Abstract
Introduction
Prior studies have demonstrated variation in surgical management of breast cancer. As randomized controlled trials have increasingly shown a limited survival advantage with surgery in stage IV breast cancer, the present study aimed to assess geographic and institutional variations in locoregional surgical management of de novo stage IV breast cancer.
Methods
We conducted a retrospective cohort analysis of 68,670 patients with de novo stage IV breast cancer from 2012 to 2021 using the National Cancer Database. The proportion of patients treated with surgical resection was compared between year of diagnosis, facility type, geographic region, and facility volume while controlling for demographic and clinical variables.
Results
Overall, 17,620 patients (25.7%) with de novo stage IV breast cancer underwent breast, axillary, or metastatic site surgical resection. There was a statistically significant decrease from 27.4 to 15.0% in the proportion of patients undergoing primary tumor resection over the included timeframe (p < 0.001). Significant variation was noted in the proportion of patients receiving primary tumor resection among geographic regions, ranging from 14.9 to 23.2% (Northeast and West, p < 0.001). Patients treated at comprehensive community cancer centers were most likely to have primary site surgery, and those treated at academic centers (including National Cancer Institute–designated comprehensive cancer centers) were least likely (22.1 vs. 16.7%, respectively, p < 0.001).
Conclusion
Significant variation exists in locoregional surgical management of de novo stage IV breast cancer. Surgical intervention has decreased over time, but this study reveals that institutional and geographic factors are associated with the surgical management for these patients.
Based on Surveillance, Epidemiology, and End Results data, approximately 6% of newly diagnosed breast cancers present with de novo stage IV disease.1,2 Historically, the principle guiding management of stage IV breast cancer has been that surgical resection of the primary tumor does not alter the course of metastatic disease. As such, the management of stage IV disease has largely been driven by systemic treatment, with surgical intervention reserved for palliation of local symptoms, such as bleeding and pain, as needed.3
Retrospective data have previously been used to assess the benefit of surgical intervention in stage IV disease, and—although these studies have shown mixed results—they have often favored locoregional treatment.4–7 However, retrospective studies in patients with stage IV breast cancer are often impacted by selection bias, stage migration bias, and immortal time bias favoring the treatment arm.4–9 To address these limitations, multiple contemporary randomized controlled trials have been conducted to assess the benefit of surgical intervention for stage IV breast cancer. Four large trials, the Tata Memorial, ECOG-ACRIN 2108, ABCSG-28 POSITIVE, and JCOG 1017 PRIM-BC trials, all found no improved overall survival with surgical intervention. Only one study, the MF07-01 trial, showed improved overall survival in the surgical group.10
Despite the randomized controlled trial data, given the heterogeneity of patients with de novo stage IV breast cancer studied, controversy on surgical management remains. Given the conflicting evidence and study limitations, there remain no specific guidelines against surgery in patients with stage IV breast cancer, and decisions for or against surgery are often made on a case-by-case basis, leaving room for significant variation in care, especially for patients with a low burden of metastatic disease. The aim of the current study was to characterize contemporary variation in the management of de novo stage IV breast cancer at the local/regional level and based on surgical volume, particularly within the timeframe following publication of multiple randomized controlled trials showing negligible survival benefit with surgical intervention.
Methods
Data Source
This retrospective cohort study used the National Cancer Database (NCDB) from the American College of Surgeons and the American Cancer Society. The NCDB includes a broad array of demographic and clinical data for the purpose of comparative benchmarking among cancer patients in the United States. Over 70% of new cancer diagnoses within the United States are represented, from over 1500 Commission on Cancer–accredited facilities. Data are abstracted at each hospital using standardized criteria and nationally standardized data coding definitions. The University of Texas Southwestern Review Board deemed that the study was exempt from informed consent requirement; all data were deidentified.
Study Population
We queried the NCDB for patients aged ≥18 years with de novo stage IV breast cancer diagnosed between 2012 and 2021. All patients had International Classification of Diseases for Oncology, Third Edition (ICD-O-3) site codes C50.0–C50.9, indicating breast primary site, and histology codes 8500, 8520, 8522, 8523, 8424, 8530, and 8575, indicating ductal, lobular, inflammatory, or metaplastic carcinoma. Only patients with American Joint Committee on Cancer clinical stage IV disease were included. We excluded patients with a previous history of cancer and cases where the reporting facility did not make any treatment decisions (class of case code 00). We also excluded those with missing sociodemographic, staging, vital status or follow-up, facility type, or location.
Exposure and Covariates
To characterize variation in the operative management of stage IV breast cancer, we examined several facility-level variables. The NCDB categorizes facility types as community cancer programs, comprehensive community cancer programs, academic/research programs (including National Cancer Institute–designated comprehensive cancer centers), and Integrated Cancer Network programs. The NCDB facility regions variable was recategorized into Northeast (New England and Middle Atlantic), South (South Atlantic, Easy South Central, West South Central), Midwest (East North Central and West North Central), and West (Mountain and Pacific) regions. Both facility-level variables are suppressed in individuals diagnosed before the age of 40 years, but were able to be imputed by matching facility identification information from older patients. The total number of patients receiving care at each facility during the study period was used to calculate the facility volume, which was then broken down into quartiles. Facilities in the first quartile treated up to 44 patients with stage IV breast cancer, those in the second quartile treated 45–78, those in the third quartile treated 79–132, and those in the fourth quartile treated ≥ 133.
Additional sociodemographic and clinical factors were abstracted from the NCDB. Sociodemographic factors included age at diagnosis, sex, race and ethnicity, travel distance and median household income. Sex was categorized as male or female. Race and ethnicity was characterized as non-Hispanic white (white), non-Hispanic Black (Black), Hispanic, Asian, other, and unknown. Each patient’s zip code at diagnosis was used to determine their median household income based on information abstracted from the American Community Survey and categorized into quartiles. Clinical characteristics included year of diagnosis (2012–2014, 2015–2017, and 2018–2021), Charlson–Deyo comorbidity score, tumor histology and grade, receptor subtype (based on hormone receptor [HR] and human epidermal growth factor receptor 2 [HER2]/neu tumor receptor positivity), clinical T and N stage, metastatic site (bone, brain, liver, lung, distant lymph nodes, and other) and burden (number of metastatic sites at presentation).
