Presented in part at the 52nd Annual Meeting of the American Society for Therapeutic Radiology and Oncology; October 31 to November 4, 2010; San Diego, California.
We thank Dr. Simon Powell for his critical insight. We acknowledge the essential work of Jane Howard in managing the Memorial Sloan-Kettering Cancer Center Breast Cancer Disease Management Team database. We also thank Lawrence A. Herman for editorial review and Tracy Iannone for assistance with medical records.
Patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer have a higher risk of locoregional recurrence (LRR), even in the setting of early stage, lymph node-negative disease. In this sequential, retrospective study, the authors evaluated whether adjuvant trastuzumab was associated with reduced LRR in women with lymph node-negative, HER2-positive disease who received breast-conservation therapy (BCT).
By using an institutional database, 197 women were identified who had lymph node-negative, HER2-positive breast cancer measuring ≤5 cm diagnosed between 2002 and 2008 and who received BCT, including whole-breast irradiation. Two cohorts were compared: 70 women who did not receive trastuzumab (the no-trastuzumab cohort) and 102 women who did receive trastuzumab (the trastuzumab cohort). Kaplan-Meier methods were used to estimate LRR-free survival.
The 2 cohorts were similar in age, tumor size, histology, and hormone receptor status. Chemotherapy was received by 73% of the no-trastuzumab cohort and by 100% of the trastuzumab cohort. In both groups, 99% of patients completed radiotherapy with a median dose of 60 Gray. The median recurrence-free follow-up was 86 months for the no-trastuzumab cohort and 47 months for the trastuzumab cohort. The 3-year LRR-free survival rate was 90% (95% confidence interval, 83%-97%) for the no-trastuzumab cohort and 99% (95% confidence interval, 97%-100%) for the trastuzumab cohort. In the no-trastuzumab cohort, LRR occurred in 7 patients (median time to LRR, 14 months). In the trastuzumab cohort, there was 1 LRR at 14 months.
Human epidermal growth factor receptor 2 (HER2) overexpression or amplification is a negative prognostic factor for breast cancer outcomes; however, trastuzumab therapy has significantly improved outcomes for patients with HER2-positive disease. In the adjuvant setting, large randomized trials have demonstrated the efficacy of trastuzumab, a humanized monoclonal antibody that binds HER2, in reducing the risk of recurrence and improving overall survival in patients with high-risk, HER2-positive disease.1-7 However, patients with lymph node-negative (presumably low-risk) disease often were excluded from those studies unless their primary tumors were relatively large. They represented only 6% of the 2 North American trials and approximately 30% of the Breast Cancer International Research Group (BCIRG) 006 and Herceptin Adjuvant (HERA) trials.1- -6 Therefore, consensus guidelines for the treatment of small, lymph node-negative breast cancers have been based on extrapolations rather than primary data and have not clearly recommended trastuzumab.8
Even when conventional factors, such as small tumor size and negative lymph nodes, suggest a good prognosis, recent retrospective studies indicate that positive HER2 status remains a negative prognostic factor.9-12 This holds true for locoregional recurrence (LRR), distant recurrence, and disease-free survival (DFS). Albert et al demonstrated an 18% LRR rate in patients who had lymph node-negative, HER2-positive breast cancer that measured <1 cm compared with a 4% LRR rate in HER2-negative patients (without adjuvant trastuzumab or other chemotherapy). In the patients who received breast-conservation therapy, positive HER2 status conferred an LRR hazard ratio of 3.8.12, 13
For this subgroup, which largely has been excluded from randomized trials, it remains unclear whether the high recurrence rates that have been observed in patients with HER2-positive disease can be reduced similarly with adjuvant trastuzumab. Subset analyses from the BCIRG and HERA trials suggest that trastuzumab has a similar magnitude of DFS benefit in lymph node-negative cancers compared with the entire cohort.3, 14 However, to our knowledge, no studies have specifically evaluated local recurrence, which is critical for many lymph node-negative patients who elect to receive breast-conservation therapy. At our institution, most patients with early HER2-positive breast cancer have been offered trastuzumab since 2005, but not before 2005. Given this policy, we elected to conduct a single-institution, sequential, retrospective study evaluating LRR in women with lymph node-negative, HER2-positive breast cancer measuring ≤5 cm who received breast-conservation therapy (BCT) with or without trastuzumab.
