Long-Term Survival Analysis of Adjuvant Chemotherapy with or without Trastuzumab in Patients with T1, Node-Negative HER2-Positive Breast Cancer Free

Overexpression of human epidermal growth factor receptor 2 (HER2) occurs in 15% of patients with breast cancer and is associated with poor prognosis (1, 2). Trastuzumab is a humanized monoclonal antibody that binds the extracellular domain of HER2 and has been shown to improve the overall survival (OS) rate in patients with early-stage and metastatic HER2-positive breast cancer (HER2⁺ BC; refs. 3, 4). Administration of adjuvant trastuzumab in combination with or following cytotoxic chemotherapy improved the OS rate compared with chemotherapy alone (5, 6). However, most randomized clinical trials included primarily patients with stage II or III HER2⁺ BC. Few pivotal randomized trastuzumab trials included patients with tumor <1 cm. The number of patients with small (≤2-cm diameter) HER2⁺ BC is increasing because of systematic mammographic screening in healthy women around the world (7). Patients with small HER2⁺ BC have a proven higher risk of relapse than do those with small tumors without HER2 amplification/overexpression. However, patients with small tumors are expected to benefit less from adjuvant systemic therapy than those with larger tumors or positive lymph nodes (8). Moreover, the decision to use chemotherapy and trastuzumab is complicated by the potential toxicity of systemic treatment (9).

Currently, HER2⁺ BC that is >1 cm is often empirically treated with chemotherapy and anti-HER2 therapy; no definitive conclusions have been made on the efficacy of the addition of trastuzumab to adjuvant chemotherapy for node-negative HER2⁺ BC ≤1 cm in diameter (10, 11). A large study containing 21,148 patients concluded that adjuvant chemotherapy provided no benefit in patients with T1a (≤0.5 cm) tumors (12).

Because the event rate of this small, low-risk tumor is so low, a placebo-controlled, prospective randomized study comparing various therapeutic strategies would require a very large trial. And physicians and patients would not feel comfortable with an arm without trastuzumab-based therapy. Instead, to address the potential role of trastuzumab and chemotherapy in these low-risk patients, we conducted a retrospective, population-based study of women with ≤2-cm diameter, node-negative HER2⁺ BC. We aimed to further explore the possibility of omitting chemotherapy in certain subgroups of these patients.

This study was conducted in accordance with the Declaration of Helsinki and was approved by The University of Texas MD Anderson Cancer Center Institutional Review Board. The institution's tumor registry system and electronic medical records were used to retrieve records of patients who had been treated at MD Anderson between January 1, 1998, and October 31, 2009. The inclusion criteria were (i) female, (ii) pathologic confirmation of invasive breast adenocarcinoma, (iii) pathologic confirmation of HER2 positivity (according to the immunohistochemistry (IHC; score of ≥3+) and equivocal cases subjected to fluorescence in situ hybridization (FISH) with positivity for HER-2/neu gene amplification defined as HER2/CEP 17 ratio >2.2; ref. 13), (iv) largest primary invasive tumor diameter of ≤2 cm, (v) no lymph node metastasis, and (vi) no distant metastasis. Patients with isolated tumors of ≤0.2 mm in lymph nodes [pN0(i+)] were considered lymph node-negative (14). The exclusion criteria were (i) did not undergo surgery, (ii) received only trastuzumab without adjuvant chemotherapy, and (iii) unknown tumor size. Patients were grouped according to therapeutic regimens into group A (no adjuvant chemotherapy), group B (adjuvant chemotherapy only), or group C (adjuvant chemotherapy plus trastuzumab).

Two trained physicians reviewed the medical records and pathologic reports to extract data on epidemiologic factors, clinical course, tumor-related characteristics, and survival. Then a senior physician-investigator checked for consistency and performed a final data check. Nuclear staining ≥10% of estrogen receptor (ER) or progesterone receptor (PR) was considered a positive result. We defined the HER2⁺:ER/PR⁺ subtype according to the immunohistochemical assessment of ER and PR, and HER2 status as described above. The HER2-enriched only subtype means those with HER2 positive but ER and PR both negative.

