Exploring the Value of Additional Primary Tumour Excision Combined with Systemic Therapy Administered in Different Sequences for Patients with de Novo Metastatic Breast Cancer

Introduction Uncertainty still remains regarding the survival improvement derived from immediate surgery or subsequent surgery in addition to systemic therapy for patients with de novo metastatic breast cancer. The current study aimed to examine the effect of combined treatment administered in different sequences on the survival of these patients. Materials and Methods We conducted a retrospective cohort study of patients with de novo stage IV breast cancer in the Surveillance, Epidemiology, and End Results (SEER) database from 2010 to 2019. Patients were categorized into 3 groups: (1) systemic therapy without primary surgery, (2) systemic therapy after primary surgery, and (3) systemic therapy before primary surgery. Cumulative incidence curves with Gray's test were used to compare breast cancer-specific death (BCSD) between groups. Kaplan–Meier curves with the log-rank test were applied to compare overall survival (OS) between groups. A competing risk model and a proportional hazards model were generated to adjust for important prognostic factors. Propensity score matching (PSM) was performed in the primary survival analysis. Stratified analysis was also performed. Results Patients who underwent systemic therapy after primary surgery and who underwent systemic therapy before primary surgery both showed a significantly reduced risk of BCSD compared to patients who received systemic therapy without primary surgery [subdistribution hazard ratio (SHR): 0.74; 95% confidence interval (CI): 0.69–0.79; and P < 0.001, and SHR: 0.62; 95% CI: 0.56–0.67; and P < 0.001, respectively]. A statistically significant disparity was also noted in OS. In the setting of single-organ metastasis, including the bone, lung, and liver, patients receiving the combination therapy showed an improved prognosis compared with patients receiving systemic therapy without primary surgery. Conclusions Additional primary tumour excision, whether before or after systemic therapy, may provide survival benefits for patients presenting with de novo metastatic breast cancer, especially for patients with single-organ disease involving the bone, lung, and liver but not the brain. Further investigations mainly focused on these carefully selected candidates are required to improve personalized treatment for metastatic breast cancer.


Introduction
Breast cancer is the most common cancer diagnosed in women, and approximately 6% of tumours have metastasized at the first presentation [1,2]. Systemic therapy is recommended by the NCCN Panel as the primary treatment approach for the management of metastatic breast cancer, while surgery after initial systemic treatment is only considered to treat specific localized problems [3].
Studies have been investigating whether primary tumour excision may provide survival benefits in the setting of de novo metastatic breast cancer over the past two decades. An early retrospective study based on the National Cancer Database (NCDB) first reported a prolonged overall survival (OS) after surgical resection of the primary tumour [4]. Since then, more retrospective studies have claimed a survival advantage associated with primary tumour excision [5,6]. More importantly, the potential survival benefit provided by primary tumour surgery reportedly increases over time [7]. ese findings have challenged the current pattern of performing surgical procedures with palliative intent in the management of de novo stage IV breast cancer. Four randomized controlled trials (RCTs) investigating this treatment modality have attempted to answer the question, but have obtained discrepant results [8][9][10][11][12]. Heterogeneity among the studies and small sample sizes increase the difficulty of reaching a definitive conclusion.
According to the design of prior prospective studies, two studies prescribed subsequent surgery following systemic therapy in the experimental arm [8,9], while the other two prescribed surgery prior to systemic therapy accordingly [10,11]. Despite a similar design in ABCSG-28 and MF07-01, conflicting results were obtained regarding the survival improvement of patients receiving upfront surgery [10,11]. e MF07-01 trial continued to show a survival advantage for patients who underwent upfront surgery at the 10-year follow-up assessment [10]. On the other hand, receipt of upfront surgery, simultaneously resulting in a delay of firstline systemic therapy and inability to monitor drug response, was reported to facilitate cancer metastasis in ABCSG-28 [11]. ESMO guidelines have proposed that only patients with a good response to initial systemic therapy may be offered primary tumour surgery [13]. erefore, uncertainty still remains regarding the timing and role of primary tumour excision. To date, a limited number of retrospective studies have focused on the effect of the combination of systemic therapy and primary tumour surgery administered in different sequences [6].
e study was thus initiated to examine the effects of upfront surgery and subsequent surgery in addition to systemic therapy compared with systemic therapy without primary surgery on de novo metastatic breast cancer.

