HER2-low status may predict poor neoadjuvant chemotherapy response in HR-negative breast cancer: a real-world multicenter study

Abstract

Objective

We aimed to investigate the impact of human epidermal growth factor receptor 2 status (human epidermal growth factor receptor 2-low versus human epidermal growth factor receptor 2-zero) on pathological response to neoadjuvant chemotherapy and survival outcomes in early-stage breast cancer.

Methods

Patients with primary invasive breast cancer received neoadjuvant chemotherapy between July 2018 and July 2021 were identified from six hospitals. The primary efficacy end-point was total pathological complete response. The second short-term efficacy end-points include breast pathological complete response, axillary lymph nodes pathological complete response and the score of Miller-Payne grade. Long-term efficacy end-point was disease-free survival.

Results

429 patients with human epidermal growth factor receptor 2 negative invasive tumors were included, 267 (62.24%) had human epidermal growth factor receptor 2-low tumors. Hormone receptor-positive patients had a higher percentage of human epidermal growth factor receptor 2-low tumors compared to hormone receptor-negative patients (71.97% versus 42.14%). The pathological response rate was significantly lower in human epidermal growth factor receptor 2-low tumors than in human epidermal growth factor receptor 2-zero tumors for total patients in univariate analysis, including the rates of total pathological complete response (5.2% versus 14.2%), breast pathological complete response (6.4% versus 17.3%), nodes pathological complete response (26.3% versus 37.7%) and MP4–5 (21.2% versus 33.8%). Subgroup analysis showed that the rates of total pathological complete response, breast pathological complete response and MP4–5 were also significantly lower in human epidermal growth factor receptor 2-low tumors versus human epidermal growth factor receptor 2-zero tumors in both univariate and multivariate analysis in hormone receptor-negative subgroup. With the median follow-up of 24 months, disease-free survival was comparable between these two subgroups (P = 0.816).

Conclusions

Our results demonstrate that human epidermal growth factor receptor 2-low tumors achieved a significantly lower pathological complete response rate with conventional chemotherapy than those with human epidermal growth factor receptor 2-zero tumors, especially for hormone receptor-negative group. Large, randomized, prospective studies are needed to confirm our data and further evaluate the prognostic value of human epidermal growth factor receptor 2-low expression.

Background

Patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer established aggressive biological behaviors and poor prognosis. It has been defined as a subgroup of breast cancer for >20 years, accounting for ~15–20% of all breast cancers (1–3). Since the development of anti-HER2 agents, the outcomes of patients with HER2-positive breast cancer have been dramatically improved (4–6). The predictor of responsiveness to anti-HER2 agents is HER2-positive, which is defined by the overexpression of HER2 protein measured using immunohistochemistry (IHC) 3+, and/or in situ hybridization (ISH) (HER2 gene copy number ≥ 6 or a HER2/CEP17 ratio ≥ 2.0) based on 2018 guidelines established by American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) (3). Thus, clinicians and pathologists have traditionally focused on the distinction of the duality between a positive and a negative result.

However, a new entity has recently emerged named HER2-low breast cancer, defined by IHC 1+, or IHC 2+/ISH-negative (3,7,8). HER2-low breast cancer represents ~45–60% of HER2-negative breast cancer. Patients with HER2-low tumors are currently reported not benefitted from conventional HER2-directed therapies (trastuzumab, pertuzumab), therefore, are not recommended (9). However, HER2-targeted antibody–drug conjugates (ADCs), such as trastuzumab deruxtecan (T-DXd, formerly DS8201a) (10) and trastuzumab duocarmazine (SYD985) (11) have shown the potential to treat patients with HER2-low expression tumors, which led to extensive attention of physicians toward the particular subset of patients with HER2-low tumors (11,12).