The primary outcome measure was disease-free survival (DFS). Other outcomes were OS, distant recurrence-free survival (DRFS), and breast cancer–specific survival (BCSS). DFS was defined as the time from the date of surgery to the date of local recurrence, distant relapse, death, or last follow-up without relapse according to Standardized Definition of Efficacy Endpoints criteria (15). OS was defined as the time from the date of diagnosis to the date of death or last follow-up. DRFS was defined as the time from the date of diagnosis to the date of distant recurrence or last follow-up. BCSS was defined as the time from the date of diagnosis to the date of death caused by breast cancer or last follow-up. The cause of death was unknown for one patient, and we censored this patient on the day before her death in the analysis for BCSS. The longest follow-up interval was 20 years.

Patient characteristics were summarized using frequency tables showing the number and percentage of patients within a particular category. The χ² test was used to identify significant differences in proportions or rates between various groups. We used the Kaplan–Meier method to estimate DFS, OS, and DRFS, and the log-rank test was used to determine statistically significant differences between groups. Cox proportional hazards regression models were fit to determine the association between therapy regimens and DFS or DRFS after adjusting for other related prognostic factors. The restricted mean survival time (RMST) at 10 years was calculated to compare OS between patients who received no adjuvant chemotherapy and those who received adjuvant chemotherapy only for the crossing OS curve. The RMST ratio was calculated to adjust for related prognostic factors and other disease characteristics (16). Cumulative incidence estimates were used to account for competing risks of death. For BCSS, Fine-Gray competing risk analysis and a multivariable regression model were used to evaluate prognostic factors.

Maximally selected rank statistics were used to find a cutoff point for tumor size with relative risk against survival outcomes. We then divided patients into two subgroups according to the tumor size cutoff and further validated the effective therapies using multivariable Cox proportional hazards models in these two subgroups.

All statistical analyses and charts of survival probability were performed with R statistical software (version 3.5.0, R Foundation for Statistical Computing; http://www.R-project.org/). Hazard ratios (HR) were reported with their 95% confidence intervals (CI). A P value of <0.05 was considered statistically significant in all analyses.

Of the 603 female patients initially identified, 16 were excluded because they had unknown tumor size (n = 10), did not undergo surgery (n = 2), or received only trastuzumab (n = 4). A total of 587 consecutive patients with pT1N0M0 HER2⁺ BC were included in our study. The median follow-up time was 123.0 months (interquartile range 85.0 to 162.0 months). Among these patients, 249 had received no adjuvant therapy (group A), 181 had received adjuvant chemotherapy only (group B), and 157 had received adjuvant chemotherapy plus trastuzumab (group C). The chemotherapy regimens included 5-fluorouracil, doxorubicin, and cyclophosphamide; doxorubicin and cyclophosphamide; taxanes; and other regimens. The percentage of patients who received adjuvant chemotherapy only was highest between 1998 and 2004; this subgroup comprised 89.5% of all group B patients in our study. From 2005 to 2009, patients who received adjuvant chemotherapy plus trastuzumab comprised 89.2% of all group C patients.

The baseline clinicopathologic characteristics for patients who received local and systemic therapy are summarized in Table 1. The proportion of invasive ductal carcinoma was higher in group B (83.4%) than in the other two groups (59.4% in group A and 77.7% in group C). The group distribution was well balanced for race, surgery type, and adjuvant radiotherapy. More patients were postmenopausal in group A than in the other groups. The most common nuclear grade was III (67.1% in group A, 76.2% in group B, and 79.0% in group C). The HER2⁺:ER/PR⁺ subtype was seen in 48.6% of group B patients, which was less than the 63.5% seen in group A and the 61.1% seen in group C. More patients had larger tumors in group B than in the other two groups; in group B, 77.3% of patients had a tumor dimension of >1 cm, whereas tumors of this size were seen in 31.7% of group A patients and 57.3% of group C patients (P < 0.001).

The P values of the comparisons, along with the 10-year survival rates for all patients with pT1N0M0 disease, is reported in Table 2. Among the three groups, patients treated with adjuvant chemotherapy with trastuzumab had the best prognosis in terms of DFS, OS, DRFS, and BCSS. However, in the OS analysis, there was no statistical difference between patients who received no adjuvant chemotherapy and patients who received adjuvant chemotherapy only (Fig. 1), suggesting that the use of adjuvant chemotherapy only in patients with pT1N0M0 HER2⁺ BC is probably not adequate to improve survival.