MATERIALS AND METHODS
From an institutional breast cancer database, we identified 197 women who were diagnosed between 2002 and 2008 with ≤5 cm, lymph node-negative, HER2-positive breast cancer and who underwent wide local excision and received adjuvant radiation therapy. Patients were included if all lymph nodes were negative or if there were isolated malignant lymph node cells measuring ≤0.2 mm (pN0[i+]). Patients who had tumors ≤1 mm (T1 mic) also were included. Positive HER2 status was defined as an expression level of 3+ by immunohistochemistry or an amplification ratio of ≥2 by fluorescence in situ hybridization. HER2 testing was performed at our institution in 87% of patients. Of the 22 patients who had testing at an outside facility, all had an expression level of 3+ by immunohistochemistry, and 5 of these patients had their results confirmed by repeat testing in another tissue sample.
Patients were excluded if they were diagnosed between May 2004 and May 2005, because trastuzumab prescribing practices varied during this period when the first adjuvant phase 3 trials were reported. Patients also were excluded if they had received partial breast irradiation or another HER2-targeted agent, such as lapatinib. Women who had a history of prior invasive breast cancer, any prior metastatic cancer, any concurrent cancer, or prior anthracycline or taxane therapy also were excluded. Therefore, comparisons were made between 2 cohorts: women diagnosed from 2002 to 2004 who did not receive trastuzumab and women diagnosed from 2005 to 2008 who received trastuzumab (Fig. 1).
Patient characteristics, tumor characteristics, and adjuvant therapies were recorded for all eligible patients. Comparisons between the 2 cohorts were made using the chi-square test. The primary endpoint of the study was LRR. Other reported outcomes included distant metastasis, breast cancer-related death, and overall death. Follow-up for LRR, distant metastasis, and breast cancer-related death was measured from the date of diagnosis to the date of an event or the last breast cancer-specific physical examination or mammogram. Follow-up for overall survival was measured from the date of diagnosis to the date of death or the last direct patient communication. Kaplan-Meier methods were used to estimate LRR-free survival (LRRFS), and the log-rank test was used for comparisons of LRRFS. The 2 cohorts were compared at a follow-up of 3 years to reduce lead-time bias.
Of 197 women who were identified with HER2-positive T1-T2N0 breast cancer and received BCT, 25 women were excluded based on the criteria listed above. Because of a change in prescribing practices in 2005, the remaining women naturally separated into 2 cohorts: 70 women who did not receive trastuzumab and were diagnosed from 2002 to 2004 and 102 women who received trastuzumab and were diagnosed from 2005 to 2008 (Fig. 1). These cohorts were well balanced with respect to patient age, tumor size, histology, and hormone receptor status (Table 1). The median tumor size was 1.5 cm in each group. In total, 80% of patients had tumors that measured <2 cm (T1), and 24% had tumors that measured <1 cm (T1a,b). This was comparable to a recent retrospective study by Perez indicating that 78% of patients who received BCT had pT1 tumors.15 The incidence of lymphovascular invasion in our study was slightly higher in the trastuzumab cohort, but this difference was not statistically significant (P = .28).
Table 1. Patient and Tumor Characteristics for Each Cohort
No. of Patients (%)
No Trastuzumab, N = 70
Trastuzumab, N = 102
Median age [range], y
Tumor size and classification
Median size [range], cm
T1a,b/ ≤1 cm
T1c/ 1-2 cm
T2/ 2-5 cm
Hormone receptor status
Almost all women received some form of adjuvant systemic therapy (chemotherapy and/or hormone therapy) (Table 2). However, 100% of patients in the trastuzumab group received additional chemotherapy, whereas only 73% of the no-trastuzumab group received chemotherapy (P < .0001). In addition, patients in the no-trastuzumab group were less likely to receive a taxane-containing regimen (P < .0001). The median duration of trastuzumab therapy was 50 weeks. One patient in the trastuzumab cohort developed shortness of breath during chemotherapy and stopped treatment. The number of patients receiving hormone therapy was similar in both cohorts, consistent with the balanced hormone receptor status.
Table 2. Treatment Summary for Each Cohort
No. of Patients (%)
No Trastuzumab, N = 70
Trastuzumab, N = 102
Abbreviations; cGy, centigray; RT, radiotherapy.
Anthracycline and taxane
Adjuvant hormone therapy
No adjuvant systemic therapy
Median RT dose [range], cGy
Boost RT to tumor bed
Adjuvant radiation to the whole breast was completed in all but 1 patient in each cohort (Table 2). One patient in the no-trastuzumab cohort developed local recurrence during the first week of radiation treatment. The median dose was 60 Gray in each cohort. Radiation details were available for 91% of patients, and an electron boost to the tumor bed was standard among these patients. A hypofractionated radiation regimen was more common in the trastuzumab cohort, probably because of mounting evidence supporting the efficacy of hypofractionated radiation, but the difference was not statistically significant (P = .20).