Outcomes
The primary outcome was receipt of any surgical therapy. Surgical variables were broken down by surgical site into types of breast/primary site surgery (none, partial mastectomy, mastectomy, unknown), axillary surgery (none, sentinel lymph node biopsy [SLNB], axillary lymph node dissection [ALND], and Unknown), and surgery of other sites (none, procedure not specified, regional site, distant lymph nodes, distant site, combination of categories, and unknown).
Statistical Analysis
Patients were grouped by receipt of any surgical therapy. Descriptive statistics were reported using median (interquartile range [IQR]) and frequency (%) for continuous and categorical variables, respectively. Differences in sociodemographic and clinical factors between groups were determined by Wilcoxon Rank Sum and chi-squared tests for continuous and categorical data, respectively. To assess variation in care across facility-level factors and time, proportions of patients receiving systemic therapy alone, systemic therapy + surgery, systemic therapy + radiation, and trimodal therapy were computed. The proportion of patients undergoing primary site surgery across the same factors were calculated and differences in these outcomes were determined using chi-squared tests. Additionally, variation in the type of primary site surgery (partial mastectomy vs. mastectomy) and axillary intervention was assessed. Multivariable logistic regression was used to determine factors associated with primary site surgery. Factors significant on univariable analysis and those of a priori clinical significance were included in the final model. Scatterplots comparing facility volume with proportion of patients receiving primary site surgery stratified by facility type were created to better visualize the variation in care. All statistical analyses were performed using STATA software (StataCorp. 2023. Stata Statistical Software: Release 18. College Station, TX, USA; StataCorp LLC.).
Results
Study Cohort
We identified a total of 68,670 patients with de novo stage IV breast cancer diagnosed between 2012 and 2021 in the NCDB who met the inclusion criteria. Over 98% of the study population was female, and the median age of patients overall was 61 years (IQR 51–71, Table 1). Most patients were non-Hispanic white (47,691 [69.5%]), followed by Black (12,190 [17.8%]), Hispanic (4288 [6.2%]), Asian (2650 [3.9%]), and other races (1232 [1.8%]). In terms of tumor characteristics, 85.9% of patients had invasive ductal carcinoma, 11.8% invasive lobular carcinoma, 1.8% inflammatory breast cancer, and 0.5% metaplastic. In total, 53.6% of patients had HR+/HER2- disease, 13.0% had HR+/HER2+ disease, 12.3% had triple-negative disease, and 7.1% had HR-/HER2+ disease. Most patients had poorly (41.0%) or moderately (37.1%) differentiated tumors. In terms of metastatic burden, 56.0% of patients who underwent surgery had one distant disease site, 25.5% had two distant disease sites, 12.5% had three sites, and 5.9% had four or more sites. Most of these patients had bone metastases (70.1%), followed by lung (30.3%), liver (25.5%), distant lymph nodes (24.3%), brain (7.7%), and other sites (11.9%).
Table 1.
Cohort demographics
| Characteristic | Overall (n = 68,670) |
No surgery (n = 51,050) |
Any surgery (n = 17,620) |
p-value |
|---|---|---|---|---|
| Age, years | 61 (51–71) | 62 (53–72) | 58 (48–67) | < 0.001 |
| Age categories, years | ||||
| < 45 | 9937 (14.5) | 6338 (12.4) | 3599 (20.4) | < 0.001 |
| 46–55 | 13,565 (19.8) | 8356 (18.3) | 4209 (23.9) | |
| 56–65 | 19,104 (27.8) | 14,278 (28.0) | 4826 (27.4) | |
| 66–75 | 14,825 (21.6) | 11,657 (22.8) | 3168 (18.0) | |
| ≥ 76 | 11,239 (16.4) | 9421 (18.5) | 1818 (10.3) | |
| Sex, female | 67,700 (98.6) | 50,399 (98.7) | 17,301 (98.2) | < 0.001 |
| Race | ||||
| White | 47,691 (69.5) | 35,359 (69.3) | 12,332 (70.0) | < 0.001 |
| Black | 12,190 (17.8) | 9327 (18.3) | 2863 (16.3) | |
| Hispanic | 4288 (6.2) | 3044 (6.0) | 1244 (7.1) | |
| Asian | 2650 (3.9) | 1917 (3.8) | 733 (4.2) | |
| Other | 1232 (1.8) | 923 (1.8) | 309 (1.8) | |
| Unknown | 619 (0.9) | 480 (0.9) | 139 (0.8) | |
| Year of diagnosis | ||||
| 2012–2014 | 18,419 (26.8) | 12,435 (24.4) | 5984 (34.0) | < 0.001 |
| 2015–2017 | 20,691 (30.1) | 15,396 (30.2) | 5295 (30.1) | |
| 2018–2021 | 29,560 (43.1) | 23,219 (45.5) | 6341 (36.0) | |
| Charlson–Deyo comorbidity score | ||||
| 0 | 55,758 (81.2) | 41,082 (80.5) | 14,676 (83.3) | < 0.001 |
| 1 | 8530 (12.4) | 6428 (12.6) | 2102 (11.9) | |
| 2 | 2521 (3.7) | 1993 (3.9) | 528 (3.0) | |
| 3 | 1861 (2.7) | 1547 (3.0) | 314 (1.8) | |
| Histology | ||||
| IDC | 59,002 (85.9) | 43,604 (85.4) | 15,398 (87.4) | < 0.001 |
| ILC | 8099 (11.8) | 6385 (12.5) | 1714 (9.7) | |
| Inflammatory | 1210 (1.8) | 880 (1.7) | 330 (1.9) | |
| Metaplastic | 359 (0.5) | 181 (0.4) | 178 (1.0) | |
| Receptor subtype | ||||
| Triple negative | 8448 (12.3) | 5850 (11.5) | 2598 (14.7) | < 0.001 |
| HR-/HER2+ | 4873 (7.1) | 3431 (6.7) | 1442 (8.2) | |
