Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 1.
Published in final edited form as: Am J Surg. 2021 Jul 22;223(1):81–93. doi: 10.1016/j.amjsurg.2021.07.018

Metastatic Breast Cancer: Who benefits from surgery?

Caitlin E Marks a, Samantha M Thomas b,c,d, Oluwadamilola M Fayanju a,b,e,f, Gayle DiLalla a,g, Sarah Sammons b,h, E Shelley Hwang a,b, Jennifer K Plichta a,b,e
PMCID: PMC8688223  NIHMSID: NIHMS1728186  PMID: 34325907

Abstract

Background:

We sought to identify characteristics of metastatic breast cancer (MBC) patients who may benefit most from primary tumor resection.

Methods:

Recursive partitioning analysis (RPA) was used to categorize non-surgical patients with de novo MBC in the NCDB (2010–2015) into 3 groups (I/II/III) based on 3-year overall survival (OS). After bootstrapping (BS), group-level profiles were applied, and the association of surgery with OS was estimated using Cox proportional hazards models.

Results:

All patients benefitted from surgery (median OS, surgery vs no surgery): 72.7 vs 42.9 months, 47.3 vs 30.4 months, 23.8 vs 14.4 months (all p<0.001) in BS-groups I, II, and III, respectively. After adjustment, surgery remained associated with improved OS (HR 0.52, 95% CI 0.50–0.55). The effect of surgery on OS differed quantitatively across groups.

Conclusion:

Prognostic groups may inform the degree of benefit from surgery, with the greatest benefit seen in those with the most favorable survival.

Keywords: metastatic breast cancer, stage IV breast cancer, breast surgery, survival

INTRODUCTION

In the United States, an estimated 154,794 women are living with metastatic breast cancer (MBC), and approximately 6% of all new breast cancer diagnoses each year are metastatic.1 The 5-year survival rate for MBC is 27%, compared to 99% and 86% for women with localized and regional disease, respectively.1 As such, contemporary management of MBC focuses primarily on extension of life, symptom management, and supportive care.2 In alignment with the guidelines published by the National Comprehensive Cancer Network (NCCN), resection of the primary tumor in this patient population is typically recommended for symptomatic and/or palliative reasons.2 Furthermore, it is recommended that surgery be performed only if the whole primary tumor can be resected, and distant metastatic sites are not currently life-threatening.2

While the palliative benefits for select patients are often clear, data exploring the potential survival benefit of resection of the primary tumor in patients with MBC is conflicting. Numerous retrospective studies have demonstrated that surgical resection is associated with a survival benefit in the MBC population, and more specifically, in those with favorable hormone receptor statuses, fewer metastases, a good response to systemic therapy, negative margins, and no evidence of disease.3–6 However, most retrospective studies have been criticized for selection bias, a criticism that is further supported by matched analyses demonstrating little to no survival benefit with surgery.7–10 Furthermore, results from prospective studies have also been conflicting. In a 2015 trial, Badwe et al reported that in those who respond to upfront chemotherapy, survival was not significantly affected by locoregional treatment,11 which is similar to findings presented recently by Khan et al.12 In contrast, the MF07-01 randomized clinical trial demonstrated that surgery was associated with a survival benefit, and factors shown to be associated with improved survival included young age, estrogen-receptor (ER)-positive and human-epidermal-growth-factor-receptor-2 (HER2)-negative disease, and bone-only metastases.13 There is clinical data to suggest that systemic therapy followed by surgical resection may lead to long term disease free intervals in de novo metastatic breast cancer patients with oligometastatic disease, including those with a limited number of metastatic tumors involving a single or few organs.14 Furthermore, the value of primary breast surgery and metastasectomy vs ablative radiotherapy to metastatic sites is currently being investigated in several ongoing clinical trials (NCT02364557, NCT03750396).15,16

Based on these conflicting findings, it remains unclear if surgical resection of the primary tumor may improve survival for some patients with MBC, and if so, which patients are most likely to benefit from surgery. Using a large national database, we sought to identify which patients with de novo MBC are most likely to benefit from surgical resection of the primary tumor and to develop patient profiles that could be used in clinical settings to guide surgical discussions.

METHODS

Patients diagnosed with de novo MBC (clinical and/or pathological M1 disease) from 2004 to 2016 were selected from the National Cancer Database (NCDB, Participant User File 2017). Patients with age less than 18y or greater than 80y, and those with histologies other than those listed in the World Health Organization (WHO) classification17 were excluded. Additionally, those with missing/unknown comorbidity score, clinical T-stage, clinical N-stage, tumor grade, ER status, progesterone-receptor (PR) status, HER2 status, and/or metastatic site data were also excluded, as were those with unknown surgery type and those with surgery listed as “surgery NOS” or “local tumor destruction”. Given the primary endpoint (survival), those with missing survival information were also excluded. Based on these inclusion/exclusion criteria, all patients diagnosed before 2010 were excluded due to unreliable reporting of HER2 status and metastatic site, and all patients diagnosed in 2016 were excluded due to administrative missingness of survival data imposed by the NCDB. Therefore, the final cohort was restricted to those diagnosed from 2010 to 2015. Importantly, patients who did not receive any systemic therapy were also excluded.

