Locoregional recurrence of triple-negative breast cancer after breast-conserving surgery and radiation


  • Presented in abstract form at the 89th Annual Meeting of the American Radium Society, Amsterdam, the Netherlands, May 5-9, 2007.

  • We thank Cindy Rosser for her collection and management of the data for the study population, and Lillian Henry for her assistance in the preparation of the article.



The results of radiation on the local control of triple receptor-negative breast cancer (negative estrogen [ER], progesterone [PR], and HER-2/neu receptors) was studied.


Conservative surgery and radiation were used in 753 patients with T1-T2 breast cancer. Three groups were defined by receptor status: Group 1: ER or PR (+); Group 2: ER and PR (−) but HER-2 (+); and Group 3: triple-negative (TN). Factors analyzed were age, menopausal status, race, stage, tumor size, lymph node status, presentation, grade, extensive in situ disease, margins, and systemic therapy. The primary endpoint was 5-year locoregional recurrence (LRR) isolated or total with distant metastases.


ER- and PR-negative patients were statistically significantly more likely to be black, have T2 disease, have tumors detectable on both mammography and physical examination, have grade 3 tumors, and receive chemotherapy. There were no significant differences noted with regard to ER− and PR− patients by HER-2 status. There was a significant difference noted in rates of first distant metastases (3%, 12%, and 7% for Groups 1, 2, and 3, respectively; P = .009). However, the isolated 5-year LRR was not significantly different (2.3%, 4.6%, and 3.2%, respectively; P = .36) between the 3 groups.


Patients with TN breast cancer do not appear to be at a significantly increased risk for isolated LRR at 5 years and therefore remain appropriate candidates for breast conservation. Cancer 2009. © 2009 American Cancer Society.

Radiotherapy after conservative surgery is a standard part of breast preservation therapy for invasive breast cancer.1, 2 The addition of radiotherapy has been found to significantly reduce the risk of local recurrence in prospective randomized trials with or without adjuvant systemic therapy.3, 4 There are relatively few contraindications to breast conservation in general or radiotherapy in particular - a history of prior therapeutic irradiation to the breast, inability to obtain negative resection margins, pregnancy, or multicentric breast cancer.1, 5 In the past, identification of factors associated with an increased risk for local recurrence after radiation has usually been limited to clinical, pathologic, or treatment-related factors. For example, there is evidence for increased rates of local recurrence in patients with young age or positive resection margins.6

Breast tumors of different gene expression profiles have been identified that are associated with different clinical behaviors. In particular, a basal subtype is associated with a more aggressive clinical behavior and worse prognosis.7-10 Without a validated clinical-use assay to identify these tumors prospectively, these different classes of tumors can be roughly distinguished more easily in clinical practice by expression of currently standard estrogen receptor (ER), progesterone receptor (PR), and HER-2 receptor testing. Luminal A and B subtypes are associated with positive ER. ER-negative tumors are divided into those positive for HER-2, or those of “basal-like” expression associated with low to absent ER, PR, and HER-2 receptor expression.11 There is an incomplete correlation between basal-like tumors and these receptor expression profiles: from 15%-45% of basal-like tumors express at least 1 of these markers, and only 85% of triple receptor-negative tumors are basal-like by expression arrays.10 These previous studies have associated these subtypes with different behaviors in overall recurrence or survival. However, to our knowledge, there have been limited series reported to date regarding the impact these gene expression profiles have on local control specifically.12-14

The purpose of the current study was to report the locoregional recurrence (LRR) rates after breast-conserving surgery and radiation for these triple-negative breast cancers to determine whether patients with this subtype remain appropriate candidates for breast conservation.