| HR+/HER2- | 36,808 (53.6) | 28,133 (55.1) | 8675 (49.2) | |
| HR+/HER2+ | 8927 (13.0) | 6442 (12.6) | 2485 (14.1) | |
| Unknown | 9714 (14.0) | 7194 (14.1) | 2420 (13.7) | |
| Grade | ||||
| Well differentiated | 4254 (6.2) | 3406 (6.7) | 848 (4.8) | < 0.001 |
| Moderately differentiated | 25,463 (37.1) | 19,503 (38.2) | 5960 (33.8) | |
| Poorly differentiated | 38,156 (41.0) | 19,331 (37.9) | 8825 (50.1) | |
| Unknown | 10,797 (15.7) | 8810 (17.3) | 1987 (11.3) | |
| Clinical T stage | ||||
| T1 | 7268 (10.6) | 5244 (10.3) | 2024 (11.5) | < 0.001 |
| T2 | 20,499 (29.9) | 14,801 (29.0) | 5698 (32.3) | |
| T3 | 11,192 (16.3) | 8175 (16.0) | 3017 (17.1) | |
| T4 | 21,919 (31.9) | 16,628 (32.6) | 5291 (30.0) | |
| Unknown | 7792 (11.4) | 6202 (12.2) | 1590 (9.0) | |
| Clinical N stage | ||||
| N0 | 13,561 (19.8) | 9898 (19.4) | 3663 (20.8) | < 0.001 |
| N1 | 31,117 (45.3) | 23,355 (45.6) | 7762 (44.1) | |
| N2 | 7668 (11.2) | 5575 (10.9) | 2093 (11.9) | |
| N3 | 9925 (14.5) | 7137 (14.0) | 2788 (15.8) | |
| Unknown | 6399 (9.3) | 5085 (10.0) | 1314 (7.5) | |
| Metastatic site | ||||
| Bone | 48,104 (70.1) | 37,348 (73.2) | 10,756 (61.0) | < 0.001 |
| Brain | 5256 (7.7) | 4248 (8.3) | 1008 (5.7) | < 0.001 |
| Liver | 17,527 (25.5) | 14,125 (27.7) | 3402 (19.3) | < 0.001 |
| Lung | 20,820 (30.3) | 16,592 (32.5) | 4228 (24.0) | < 0.001 |
| Distant lymph nodes | 16,701 (24.3) | 12,488 (24.5) | 4213 (23.9) | 0.141 |
| Other sites | 8160 (11.9) | 6625 (13.0) | 1535 (8.7) | < 0.001 |
| Metastatic burden | ||||
| 1 site | 38,482 (56.0) | 26,052 (51.0) | 12,430 (70.5) | < 0.001 |
| 2 sites | 17,536 (25.5) | 14,093 (27.6) | 3443 (19.5) | |
| 3 sites | 8613 (12.5) | 7342 (14.4) | 1271 (7.2) | |
| 4 or more sites | 4039 (5.9) | 3563 (7.0) | 476 (2.7) | |
| Breast surgery type | ||||
| None | 55,021 (80.1) | 51,050 (100) | 3971 (22.5) | < 0.001 |
| Partial mastectomy | 3769 (5.5) | 0 (0) | 3769 (21.4) | |
| Mastectomy | 9839 (14.3) | 0 (0) | 9839 (55.8) | |
| Unknown | 41 (0.1) | 0 (0) | 41 (0.2) | |
| Axillary surgery type | ||||
| None | 56,587 (82.4) | 51,050 (100) | 5537 (31.4) | < 0.001 |
| SLNB | 2641 (3.9) | 0 (0) | 2641 (15.0) | |
| ALND | 9155 (13.4) | 0 (0) | 9166 (52.0) | |
| Unknown | 276 (0.4) | 0 (0) | 276 (1.6) | |
| Other site surgery | < 0.001 | |||
| None | 64,433 (93.8) | 51,050 (100) | 13,383 (76.0) | |
| Procedure not specified | 600 (0.9) | 0 (0) | 600 (3.4) | |
| Regional site | 516 (0.8) | 0 (0) | 516 (2.9) | |
| Distant lymph nodes | 475 (0.7) | 0 (0) | 475 (2.7) | |
| Distant site | 2490 (3.6) | 0 (0) | 2490 (14.1) | |
| Combination | 76 (0.1) | 0 (0) | 76 (0.4) | |
| Unknown | 80 (0.1) | 0 (0) | 80 (0.5) | |
| Systemic therapy | < 0.001 | |||
| No | 5091 (7.4) | 4423 (8.7) | 668 (3.8) | |
| Yes | 59,830 (87.1) | 43,414 (85.0) | 16,416 (93.2) | |
| Unknown | 3749 (5.5) | 3213 (6.3) | 536 (3.0) | |
| Distance traveled, miles | 9.0 (4.2–19.8) | 8.7 (4.1–19.1) | 9.9 (4.6–21.8) | < 0.001 |
| < 5 | 17,873 (26.0) | 13,782 (27.0) | 4091 (23.2) | < 0.001 |
| 5–10 | 14,022 (20.4) | 10,528 (20.6) | 3494 (19.8) | |
| 10–20 | 12,847 (18.7) | 9429 (18.5) | 3148 (19.4) | |
| > 20 | 14,792 (21.5) | 10,635 (20.8) | 4157 (23.6) | |
| Unknown | 9136 (13.3) | 6676 (13.1) | 2460 (14.0) |
Data are presented as median (interquartile range) or n (%) unless otherwise indicated
ALND Axillary lymph node dissection, HER2 Human epidermal growth factor receptor 2, HR Hormone receptor, IDC Invasive ductal carcinoma, ILC Invasive lobular carcinoma, SLNB Sentinel lymph node biopsy
The majority of patients (51,050 [74.3%]) did not undergo surgery as a part of their stage IV breast cancer treatment, whereas a select minority (17,620 [25.7%]) underwent surgery either for the primary breast and/or axillary lesion or at a metastatic site (Table 1). Patients who underwent surgery were younger (median age 58 years [IQR 48–67] vs. median age 62 years [IQR 53–72], p < 0.001), more likely to have HER2+ or triple-negative disease (37.0% vs. 30.8%, p < 0.001), and more likely to have a lower metastatic burden (70.5% vs. 50.0% with only one metastatic site, p < 0.001). Among the overall cohort, 19.8% of patients underwent breast surgery (partial mastectomy or mastectomy), 17.3% of patients underwent axillary surgery (sentinel lymph node biopsy or axillary lymph node dissection), and 4.4% of patients underwent surgery at a metastatic site (distant disease site or distant lymph node resection). Among those who underwent surgery of any type, 55.8% had a mastectomy, 21.4% had a lumpectomy, and 22.5% had no primary breast surgery (had axillary or metastatic site resection alone). In terms of axillary management in patients who underwent surgery of any type, 52.0% had an ALND, 15.0% had an SLNB, and 31.4% had no axillary surgery. Among patients who received surgery, 24.0% underwent surgical resection of a metastatic site. The vast majority (93.2%) of patients who underwent surgery also received systemic therapy, whereas only 85.0% of the ‘no surgery’ cohort received systemic therapy (p < 0.001).