Overall survival (OS) was defined as the time from date of diagnosis to death. Patients who did not die were censored at the date of last follow-up. ER, PR, and HER2 status recorded as “borderline” were redefined as negative for all analyses. Hormone-receptor-positive (HR+) was defined as having ER+ and/or PR+ disease, while triple-negative disease was defined by having ER−/PR−/HER2− tumors. Study cohorts were defined by resection of the primary tumor in the breast, referred to as ‘surgical’ (lumpectomy or mastectomy) and ‘non-surgical’.

Recursive partitioning analysis (RPA) was used to categorize patients into prognostic groups based on OS. RPA is a decision-tree-style classification technique that selects the covariate that would result in the most significant difference and splits the sample based on that covariate. After the initial split, each covariate is re-evaluated with each new subgroup, and the most significant split is again selected. This process continues until a pre-specified stopping criteria is achieved. Potential variables included in the recursive partitioning models included clinical T-stage (cT), clinical N-stage (cN), tumor grade, ER status, PR status, HER2 status, bone-only metastases, and number of metastatic sites. This analysis was conducted using the non-surgical cohort only. A p-value of 0.10 was set as the maximum splitting criteria to conduct the recursive partitioning. The full recursive partitioning tree was used to define patient profiles for each terminal node. Terminal nodes were then amalgamated based on estimated 3-year survival rate into 3 groups: >50% 3-year survival rate (group I), 30%−50% 3-year survival rate (group II), and <30% survival rate (group III). This process was repeated using a bootstrap simulation with 1000 iterations, thus providing some internal validation. After bootstrapping, each characteristic profile (based on cT/cN/grade/ER/PR/HER2/number of metastatic sites) was assigned to a final bootstrap (BS)-RPA group I/II/III based on the group it was most commonly assigned to out of 1000 iterations.

Patient characteristics were summarized for surgical vs. non-surgical patients using N (%) for categorical variables and median (interquartile range, IQR) for continuous variables. Differences between patient cohorts were tested using the Chi-square or Fisher’s exact test for categorical variables, and t-test for continuous variables. Unadjusted OS was estimated using the Kaplan-Meier method, and the log-rank test was used to test for differences in OS. Unadjusted OS analyses were conducted for surgical vs. non-surgical groups stratified by the original RPA groups I/II/III, and BS-RPA groups I/II/III separately. Cox proportional hazards models were used to estimate the association of receiving surgery (vs no surgery) and RPA group I/II/III or BS-RPA group I/II/III (modeled separately) with OS after adjustment for available covariates including age, gender, year of diagnosis, facility type and location, Charlson/Deyo comorbidity score, insurance status, race/ethnicity, income level, education level, and treatment with chemotherapy, radiation, and endocrine therapy. Additional adjusted models including study group*RPA group I/II/III, and study group*BS-RPA group I/II/III interaction terms were also conducted to determine if the association of surgery vs. no surgery with OS differed by RPA group I/II/III, or BS-RPA I/II/III. Additional models were conducted stratifying by RPA group I/II/III, and BS-RPA group I/II/III. All adjusted survival models included a robust sandwich covariance estimator to account for the correlation of patients treated at the same facility.

No adjustments were made for multiple comparisons. Only patients with complete data for all covariates were included in each analysis, and effective sample sizes are included for all tables and figures. All statistical analyses were conducted with SAS version 9.4 (SAS Institute, Cary, NC) or R version 3.5.1 (The R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Study Population and Disease Characteristics

After applying the inclusion/exclusion criteria, 24,808 patients with MBC were identified: 9,735 (39.2%) underwent surgery of the primary tumor and 15,073 (60.8%) did not undergo surgery (Figure 1). The median age of the overall population was 58y (IQR 50–66), and the median follow-up was 45.5 months (95% confidence interval (CI) 44.7–46.2). Compared to those who did not undergo surgery, the surgical cohort was slightly younger (median age 57y vs 59y; p<0.001) and had lower Charlson/Deyo Comorbidity scores (score ≥2: surgery cohort 3.5% vs no surgery cohort 4.2%; p=0.04). The surgical cohort included a higher proportion of patients with HER2+ (27.5% vs 26.3%) and triple-negative (16.5% vs 12.8%; overall p<0.001) disease. Not surprisingly, the surgical cohort was more likely to have only 1 site of metastatic disease (84.3% vs 62.8%; overall p<0.001) and bone-only metastases (40.3% vs 38%; p<0.001). Furthermore, the surgical cohort was more likely to receive chemotherapy (78.5% vs 64.3%; p<0.001), endocrine therapy if HR+ (80.9% vs 76.1%; p<0.001), and radiation therapy (50.2% vs 29.4%; p<0.001). Of the 5525 patients who received systemic therapy (chemotherapy and/or endocrine therapy) prior to surgery, the median time from first treatment to surgery was 161 days (IQR 132–206). Within the surgical cohort, the majority (73%) underwent mastectomy. (Table 1)

Figure 1.

Figure 1.

Open in a new tab

Patient selection based on the defined inclusion and exclusion criteria.

Table 1.

Patient characteristics for the entire cohort and stratified by receipt of surgery. Analysis based on a cohort of patients with de novo metastatic breast cancer in the National Cancer Database (diagnosed 2010–2015). Data reported as N (%) unless otherwise specified. Percentages may not add up to 100 due to rounding or missing values.