Data Collection.
Patients with malignant breast tumours from 2010 to 2019 were identified in the Surveillance, Epidemiology, and End Results (SEER) database: Incidence-SEER Research Plus Data, 17 Registries, Nov 2021Sub (2000 using SEER * Stat software (version 8.3.9.2) [14]. Patients with metastatic disease were further selected from the patients above (N � 35,517). All included patients were aged over 18 years and under 100 years. Patients with noninvasive disease, an unknown sequence of systemic therapy and surgery, an unknown type of surgical procedure, missing follow-up data, unknown cause of death, or survival time less than 1 month after diagnosis were excluded from the analysis. Patients who received systemic therapy both before and after surgery, who received surgery both before and after systemic therapy, who received intraoperative systemic therapy, who received surgery of the primary site without systemic therapy, and who did not receive any systemic therapy or surgery of the primary site were also excluded. Moreover, patients with unknown metastatic status of the bone, brain, liver, and lung and patients who had metastasis to other distant organs rather than the bone, brain, liver, and lung were excluded. Ultimately, patients were categorized into 3 groups: (1) systemic therapy without primary surgery, (2) systemic therapy after primary surgery, and (3) systemic therapy before primary surgery. e flowchart of the patient selection process is presented in Figure 1.

Statistical Analysis.
Baseline characteristics were compared across treatment groups using the Pearson's chisquared test for categorical variables. e percentage of patients receiving each treatment modality was plotted against the year of diagnosis to reflect the overall trend of adopting each modality. A multivariate logistic regression model was generated to evaluate the correlation between treatment sequence and type of surgical procedure among patients undergoing surgery of the primary site. e survival outcomes measured in this study were breast cancer-specific survival (BCSS) and OS. Cumulative incidence curves of breast cancer-specific death (BCSD) were plotted and compared using Gray's test. Kaplan-Meier curves of OS were plotted and compared using the log-rank test. In multivariate analyses of BCSS and OS, Fine-Gray regression models and Cox regression models were implemented to identify the independent prognostic variables. In the primary survival analysis, propensity score matching (PSM) was employed to reduce confounding biases. In the subset analysis, the patients were stratified according to the metastatic site, age, tumour subtype, grade, and disease stage. A two-sided P-value <0.05 was considered statistically significant. All statistical analyses were performed using R (version 4.1.1).

Results
Of the 35,517 patients diagnosed with stage IV breast cancer from 2010 to 2019, 15,012 met the inclusion criteria and were analysed in the study, including 10,774 patients receiving systemic therapy without primary surgery, and the remaining 2,948 and 1,290 patients receiving systemic therapy after and before primary surgery, respectively.