Variation in Treatment
Variation Across Study Timeframe
There was a statistically significant decrease from 27.4 to 15.0% in the proportion of patients undergoing primary-site surgery over the included timeframe (p < 0.001, Table 2). When divided across three timeframes (2012–2014, 2015–2017, and 2018–2021), there was a significant decrease in trimodal therapy, from 15.4 to 10.9% (p < 0.001), a significant decrease in the use of systemic therapy plus surgery, from 16.4 to 9.7% (p < 0.001), and an increase from 42.0 to 50.1% in systemic therapy alone as treatment for patients with de novo stage IV breast cancer (Table 3).
Table 2.
Variation in primary site surgery by institutional factors
| Facility type | Number of patients with metastases seen | Number of patients with metastases undergoing surgery | Proportion of patients with metastases undergoing surgery (%) | p-value |
|---|---|---|---|---|
| Facility type | < 0.001 | |||
| Community | 5520 | 1200 | 21.7 | |
| Comprehensive community | 26,139 | 5782 | 22.1 | |
| Academic/research | 23,369 | 3903 | 16.7 | |
| Integrated | 13,601 | 2723 | 20.0 | |
| Facility region | < 0.001 | |||
| Northeast | 14,141 | 2112 | 14.9 | |
| South | 26,525 | 5769 | 21.8 | |
| Midwest | 16,695 | 3113 | 18.7 | |
| West | 11,268 | 2614 | 23.2 | |
| Diagnosis year | < 0.001 | |||
| 2012–2014 | 18,406 | 5048 | 27.4 | |
| 2015–2017 | 20,684 | 4137 | 20.0 | |
| 2018– 2021 | 29,539 | 4423 | 15.0 | |
| Facility volume | < 0.001 | |||
| First quartile | 17,903 | 3955 | 22.1 | |
| Second quartile | 17,645 | 3723 | 21.1 | |
| Third quartile | 16,667 | 3076 | 18.5 | |
| Fourth quartile | 16,414 | 2854 | 17.4 |
Table 3.
Variation in multidisciplinary management by institutional factors
| Variable | No therapy (n = 6406) |
Systemic therapy alone (n = 31,318) |
Systemic therapy + surgery (n = 8214) |
Systemic therapy + radiation (n = 12,096) |
Trimodal therapy (n = 8202) |
p-value |
|---|---|---|---|---|---|---|
| Year of diagnosis | < 0.001 | |||||
| 2012–2014 | 1608 (9.1) | 7433 (42.0) | 2896 (16.4) | 3049 (17.2) | 2718 (15.4) | |
| 2015–2017 | 1874 (9.4) | 8572 (48.0) | 2540 (12.7) | 3579 (18.0) | 3378 (11.9) | |
| 2018-2021 | 2924 (10.2) | 14,313 (50.1) | 2778 (9.7) | 5468 (19.1) | 3106 (10.9) | |
| Facility type | < 0.001 | |||||
| Community | 680 (12.9) | 2313 (43.8) | 772 (14.6) | 973 (18.4) | 540 (10.2) | |
| Comprehensive community center | 2594 (10.3) | 11,424 (45.4) | 3330 (13.2) | 4578 (18.2) | 3236 (12.9) | |
| Academic/research | 1751 (7.7) | 11,463 (50.6) | 2576 (11.4) | 4103 (18.1) | 2781 (12.3) | |
| Integrated | 1381 (10.5) | 6118 (46.6) | 1536 (11.7) | 2442 (18.6) | 1645 (12.5) | |
| Region | < 0.001 | |||||
| Northeast | 1368 (10.0) | 7030 (51.5) | 1387 (10.2) | 2505 (18.4) | 1357 (9.9) | |
| South | 2580 (10.1) | 11,690 (45.9) | 3522 (13.8) | 4447 (17.5) | 3243 (12.7) | |
| Midwest | 1433 (8.9) | 7762 (47.9) | 1920 (11.9) | 3092 (19.1) | 1994 (12.3) | |
| West | 1025 (9.4) | 4836 (44.3) | 1385 (12.7) | 2052 (18.8) | 1608 (14.7) | |
| Facility volume | < 0.001 | |||||
| First quartile | 2248 (13.2) | 7540 (44.1) | 2338 (13.7) | 2997 (17.5) | 1963 (11.5) | |
| Second quartile | 1731 (10.2) | 7760 (45.6) | 2151 (12.6) | 3198 (18.8) | 2191 (12.9) | |
| Third quartile | 1443 (9.0) | 7709 (47.9) | 1839 (11.4) | 3077 (19.1) | 2029 (12.6) | |
| Fourth quartile | 984 (6.1) | 8309 (51.9) | 1886 (11.8) | 2824 (17.6) | 2020 (12.6) |
Data are presented as n (%) unless otherwise indicated
Variation by Geographic Region
Significant variation was noted in the proportion of patients receiving primary-site surgical intervention among geographic regions, ranging from 14.9 to 23.2% (Northeast and West, respectively; Midwest 18.7%, South 21.8%; p < 0.001, Table 2). Rates of systemic therapy alone were highest in the Northeast (51.5 vs. 47.9% Midwest, 45.9% South, and 44.3% West; p < 0.001), and consequently rates of systemic therapy + surgery were lowest in the Northeast (10.2 vs. 11.9% Midwest, 12.7% West, and 13.8% South; p < 0.001). Rates of trimodal therapy were also lowest in the Northeast (9.9 vs. 12.3% in the Midwest, 12.7% in the South, and 14.7% in the West; p < 0.001, Table 3).