All Patients (N=24,808) Did Not Undergo Surgery (N=15,073) Underwent Surgery (N=9,735) P-Value
Age (Years) <0.001
 <40 2092 (8.4%) 1076 (7.1%) 1016 (10.4%)
 40–60 11891 (47.9%) 6946 (46.1%) 4945 (50.8%)
 >60 10825 (43.6%) 7051 (46.8%) 3774 (38.8%)
 Median (IQR) 58 (50 – 66) 59 (51 – 67) 57 (48 – 65) <0.001
Gender 0.03
 Female 24437 (98.5%) 14868 (98.6%) 9569 (98.3%)
 Male 371 (1.5%) 205 (1.4%) 166 (1.7%)
Race/Ethnicity <0.001
 Non-Hispanic White 17190 (69.3%) 10319 (68.5%) 6871 (70.6%)
 Non-Hispanic Black 4426 (17.8%) 2842 (18.9%) 1584 (16.3%)
 Non-Hispanic Other 1020 (4.1%) 621 (4.1%) 399 (4.1%)
 Hispanic 1379 (5.6%) 826 (5.5%) 553 (5.7%)
Insurance Status <0.001
 Private 11670 (47%) 6487 (43%) 5183 (53.2%)
 Government 11445 (46.1%) 7377 (48.9%) 4068 (41.8%)
 Not Insured 1386 (5.6%) 1011 (6.7%) 375 (3.9%)
Charlson/Deyo Comorbidity Score 0.04
 0 20551 (82.8%) 12455 (82.6%) 8096 (83.2%)
 1 3281 (13.2%) 1987 (13.2%) 1294 (13.3%)
 ≥2 976 (3.9%) 631 (4.2%) 345 (3.5%)
Grade <0.001
 1 1834 (7.4%) 1227 (8.1%) 607 (6.2%)
 2 10418 (42%) 6841 (45.4%) 3577 (36.7%)
 3 12556 (50.6%) 7005 (46.5%) 5551 (57%)
Clinical T-Stage <0.001
 cT0 86 (0.3%) 51 (0.3%) 35 (0.4%)
 cT1 3845 (15.5%) 2223 (14.7%) 1622 (16.7%)
 cT2 8245 (33.2%) 4722 (31.3%) 3523 (36.2%)
 cT3 4161 (16.8%) 2545 (16.9%) 1616 (16.6%)
 cT4 8471 (34.1%) 5532 (36.7%) 2939 (30.2%)
Clinical N-Stage <0.001
 cN0 6440 (26%) 3480 (23.1%) 2960 (30.4%)
 cN1 11385 (45.9%) 7406 (49.1%) 3979 (40.9%)
 cN2 3239 (13.1%) 1909 (12.7%) 1330 (13.7%)
 cN3 3744 (15.1%) 2278 (15.1%) 1466 (15.1%)
Tumor Phenotype <0.001
 HER2+ 6639 (26.8%) 3966 (26.3%) 2673 (27.5%)
 HR+/HER− 14633 (59%) 9176 (60.9%) 5457 (56.1%)
 TNBC 3536 (14.3%) 1931 (12.8%) 1605 (16.5%)
Surgery Type <0.001
 Lumpectomy 2628 (10.6%) 0 (0%) 2628 (27%)
 Mastectomy 7107 (28.6%) 0 (0%) 7107 (73%)
 No Surgery 15073 (60.8%) 15073 (100%) 0 (0%)
Treatment with Chemotherapy 17338 (69.9%) 9695 (64.3%) 7643 (78.5%) <0.001
Treatment with Radiation 9326 (37.6%) 4437 (29.4%) 4889 (50.2%) <0.001
Treatment with Endocrine Therapy
 Among HR+ Patients 14732 (77.9%) 8945 (76.1%) 5787 (80.9%) <0.001
Metastatic Site
 Bone-only 9658 (38.9%) 5732 (38%) 3926 (40.3%)
 Brain-only 265 (1.1%) 163 (1.1%) 102 (1%)
 Liver-only 1892 (7.6%) 990 (6.6%) 902 (9.3%)
 Lung-only 2464 (9.9%) 1317 (8.7%) 1147 (11.8%)
Number of Metastatic Sites <0.001
 1 17666 (71.2%) 9460 (62.8%) 8206 (84.3%)
 2 5239 (21.1%) 3981 (26.4%) 1258 (12.9%)
 3 1668 (6.7%) 1417 (9.4%) 251 (2.6%)
 4 235 (0.9%) 215 (1.4%) 20 (0.2%)
Open in a new tab

IQR: interquartile range, ER: estrogen receptor; PR: progesterone receptor; HER2: human epithelial grown factor-2; HR: hormone receptor (defined as ER+ and/or PR+); TNBC: triple negative breast cancer (defined as ER−/PR−/HER2−); RPA: recursive partitioning analysis.

Recursive Partitioning Analysis

The initial RPA stratified the non-surgical cohort into 25 groups based on OS (Figure 2). Nearly all of the potential variables included in the RPA were found to be significantly associated with survival and determining these groups (clinical T-stage, clinical N-stage, tumor grade, ER status, PR status, HER2 status, and number of metastatic sites), except for the presence of bone-only metastases. Of the 25 groups created by the RPA, the 3-year survival rates ranged from 5.3% to 61.0% (Supplemental Table 1). Based on the distribution of the survival rates, the 25 smaller groups from the RPA were amalgamated into 3 larger groups (I/II/III). The final 3 RPA groups were defined by 3-year survival rates as follows: group I, >50%; group II, 30–50%; group III, <30% (Supplemental Table 1). In general, group I was characterized by fewer metastatic sites, ER+ status, and/or HER2+ status, while all patients with ≥3 sites of metastatic disease and/or triple negative disease were assigned to group III. Unadjusted OS was significantly different for each of the 3 groups (I/II/III, log-rank p<0.001; Supplemental Figure 1).