Several studies indicated that HER2-low and HER2-zero breast cancer may be biological different; and HER2-low breast cancer is significantly different from HER2-zero breast cancer with regard to histology, stage at diagnosis, hormone receptor (HR) status, tumor proliferation and grading (13,14). But there were contradictory results of the impact of HER2-low expression in response to conventional chemotherapy and survival for patients with early-stage breast cancer (15,16). Several retrospective studies did not support HER2-low as a biologically distinct breast cancer subtype, with no predictive effect for pathological complete response (pCR) after conventional neoadjuvant chemotherapy (NAC) and no prognostic value on survival outcome (17–20). But a recent pooled analysis of 2310 patients from four prospective neoadjuvant trials indicated that HER2-low breast cancer is significantly different from HER2-zero breast cancer, with significantly lower pCR rate and longer survival (13). A potential explanation for the discordance among different studies could be different patient populations with different biological characteristics in each study. More studies are needed before defining the separate subtype of HER2-low breast cancer.

The role of NAC for localized breast cancer has evolved tremendously over the past several years. Patients with residual disease after NAC in the breast and/or axilla are at increased risk of recurrence or death (21), emphasizing the need to improve pCR rates and to investigate the role of HER2-low in the efficacy of conventional NAC and patients’ survival. The aim of the present retrospective multi-center cohort study was to investigate the impact of HER2 status (HER2-low vs. HER2-zero) on the pathological response of NAC for breast cancer patients, and to evaluate the HER2 status on survival outcomes in early-stage breast cancer.

Patients and methods

Patients selection and study design

This is a retrospective multi-center cohort study. Patients with primary invasive breast cancer received NAC between July 2018 and July 2021 were identified from six hospitals, The First Affiliated Hospital with Nanjing Medical University, The Affiliated Changzhou No. 2 People’s Hospital of Nanjing Medical University, Xuzhou Central Hospital, Affiliated Hospital of Xuzhou Medical University, The Affiliated Hospital of Nantong University, and The Second Affiliated Hospital of Soochow University. This study was conducted with the approval of the institutional ethics committee of our institution, and was in compliance with the Helsinki declaration. All of the patients provided written informed consent to participate.

The inclusion criteria included the following: (i) female patients with the age over 18 years old diagnosed with HER2-negative invasive breast cancer (HER2 0 or 1+ in IHC or HER2 2+ in IHC with fluorescence in situ hybridization [FISH] negative); (ii) Tany (any tumor stage), Nany (any lymph node stage), M0 staging according to UICC criteria; (iii) unilateral breast cancer; (iv) received neoadjuvant chemotherapy (6–8 cycles) and underwent curative surgery. Exclusion criteria were: (i) male patients; (ii) patients who received any previous surgery, radiotherapy or endocrine therapy before enrollment; (iii) patients with metastatic or relapsed disease; (iv) patients had previous of concomitant other malignancies.

Estrogen receptor (ER) status, progesterone receptor (PR), HER2 status, Ki-67 were available on baseline core needle biopsy. Based on the guidelines established by ASCO/CAP, HR-positivity was defined as ER and/or PR ≥ 1% stained cells. HER2 status was assessed by IHC and/or FISH. HER2-negative was defined as IHC scoring 0–1+ or ICH 2+, but without FISH amplified. Besides, HER2-low status was defined IHC 1+ or 2+/FISH non-amplified.

The patients flow diagram for the study is shown in Figure 1. A total of 781 consecutive breast cancer patients were screened, 322 patients with HER2-positive tumors and 30 patients with other reasons were excluded. Baseline clinicopathological characteristic were collected including age at diagnosis, gender, histologic subtype, ER status, PR status, HER2 status, Ki-67, clinical TNM staging. We also collected treatment data like chemotherapy and surgery.

Chemotherapy agents included the use of either conventionally scheduled or intensified dose-dense anthracyclines and taxanes for most patients, carboplatin for some patients with triple-negative breast cancer (TNBC), and some other agents. None of the patients included in our analysis had received an anti-HER2 therapy.