The study population was comprised of 753 women with American Joint Committee on Cancer stages I-III15 disease who were treated with breast–conserving surgery and radiotherapy between October 1990 and July 2006. All patients had known ER, PR, and HER-2/neu receptor status. Positive HER-2/neu was either immunohistochemical 3 + staining or amplication by fluorescence in situ hybridization. There was no central review or retrospective retesting of receptor status. Exclusion criteria for this study included male breast cancer, T3 to T4 disease, stage IV disease, mastectomy, or patients treated without radiotherapy. Of a total of 2315 patients in the database from October 1990 to July 2006, there were 1862 with T1-2 breast cancer, breast-conserving surgery, and axillary lymph node dissection. Of these, 1109 were excluded for having an unknown ER, PR, or HER-2/neu status. The remaining 753 patients comprised the study population. Patient information is maintained and updated in a confidential database by a single data manager. The protocols for collection, storage, and data retrieval are in compliance with the Hospital Institutional Review Board and Health Insurance Portability and Privacy Act regulations.

All patients underwent breast-conserving surgery and postoperative whole breast irradiation. The median dose of to the whole breast volume was 46 gray (Gy), and the primary tumor bed was boosted in 99% of patients. The total tumor bed dose with the boost was generally determined by the extent of surgery and final resection margin status. The total dose combining the whole breast and boost doses was a median of 60 Gy for all patients (range, 22-66 Gy). Patients with a negative resection margin were treated to a median 60 Gy (range, 2200-6600 Gy) and those with close or positive margins were treated to a median of 64 Gy (range, 5000-6600 Gy). There were only 8 patients treated to a dose <60 Gy.

Analysis for significant differences was performed using global and pairwise comparisons. Factors analyzed included clinical (age, menopausal status, race, tumor stage, tumor size, method of detection, and use of prior tamoxifen or estrogen replacement therapy before diagnosis); pathologic (lymph node stage, Modified Scarff-Bloom-Richardson grade, necrosis, lymphovascular invasion, extensive intraductal component [EIC], and final resection margin status); and treatment (use of systemic therapy). Differences in patient characteristics between subgroups were compared using the chi-square test. The endpoints of the study were 5-year LRR as an isolated event or with distant metastases. Only first patterns of recurrence were included, not LRR after distant metastases. Other endpoints were disease-free survival and overall survival. Univariate analyses of 5-year outcomes were conducted using Kaplan-Meier methodology and the log-rank test.


The initial analysis of patient characteristics and outcomes demonstrated no significant differences between patients who were ER positive or PR positive by HER-2 status, so they were combined into a single group (Group 1). Patients who were ER negative and PR negative and HER-2/neu positive were in Group 2, and patients who were negative for all 3 receptors were in Group 3. Patient characteristics for each group and the results of statistical comparisons are shown in Table 1.

Table 1. Patient Characteristics by Results of ER, PR and HER-2/neu Receptor Testing*
 ER+ or PR+ER– PR– HER-2+ER– PR– HER-2–P
  • ER indicates estrogen receptor; +, positive; PR, progesterone receptor; −, negative; EIC, extensive intraductal component; LVI, lymphovascular invasion.

  • *

    Numbers are shown as percentages unless otherwise stated.

No. of patients6005598 
Follow-up, mo484144.23
Median age, y595454.2
Race   .03
Tumor size   .0015
Lymph node status   .47
Detection   <.0001
 Physical examination13187 
Grade   <.0001
EIC   .14
Necrosis   .01
LVI   .66
Resection margins   .96
Systemic therapy   <.0001

Statistically significant differences between ER and PR patients in Groups 2 and 3 versus Group 1 were they were more likely to have T2 disease, disease that was detectable on both mammography and physical examination, grade 3 disease, presence of necrosis, black race, and the use of chemotherapy (grading determined according to the Modified Scarff-Bloom-Richardson grading system). Trastuzumab was administered to 14 patients (2%) in Group 1 and to 12 patients (22%) in Group 2 who were HER-2/neu positive in later years of the study period. There were no significant differences between Group 2 versus Group 3 patients. There were no significant differences between patients by median age, menopausal status, use of tamoxifen or estrogen replacement therapy before diagnosis, lymph node status, EIC, lymphovascular invasion, or resection margin status.

