The role of surgical primary tumor extirpation in de novo stage IV breast cancer in the era of targeted treatment


Surgical Breast Oncology

Center for Urban Population Health

Aurora Research Institute


OBJECTIVE: Previous reports evaluating primary tumor extirpation (hereafter, surgery) in patients presenting with de novo stage IV breast cancer describe mixed results regarding overall survival (OS). In this modern era of treatment, the impact of surgery was assessed, both controlling and adjusting for potential confounders, including comorbidities, tumor burden, vitality impact of distant metastatic site, hormonal therapy of ER/PR+ disease, and targeted therapy of HER-2+ disease.

METHODS: Women presenting with de novo stage IV breast cancer during 2000–2015 were retrospectively studied using a single institution’s cancer registry data. Patients with severe competing comorbidities (heart failure, chronic kidney disease) were excluded, as well as those missing data for patient, tumor, or treatment variables used in matching or analysis. As primary tumor extirpation was of principal interest, patients who underwent surgery as a first course of treatment were 1:1 matched with those treated without surgery by patient age (within ± 20 years), number of cardiovascular risk factors (smoking, hypertension, dyslipidemia, diabetes mellitus, obesity; within ± 1 factor), coronary artery disease, HER-2/neu and ER/PR, tumor grade, number of metastatic sites (tumor burden within ± 1 site), vitality impact of metastatic sites (CNS, visceral, bone), and first-course systemic and site-specific radiation (breast/chest, metastatic site) therapies received. The adjusted effects of surgery and other patient, tumor, and treatment characteristics on OS were quantified using hazard ratios (HR) derived from marginal Cox proportional hazards models, all containing surgery. Through estimation of the survivor function, OS rates were computed per study group.

RESULTS: Of 609 total patients identified, 280 entered the matching algorithm. Women who underwent surgery (n = 58) vs those who did not undergo surgery (n = 58) within the matched-pairs population did not differ by age (mean, 62 yr) or other matched characteristics, but did significantly differ by length of follow-up (3.03 vs 1.97 yr, respectively). Single-variable adjustment led to detection of a significant surgery effect (P < 0.04) in 4 of 10 models of OS (table). Across models of nonsignificant surgery effects (P = 0.06-0.08), HRs were within the range of values produced by models revealing significance. All models suggested a 40% reduction in risk for patients receiving surgery, and 9 of the 10 models suggested 3-yr OS rates of approximately 60% for patients undergoing surgery vs. 45% for patients treated without surgery. Age, number of risk factors, ER/PR, and vitality impact of metastatic sites impacted OS.

Hazard Ratios and Adjusted 3-Year Overall Survival Rates Derived From Cox Proportional Hazards Models of Overall Survival in Women Who Presented With Stage IV Breast Cancer During 2000– 2015 and Were Matched by Primary Tumor Extirpation (Surgery, N = 116)

Model No. Model Variable 1 HR (95% CI) 3-year OS rate (95% CI) Model Variable 2 HR (95% CI)

1 Surgery Patient agea 1.50 (1.24–1.84)* Performed 0.65 (0.44–1.03) 0.60 (0.48–0.75) Not performed Reference 0.46 (0.32–0.67)

2 Surgery Number of risk factors Performed 0.61 (0.40–0.95)* 0.59 (0.46–0.74) b 1.33 (1.06–1.66)* Not performed Reference 0.43 (0.31–0.61)

3 Surgery Tumor sizec Performed 0.65 (0.41–1.02) 0.58 (0.460.74) 0.98 (0.91–1.06) Not performed Reference 0.43 (0.31–0.61)

4 Surgery HER2neu expression Performed 0.65 (0.41–1.02) 0.58 (0.45–0.74) Positive 0.89 (0.54–1.47) Not performed Reference 0.43 (0.30–0.62) Negative Reference

5 Surgery ER/PR expressiond Performed 0.59 (0.36–0.97)* 0.17 (0.05–0.58) Positive 0.23 (0.12–0.46)* Not performed Reference 0.05 (0.01–0.53) Negative Reference

6 Surgery Grade Performed 0.61 (0.39–0.97)* 0.63 (0.52–0.79) I or II Reference Not performed Reference 0.49 (0.34–0.70) III or IV 1.51 (0.89–2.56)

7 Surgery Tumor burdene Performed 0.66 (0.42–1.05) 0.58 (0.46–0.74) 1.30 (0.90–1.86) Not performed Reference 0.43 (0.31–0.61)

8 Surgery Metastatic site impact Performed 0.59 (0.38–0.91)* 0.58 (0.47–0.74) Visceral 1.93 (1.15–3.25)* Not performed Reference 0.42 (0.29–0.61) Bone Reference

9 Surgery Chemotherapy Performed 0.68 (0.43–1.08) 0.52 (0.37–0.73) Performed 0.70 (0.40–1.23) Not performed Reference 0.38 (0.25–0.59) Not performed Reference

10 Surgery Radiation therapy Performed 0.65 (0.41–1.03) 0.55 (0.42–0.72) Performed 0.66 (0.35–1.25) Not Performed Reference 0.40 (0.26–0.60) Not performed Reference

HR indicates hazard ratio; OS, overall survival; CI, confidence interval; HER2neu, human epidermal growth factor receptor 2; ER, estrogen receptor; PR, progesterone receptor; CNS, central nervous system.

*Hazard ratio significantly differs from 1.

a Number of times the hazard increases per 10-year increase in age.

b Number of times the hazard increases per 1-factor increase in cardiovascular risk.

c Number of times the hazard increases per 10-mm increase in tumor size.

d Classified as positive when either ER or PR or both are overexpressed and negative when neither ER or PR are overexpressed.

e Number of times the hazard increases per 1-metastatic site increase in tumor burden.

CONCLUSION: Even after accounting for hormonal therapy, targeted therapy, and radiation to local and distant metastatic sites, surgical extirpation of the primary tumor remains associated with an OS improvement in patients with de novo stage IV breast cancer.

Document Type


Link to Full Text