Primary Survival
Analysis. e cumulative incidence curve shows the lowest unadjusted BCSD in patients who underwent combined treatment with systemic therapy and primary tumour surgery ( Figure 3). In the adjusted competing risk model, combination therapy remained a favourable prognostic factor compared with systemic therapy only (systemic therapy after primary surgery: SHR: 0.74; 95% CI: 0.69-0.79; and P < 0.001, and systemic therapy before primary surgery: SHR: 0.62; 95% CI: 0.56-0.67; and P < 0.001) ( Table 3). Moreover, patients receiving systemic therapy before primary surgery had a significantly reduced risk of BCSD than those undergoing upfront surgery followed by systemic therapy (SHR: 0.83; 95% CI: 0.76-0.62; and P < 0.001).
e Kaplan-Meier curve revealed the greatest improvement in OS for patients who underwent systemic therapy before primary surgery, followed by patients who underwent systemic therapy after primary surgery and those who underwent systemic therapy without primary surgery in sequence (5-year OS: 41.8%, 38.6%, and 26.9%, respectively) ( Figure 4). e results were confirmed by the multivariate analysis using a Cox proportional hazards model (Supplementary Table 1). e multivariate analysis of BCSS and OS in the whole study population also revealed that patients who received non-primary surgical procedure to distant site had a significantly reduced risk of BCSD (SHR: 0.80; 95% CI: 0.69-0.93; and P � 0.003) and all-cause death (HR: 0.76; 95% CI: 0.66-0.87; and P < 0.001) compared to those who did not (Table 3 and Supplementary Table 1).
Propensity score matching was performed to exclude the imbalance of baseline characteristics between different treatment groups. After 1 : 1 matching for patient, tumour, metastatic, and treatment characteristics, 2609 and 1283 patients receiving systemic therapy without primary surgery were identified to be compared with patients in the groups receiving systemic therapy after and before surgery, respectively. e distribution of baseline characteristics was well balanced as shown in Supplementary Tables 2 and 3. e results of the survival analysis after PSM were consistent with those obtained before PSM. Compared with patients in the nonsurgery group, patients undergoing systemic therapy after surgery showed a significantly reduced risk of BCSD (SHR: 0.75; 95% CI: Patients with any of the following: (N=20,505) Aged 100 years and over or aged under 18 years Noninvasive disease    Tables 4 and 5 Table 4 and Supplementary Table 6). Similar results were obtained for patients with liver-only and lungonly metastases. No difference was observed between combination therapy and monotherapy for patients with brain-only metastases ( Figure 5 and Supplementary Figure 1). After comparing the two treatment modalities administered in different sequences, systemic therapy before surgery was superior to systemic therapy after surgery in the whole study population and the subgroup of patients with lung-only metastasis in terms of BCSS.  In terms of the effect of distant site surgery on patients with single-organ metastasis, multivariate analysis in patients with lung-only metastasis showed that receipt of distant site surgery approached the borderline of significance in reducing the risk of BCSD (SHR: 0.61; 95% CI: 0.36-1.03; and P � 0.07) and reached the level of significance in reducing the risk of all-cause death (HR: 0.47; 95% CI: 0.29-0.75; and P � 0.002). For patients with single-organ metastasis involving the bone, liver, or brain, the impact of distant site surgery on survival was not statistically significant.
In the subgroup of patients younger than 40 years old, a numerical trend was observed in favour of systemic therapy after primary surgery compared with systemic therapy alone in terms of BCSS (SHR: 0.81; 95% CI: 0.64-1.03; and P � 0.09). Except for this young subgroup, in any subgroup based on age, tumour subtype, grade, and disease stage, the addition of primary tumour excision was associated with significantly prolonged BCSS and OS, regardless of the sequence of primary surgery and systemic therapy (Supplementary Figures 2 and 3). In the setting of multiple-organ involvement, combined treatment brought both significant BCSS and OS advantages to patients with bone plus liver involvement compared to monotherapy.

Discussion
We performed a retrospective population-based cohort study to assess the value of primary tumour excision before and after systemic therapy in patients with de novo stage IV breast cancer. Our results revealed a statistically significant association between primary tumour excision before or after systemic therapy and prolonged BCSS and OS. A subgroup analysis based on the metastatic site found that patients with single-organ-involved diseases, including bone, lung, and liver metastasis, may derive a survival benefit. In any subgroup based on age, tumour subtype, tumour differentiation, and disease stage, a combination of systemic therapy and primary tumour surgery resulted in superior survival compared to systemic therapy alone in terms of both BCSS and OS.
In the pooled analysis of three related RCTs comprising 880 patients, their overall survival did not differ by the addition of surgical treatment in the primary breast site [15]. Pooling reported data on locoregional progression and distant progression showed that locoregional treatment exerted a favourable effect on time to locoregional progression and an unfavourable effect on time to distant progression [9,11,15]. No prospective studies have reported results on BCSS, while our study lacks data on locoregional or distant progression, and thus a comparison is unrealistic.
As the first RCT published on this topic, the results derived from the Indian population suggested a similar survival benefit derived from surgery after systemic therapy and nonsurgery treatment. e patients were believed to be diagnosed late and undertreated because most of the patients had symptomatic disease, and 92% of HER2+ patients did not receive HER2-targeted therapy. e 2-year survival of the nonsurgery group in their study was also absolutely lower than that in our study (43.0% vs. 58.4%) [9]. e latest update of data from the MF07-01 trial showed prolonged survival with additional upfront surgery after 10 years of follow-up, consistent with the results obtained after a median follow-up period of 40 months. In contrast, no  Only patients who underwent combined treatment with systemic therapy and surgery of the primary site were included in this model. e logistic regression model was adjusted for year of diagnosis, age, grade, T category, N category, and number of sites of metastasis. e variables above were selected using the stepwise AIC method in both directions. e reference category of surgical procedure is partial mastectomy. Abbreviations: OR, odds ratio; CI, confidence interval. difference was noted in 36-month survival. e weakness of the MF07-01 trial was poor randomization by biomarker status, specifically referring to more patients with HR + tumours and fewer patients with triple-negative disease in the surgery group [10,12]. e ABCSG-28 trial with a similar study design to MF07-01 yielded no survival benefits with primary tumour surgery, and its insufficient recruitment made it underpowered to provide firm conclusions [11]. e most recently published trial reported that locoregional therapy had no effect on overall survival but had a favourable effect on locoregional control [8].