Study end points

The primary efficacy end-point was total pCR (tpCR) defined as no residual invasive tumor cells in any resected specimens of the breast and axillary nodes (ypT0/is ypN0), stratified by HER2-zero and HER2-low. The second short-term efficacy end-points include breast pCR (bpCR) defined as no residual invasive tumor cells in breast (ypT0 yPN0/+), axillary lymph nodes pCR (npCR) defined as no residual invasive tumor cells in axillary lymph nodes (ypTany ypN0) and the score of Miller-Payne (MP) grade. MP assessment was evaluated based on reduced tumor cellularity of resection samples and comparison with core needle biopsy samples (22). Grades 4–5 are categorized as good pathological response. pCR and MP assessment were determined by two pathologists. Long-term efficacy end-point was disease-free survival (DFS) and the proportion of patients who had tumor recurrence. DFS was defined as the time from breast cancer diagnosis to the earliest locoregional or contralateral relapse, distant metastasis, or death from any cause. Additional stratification was performed by HR status (HR-negative and HR-positive).

Statistical analysis

Patients included in the study were groups according to their HER2 status (HER2-low and HER2-zero). The ages of patients were represented by means and ranges. Other categorical clinical and pathological variables were summarized and compared across groups using chi-square or Fisher’s exact test. The differences in pathological response rates were evaluated as rates difference with 95% confidence intervals (CI) between the groups of patients with HER2-zero and HER2-low tumors. Multivariable logistic regression models including age, clinical T stage, clinical N stage, HR status, Ki-67 and chemo-regimen were used to report odds ratios (OR) with 95% CI. In logistic regression analysis, cases where data of npCR or MP grade were not available were excluded. Survival curves for DFS were constructed using the Kaplan–Meier method. Baseline clinicopathological variables associated with DFS with Wald’s P-value ≤ 0.2 in the univariate analysis were selected for multivariate analysis. The proportion of tumor recurrence was calculated as the ratio of patients with radiologically or pathologically confirmed tumor recurrence to the total number of patients. All P-values were two-tailed with 5% significance levels. All statistical analyses were performed by using software STATA version 16.0 (StataCorp, College Station, TX).

Results

Patients and tumor characteristics

In the present study, we finally identified 429 patients with HER2 negative invasive breast cancer between July 2018 and July 2021. The main clinical and pathological characteristics are summarized in Table 1. Median age of the patients at study entry was 49.56 (range: 24–87) years. Of the 429 patients, 289 (67.37%) had HR-positive tumors (luminal-like breast cancer) and 140 (32.63%) had HR-negative tumors (TNBC). According to HER2 status, 162 (37.76%) had HER2-zero tumors and 267 (62.24%) had HER2-low tumors. Of the 267 patients with HER2-low tumors, 129 (48.31%) were HER2 1+ and 138 (51.69%) were HER2 2+. HR-positive patients had a higher percentage of HER2-low tumors compared to those with HR-negative patients (71.97% vs. 42.14%, P < 0.001). Statistically significant differences were also detected between HER2-zero and HER2-low tumors for clinical stage at diagnosis (P = 0.031) and Ki-67 status (P = 0.013). Patients with HER2-low tumors had slightly less conserving surgery when compared to those with HER2-zero tumors. Most of the patients (85.08%) received anthracyclines plus taxanes as preferred chemotherapy regimen.

Predictive value of HER2 status on pathological response

Overall, 37 (8.62%) patients achieved tpCR, with 9 (3.11%) in patients with HR-positive tumors and 28 (20.00%) in patients with HR-negative tumors. HER2-low tumors had a significantly lower pCR rates compared with HER2-zero tumors, including the rates of tpCR (5.2% vs. 14.2%, P = 0.002), bpCR (6.4% vs. 17.3%, P = 0.001) and npCR (26.3% vs. 37.7%, P = 0.014) (Table 2 and Fig. 2A). But no significant differences in pCR rates were seen in multivariable logistic regression analysis. The results of multivariate analysis including all variables were shown in Supplementary Table 1.