Variation by Facility Type
Significant variation in treatment was noted by facility type and volume. Patients treated at comprehensive community centers were most likely to have primary-site surgical intervention, and those treated at academic/research centers (including NCI-designated comprehensive cancer centers) were least likely (22.1 vs. 16.7%%, respectively, p < 0.001) (Table 2). Patients treated at community centers had the highest rates of no treatment (surgical, systemic, or radiation), and patients at academic/research centers were least likely to forgo treatment altogether (12.9 vs. 7.7%, p < 0.001). Patients treated at community centers also had the lowest rates of systemic therapy alone (44.8 vs. 45.4% at comprehensive cancer centers, 46.6% at integrated cancer centers, 50.6% at academic/research centers, p < 0.001). Patients treated at community centers also had the highest rates of systemic therapy + surgery (14.6 vs. 13.2% at comprehensive cancer centers, 11.7% at integrated cancer centers, and 11.4% at academic/research centers, p < 0.001) (Table 3). Community centers saw the lowest volumes of patients with stage IV breast cancer but had the highest variation in proportion of patients with stage IV disease receiving surgery (Fig. 1). Conversely, despite seeing a wide range of volumes of patients with stage IV breast cancer, academic centers had consistently lower proportions of those patients receiving surgery (Fig. 1).
Fig. 1.
Scatterplots of total volume versus proportion undergoing surgery by facility type
Variation by Facility Volume
Facilities with the lowest volume of patients with stage IV disease were significantly more likely to perform primary-site surgical intervention, with 22.1% undergoing primary-site surgery at facilities in the lowest quartile of stage IV patient volume and 17.4% undergoing primary-site surgery at facilities in the highest quartile of stage IV patient volume (p < 0.001, Table 2). Rates of forgoing treatment altogether decreased significantly with stage IV patient volume; facilities seeing the lowest volumes of patients with stage IV breast cancer had the highest rates of no treatment, whereas facilities seeing the highest volumes of patients with stage IV breast cancer had the lowest rates of no treatment (13.2 vs. 6.1%, p < 0.001). Rates of systemic therapy alone increased significantly with patient volume; facilities within the first quartile of stage IV patient volume had the lowest rates of systemic therapy alone, whereas facilities within the fourth quartile of patient volume had the highest rates of systemic therapy alone (44.1 vs. 51.9%, p < 0.001). Facilities within the first quartile of stage IV patient volume also had the highest rates of systemic therapy + surgery (13.7 vs. 12.6% at second quartile facilities, 11.4% at third quartile facilities, and 11.8% at fourth quartile facilities, p < 0.001). The greatest range of variation was seen at the individual facility level, with proportions of patients treated with surgery ranging from 0 to 100% (median 18.8% [IQR 13.1–24.4], Table 3).
Variation in Type of Primary Site (Breast and/or Axillary) Surgery
Looking more granularly at the type of primary-site surgery, rates of mastectomy decreased over time from 20.1% in 2012–2014 to 10.7% in 2018–2021 but remained higher than rates of lumpectomy in each of the timeframes (7.4 and 4.3%, respectively) (p < 0.001). Rates of ALND also decreased over time, from 19.7% in 2012–2014 to 9.3% in 2018–2021 (p < 0.001), but were higher than rates of SLNB in each of the timeframes. Rates of mastectomy were highest at comprehensive community centers and lowest at academic centers (16.1 vs. 12.1%, p < 0.001). Similarly, rates of ALND were highest at comprehensive community centers and lowest at academic centers (14.6 vs. 12.0%, p < 0.001). Rates of mastectomy were also highest at centers within the first quartile of stage IV patient volume and lowest at centers within the fourth quartile (15.7 vs. 12.7%, p < 0.001). Similarly, rates of ALND also decreased with increasing stage IV patient volume (14.2% at centers within the first quartile of stage IV patient volume vs. 12.3% at centers within the fourth quartile, p < 0.001) (Table 4).
Table 4.
Variation in type of primary site surgery
| Variable | Breast surgery | Axillary surgery | ||||||
|---|---|---|---|---|---|---|---|---|
| None (n = 55,062) |
Partial mastectomy (n = 3769) |
Mastectomy (n = 9839) |
p-Value | None (n = 56,863) |
SLNB (n = 2641) |
ALND (n = 9166) |
p-Value | |
| Year of diagnosis | ||||||||
| 2012–2014 | 13,371 (72.6) | 1354 (7.4) | 3694 (20.1) | < 0.001 | 14,189 (77.0) | 607 (3.3) | 3623 (19.7) | < 0.001 |
| 2015–2017 | 16,554 (80.0) | 1154 (5.6) | 2983 (14.4) | 17,171 (83.0) | 730 (3.5) | 2790 (13.5) | ||
| 2018–2021 | 25,137 (85.0) | 1261 (4.3) | 3162 (10.7) | 25,503 (86.3) | 1304 (4.4) | 2753 (9.3) | ||
| Facility type | ||||||||
| Community | 4327 (78.3) | 332 (6.0) | 868 (15.7) | < 0.001 | 4541 (82.2) | 208 (3.8) | 778 (14.1) | < 0.001 |
| Comprehensive community center | 20,363 (77.9) | 1583 (6.1) | 4199 (16.1) | 21,212 (81.1) | 1110 (4.3) | 3823 (14.6) | ||
| Academic/research | 19,485 (83.3) | 1075 (4.6) | 2828 (12.1) | 19,812 (84.7) | 769 (3.3) | 2807 (12.0) | ||
| Integrated | 10,887 (80.0) | 779 (5.7) | 1944 (14.3) | 11,298 (83.0) | 554 (4.1) | 1758 (12.9) | ||
| Region | ||||||||
| Northeast | 12,033 (85.1) | 709 (5.0) | 1403 (9.9) | 12,326 (87.1) | 387 (2.7) | 1432 (10.1) | ||
| South | 20,780 (78.3) | 1453 (5.5) | 4316 (16.3) | 21,539 (81.1) | 1064 (4.0) | 3946 (14.9) | ||
| Midwest | 13,588 (81.4) | 836 (5.0) | 2277 (13.6) | 13,962 (83.6) | 666 (4.0) | 2073 (12.4) | ||
| West | 8661 (76.8) | 771 (6.8) | 1843 (16.4) | 9036 (80.1) | 524 (4.7) | 1715 (15.2) | ||
| Facility volume | ||||||||
| First quartile | 13,960 (77.9) | 1112 (6.2) | 2843 (15.7) | < 0.001 | 14,662 (81.8) | 714 (4.0) | 2539 (14.2) | < 0.001 |
| Second quartile | 13,931 (78.9) | 1041 (5.9) | 2682 (15.2) | 14,469 (82.0) | 680 (3.9) | 2505 (14.2) | ||
| Third quartile | 13,600 (81.6) | 842 (5.1) | 2234 (13.4) | 13,951 (83.7) | 621 (3.7) | 2104 (12.6) | ||
| Fourth quartile | 13,571 (82.6) | 774 (4.7) | 2080 (12.7) | 13,781 (83.9) | 626 (3.8) | 2018 (12.3) | ||
Data are presented as n (%) unless otherwise indicated.