Figure 2.

Figure 2.

Figure 2.

Figure 2.

Figure 2.

Figure 2.

Open in a new tab

Recursive partitioning tree based on patients with de novo metastatic breast cancer in the National Cancer Database (diagnosed 2010–2015) who did not undergo surgery (N=15,073). Covariates included in the original data cohort: clinical T category (0/1/2/3/4), clinical N category (0/1/2/3), tumor grade (1/2/3), ER (+/−), PR (+/−), HER2 (+/−), number of metastatic sites (1/2/3/4), and bone-only metastases (yes/no). The initial partitioning variable (number of metastatic sites) is shown in (A), while the continued partitioning for those with 1 metastatic site (B/C) and those with 2–4 metastatic site (C/D) are shown separately. Black boxes represent decision points, while white and gray shapes represent potential outcomes for the decision. Gray boxes also represent terminal nodes. S indicates number of metastatic sites, G indicates grade. ER: estrogen receptor. PR: progesterone receptor. HER2: human epidermal growth factor receptor 2.

In order to improve accuracy and generalizability, this process was repeated using a bootstrap simulation with 1000 iterations, which yielded similar stratification of the non-surgical cohort into 3 distinct groups (log-rank p<0.001; Figure 3). The median OS was 53.7 months for group I, compared to 36.7 months for group II and 18.0 months for group 3. After bootstrapping, each characteristic profile (based on cT/cN/grade/ER/PR/HER2/number of metastatic sites) was assigned to a final bootstrap (BS)-RPA group I/II/III based on the group that it was most commonly assigned to out of 1000 iterations. These same characteristic profiles (determined the by RPA) were then applied to the surgical cohort.

Figure 3.

Figure 3.

Open in a new tab

Unadjusted overall survival for bootstrap recursive partitioning analysis (RPA) groups (I/II/III). Analysis based on a cohort of patients with de novo metastatic breast cancer in the National Cancer Database (diagnosed 2010–2015). Median Survival reported in months. CI: confidence interval.

Comparison of the Surgical and Non-Surgical Cohorts

Within each group (I/II/III), unadjusted OS was significantly better for patients in the surgical cohort (bootstrapped data represented in Figure 4, all log-rank p<0.001). Among patients assigned to BS-RPA group I – which had the most favorable disease characteristics and outcomes – median OS was 72.7 months for patients in the surgical cohort vs 42.9 months for surgery non-recipients, yielding an overall improvement in survival of 29.8 months. For patients assigned to BS-RPA group II (bootstrapped data), median OS was 47.3 months for patients in the surgical cohort, compared to 30.4 months for patients in the non-surgical cohort with an overall improvement in survival of 16.9 months. For patients assigned to BS-RPA group III (bootstrapped data) who had the most advanced/aggressive disease, median OS was 23.8 months for patients in the surgical cohort vs only 14.4 months for patients who did not undergo surgery (absolute benefit from surgery was 9.4 months).

Figure 4.

Figure 4.

Figure 4.

Figure 4.

Open in a new tab

Unadjusted overall survival for patients with de novo metastatic breast cancer in the National Cancer Database (diagnosed 2010–2015), stratified by surgery vs. no surgery for the 3 subgroups created by the RPA: (A) Bootstrap RPA Group I; (B) Bootstrap RPA Group II; (C) Bootstrap RPA Group III. Median Survival reported in months. CI: confidence interval, RPA: recursive partitioning analysis.

After adjustment for relevant patient, facility, and treatment characteristics, surgery remained significantly associated with an improved OS (bootstrapped data: HR 0.52, 95% CI 0.49–0.54; Supplemental Table 2). However, the effect of surgery on OS varied by the assigned BS-RPA grouping (I/II/III; interaction p=0.01; Supplemental Table 3). After stratifying patients by their BS-RPA group (I/II/III), patients in all BS-RPA groups benefitted from surgery (I: HR 0.49, 95% CI 0.45–0.53; II: HR 0.55, 95% CI 0.51–0.59; III: HR 0.52, 95% CI 0.49–0.56) (Table 2 highlights the effect of surgery, after adjusting for other covariates; Supplemental Table 4 presents the findings from the complete adjusted model, including surgery and all other covariates in the model).

Table 2. Adjusted Overall Survival Stratified by Bootstrap RPA Group.

Model adjusted for age, gender, race/ethnicity, insurance status, income level, education level, Charlson/Deyo comorbidity score, facility type, facility location, year of diagnosis, treatment with chemotherapy, treatment with endocrine therapy, and treatment with radiation therapy.

I (N=8622) II (N=8031) III (N=5903)
Hazard Ratio (95% CI) P-Value Overall
P-Value		 Hazard Ratio (95% CI) P-Value Overall
P-Value		 Hazard Ratio (95% CI) P-Value Overall
P-Value
Study Group <0.001 <0.001 <0.001
 No Surgery REF REF REF
 Surgery 0.487 (0.446–0.532) <0.001 0.546 (0.510–0.585) <0.001 0.521 (0.486–0.557) <0.001
Open in a new tab

CI: confidence interval, REF: reference level, RPA: recursive partitioning analysis.