e Breast Journal
Although a substantial number of observational studies have confirmed that primary tumour excision might provide additional survival benefits, they had different opinions on potential beneficiaries of the surgical intervention [16][17][18][19].
Other matched-pair analyses did not identify any survival difference between the surgery group and the nonsurgery group and thus attributed the survival advantage of primary tumour excision to case selection bias [20,21]. A prior study accounting for the treatment sequence of locoregional surgery and systemic therapy reported a noticeable survival improvement with primary tumour excision, regardless of whether surgery was performed before or after systemic therapy, but the authors did not compare the survival outcomes of these two groups or perform any subgroup analyses [6]. Another study also confirmed the advantageous effect of primary tumour resection on survival and found no survival difference between the two treatment groups according to the timing of surgery [22]. e current study suggested that systemic therapy before surgery was superior to systemic therapy after surgery in the whole study population and the subgroup of patients with lung-only metastasis, which deserves further research.
Patients undergoing systemic therapy before surgery may represent a subpopulation with a favourable prognosis since they putatively respond well to systemic therapy. Moreover, in consideration of the foregoing limitations of upfront surgery, patients receiving systemic therapy before surgery were distinguished from patients undergoing upfront surgery. In two prospective trials where patients received immediate systemic therapy before randomization, only patients without disease progression were admitted to subsequent trials [8,9]. In the current study, patients with rapid progression after initial systemic therapy were most likely to be included in the group receiving systemic therapy alone, which may exaggerate the effect of systemic therapy before surgery compared with nonsurgery on survival. On the other hand, patients undergoing upfront surgery also included patients who responded poorly to systemic therapy. Hence, the results suggesting that systemic therapy plus primary tumour excision is superior to systemic therapy alone are convincing to some extent.
Regarding the metastatic pattern, the current study focused on a more limited disease, which made it easier to interpret the potential benefits of surgery. Moreover, patients with metastasis to a single distant organ accounted for over half of the study population. In the initial analysis   e Breast Journal before adopting the present stricter exclusion criteria, distant lymph node metastasis (DLNM) was shown to be unrelated to the survival outcomes of patients with de novo stage IV breast cancer. A prior SEER-based study also concluded that patients with de novo stage IV breast cancer along with DLNM only had similar survival to patients with N3C disease [23]. erefore, the status of DLNM was not considered. Here, better survival outcomes were obtained from a combination of primary tumour surgery and systemic therapy in patients with single distant organ metastasis including the bone, liver, and lung, but not the brain, which was similar to another SEER-based analysis [17]. e MF07-01 trial indicated the protective effect of initial surgery on the subgroup with solitary bone-only metastasis but not on the subgroup with multiple bone-only metastases [12]. Because the number of lesions in a single organ was not available, the effect of the metastatic burden on the potential effectiveness of surgical procedures in patients with single-organ metastasis was unable to be determined. Pooling data for the subgroup analysis of patients stratified by tumour subtype from prospective trials showed that primary tumour surgery did not provide a survival benefit for patients with any tumour subtype, which was the exact opposite of the results in the current study [15]. e general availability of HER2-directed therapy may decrease the magnitude of the effect on locoregional progression [15]. However, the claim was not substantiated here since patients with HER2+ disease experienced longer survival than patients diagnosed with luminal A (HER2−) disease, which indicated the general availability of HER2-directed therapy. Moreover, retrospective studies have documented prolonged survival with surgical interventions for the primary breast tumour site in the presence of effective HER2-directed therapy [24][25][26][27]. Tremendous advances in systemic treatment may also play a large role in the beneficial effect of primary tumour surgery on triple-negative disease. A subgroup analysis based on tumour differentiation proved increased BCSS and OS with surgical treatment in all subgroups, consistent with the results based on the tumour subtype. e effect of distant site surgery on survival was also evaluated in patients with single-organ involvement. Results suggested that patients with lung-only metastasis had the potential to benefit from local therapy of the metastatic lesions to the lung. But the proportion of patients undergoing distant site surgery was only 3.2% in the subgroup of lung-only involvement, which weakened the credibility of the results. e concept of breast cancer stem cells (CSCs) may partially explain why primary tumour excision improves the oncologic outcomes of patients presenting with de novo metastatic breast cancer. Breast CSCs, constituting up to 35% of the cancer cells in a tumour [28], are considered the root of tumour relapse and drug resistance [29,30]. Primary tumour excision rapidly eliminates breast CSCs. e assumption is strengthened by early evidence suggesting that the survival of women who received surgery with positive margins was comparable to that of women who received nonsurgical treatment [19]. Other underlying mechanisms include restoration of suppressed immunocompetence [31] and a decrease in the overall tumour burden [32].
is study represents the only analysis comparing the effect of combination therapy administered in different sequences with monotherapy on de novo metastatic breast cancer using the SEER database. Patients were collected in the recent period when more effective novel systemic therapies were available, which may be supported by the growing trend of prescribing systemic therapy alone, as shown in Figure 2. However, limitations still exist. First, selection bias exists in this study due to its retrospective nature. Adjustments for main prognostic factors and PSM were applied to avoid the imbalance of baseline characteristics. Second, treatment details, such as the type of systemic therapy and response to medication, were not considered due to the lack of this information in the SEER database.