ALND Axillary lymph node dissection, SLNB Sentinel lymph node biopsy
Factors Associated with Primary Site Surgery
On univariate analysis, patient age, sex, race/ethnicity, year of diagnosis, income quartile, Charlson–Deyo comorbidity score, histology, receptor status, grade, metastatic site and burden, geographic region, facility type and volume, and distance traveled were all factors significantly associated with primary site surgery. On multivariate analysis, not only did the patient demographic and clinical variables remain significant, but also the facility-related variables, such as geographic region, facility type, facility volume, and distance traveled also remained significantly associated with primary site surgery. Patients treated at facilities in the Northeast were least likely to receive primary site surgery (compared with odds ratio [OR] 1.27 [95% confidence interval (CI) 1.18–1.36] in the Midwest; OR 1.55 [95% CI 1.46–1.66] in the South; and OR 1.62 [95% CI 1.50–1.74] in the West). Academic/research centers had lower odds of primary site surgery than did community centers (OR 0.79 [95% CI 0.71–0.87]), and the odds of primary site surgery decreased significantly with increasing patient volume (Table 3). Greater distance traveled to the treatment facility was also associated with significantly increased odds of primary site surgery (> 20miles traveled, OR 1.23 [95% CI 1.16–1.31]) (Table 5).
Table 5.
Factors associated with primary site surgery
| Factor | Univariate | Multivariate | ||
|---|---|---|---|---|
| OR (95% CI) | p-Value | OR (95% CI) | p-Value | |
| Age category, years | ||||
| < 45 | Ref | Ref | Ref | Ref |
| 46–55 | 0.74 | 0.70–0.79 | 0.72 | 0.67–0.77 |
| 56–65 | 0.54 | 0.51–0.58 | 0.56 | 0.53–0.60 |
| 66–75 | 0.44 | 0.42–0.47 | 0.46 | 0.43–0.49 |
| ≥ 76 | 0.33 | 0.30–0.35 | 0.32 | 0.29–0.34 |
| Sex | ||||
| Female | Ref | Ref | Ref | Ref |
| Male | 1.44 | 1.25–1.67 | 1.73 | 1.46–2.04 |
| Race and ethnicity | ||||
| White | Ref | Ref | Ref | Ref |
| Black | 0.87 | 0.83–0.92 | 0.79 | 0.74–0.84 |
| Hispanic | 1.22 | 1.13–1.31 | 1.11 | 1.02–1.22 |
| Asian | 1.15 | 1.05–1.27 | 0.96 | 0.86–1.08 |
| Other | 0.97 | 0.83–1.12 | 0.91 | 0.77–1.08 |
| Unknown | 0.78 | 0.63–0.97 | 0.75 | 0.59–0.96 |
| Income quartiles | ||||
| First | Ref | Ref | Ref | Ref |
| Second | 1.12 | 1.03–1.21 | 1.07 | 0.98–1.17 |
| Third | 1.09 | 1.02–1.18 | 1.08 | 0.99–1.17 |
| Fourth | 1.10 | 1.02–1.18 | 1.12 | 1.03–1.21 |
| Year of diagnosis | ||||
| 2012–2014 | Ref | Ref | Ref | Ref |
| 2015–2017 | 0.66 | 0.63–0.69 | 0.68 | 0.63–0.73 |
| 2018–2021 | 0.47 | 0.45–0.49 | 0.49 | 0.45–0.54 |
| Charlson–Deyo comorbidity score | ||||
| 0 | Ref | Ref | Ref | Ref |
| 1 | 0.91 | 0.86–0.97 | 0.98 | 0.91–1.05 |
| 2 | 0.67 | 0.60–0.75 | 0.85 | 0.74–0.96 |
| 3 | 0.49 | 0.73–0.57 | 0.70 | 0.59–0.83 |
| Histology | ||||
| IDC | Ref | Ref | Ref | Ref |
| ILC | 0.61 | 0.57–0.65 | 0.85 | 0.78–0.92 |
| Inflammatory | 1.14 | 1.00–1.31 | 1.02 | 0.87–1.20 |
| Metaplastic | 3.51 | 2.85–4.33 | 3.36 | 2.61–4.31 |
| Receptor subtype | ||||
| Triple negative | Ref | Ref | Ref | Ref |
| HR-/HER2+ | 0.96 | 0.88–1.04 | 1.07 | 0.97–1.18 |
| HR+/HER2- | 0.64 | 0.60–0.67 | 0.84 | 0.78–0.90 |
| HR+/HER2+ | 0.86 | 0.80–0.93 | 0.98 | 0.90–1.07 |
| Unknown | 0.68 | 0.63–0.73 | 0.76 | 0.69–0.82 |
| Grade | ||||
| Well differentiated | Ref | Ref | Ref | Ref |
| Moderately differentiated | 1.28 | 1.17–1.41 | 1.33 | 1.20–1.47 |
| Poorly differentiated | 2.09 | 1.91–2.29 | 2.04 | 1.83–2.26 |
| Unknown | 0.59 | 0.53–0.65 | 0.68 | 0.60–0.77 |
| Clinical T stage | ||||
| T1 | Ref | Ref | Ref | Ref |
| T2 | 1.09 | 1.02–1.16 | 1.05 | 0.97–1.13 |
| T3 | 1.08 | 1.01–1.16 | 1.02 | 0.93–1.11 |
| T4 | 0.96 | 0.90–1.03 | 0.99 | 0.92–1.08 |
| Unknown | 0.48 | 0.44–0.53 | 0.65 | 0.58–0.73 |
| Clinical N stage | ||||
| N0 | Ref | Ref | Ref | Ref |
| N1 | 0.82 | 0.78–0.86 | 0.75 | 0.71–0.80 |
| N2 | 1.01 | 0.95–1.09 | 0.93 | 0.86–1.01 |
| N3 | 1.05 | 0.99–1.12 | 0.88 | 0.81–0.95 |
| Unknown | 0.58 | 0.53–0.63 | 0.82 | 0.74–0.91 |
| Metastatic site | ||||
| Bone | 0.53 | 0.51–0.55 | 0.74 | 0.83–1.03 |
| Brain | 0.38 | 0.34–0.41 | 0.63 | 0.45–0.88 |
| Liver | 0.64 | 0.61–0.67 | 0.78 | 0.56–1.08 |
| Lung | 0.64 | 0.61–0.67 | 0.93 | 0.67–1.30 |
| Distant lymph nodes | 0.92 | 0.88–0.96 | 1.38 | 0.99–1.92 |
| Other sites | 0.54 | 0.50–0.58 | 0.82 | 0.32–0.62 |
| Metastatic burden | ||||
| 1 site | Ref | Ref | Ref | Ref |
| 2 sites | 0.43 | 0.41–0.45 | 0.44 | 0.32–0.62 |
| 3 sites | 0.27 | 0.25–0.29 | 0.30 | 0.15–0.58 |
| ≥4 sites | 0.17 | 0.15–0.19 | 0.24 | 0.08–0.70 |
| Facility type | ||||
| Community | Ref | Ref | Ref | Ref |
| Comprehensive community | 1.02 | 0.95–1.10 | 1.07 | 0.98–1.17 |