In subgroup analysis according to HR status (Table 2), HER2-status was not associated with the rates of pCR (tpCR, bpCR and npCR) in patients with HR-positive breast cancer in both univariate and multivariable logistic regression analysis (all P > 0.05). For patients in HR-negative group, HER2-low tumors had significantly lower pCR rates than HER2-zero tumors, including both the tpCR (11.9% vs. 25.9%, P = 0.045) and bpCR (15.3% vs. 30.9%, P = 0.037). Similar results were seen in multivariable logistic regression analysis, for both the tpCR and bpCR (OR = 0.344, 95% CI = 0.127–0.927, P = 0.035; OR = 0.367, 95% CI = 0.149–0.907, P = 0.030, respectively). Differences of nCR rates were not seen by HER2 status in HR-negative tumors (42.4% vs. 54.3%, P = 0.193) (Fig. 2B).

The pathological responses were also assessed by MP grades. Of the 409 patients assessed with MP grading system, 106 (25.92%) patients had an MP grade 4–5 response. The rate of MP4–5 was significantly different between patients with HER2-low tumors and those with HER2-zero tumors (21.2% vs. 33.8%, P = 0.005) in univariate analysis. But multivariable logistic regression analysis confirmed that HER2-low was associated with significantly lower MP4–5 rate only in HR-negative subgroup (OR = 0.424, 95% CI = 0.197–0.913, P = 0.028), not in HR-positive subgroup (Table 2 and Fig. 2).

Predictive value of HER2 status on patient survival

With the median follow-up time of 24 months (9–44 months) for all the patients included in the study, 40 DFS events were registered. The difference was not statistically significant between patients with HER2-low breast cancer and with HER2-zero breast cancer (hazard ratio = 0.915, 95% CI = 0.477–1.792, P = 0.816) (Fig. 3A). We should note that the DFS data were only 9.32% (40/429) mature and results might be inconclusive. Subgroup analyses revealed that the rates of DFS were also not significantly different in patients with HR-positive tumors nor with HR-negative tumors (hazard ratio = 0.741, 95% CI = 0.290–1.897, P = 0.530; hazard ratio = 1.549, 95% CI = 0.596–4.025, P = 0.362, respectively) (Fig. 3B and C).

We also performed further analysis to compare DFS of patients who experienced disease relapse (n = 40) according to HER2 status. The mean DFS time of patients with HER2-low tumors (28.28 months) was similar to those with HER2-zero tumors (28.93 months) (P = 0.812). Among HR-negative cases who experienced disease relapse, the mean DFS time of patients with HER-low tumors was shorter than that of patients with HER-zero tumors, although the difference was not statistically significant (24.78 months vs. 28.75 months, P = 0.322).

Besides, sites of recurrence were not significantly different between patients with HER2-low and HER2-zero tumors (Supplementary Fig. 1).

Discussion

It is proposed that HER2-low breast cancer could be distinguished as a distinct subtype on the basis of molecular characteristics, response to treatment and prognosis. But previous studies still have contradictory results (13,14,17–20,23). To the best of our knowledge, this is the first study with series of real-world data in China that specifically investigate the role of HER2-low in early-stage breast cancer patients undergoing NAC. We have detected the differences between HER2-low and HER2-zero tumors among 429 patients with 62.24% of whom were classifies as HER2-low, comparable with that reported in previous studies with the prevalence of this emerging subtype ranged between 45 and 60% (7,16). And we have confirmed the strong relationship between HER2-low status and HR-positive status, which is consistent with previous studies (16).