Among all patients in the cohort, the 10-year DFS rate was 81.0% in group A (42 events in 249 patients), 65.4% in group B (59 events in 181 patients), and 97.3% in group C (4 events in 157 patients; P < 0.001). For OS analysis, group C patients had the best results among the three groups as well. The DRFS rate was highest in group C as well (97.2% compared with 86.7% in group A and 69.7% in group B; P < 0.001). The 10-year BCSS rate was 91.1% in group A, 77.1% in group B, and 98.6% in group C (P < 0.001). Invasive ductal carcinoma was associated with poorer DFS, OS, DRFS, and BCSS compared with other histologic subtypes (P < 0.05 for all). Other factors associated with better DFS included age <65 years, postmenopausal status, tumor size ≤0.5 cm, and ER-positive disease. Additional factors associated with relatively better OS were age <65 years and tumor size ≤0.5 cm (Table 2). In DFS, DRFS, and BCSS analyses, the prognoses of patients treated with adjuvant chemotherapy only were the poorest. Patients who had ER⁺ disease had a significantly better prognosis in terms of DFS, DRFS, and BCSS.

In Kaplan–Meier analysis, the therapy regimen was associated with prognosis in DFS, OS, DRFS, and BCSS analysis in patients with T1N0M0 disease (Fig. 1). Patients who received adjuvant chemotherapy plus trastuzumab (compared with no adjuvant therapy; HR, 0.071; 95% CI, 0.025–0.204; P < 0.001), were ages 65 years or older (compared with <65 years; HR, 0.341; 95% CI, 0.167–0.695; P = 0.003), and had HER2⁺:ER/PR⁺ subtype (compared with HER2-enriched only subtype; HR, 0.622; 95% CI, 0.415–0.931; P = 0.021) had significantly better DFS (Table 3). In addition, larger tumor size (HR, 2.384; 95% CI, 1.549–3.056; P < 0.001) was associated with worse DFS. Patients who received adjuvant chemotherapy plus trastuzumab (compared with no adjuvant chemotherapy) had better DRFS (HR, 0.106; 95% CI, 0.037–0.310; P < 0.001) and BCSS (HR, 0.116; 95% CI, 0.035–0.385; P < 0.001). There were no statistical differences in DFS, DRFS, and BCSS between patients who received adjuvant chemotherapy only and those who received no adjuvant chemotherapy (Table 3).

A crossing OS curve was observed between group A and group B patients. In these patients, the mean OS time (as of 10 years from diagnosis) was 111.70 months for group A and 105.33 months for group B. The RMST was 112.25 months in group A and 103.79 months in group B (Supplementary Fig. S1). However, the P value for this analysis was not statistically significant. The RMST ratio in group B patients was not significantly different from that of group A patients (RMST ratio = 0.982; 95% CI, 0.930–1.036; P = 0.498), after adjusting for invasive ductal carcinoma pathologic findings, age, nuclear grade, ER status, and tumor size (Supplementary Table). These results suggest that stage I HER2⁺ BC may not have a significantly demonstrable survival benefit from receiving adjuvant chemotherapy only.

The standardized Wilcoxon statistic for the tumor cutoff value is shown in Supplementary Fig. S2. The optimal tumor cutoff size identified for predicting DFS, OS, BCSS, and DRFS was 0.8 cm.

We found that after adjusting for related prognostic factors, receiving adjuvant chemotherapy plus trastuzumab was not associated with better DFS, OS, or DRFS in patients with tumors <0.8 cm. For OS analysis, age ≥65 years (compared with <65 years; HR, 4.472; 95% CI, 1.851–10.800; P = 0.001) and nuclear grade III (compared with other grades; HR, 3.062; 95% CI, 1.248–7.510; P = 0.015) were associated with increased all-cause mortality (Table 4). However, age ≥65 years was associated with better DFS and DRFS but worse OS in patients with tumors ≥0.8 cm. Adjuvant chemotherapy with trastuzumab was associated with better DFS, OS, DRFS, and BCSS in patients with tumors ≥0.8 cm (Table 5).

The role of trastuzumab-based chemotherapy remains controversial in small, node-negative HER2⁺ BC. The 2018 National Comprehensive Cancer Network guidelines recommend consideration of adjuvant trastuzumab-based chemotherapy in patients with T1b, node-negative HER2⁺ BC (17). In our study, patients with HER2⁺ tumors smaller than 0.8 cm at maximum diameter had excellent prognosis without trastuzumab-based chemotherapy, raising the question of whether these therapies are necessary for these patients.