| Academic/research | 0.72 | 0.67–0.78 | 0.79 | 0.71–0.87 |
| Integrated | 0.90 | 0.83–0.97 | 0.93 | 0.85–1.03 |
| Facility region | ||||
| Northeast | Ref | Ref | Ref | Ref |
| South | 1.58 | 1.50–1.67 | 1.55 | 1.46–1.66 |
| Midwest | 1.31 | 1.23–1.39 | 1.27 | 1.18–1.36 |
| West | 1.72 | 1.61–1.83 | 1.62 | 1.50–1.74 |
| Distance traveled, miles | ||||
| < 5 | Ref | Ref | Ref | Ref |
| 5–10 | 1.12 | 1.06–1.19 | 1.08 | 1.02–1.15 |
| 10–20 | 1.22 | 1.15–1.29 | 1.14 | 1.07–1.21 |
| > 20 | 1.24 | 1.18–1.31 | 1.23 | 1.16–1.31 |
| Unknown | 1.16 | 1.09–1.24 | 5.80 | 0.35–96.00 |
| Facility volume | ||||
| First quartile | Ref | Ref | Ref | Ref |
| Second quartile | 0.94 | 0.90–0.99 | 0.89 | 0.84–0.95 |
| Third quartile | 0.80 | 0.76–0.84 | 0.77 | 0.72–0.83 |
| Fourth quartile | 0.74 | 0.70–0.78 | 0.70 | 0.65–0.76 |
CI Confidence interval, HER2 Human epidermal growth factor receptor 2, HR Hormone receptor, IDC Invasive ductal carcinoma, ILC Invasive lobular carcinoma, OR Odds ratio, ref reference.
Discussion
In this large-scale, nationally representative retrospective cohort study, we found significant variation in locoregional management of patients with de novo stage IV breast cancer. The paradigm for management of stage IV breast cancer has predominantly been treatment with systemic therapy alone, with surgery reserved only for palliation of symptoms under the guiding principle that primary-site locoregional surgery would not alter the course of metastatic disease. In recent years, several retrospective studies and five randomized controlled trials have re-examined this paradigm, with four of them (the Tata Memorial trial in India, the Austrian POSYTIVE trial, E2108, JCOG 1017 PRIM-BC) showing no improved outcomes with primary-site surgery.11–14 Despite conflicting evidence on the role of locoregional therapy in patients with an intact breast primary, studies show an 18–44% rate of surgery in patients with de novo stage IV breast cancer.3,8,15
In this study, we found that 25.7% of patients with de novo stage IV breast cancer underwent surgery at centers captured within the NCDB between 2012 and 2021; 19.8% of patients underwent breast surgery, 17.3% of patients underwent axillary surgery, and 4.4% of patients underwent surgery at a metastatic site. As noted in previous studies, patients with stage IV disease who underwent surgical resection were younger, more likely to have HER2+ or triple-negative disease, and more likely to have a lower burden of metastatic disease, but this study showed significant variation in the surgical management of de novo stage IV breast cancer by facility-level factors such as geographic region and facility type. With regard to geography, the Northeast had the highest rates of systemic therapy alone and the lowest rates of primary site surgery, whereas the West had lower rates of systemic therapy and the highest rates of primary site surgery. With regard to variation by facility type, comprehensive community cancer centers had the highest rates of primary site surgery, whereas academic centers had the lowest rates. There may be interaction between the variation in surgical patterns based on geography and facility type, potentially driven by varying concentrations of academic versus community centers in each geographic region; however, both factors remained significantly associated with surgery in multivariate analyses. Most importantly, patients receiving treatment at centers with lower volumes of stage IV breast cancer were far more likely to undergo primary tumor resection despite having distant metastatic disease, and—among those patients with stage IV disease receiving surgical intervention—centers with lower volumes are more likely to recommend mastectomy and an ALND.
Variation has previously been noted in the surgical management of breast cancer, with differences in management noted based on a range of factors, including patient demographic and clinical factors, and surgeon, institutional, and geographic factors.16–19 Importantly, institution-level factors, such as availability of plastic surgery, have been shown to account for a significant amount of surgical variation. Specific to management of stage IV breast cancer, variation in surgical management was previously considered using the NCDB from 2003 to 2012, where 44% of patients underwent surgery, with a very slight decrease in surgical intervention noted during the study timeframe.3 More recently, in a study considering the years 2013–2016, 24% of patients underwent surgical resection.9 Our study does show a further decrease in the rates of primary site surgery in stage IV breast cancer, but this decrease is not seen uniformly across institutions.