DISCUSSION

In this study, we used a large national database (NCDB) of patients with de novo MBC and an advanced statistical method (RPA) to create patient profiles and identify those that may benefit the most from resection of the primary tumor. All patients who underwent surgery experienced improved OS compared to those who did not undergo surgery, with patients in RPA group I having the longest survival and the most benefit from surgery (nearly 30 months for group I vs 9.4 months for group III). In general, the patients in RPA group I had fewer metastatic sites and more favorable receptor status (ER+ and/or HER2+), compared to RPA groups II and III. These results support ongoing clinical trials evaluating the value of primary breast surgery and metastasectomy or ablative radiotherapy in oligo-metastatic breast cancer. In the interim, it may be reasonable to consider local-regional therapy, such as surgery, in these particular patients that appear to already have more favorable outcomes.

The findings from our work parallel those of other retrospective studies where surgical resection of the primary tumor was associated with a survival benefit in the MBC population.18–22 However, findings from subsequent prospective trials have been conflicting, and despite NCCN guidelines suggesting that surgery be performed only for symptomatic and/or palliative reasons, resection of the primary tumor is still being performed for many patients.5,13,23–25 For example, a retrospective study of 24,015 women with MBC in the NCDB (2003–2012) reported that among patients who were alive 1 year after diagnosis, 43.8% underwent resection of the primary tumor.5 In the current study, we similarly noted that 39.2% of patients received surgery. Proponents for surgery argue that resection of the primary tumor reduces tumor burden and activates the immune system to better aid systemic therapy, while those who advocate against surgical resection suggest that surgery of the primary tumor does not improve survival, and may actually promote the formation of distant metastases by triggering processes that promote cancer growth.26

Given the consistent findings of most retrospective studies suggesting a survival benefit to surgery, multiple prospective randomized clinical trials have been conducted, which have not conclusively demonstrated a survival benefit. Many would suggest that the conflicting findings from the retrospective vs. the prospective studies relate to the non-randomization and inherent selection bias in retrospective studies, which is inadequately addressed in most analyses. Therefore, some speculate that the patients who underwent surgery in the retrospective studies likely had a lower burden of disease than those who did not undergo surgery, thus conferring an additional survival benefit.22 Indeed, we did note similar differences in the current study with surgical patients being significantly more likely to have only 1 site of metastatic disease. Given the conflicting findings from the retrospective and prospective studies, we sought to determine if there may be a subgroup of patients who may benefit from surgery that has not yet been clearly defined in the literature. While we acknowledge that many patients with a substantial disease burden (≥3 sites of metastatic disease) are unlikely to benefit from resection of the primary tumor, prospective studies to identify the subgroup of patients with MBC that may benefit from surgery are unrealistic, as the number of patients required with any given disease subtype would be too high to achieve statistical significance. Therefore, we used the NCDB, a large national database with a significant number of patients in nearly all subcategories, to identify those patients who are most likely to benefit from resection of the primary tumor. In contrast to other retrospective studies, our unique methodology and statistical approach allowed us to build patient profiles that can be used to more easily identify patients that are most likely to benefit from surgery.

Similar to our work, others have also sought to identify a subgroup of patients within the overall population of patients with MBC who would receive some survival benefit from primary resection of the tumor. For example, Xiao et al.’s meta-analysis published in 2018, which considered both retrospective and prospective trials, concluded that primary resection of the tumor should not be routinely performed in patients with MBC, but may be an option for those with clear margins and few metastatic sites.27 More recently, a retrospective study by Li et al. using the SEER database primarily concluded that surgery reduced the risk of mortality by 40–50% (similar to our findings) and was an independent prognostic factor; in addition, they also noted that those patients who were younger, had a small primary tumor, only 1 distant metastasis, and negative surgical margins were more likely to benefit from surgery.28 Stahl et al.’s retrospective study suggested that surgery be considered for patients with ER+, PR+, or HER2+ disease after neoadjuvant chemotherapy,29 who are generally similar to the group I patients in the current study (ER+ and/or HER2+). Co et al. published a smaller study in 2019 of 1,769 patients from a prospectively-maintained database and reported a statistically significant difference in 5-year survival between the surgical and non-surgical group; in addition, the multivariate analysis revealed that old age and visceral metastases were associated with a worse survival, while ER+ status was associated with an improved survival.30

In comparison to our study, the randomized control trail, MF07-01, drew similar conclusions about which patient characteristics are associated with improved OS after surgery.13 In an unplanned subgroup analysis, Soran et al. demonstrated that patients who were younger (<55y), with HR+/HER2− disease, and bone-only metastases experienced a lower risk of death.13 In the United States, many have awaited the results of the ECOG-ACRIN 2108 trial, and early analyses suggest that there is no survival benefit to surgical resection of an intact primary tumor when compared to systemic therapy alone.12,31 Furthermore, they demonstrated that those with triple-negative breast cancer who received surgery had a worse survival. However, the patients in this trial were only stratified at study entry by HR status, which is only one of the clinicopathological features that may be important in determining who is most likely to benefit from surgery. These findings, and those from the current study, suggest that many patients derive a survival benefit from resecting the primary tumor. However, in addition to the potential effect of surgery on survival, it is also important to recognize the impact of surgery on one’s health-related quality of life, which may be temporarily worsened by local-regional therapy.31

However, all of the previously published studies focused on individual factors that may be associated with survival and potential benefits from surgery, which make them difficult to apply to one’s clinical practice as each disease profile is shaped by multiple factors. As opposed to listing positive/negative associations with each individual variable, we constructed patient profiles, based on key disease characteristics (cT/cN/grade/ER/PR/HER2/number of metastatic sites), that can be applied to any patient. Thus, if a patient has ER+ disease (generally favorable) but 3 sites of metastatic disease (generally less favorable), our groupings provide a more accurate prognostic estimate and may enhance shared decision-making and patient-surgeon discussions.