Our data revealed significantly improved local control for patients who received trastuzumab compared with those who did not receive trastuzumab (Table 3). Because of the difference in diagnosis dates for the 2 cohorts, the median recurrence-free follow-up was 86 months for the no-trastuzumab group and 47 months for the trastuzumab group. To address the lead-time bias associated with longer follow-up in 1 cohort, we specifically compared events with matched follow-up of 3 years in all patients. At this time point, follow-up data were available for 95% of the no-trastuzumab cohort and 74% of the trastuzumab cohort. Among patients in the no-trastuzumab group, there were 7 local recurrences (10%). The median time to recurrence was 14 months, and no recurrences were observed after 3 years. This early recurrence pattern supported the choice of 3 years as a comparison time point. In the trastuzumab group, there was 1 local recurrence (1%) at 14 months. We used the Kaplan-Meier method to estimate LRRFS (Fig. 2). At 3 years, the no-trastuzumab group had a 90% LRRFS rate (95% confidence interval [CI], 83%-97%), whereas the trastuzumab group had a 99% LRRFS rate (95% CI, 97%-100%). This difference was statistically significant (P = .01; log-rank test).
LRR-free survival was statistically different in these cohorts (P = .01; log-rank test).
Median follow-up (range), mo
3-Year LRR-free survival [95% CI], %
No. with LRR at 3 y
Median time to LRR (range), mo
No. with distant metastases at 3 y
No. of breast cancer-related deaths at 3 y
No. of overall deaths at 3 y
To assess the effect of other chemotherapy on LRRFS, we also used the Kaplan-Meier method to estimate LRRFS in the subgroups of patients who received trastuzumab plus chemotherapy (n = 102), chemotherapy alone (n = 51), or neither (n = 19) (Fig. 3). There were no patients who received trastuzumab alone. At 3 years, these subgroups had LRRFS rates of 99% (95% CI, 97%-100%), 92% (95% CI, 85%-100%), and 84% (95% CI, 68%-100%), respectively. The difference in these results was statistically significant (P = .01; log-rank test), indicating the importance of other chemotherapeutic agents as well as trastuzumab.
The patterns of failure in this study also suggested that hormone receptor status and hormone therapy were key factors in local control (Table 4). Of the 8 local failures, only 1 occurred in a patient who received hormone therapy. Four failures occurred in patients who received chemotherapy alone (including 3 who received both anthracyclines and taxanes). Among the patients who received no systemic therapy at all, 3 of 4 patients developed local failures. All failures occurred in patients who had T1N0 disease (2 patients with T1 mic tumors and 6 patients with T1c tumors). Only 1 failure involved the regional lymph nodes in a patient who also had a local recurrence in the breast.
Table 4. Local Failure Patterns According to Hormone Receptor Status and Adjuvant Systemic Therapya
The number of failures is indicated over the total number of patients in each subgroup.
All of the observed distant metastases and breast cancer-related deaths occurred in patients who did not receive trastuzumab (Table 3). Two of the patients with LRR subsequently developed distant metastases and died of breast cancer. Another patient in the no-trastuzumab cohort developed an isolated liver recurrence and underwent surgery followed by chemotherapy. At the time of this report, she remained alive without evidence of disease. There were 2 additional deaths in the no-trastuzumab cohort and 1 in the trastuzumab cohort that we believe were unrelated to breast cancer.
Our data demonstrate that the high local failure rates that have been observed in early, HER2-positive breast cancer may be lowered with adjuvant trastuzumab chemotherapy. In women with T1-T2N0, HER2-positive breast cancer who received BCT, we observed a 10% local recurrence rate within 3 years without trastuzumab and a 1% rate with trastuzumab. To our knowledge, this is the first published report demonstrating the benefit of trastuzumab for local control in the setting of BCT. We also observed that trastuzumab chemotherapy resulted in consistent benefits across other endpoints, including distant metastasis, breast cancer-related death, and overall death.
Our group recently published a corollary study investigating the effects of trastuzumab in all patients with T1N0, HER2-positive breast cancer.16 That study included women who underwent mastectomy (45%) as well as those who received BCT (55%). It similarly compared an earlier no-trastuzumab cohort (n = 106) with a more recent trastuzumab cohort (n = 155). Patients with T2N0 disease were not included (whereas they comprised 20% of the current study). In that study, again, the patients who received trastuzumab had an improved LRRFS rate (98%; P = .014) and an improved distant recurrence-free survival rate (100%; P = .007) compared with the no-trastuzumab cohort (LRRFS rate, 92%; distant recurrence-free survival, 95%).16 Those results are consistent with our current data and suggest that the impact of trastuzumab is not confined to distant or local sites of potential relapse.