In our study, larger tumor size was a relevant negative prognostic factor for patients with pT1N0M0 disease, which is consistent with previous studies (18, 19). The tumor size was determined meticulously by expert pathologists at MD Anderson. Because adjuvant treatment recommendations for pT1N0M0 HER2⁺ BC were lacking before 2009 (20), the therapeutic regimens given to patients were not uniform across the three groups under investigation. For example, a substantial number of patients with T1c tumors did not receive adjuvant chemotherapy plus trastuzumab, whereas many patients with T1a tumors received the combination therapy. Therefore, there seemed to be equipoise among medical oncologists.

Using tumor size as a continuous variable in the setting of non-uniform distribution of treatment regimens conferred enough statistical power to find the accurate tumor cutoff size to define the population for whom combination therapy may be omitted without adverse influence on DFS, OS, DRFS, and BCSS. Previous studies discussed treatment according to the traditional American Joint Committee on Cancer T category only but did not consider a subgroup of smaller-sized tumors. We hope that our study can help select the optimal patient population with small, node-negative HER2⁺ BC to receive and benefit from adjuvant chemotherapy and anti-HER2 therapy.

The World Health Organization defined curable cancers as those in patients who have a high potential for DFS in the 10 years after cessation of treatment (21). We defined 10-year DFS as the primary outcome for our study because long-term outcomes data are not readily available in most reports of small HER2⁺ BC. Furthermore, it is important to keep in mind competing risks. In one study, the risks of death from early-stage breast cancer and cardiovascular disease were equivalent for the first 5 years, whereas the risk of death from cardiovascular disease was higher after the first 5 years (22). Fehrenbacher and colleagues (23) reported that the 5-year recurrence-free rate was 96.2% in T1a and 91.1% in T1b HER2⁺ BC. The risks of recurrence were too low for patients to receive adjuvant chemotherapy with or without trastuzumab.

Because of potential adverse events caused by trastuzumab and chemotherapy, the efficacy and potential risk should be balanced. Most patients with early-stage breast cancer are cured, and therefore survival analysis of 10 years or longer is essential for evaluating the efficacy of different therapies. Our median follow-up time was longer than 10 years, allowing us to evaluate the long-term efficacy and toxicity of various therapies.

In our study, age ≥65 years was a positive prognostic factor for DFS and DRFS, but a negative prognostic factor for OS. Younger patients tend to have cancers that exhibit more aggressive features and are more likely to have relapse. In contrast, older patients are more likely to present with comorbidities and die of cardiovascular diseases (24). Our study showed that age was not a prognostic factor for BCSS, which was in accordance with the study conducted by Jenkins and colleagues (25). Older patients may be more likely to die from intercurrent disease or comorbidities than younger patients.

Most reported clinical trials of T1 HER2⁺ BC were single-arm studies. We included 249 patients treated with no adjuvant chemotherapy as a control group, even including those with T1c disease, which can help us better compare the various therapeutic regimens used in clinical practice. Nowadays it would be impossible to conduct a trial with such a control group for ethical reasons. In patients who received no adjuvant chemotherapy, 10-year BCSS was 91.1% and 10-year OS was just 79.0%. This means that a large percentage of patients who received no adjuvant chemotherapy died of other diseases.

In our study, the 10-year DFS rate was 97.3% in those who received adjuvant chemotherapy plus trastuzumab. This indicates that the addition of trastuzumab to adjuvant chemotherapy can improve the prognosis of patients with small HER2⁺ BC. We found that not only 10-year DFS but also 10-year OS, BCSS, and DRFS were higher in patients who received no adjuvant chemotherapy than in those who received adjuvant chemotherapy only. In short, patients treated with both trastuzumab and chemotherapy fared better than those in the other two groups in this cohort. However, the percentages of invasive ductal carcinoma and larger tumors were higher in patients who received adjuvant chemotherapy only than in those who received no chemotherapy, which may have affected the survival outcomes in our study. The differences of tumor characteristic distributions in different therapy groups can affect our results to a certain degree although we take the associated prognostic factors into account in multivariate analysis to balance the bias and control for these covariates. Additional analysis with data from other centers and meta-analysis is needed to validate these conclusions.

Patients with HER2⁺:ER/PR⁺ breast cancer had a good prognosis in our study, which was consistent with another study (26). However, Vaz-Luis and colleagues reported that 5-year DRFS was 93% for patients with hormone receptor-positive subtypes and 96% for those with hormone receptor-negative subtypes in a study of patients with pT1aN0M0 HER2⁺ BC who did not receive chemotherapy or trastuzumab (27). This implies that HER2⁺ BC has distinct subtypes: hormone receptor-positive versus hormone receptor-negative. The use of endocrine therapy may help improve the survival of those patients with hormone receptor-positive BC.