Similar to the current study, our group recently proposed a staging system for patients with de novo MBC that also stratifies patients into 3 unique subgroups, each associated with a specific prognosis.32 As such, it may be reasonable to apply those stage groups to MBC patients in order to identify those that are most likely to benefit from surgery. Similarly, Lin et al. has also proposed dividing patients with MBC into 3 subgroups: M1a, M1b, and M1c.33 In this proposed stratification system, group M1a consisted of those with a single site of metastases that did not include brain or liver, M1b consisted of liver-only metastasis or multiple metastases except those with brain or liver involvement, and M1c consisted of any brain involvement or liver plus one other site besides the brain. In summary, their findings suggest that patients in the M1a group who underwent surgical resection of the primary tumor experienced a significant survival benefit compared to chemotherapy alone.33 A recent study evaluated hepatectomy plus systemic therapy (n = 136) versus systemic therapy alone (n=763) for isolated breast cancer liver metastases (n = 763).34 Only those patients with ER−, PR−, HER2+ breast cancer had PFS and OS benefit from hepatectomy, highlighting the role of intrinsic breast cancer subtypes in metastatic surgical resection. Taken together, multiple studies – including ours – have now demonstrated that it is possible to create subgroups of patients with MBC that may have a better prognosis and are more likely to benefit from surgical resection of the primary tumor.

Other researchers have used advanced statistical methods to predict who might benefit from surgery. For example, Kommalapati et al. developed and validated a prognostic scoring model using the NCDB (2004–2015) to predict the survival of patients with de novo MBC treated with surgery and radiotherapy.35 Overall, those with lower prognostic scores benefited the most from local-regional therapy. Similar to our findings, they demonstrated that factors such as low tumor grade, smaller tumor size, and a favorable receptor status (i.e., not triple-negative) led to lower prognostic scores and thus better overall survival.35 Zheng et al. also developed and validated a preoperative prognostic nomogram using the SEER database for patients who are eligible for surgery to assess both OS and breast cancer specific survival.36 Similarly to our study, factors such as low tumor grade, lower T-stage, having a single site of metastasis, and HR+ status were again found to be favorable clinicopathological factors on the nomogram.36 Thus, these select clinicopathological features, if present in patients with MBC, may indicate that surgical resection of the primary tumor may offer a survival benefit. Regardless of the method used, multiple authors have now demonstrated that there are several ways to identify subgroups of patients that may benefit from surgery, including the strategy outlined in the current study.

The limitations of our study accrue from its retrospective approach, as well as the overall design of the NCDB.37 One of the most signification limitations is the inherent selection bias present in retrospective studies, which is likely inadequately addressed by the available data and statistical methods employed. Although we attempted to mitigate the selection bias as much as possible through various approaches, residual bias and/or confounding that cannot be quantified and/or identified is probable. However, while the differences between our findings and those of prospective studies11,12,31 on this topic appear to be conflicting, it is unlikely that our findings can be fully explained by bias alone, and therefore, it is most likely that there is truly some degree of survival benefit for select patients with metastatic breast cancer. In addition, some patients in the NCDB may have been diagnosed with metastatic disease shortly after having breast surgery for presumed non-metastatic disease (staging scans may have been performed after surgery). Furthermore, it is unclear if or how the order of treatments may impact survival outcomes. Unfortunately, others have shown that the data on systemic therapies may not be accurately entered in NCDB,38 although this is less likely for patients with MBC, for whom systemic therapy is the main treatment. While breast cancer-specific survival may have been a better endpoint, patients with MBC are most likely to die from their cancer, and thus, OS is considered a reasonable alternative, particularly since cancer-specific-survival data is not available in NCDB. Despite these limitations, the NCDB provides data on a large population of patients with MBC (>24,000), thus allowing for multiple subgroup comparisons and detailed adjusted analyses, which are rarely possible with smaller prospective studies.

CONCLUSIONS

Select clinicopathological features can stratify patients with de novo MBC into subgroups and may inform the degree of benefit from surgery. Although patients with already favorable features, such as fewer metastatic sites and favorable receptor status (ER+ and/or HER2+), may experience the longest survival when undergoing surgery, other MBC patients may also benefit from surgical intervention.

Supplementary Material

1
2

HIGHLIGHTS.

  • Recursive partitioning placed metastatic breast cancer patients into groups I/II/III.

  • Group I patients had few metastatic sites and more favorable receptor status.

  • Group I patients had the highest overall survival, although all patients benefitted from surgery.