Regarding the HER2-low status impact on conventional chemotherapy response, the differences of pathological response after NAC between HER2-low and HER2-zero tumors varied substantially across studies (Table 3). Up to data, there were seven studies published, investigating the role of HER2-low in pathological response after NAC (13,17–20,23,24). Four of the seven studies did not support HER2-low as a biological distinct breast cancer subtype (17–20). With the largest sample size, a pooled analysis of 2310 patients from four prospective neoadjuvant trials indicated that patients with HER2-low tumors exhibited significantly lower rate of total pCR after NAC compared to patients with HER2-zero tumors (17.5% vs. 23.6%, P = 0.024), but better survival. Besides, the study reported that the lower rate of pCR for HER2-low tumors was seen only in the HR-positive subgroup (17.5% vs. 23.6%, P = 0.024), but not in HR-negative group (13). Of the other six retrospective studies, a retrospective series of 331 patients presented at San Antonio 2020 also reported that patients with HER2-low tumors experienced significantly lower pCR than those with HER-zero tumors in the overall cohort, but no differences were found in subgroups according to HR status possibly due to small sample size (24). What is more, most of the studies are limited to patients in western countries, and only one study was carried out in China that included just 87 patients receiving NAC (23). The study reported that HER2-low tumors had significantly lower pCR rate than HER2-zero tumors but a higher proportion of disease-free patients across follow-up time points. No differences of pCR rates were determined in subgroup analysis according to HR status. Our study validated and added on to this work based on larger sample size in multi-center study in China. We observed significantly lower tpCR rate, bpCR rate and MP4–5 rate in patients with HER2-low tumors compared to HER2-zero tumors, and multivariable analysis confirmed the differences in HR-negative subgroup.

Though the prognostic value of HER2-low expression warrants further evaluation, the pCR rate in our study differed according to HER2 status in HR-negative patients, different from previous reports from western countries. A multiomic early-stage breast cancer cohort that contains 411 Chinese HER2-low patients also hinted the potential ethnic difference of HER2-low breast cancer. This single cohort study showed that HER2-low tumors seemed to be more distinguished from HER2-zero tumors in the HR-negative subgroup compared with the HR-positive subgroup. HR-negative HER2-low tumors have higher internal molecular heterogeneity, which is the ethnic specificity of the East Asian population (25). A single cell analysis showed that TNBC with intratumoral heterogeneity may be associated with a decreased likelihood of achieving a pCR. It was hypothesized that HR-negative breast cancer with HER2-low status is a heterogeneous entity, with HER2 expression present a variable profile (26,27). Specifically, 37.9% of HER2-low patients in HR-negative group were non-basal-like disease, which had unique driving mechanisms including PI3KCA mutation enrichment, FGFR4/PTK6 overexpression and PI3K/AKT/mTOR pathway activation (25). And upregulated PI3K/AKT pathway serves as a potential indicator of lack of response to NAC in breast cancer patients (28). Additionally, further basic research are needed to clarify the relationship between chemo-effect and HER2 low expression in HR-negative patients. And this population may need additional treatment such as deruxtecan in the future.

Several limitations existed in the current study. Firstly, it is a retrospective study, which may have been responsible for both selection and information bias. Secondly, it is a multi-center cohort study with patients from six centers. There was a lack of HER2 assessment by a central laboratory, but all of the samples were reviewed by two expert pathologists in blinded fashion and obtained a good agreement. Thirdly, it is worth mentioning that the DFS data were immature and OS was hard to analyze due to the relatively short follow-up time and limited incidence of death events. Thus, a longer follow-up period is necessary for definitive analysis. Prospective studies with large sample size would be helpful for confirming our data.

Conclusions

To our knowledge, the present study is the first retrospective multi-center study in China with real-world data, specifically investigating the potential effect of HER2-low expression in pathological response after NAC in patients with HER2-low tumors. We found that patients with HER2-low tumors achieved a significantly lower pCR rate with conventional chemotherapy than those with HER2-zero tumors, especially for HR-negative group. Large, randomized, prospective studies are needed to confirm our data and further evaluate the prognostic value of HER2-low expression. Besides, basic research might be important to clarify the relationship between chemo-effect and HER2 low expression.

Conflict of interest statement

The authors have declared that no competing interests exist.

Funding

This work was supported in part by the National Natural Science Foundation of China (81771953, 82172683), the Natural Science Foundation of Jiangsu Province (BK20180108) and A project Funded by the Priority Academic Program Development of Jiangsu higher Education Institutions (PAPD).

Availability of data and material

All data supporting the results reported in the article are available from the corresponding authors.

References

Author notes