Conclusions
In conclusion, the current study suggests that receipt of additional primary tumour excision before or after systemic therapy may provide survival benefits for patients presenting e reference category of treatment modality on primary site varied within different models ("systemic therapy without primary surgery" for model 1 and "systemic therapy after primary surgery" for model 2). e competing risk model was adjusted for age, race, marital status, year of diagnosis, histologic type, grade, molecular subtype, AJCC T category, AJCC N category, receipt of radiotherapy, and non-primary surgical procedure to distant site. Abbreviations: SHR, subdistribution hazard ratio; CI, confidence interval; NA, not applicable. 10 e Breast Journal with de novo metastatic breast cancer, especially for those with single-organ involvement, including the bone, lung, and liver. Additional randomized clinical trials and highquality observational studies are required to evaluate the value of primary tumour excision for these specific candidates.
Data Availability e cases were abstracted from the SEER database. Access can be obtained after submitting a request. SEER datasets and Software (RRID:SCR_003293) are available at https:// seer.cancer.gov/resources/.
Ethical Approval e study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). e study was approved by the institutional ethics board of Shanghai Ninth People's Hospital, and individual consent for this retrospective analysis was waived.

Conflicts of Interest
e author has no conflicts of interest to declare. Acknowledgments e author thanks the SEER database for providing a large amount of important data.

Supplementary Materials
Supplementary Table 1: Multivariate analysis of all-cause death: a Cox proportional hazards model. Supplementary Table 2: Baseline characteristics of patients in the group of systemic therapy without primary surgery and the group of systemic therapy after primary surgery after propensity score matching. Supplementary Table 3: Baseline characteristics of patients in the group of systemic therapy without primary surgery and the group of systemic therapy before primary surgery after propensity score matching. Supplementary Table 4: Multivariate analysis of BCSS and OS after 1 : 1 matching the cases from the group of systemic therapy without primary surgery and the group of systemic therapy after primary surgery. Supplementary Table 5: Multivariate analysis of BCSS and OS after 1 : 1 matching the cases from the group of systemic therapy without primary surgery and the group of systemic therapy before primary surgery. Supplementary Table 6: Multivariate analysis of all-cause death according to the metastatic site in patients with single-organ involvement. Supplementary Figure 1