ACKNOWLEDGEMENTS

The National Cancer Data Base (NCDB) is a joint project of the Commission on Cancer (CoC) of the American College of Surgeons and the American Cancer Society. The CoC’s NCDB and the hospitals participating in the CoC NCDB are the source of the de-identified data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.

DISCLOSURES

  • The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.

  • Dr. J. Plichta is a recipient of research funding by the Color Foundation (PI: Plichta). She serves on the NCCN Breast Cancer Screening Committee and the ASCO Clinical Practice Guideline Committee for the Management of Male Breast Cancer.

  • Dr. E.S. Hwang serves on the NCI Breast Cancer Steering Committee and the NCCN Breast Cancer Prevention Committee.

  • The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

FUNDING SOURCES

  • Dr. O. Fayanju is supported by the National Institutes of Health (NIH) under Award Number 1K08CA241390 (PI: Fayanju).

  • This work was in part supported by Duke Cancer Institute through NIH grant P30CA014236 (PI: Kastan) for the Biostatistics Core.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  • 1.American Cancer Society: Breast Cancer Facts & Figures 2019–2020. American Cancer Society. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/breast-cancer-facts-and-figures/breast-cancer-facts-and-figures-2019-2020.pdf. Published 2019. Accessed 10/30/2019. [Google Scholar]
  • 2.Gradishar WJ, Anderson BO, Abraham J, et al. NCCN Clinical Practice Guidelines in Oncology: Breast Cancer. Online 9/8/2020 2020.
  • 3.Thomas A, Khan SA, Chrischilles EA, Schroeder MC. Initial Surgery and Survival in Stage IV Breast Cancer in the United States, 1988–2011. JAMA surgery. 2016;151(5):424–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Pons-Tostivint E, Kirova Y, Lusque A, et al. Survival Impact of Locoregional Treatment of the Primary Tumor in De Novo Metastatic Breast Cancers in a Large Multicentric Cohort Study: A Propensity Score-Matched Analysis. Ann Surg Oncol. 2019;26(2):356–365. [DOI] [PubMed] [Google Scholar]
  • 5.Lane WO, Thomas SM, Blitzblau RC, et al. Surgical Resection of the Primary Tumor in Women With De Novo Stage IV Breast Cancer: Contemporary Practice Patterns and Survival Analysis. Ann Surg. 2019;269(3):537–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Lin Y, Huang K, Zeng Q, Zhang J, Song C. Impact of breast surgery on survival of patients with stage IV breast cancer: a SEER population-based propensity score matching analysis. PeerJ. 2020;8:e8694. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Lee JS, Toktas O, Soran A. Role of Locoregional Treatment in De Novo Stage IV Breast Cancer. Clin Med Insights Oncol. 2020;14:1179554920942440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Teshome M Role of Operative Management in Stage IV Breast Cancer. Surg Clin North Am. 2018;98(4):859–868. [DOI] [PubMed] [Google Scholar]
  • 9.Cady B, Nathan NR, Michaelson JS, Golshan M, Smith BL. Matched pair analyses of stage IV breast cancer with or without resection of primary breast site. Ann Surg Oncol. 2008;15(12):3384–3395. [DOI] [PubMed] [Google Scholar]
  • 10.Dominici L, Najita J, Hughes M, et al. Surgery of the primary tumor does not improve survival in stage IV breast cancer. Breast Cancer Res Treat. 2011;129(2):459–465. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Badwe R, Hawaldar R, Nair N, et al. Locoregional treatment versus no treatment of the primary tumour in metastatic breast cancer: an open-label randomised controlled trial. The LancetOncology. 2015;16(13):1380–1388. [DOI] [PubMed] [Google Scholar]
  • 12.Soran A, Özbaş S, Doğan L, et al. Loco-Regional Treatment for Intact Primary Tumor in Patient with De Novo Metastatic Breast Cancer; Comments andConcerns of ECOG-ACRIN 2108 Trial. Eur J Breast Health. 2020;16(3):158–159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Soran A, Ozmen V, Ozbas S, et al. Randomized Trial Comparing Resection of Primary Tumor with No Surgery in Stage IV Breast Cancer at Presentation: Protocol MF07–01. Ann Surg Oncol. 2018;25(11):3141–3149. [DOI] [PubMed] [Google Scholar]
  • 14.Hortobagyi GN. Can We Cure Limited Metastatic Breast Cancer? Journal of Clinical Oncology. 2002;20(3):620–623. [DOI] [PubMed] [Google Scholar]
  • 15.A Phase IIR/III Trial of Standard of Care Therapy With or Without Stereotactic Body Radiotherapy (SBRT) and/or Surgical Ablation for Newly Oligometastatic Breast Cancer. ClinicalTrials.gov. ClinicalTrials.gov Identifier: NCT02364557 Web site. https://clinicaltrials.gov/ct2/show/NCT02364557. Updated 3/1/2021. Accessed 3/25/2021.
  • 16.Local Treatment in Addition to Endocrine Therapy in ER-positive/HER2-negative Oligo-metastatic Breast Cancer (CLEAR): a Multicentre, Single -Arm, Phase 2 Trial. ClinicalTrials.gov. ClinicalTrials.gov Identifier: NCT03750396 Web site. https://clinicaltrials.gov/ct2/show/NCT03750396. Updated 2/15/2019. Accessed 3/25/2021. [Google Scholar]
