-Meier analysis and compared with the log-rank test. Propensity score matching (PSM) was performed using 1:1 nearest neighbor matching to balance the distribution of most demographic and clinical characteristics including age at diagnosis, race/ethnicity, histology, grade, T stage, N stage, ER status, and PR status. Cox proportional hazards multivariable regression was used to calculate the independent risk factors of BCSS. All statistical results were conducted using SPSS statistical software (version 22.0, IBM Corporation, Armonk, NY, USA). A p < 0.05 (two-tailed) was considered statistically significant.
Patients with IDC (p < 0.001) and early T and N stage (p < 0.001) were more likely to receive BCS, while patients with poorly differentiated/undifferentiated (p = 0.001) and advanced T/N stage (p < 0.001) had more possibility to receive MAST. In addition, patients with younger age, advanced T/N stage, and ER/PR-positive diseases were more likely to receive postoperative RT (Table 1).

| Survival and prognostic analyses after PSM
When performing PSM, a total of 1227 pairs were completely matched between BCS and MAST cohorts. The patient characteristics for the whole patients after PSM are shown in Table S1. After adjusting age, ethnicity, grade, histology, T/N stage, ER/PR status, patients who received BCS were associated with worse BCSS compared with those who received MAST in the whole cohort (HR: 0.82, 95% CI: 0.736-0.915, p < 0.001, Table S2). The survival curve is shown in Figure 3 (p < 0.001).

| DISCUSSION
In the present study, we aimed to assess the effect of surgical approaches and additional RT on patients with de novo stage IV breast cancer, and we found that BCS alone was significantly associated with poorer survival than BCS+RT, while additional RT to BCS achieved comparable BCSS compared with RT to MAST and MAST alone.
Traditionally, the primary therapeutic tactics in de novo stage IV breast cancer focus on systemic therapy; however, the application status of LRT including local surgery and RT remains unclear. In previous studies, 35-77.6% of the patients   7 In another two studies, 10.8% and 20.7% of the patients received post-surgery RT, respectively. However, the details of the relationship between surgical procedures and postoperative RT were not recorded in these two studies. 14,17 Therefore, there was no consensus regarding the operation patterns and RT selection in de novo stage IV breast cancer, and decisions to administrate LRT were generally made by each institution according to their treatment protocols. In our study, the probability of receiving MAST (72.2%) was similar to previous studies. 7,13,14,17 However, significant differences were found in RT administration with 13.2% and 31.9% of patients treated with BCS+RT and MAST+RT, respectively. It is noteworthy that patients who received BCS decreased by 9.7%, while the patients treated with MAST increased by 10.7% over time. However, there was no significance in the probability of receiving post-surgery RT. The reasons for the changing trends are unclear, and a possible explanation is that systemic therapy has achieved the best long-term survival, resulting in decreased BCS administration and stable post-surgery RT receipt. In addition, patients receiving MAST had more advanced T/N stage and poorer grade, leading to increased MAST selection for improving BCSS rate. 18 The effect of surgical procedures on survival outcomes in de novo stage IV breast cancer remains unclear. Numerous previous studies with large cohorts have suggested that LRT could improve survival in this disease. [13][14][15][16] However, all of them were retrospective studies, and few regarded the survival difference between BCS and MAST. A retrospective study identified 566 patients with metastatic breast cancer who received surgery and no surgery therapy, and patients treated with MAST were associated with an improved OS compared with those who received BCS (37% vs. 20%, p = 0.04). 24 However, 34% of the MAST group had preoperative chemotherapy, while only 15% of the BCS cohort received chemotherapy in their study. Another study from the National Cancer Database showed that patients receiving BCS were associated with poorer 3-year overall survival (27.7% vs. 31.8%) than those receiving MAST. 2 Interestingly, a similar result was found in the above studies that both BCS groups were more likely to have positive margins than MAST groups (55% vs. 27% by Khan et  with positive margins had significantly unfavorable survival than those with negative margins (p < 0.001), 2 which might be a relatively reasonable explanation that BCS had poor survival than MAST. The same trend was found in our study. Despite this, the result of our study is more convincing for a large sample size of 5798 patients and chemotherapy receipt in the whole cohort.
For LRT in de novo stage IV breast cancer, most studies focused on surgical treatment, while rare studies  25 However, the patients receiving RT had smaller tumor size, lower nodal burden, more bone-only metastases, less visceral and brain metastases, and more chemotherapy and endocrine therapy receipt. 25 Another retrospective analysis have found that patients treated with surgery alone had comparable local recurrence-free survival or OS compared with surgery+RT. 26 However, the details of lymph node (LN) status were not recorded in the surgery alone and surgery+RT cohort in their study, which could affect RT administration. In our population-based study with a large sample size, additional RT to BCS or MAST both achieved better survival than surgery alone in the premise of receiving chemotherapy, which was similar to previous results. 18,27 The most important is that our data had detailed information of LN status with 80.5% and 86.3% had positive LNs in the surgery alone and surgery+RT cohort, respectively. In non-metastatic breast cancer with positive LNs, post-surgery RT could reduce the locoregional recurrence rate and improve survival. 28,29 Therefore, RT may also play an important role in LRT of metastatic breast cancer with positive LNs.
Advances in systemic management including taxanebased chemotherapy, targeted therapy, and endocrine therapy allow patients to live longer with their disease. [5][6][7] Quality of life in cancer patients is increasingly regarded as a clinically relevant goal that warrant consideration, and the advantage of smaller surgery and longer survival is gradually highlighted. As our study showed, BCS+RT had non-inferior survival than MAST+RT or surgery alone; therefore, additional RT to BCS could be the optimal choice in de novo stage IV breast cancer. There are many ascendancies regarding BCS+RT as the LRT patterns in stage IV breast cancer. Firstly, patients with BCS suffer less treatment-related side effects such as inflammatory response and tissue damage than those with MAST. 30 Secondly, patients treated with BCS have a better self-image and sexual well-being, leading to better psychological health and life satisfaction than those treated with MAST. 31,32 Thirdly, BCS procedures are more likely to be carried out by experienced surgeons in teaching hospitals with detailed discussion, which is associated with a better outcome. 33 Fourthly, patients receiving BCS have better treatment compliance and tolerance, and they may have more possibilities of receiving post-surgery medical surveillance. Therefore, cooperative multidisciplinary care is highly significant in the management of de novo stage IV breast carcinoma, and the addition of RT to BCS might be the optimal treatment mode. There are several limitations to acknowledge in our study. Firstly, the data were obtained from the SEER database, and selection biases that retrospective study inherently exists could not be ruled out, even though PSM analysis was used. Secondly, detailed treatment information including chemotherapy regimen, endocrine therapy, anti-HER2 targeted therapies, RT technique, RT dose, target volume, and sequential surgery, chemotherapy, and RT data was unavailable in the SEER database. Fourthly, metastasis sites and treatment patterns after relapse were not recorded, which could also interfere with survival outcomes. The strong point of our study is that our study detailedly assessed the effect of BCS, MAST, and additional RT in de novo stage IV breast cancer based on a large population in the premise of chemotherapy, especially few scholars to explore the effect of surgical approach in this patient subset.
In conclusion, our study suggests that postoperative RT improves BCSS in patients with de novo stage IV breast cancer, and BCS+RT shows a non-inferior outcome compared to MAST+RT. BCS+RT may be the optimal local management of de novo stage IV breast cancer. More studies are needed to confirm our results.