The high LRR rate in our no-trastuzumab group (10%) supports the conclusion that patients with early, lymph node-negative, HER2-positive breast cancer should not be categorized as low risk. Several other studies also have indicated that positive HER2 status confers a considerable risk of recurrence in these patients. A Canadian population-based study reported a 66% 10-year recurrence-free survival rate for patients with lymph node-negative, HER2-positive disease compared with a 76% recurrence-free survival rate for HER2-negative patients.17 Studies from Scotland and Finland also reported decreased DFS and breast cancer-specific survival in patients with T1N0, HER2-positive disease.10, 11 Most relevant to our study, data from The University of Texas M. D. Anderson Cancer Center indicated an 18% LRR rate at 8 years in patients with HER2-positive, T1a,bN0 disease compared with a 4% LRR rate in patients with HER2-negative disease.12, 13
Our retrospective data are consistent with mounting evidence that trastuzumab provides significant benefit for women with small, lymph node-negative breast cancers. Unfortunately, to our knowledge, no randomized studies have been conducted to specifically address the role of trastuzumab in this group of patients. However, subset analyses from the randomized BCIRG and HERA trials offer valuable evidence regarding lymph node-negative patients. In the HERA trial, 32% of enrolled patients were lymph node-negative, and these patients had a DFS benefit with trastuzumab (hazard ratio, 0.59) that was similar to that for the entire cohort (hazard ratio, 0.64).14 In the BCIRG trial, lymph node-negative patients (29% of total) had a 7% improvement in DFS with trastuzumab compared with a 6% improvement in DFS for the entire cohort. Local recurrence was not evaluated for these subgroups.
One potential limitation of this study was the confounding effect of chemotherapy on local control, because all patients in the trastuzumab cohort received additional chemotherapy compared with only 73% of patients in the no-trastuzumab cohort. Our subgroup analysis revealed significant differences in LRRFS among patients who received trastuzumab plus chemotherapy, chemotherapy alone, or no chemotherapy, suggesting that both trastuzumab and chemotherapy were important for locoregional control. However, these subgroup analyses were limited by a low number of events. Furthermore, the effect of hormone therapy may be even more significant than that of chemotherapy for patients with estrogen receptor-positive/HER2-positive disease. Indeed, in our current study, 5 of 8 patients who had local failure received chemotherapy, whereas only 1 received hormone therapy.
Another potential source of bias was the difference in follow-up between the 2 cohorts, because 1 group was diagnosed in the pre-trastuzumab era (2004 and earlier), and the other was diagnosed in the post-trastuzumab era (2005 and after). We specifically compared events with matched follow-up of 3 years to address the lead-time bias, as noted above. At the 3-year time point, data were available for 84% of all patients. Further follow-up is needed to assess for late recurrences in the trastuzumab cohort. However, the recurrences in the no-trastuzumab cohort occurred early, as expected in HER2-positive patients, and no recurrences were observed after 3 years.12, 18
In the United States, 10% of all small, lymph node-negative breast cancers are HER2-positive, representing approximately 15,000 women per year.12, 19 Our data add to the evidence that trastuzumab provides significant benefit for these women. In clinical practice, trastuzumab already is being received by most women who have lymph node-negative, HER2-positive tumors measuring >1 cm, making it difficult to initiate a randomized trial with a no-trastuzumab arm.8, 20 However, several important questions remain unanswered. First, it is unknown whether there are low-risk subgroups that truly do not benefit from trastuzumab. In our study, there was only 1 recurrence among the estrogen receptor-positive/HER2-positive patients who received hormone therapy, suggesting that this subgroup may have a lower risk. However, HER2 overexpression is associated with resistance to hormone therapy, and trastuzumab may improve the risk of late relapse in these patients.21, 22 Second, it is unclear which chemotherapy agents or regimens are optimal in combination with trastuzumab. All studies that demonstrated a benefit of trastuzumab in early stage breast cancer used additional adjuvant chemotherapy. However, the risks of long-term side effects (such as cardiomyopathy for those who receive anthracyclines), may be unnecessarily high for patients with small, lymph node-negative disease. Currently, prospective studies are underway that hopefully will address these questions about the relative benefits and risks of trastuzumab and chemotherapy in this patient population.
In conclusion, in this retrospective study, we observed a 10% LRR rate within 3 years without trastuzumab and a 1% LRR rate within 3 years with trastuzumab in patients who had lymph node-negative, HER2-positive breast cancer who received BCT. Our results demonstrate that the high local failure rates that have been observed in early, HER2-positive breast cancer can be reduced markedly with adjuvant trastuzumab. Lacking prospective randomized data, our results may inform patient care decisions.
This research was supported by the Memorial Sloan-Kettering Cancer Center Breast Cancer Fellowship Support Fund and by the Julie Laub Fund.