  • 17.WHO/IARC Classification of Tumours. Vol 4. 4 ed: World Health Organization; 2012. [Google Scholar]
  • 18.Lang JE, Tereffe W, Mitchell MP, et al. Primary tumor extirpation in breast cancer patients who present with stage IV disease is associated with improved survival. Ann Surg Oncol. 2013;20(6):1893–1899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Thomas A, Khan SA, Chrischilles EA, Schroeder MC. Initial Surgery and Survival in Stage IV Breast Cancer in the United States, 1988–2011. JAMA Surg. 2016;151(5):424–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Pérez-Fidalgo JA, Pimentel P, Caballero A, et al. Removal of primary tumor improves survival in metastatic breast cancer. Does timing of surgery influence outcomes? Breast. 2011;20(6):548–554. [DOI] [PubMed] [Google Scholar]
  • 21.Chen W, Huang Y, Lewis GD, et al. Treatment Outcomes and Prognostic Factors in Male Patients With Stage IV Breast Cancer: A Population-based Study. Clin Breast Cancer. 2018;18(1):e97–e105. [DOI] [PubMed] [Google Scholar]
  • 22.Pockaj BA, Wasif N, Dueck AC, et al. Metastasectomy and surgical resection of the primary tumor in patients with stage IV breast cancer: time for a second look? Ann Surg Oncol. 2010;17(9):2419–2426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Badwe R, Hawaldar R, Nair N, et al. Locoregional treatment versus no treatment of the primary tumour in metastatic breast cancer: an open-label randomised controlled trial. Lancet Oncol. 2015;16(13):1380–1388. [DOI] [PubMed] [Google Scholar]
  • 24.Tosello G, Torloni MR, Mota BS, Neeman T, Riera R. Breast surgery for metastatic breast cancer. Cochrane Database of Systematic Reviews. 2018(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Breast Cancer (Version 4.2020). National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Web site. https://www.nccn.org/professionals/physician_gls/pdf/breast.pdf. Accessed May 25, 2020.
  • 26.Hu Q, Zhong X, Liu X, et al. Resection of primary lesion for patients with metastatic breast cancer: where are we now? Chin Clin Oncol. 2018;7(3):24. [DOI] [PubMed] [Google Scholar]
  • 27.Xiao W, Zou Y, Zheng S, et al. Primary tumor resection in stage IV breast cancer: A systematic review and meta-analysis. Eur J Surg Oncol. 2018;44(10):1504– 1512. [DOI] [PubMed] [Google Scholar]
  • 28.Li X, Huang R, Ma L, Liu S, Zong X. Locoregional surgical treatment improves the prognosis in primary metastatic breast cancer patients with a single distant metastasis except for brain metastasis. Breast. 2019;45:104–112. [DOI] [PubMed] [Google Scholar]
  • 29.Stahl K, Wong W, Dodge D, et al. Benefits of Surgical Treatment of Stage IV Breast Cancer for Patients With Known Hormone Receptor and HER2 Status. Ann Surg Oncol. 2020. [DOI] [PubMed] [Google Scholar]
  • 30.Co M, Ng J, Kwong A. De-novo metastatic breast cancers with or without primary tumor resection - A 10-year study. Cancer Treat Res Commun. 2019;19:100118. [DOI] [PubMed] [Google Scholar]
  • 31.Khan SA, Zhao F, Solin LJ, et al. A randomized phase III trial of systemic therapy plus early local therapy versus systemic therapy alone in women with de novo stage IV breast cancer: A trial of the ECOG-ACRIN Research Group (E2108). Paper presented at: 2020 ASCO Virtual Scientific Program2020. [Google Scholar]
  • 32.Plichta JK, Thomas SM, Sergesketter AR, et al. A Novel Staging System for De Novo Metastatic Breast Cancer Refines Prognostic Estimates. Ann Surg. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Lin C, Wu J, Ding S, et al. Subdivision of M1 Stage for De Novo Metastatic Breast Cancer to Better Predict Prognosis and Response to Primary Tumor Surgery. J Natl Compr Canc Netw. 2019;17(12):1521–1528. [DOI] [PubMed] [Google Scholar]
  • 34.Chun YS, Mizuno T, Cloyd JM, et al. Hepatic resection for breast cancer liver metastases: Impact of intrinsic subtypes. Eur J Surg Oncol. 2020;46(9):1588– 1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Kommalapati A, Tella SH, Goyal G, Ganti AK, Krishnamurthy J, Tandra PK. A prognostic scoring model for survival after locoregional therapy in de novo stage IV breast cancer. Breast Cancer Res Treat. 2018;170(3):677–685. [DOI] [PubMed] [Google Scholar]
  • 36.Zheng Y, Zhong G, Yu K, Lei K, Yang Q. Individualized Prediction of Survival Benefit From Locoregional Surgical Treatment for Patients With Metastatic Breast Cancer. Front Oncol. 2020;10:148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Boffa DJ, Rosen JE, Mallin K, et al. Using the National Cancer Database for Outcomes Research: A Review. JAMA Oncol. 2017;3(12):1722–1728. [DOI] [PubMed] [Google Scholar]
  • 38.Mallin K, Palis BE, Watroba N, et al. Completeness of American Cancer Registry Treatment Data: implications for quality of care research. J Am Coll Surg. 2013;216(3):428–437. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1728186-supplement-1.docx (71.8KB, docx)
2
NIHMS1728186-supplement-2.docx (82.5KB, docx)

RESOURCES