Pertuzumab is another humanized monoclonal antibody that was shown to improve DFS in patients with early HER2⁺ BC when the pertuzumab was combined with trastuzumab and chemotherapy (28). The APHINITY trial have found that the addition of pertuzumab to the chemotherapy cannot significantly improve the node-negative early HER2⁺ BC (29). Pertuzumab plus trastuzumab may not be effective in patients with node-negative HER2⁺ BC. Ado-trastuzumab emtansine (T-DM1) is effective in extending survival in patients with metastatic HER2⁺ BC (30). The KATHERINE trial explored the role of T-DM1 in adjuvant chemotherapy for patients with residual disease after neoadjuvant chemotherapy (31). The phase II ATEMPT clinical trial is designed to compared the toxicity of T-DM1 and the traditional combination therapy of trastuzumab and paclitaxel in patients with stage I HER2⁺ BC. And the investigators are also hoping to learn about the long-term benefits and DFS between T-DM1 and the combination of trastuzumab and paclitaxel (32). Careful patient selection is important when choosing a certain treatment regimen to avoid unnecessary toxicity. More studies are needed to explore the possibility of T-DM1 or dual anti-HER2 agents without chemotherapy such as trastuzumab plus Neratinib in adjuvant therapy for patients with high-risk HER2⁺ BC that is <8 mm.

Therefore, we would suggest chemotherapy combined with trastuzumab for HER2⁺ BC ≥8 mm. Furthermore, it would be reasonable to individualize treatment recommendations for patients with tumor less than 8 mm. For example, for those low risk patients with tumor <8 mm, chemotherapy and trastuzumab may be omitted. And for high risk patients with tumor <8 mm such as those with histologic grade III and/or high Ki-67 index, T-DM1 or dual anti-HER2 agents without chemotherapy may be considered.

Our study is limited by its retrospective nature, resulting in unbalanced baseline characteristics within the groups and lack of randomized distribution of patients to the different treatment groups. Moreover, the study was based on a single institution analysis with the possibility of difference in local oncology practice patterns. However, we used multivariable analysis containing related prognostic factors to balance the results. Because of the limited number of patients, we did not compare individual cytotoxic chemotherapy regimens in our study. However, a non-randomized clinical trial recently showed excellent outcomes in patients with node-negative HER2⁺ BC that was <3 cm when treated with adjuvant paclitaxel and trastuzumab (33).

Despite these limitations, our results offer convincing evidence that patients with pT1, node-negative HER2⁺ BC–treated adjuvant chemotherapy plus trastuzumab fared better than patients who received no chemotherapy or adjuvant chemotherapy only. Patients with node-negative, HER2⁺ BC with tumors ≥0.8 cm should receive a combination of chemotherapy and trastuzumab; however, the combination therapy of trastuzumab and chemotherapy may not be necessary for patients with tumors <0.8 cm. Considering the retrospective nature of our study, related meta-analysis may be needed to validate the optimal systemic therapy for patients with node-negative HER2⁺ BC with tumors <0.8 cm.

S.-C.J. Yeung reports receiving commercial research grants from Bristol-Myers Squibb and DepoMed. No potential conflicts of interest were disclosed by the other authors.

Conception and design: X. He, M. Tian, F.J. Esteva, G.N. Hortobagyi, S.-C.J. Yeung

Development of methodology: X. He, J. Ji, S.-C.J. Yeung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. He, G.N. Hortobagyi, S.-C.J. Yeung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. He, J. Ji, F.J. Esteva, G.N. Hortobagyi, S.-C.J. Yeung

Writing, review, and/or revision of the manuscript: X. He, J. Ji, F.J. Esteva, G.N. Hortobagyi, S.-C.J. Yeung

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. He, S.-C.J. Yeung

Study supervision: X. He, G.N. Hortobagyi, S.-C.J. Yeung

The authors wish to acknowledge Tamara K. Locke from the Department of Scientific Publications at MD Anderson Cancer Center for editing this article. No specific funding for this research. MD Anderson Cancer Center was supported by the National Cancer Institute [MD Anderson Cancer Center Support Grant number P30 CA016672] of the National Institutes of Health.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.