In an era of novel systemic therapies and improved durable response to systemic therapy, the role of primary-site surgery for stage IV breast cancer is still evolving. In patients with oligometastatic disease or a very limited burden of metastatic disease, surgical resection of the primary tumor is sometimes recommended by the multidisciplinary team as a part of curative-intent treatment. Our data show that more patients who underwent surgery also received systemic therapy (93.2%), whereas only 85.0% of the ‘no surgery’ cohort received systemic therapy (p < 0.001), probably because patients undergoing surgery were doing so with a plan for curative-intent treatment. If surgical intervention is to be pursued, it must be done in the context of complete evaluation of the metastatic burden and with multidisciplinary consensus. The finding of higher rates of surgery at centers with lower stage IV patient volumes raises concern for potential mismanagement or may indicate patient-level factors driving these decisions. Institutions seeing lower volumes of patients with stage IV disease may perform surgery in cases where institutions with greater experience with stage IV breast cancer may deem surgery futile. Recognizing these differences in practice patterns between institutions raises awareness of the need for further standardization of care.
The findings of this study must be considered within the context of its limitations. Although the NCDB is subject to rigorous quality assurance measures, the data are collected retrospectively and are susceptible to potential confounding variables. Given the retrospective nature of the study, our data can only demonstrate an association between the geographic or institutional factors and surgical management rather than infer causality. Furthermore, errors in the documentation and data procurement are possible. The data do not allow for the granularity required to assess specific reasons for surgical intervention in any given case that could impact the surgical decision-making. The reason for surgery (palliative vs. curative intent) cannot be determined. The data also cannot be used to suggest a reason for the variation in surgical patterns. For instance, we cannot ascertain whether surgery was more frequently offered by providers or requested by patients at lower-volume centers. The NCDB only includes data from participating Commission on Cancer–accredited institutions. As such, this study only includes hospitals that participate in reporting to the NCDB and does not fully capture the extent of primary site surgery in stage IV breast cancer. True rates of primary site surgery and variation in care may be higher at non-reporting centers.
Conclusion
This retrospective cohort analysis of patients with de novo stage IV breast cancer shows that significant variation exists in locoregional surgical management of de novo stage IV breast cancer. Surgical intervention has decreased over time, likely because evidence has demonstrated limited survival advantage and improving systemic therapy. However, this study reveals that institutional and geographic factors are significantly associated with variations in the surgical management for these patients. Initiatives to standardize breast cancer care for patients with de novo stage IV disease will need to address specific local and regional factors for variation.
Funding
No funding was obtained for this project.
Disclosure
No conflicts of interest were reported by any of the study’s authors.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.O’Shaughnessy J. Extending survival with chemotherapy in metastatic breast cancer. Oncologist. 2005;10(S3):20–9. [DOI] [PubMed] [Google Scholar]
- 2.Surveillance. Epidemiology aERSPwscgSSDISRP-U, Nov 2004 Sub (1973–2002), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2005, based on the November 2004 submission. Surveillance, Epidemiology, and End Results Program
- 3.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):537–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gera R, Chehade HELH, Wazir U, Tayeh S, Kasem A, Mokbel K. Locoregional therapy of the primary tumour in de novo stage IV breast cancer in 216 066 patients: a meta-analysis. Sci Rep. 2020;10(1):2952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gnerlich J, Jeffe DB, Deshpande AD, Beers C, Zander C, Margenthaler JA. Surgical removal of the primary tumor increases overall survival in patients with metastatic breast cancer: analysis of the 1988–2003 SEER data. Ann Surg Oncol. 2007;14(8):2187–94. [DOI] [PubMed] [Google Scholar]
- 6.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):1893–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Vohra NA, Brinkley J, Kachare S, Muzaffar M. Primary tumor resection in metastatic breast cancer: a propensity-matched analysis, 1988–2011 SEER data base. Breast J. 2018;24(4):549–54. [DOI] [PubMed] [Google Scholar]
- 8.Arciero C, Liu Y, Gillespie T, Subhedar P. Surgery and survival in patients with stage IV breast cancer. Breast J. 2019;25(4):644–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Duchesneau ED, Jackson BE, Webster-Clark M, et al. The timing, the treatment, the question: comparison of epidemiologic approaches to minimize immortal time bias in real-world data using a surgical oncology example. Cancer Epidemiol Biomarkers Prev. 2022;31(11):2079–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Soran A, Ozmen V, Ozbas S, et al. Primary surgery with systemic therapy in patients with de novo stage IV breast cancer: 10-year follow-up; protocol MF07-01 randomized clinical trial. J Am Coll Surg. 2021;233(6):742-751.e745. [DOI] [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. Lancet Oncol. 2015;16(13):1380–8. [DOI] [PubMed] [Google Scholar]
- 12.Fitzal F, Bjelic-Radisic V, Knauer M, et al. Impact of breast surgery in primary metastasized breast cancer: outcomes of the prospective randomized phase III ABCSG-28 POSYTIVE trial. Ann Surg. 2019;269(6):1163–9. [DOI] [PubMed] [Google Scholar]
- 13.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). J Clin Oncol. 2020;38(18_suppl):LBA2–LBA2. [Google Scholar]
- 14.Shien T, Hara F, Aogi K, et al. Primary tumour resection plus systemic therapy versus systemic therapy alone in metastatic breast cancer (JCOG1017, PRIM-BC): a randomised clinical trial. Br J Cancer. 2025;133(5):625–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Teshome M. Role of operative management in stage IV breast cancer. Surg Clin North Am. 2018;98(4):859–68. [DOI] [PubMed] [Google Scholar]
- 16.Katz SJ, Lantz PM, Janz NK, et al. Patient involvement in surgery treatment decisions for breast cancer. J Clin Oncol. 2005;23(24):5526–33. [DOI] [PubMed] [Google Scholar]
- 17.Hawley ST, Hofer TP, Janz NK, et al. Correlates of between-surgeon variation in breast cancer treatments. Med Care. 2006;44(7):609–16. [DOI] [PubMed] [Google Scholar]
- 18.Chiu AS, Thomas P, Killelea BK, Horowitz N, Chagpar AB, Lannin DR. Regional variation in breast cancer surgery: results from the national cancer database (NCDB). Am J Surg. 2017;214(5):907–13. [DOI] [PubMed] [Google Scholar]
- 19.Dankwa-Mullan I, George J, Roebuck MC, et al. Variations in breast cancer surgical treatment and timing: determinants and disparities. Breast Cancer Res Treat. 2021;188(1):259–72. [DOI] [PMC free article] [PubMed] [Google Scholar]