Outcomes are shown in Table 2. The isolated 5-year LRR was not significantly different between Groups 1 (2.3%), 2 (4.6%), or 3 (3.2%) (P = .36). Figure 1 shows the actuarial total LRR rate between the 3 groups. There was no overall difference between the 3 groups (P = .13). There was a higher observed rate of distant metastases in Group 2 (11.9%) that translated into a lower recurrence-free survival rate (84%). However, the total LRR (5.3%) for triple–negative patients in Group 3 was of borderline significance compared with Group 1 patients (2.6%) (P = .05) because of the greater numbers of patients with simultaneous local and distant metastases.

Figure 1.

Total locoregional recurrence as a first site of disease recurrence with or without simultaneous distant metastases. ER indicates estrogen receptor; +, positive; PR, progesterone receptor; −, negative.

Table 2. Five-Year Outcomes*
 ER+ or PR+ER– PR– HER-2+ER– PR– HER-2–P
  • ER indicates estrogen receptor; +, positive; PR, progesterone receptor; LRR, locoregional recurrence.

  • *

    Numbers are shown as percentages.

LRR isolated2.
Distant metastases311.96.5.009
LRR total (+distant)
Recurrence-free survival948490.005
Overall survival948890.15


Patients with triple receptor-negative breast cancers in the current study were not found to be at an increased risk for isolated LRR at 5 years. This finding is consistent with our finding in the current study of no significant differences in clinical or pathologic factors with this subtype that have been found to be predictive of a higher rate of local recurrence in the past, factors such as young age, positive resection margins, or EIC.6 Of what to our knowledge are the 3 other largest studies providing information on LRR by subtype, 2 of the 3 demonstrated an increased risk with basal-like expression.12–14

Haffty et al12 reported a series of 117 triple receptor-negative patients treated with breast-conserving surgery and radiation and compared them with 365 nontriple-negative controls. Our study is concordant with their findings in that patients with triple-negative breast cancer were more likely to be black, have T2 disease, and receive chemotherapy. There was no difference noted in breast recurrence-free survival at 5 years.

Nguyen et al13 reported a series of 89 of 793 patients with basal-like receptor expression who were treated with breast-conserving surgery and radiation. Basal subtype was associated with younger patient age, high grade, and larger size. The 5-year local recurrence rate was 7.1% in patients with a basal-like subtype. On multivariate analysis, basal-like subtype was associated with increased local recurrence compared with luminal A type patients. There was also a statistically significant increase in distant disease recurrence in basal-like tumors on multivariate analysis.

Kyndi et al14 reported a series of 152 patients with triple-negative breast cancer treated on prospective randomized trials of mastectomy with or without radiotherapy. These randomized trials had shown improvements with radiation in LRR and overall survival in general. These patients with basal-like subtype represented only 15% of all patients tested in their analysis. There was a statistically significant increase in LRR and overall mortality but not distant metastases found on multivariate analysis. When the analysis was restricted to those patients randomized to no radiation, patients with triple-negative status were found to have a higher LRR rate, distant metastasis, and overall mortality. In contrast, among those patients randomized to radiation, triple-negative status was only found to be associated with increased LRR. In an analysis testing for interaction between postmastectomy radiation and receptor subtypes, there was no improvement in overall survival noted with radiation in the triple-negative subgroup.

In the current series, patients with ER/PR-negative and HER–2/neu-positive tumors had the highest risk of isolated distant metastases and lowest recurrence-free survival. This is most likely reflective of the study period before the routine availability of trastuzumab. Trastuzumab was given to only 14 patients (2%) in Group 1 and 12 patients (22%) in Group 2 in later years of the study period. Whereas the HER-2-positive patients comprising Group 2 had a higher risk of distant metastases and lower recurrence-free survival, this finding was not associated with a higher rate of isolated or total LRR compared with other patients. The addition of trastuzumab is now part of the routine management of these patients but would require study in a very large population to determine whether it can significantly reduce further the already low rates of LRR observed in the current study.

In conclusion, our current policy, based on the outcomes data provided in this study, is to routinely continue to offer breast-conserving surgery and radiation for triple receptor-negative patients.

Conflict of Interest Disclosures

The authors made no disclosures.