In traditional clinical practice, HER2-low BC is regarded as HER2-negative BC without any effective anti-HER2 targeted therapies. The results of the Destiny-Breast 04 trial illustrate that these patients could benefit from T-DXd, with considerably longer PFS and OS. Subsequently, an increasing number of clinical trials are being conducted, focusing on the strategy of T-DXd used in HER2-low BC and growing interest in HER2-low in the field (Schmid et al. 2021). In our retrospective study, we found that the clinicopathological characteristics of HER2-low were mainly associated with HR status; HER2-low and HER2 IHC0 BC tended to be similar in HR-positive cases, whereas more differences were observed in HR-negative cases. Regarding the prognostic role, the HER2-low group showed a longer OS, but not PFS, compared with the HER2 IHC0 group in the entire population. The same trend was observed in both the HR-positive and HR-negative groups.
In this single-institution study, 86.7% of patients with HER2-negative diagnosis were defined as HER2-low. The proportions of HER2-low in HR positive and HR-negative cases were 86.54% and 72.14%, respectively. The proportions of HER2-low in our cohort was slightly higher than that reported in published studies. A large retrospective study enrolled 1,136,016 BC patients from the National Cancer Database and reported that 70.0% of classical HER2-negative BC in their cohort could be defined as HER2-low in the Native American population (Peiffer et al. 2023). Denkert et al. showed 47.5% HER2-low BC in their pooled analysis, which included data from four clinical trials (Denkert et al. 2021). Previous studies in the Chinese population reported rates ranging from 43.1–75.9% (Xu et al. 2023; Li et al. 2021). We believe that the discordance in the HER2-low rates in different studies is caused by several reasons. First, the patients in the previous pooled analysis were from several hospitals or clinical trials, and heterogeneity of HER2 evaluation among laboratories was inevitable. In addition, the methodological limitations of IHC as the primary technique for identify HER2-low population, which may be influenced by several factors (Tarantino et al. 2020). To avoid the heterogeneity of HER2 IHC scores, we retested HER2 status following the 2018 ASCO/CAP guidelines with a group of well-qualified pathologists at our institution. Regarding clinicopathological features, HER2-low is heterogeneous, with mixed results associated with HR status (Tarantino, Jin, et al. 2022; Alhamar et al. 2021; Horisawa et al. 2022; Xu et al. 2023). In line with prior studies, HER2-low BC in our cohort was more often HR-positive compared with HER2 IHC0 BC in the entire cohort, with a higher proportion of ER level > 10% (76.1% vs. 60.1%, p < 0.001) and PgR-positive (76.7% vs. 62.6%, p < 0.001) and more HER2-low patients received adjuvant endocrine therapy (52.0% vs. 33.3%, p = 0.002). However, subgroup analysis according to HR status is more complex. HER2-low BC had similar clinicopathological characteristics to HER2 IHC0 in the HR-positive population. However, the HER2-low group tended to have more patients with the invasive ductal type compared with the HER2-positive group (94.5% vs. 83.9%) among the HR-negative population, while other features were balanced between the HER2 IHC0 and HER2-low groups. Mutai et al. investigated 608 women with HER2-low HR positive BC and found no differences in any of the prognostic pathological features between HER2–0 and HER2-low, which is consistent with the results in our HR-positive subgroup (Mutai et al. 2021). However, a study that enrolled 777 non-HER2-positive early stage BC patients in a Chinese institution indicated that the clinicopathological features of HER2 IHC0 and HER2-low were balanced in the HR-negative group, whereas the rate of PgR positivity in the HER2-low group was higher than that in the HER2 IHC0 group in HR-positive cases. In the entire cohort, the ki-67 index and TOP2A expression appeared to be lower in the HER2-low group; however, the differences disappeared when patients were divided according to their HR status (Xu et al. 2022). A Korean study based on a national database reported more differences in clinicopathological features in their cohort (Won et al. 2022). HER2-low BC was more frequent in premenopausal patients and was associated with fewer T4 tumors, higher histological grade, and negative lymphatic invasion compared with HER2 IHC0 BC in the HR-positive group, whereas older patients with a higher lymph node ratio and positive lymphatic invasion were observed in the HR-negative group. Across the two groups, they also reported a lower ki-67 index in the HER2-low index. Taken together, although no clear conclusion can be drawn, one common finding was that HER2-low BC was more likely to be HR-positive and probably have a lower ki-67 index compared with HER2 IHC0. The survival analysis in our study suggested that HER2-low BC had a longer OS in the entire cohort. We further analyzed the results under different HR statuses, and the results showed that OS improved in both the HR-positive and HR-negative groups. No statistically better DFS was observed between the HER2-low and HER2 IHC0 groups, regardless of the HR status. Although HER2 expression is a negative prognostic factor (Moasser 2007), the results in the field of survival of HER2-low BC compared to HER2 IHC0 BC are still disputable. Consistent with the results of the present study, Mutai et al.. reported that compared to HER2 IHC0 disease, HER2-low was associated with significantly improved OS, DFS (Mutai et al. 2021). Data from the US National Cancer Database reported small improvements in OS in patients with stage III HER2-low BC (Peiffer et al. 2023). Several studies found no differences in OS or DFS (Tarantino, Jin, et al. 2022; Tan et al. 2022); however, a study in Japan showed HER2-low was a negative prognostic factor for stage I triple-negative BC (Sanomachi et al. 2023). Another study enrolled 5907 primary non-metastatic invasive BC cases pointed out that HER2-low may be a poor prognostic factor in the HR-positive group. Given the unignorable influence of treatment strategies on survival, a study enrolled 410 patients with systemic treatment-naive node-negative BC with a long-term follow-up time of over 10 years and found a substantially longer 3-year DFS (83.4% vs.76.1%) and 3-year OS (91.6% vs. 85.8%) compared with HER2 IHC0 BC, proving the pure prognostic impact of HER2-low versus HER2 IHC0. However, differences in OS and DFS were detected in the HR-positive group, and no significant difference in OS was detected in the HR-negative group (Almstedt et al. 2022). Overall, despite the inconsistent prognostic results, the majority of studies reported an improved trend of both OS and DFS in the HER2-low group in both the HR-positive and HR-negative groups, with or without statistical significance.
Consistent with prior reports (Tarantino, Jin, et al. 2022), we found a positive correlation between ER status and the HER2-low rate, which means that the rate of HER2-low increased with increasing ER levels. Recent data at the genetic and transcription levels have confirmed these findings. Schettini et al. analyzed the PAM50 raw gene data of 1320 BC patients from several datasets and found that HER2-low tumors had higher expression of luminal- and proliferation-related genes, and tyrosine kinase receptor genes were more highly expressed in HER2 IHC0 tumors than in HER2-low tumors (Schettini et al. 2021; Massarweh and Schiff 2006), which can also explain the lower ki-67 index in HER2-low BC. Agostinetto et al. evaluated ERBB2 and ESR1 expression, and their results were consistent with those reported by Schettini. Moreover, they reported a positive correlation between ERBB2 and ESR1 expression in both the HER2-low and HER2 IHC0 groups. However, when further analyzed according to HR status, the differences disappeared in the negative group (Agostinetto et al. 2021). The positive correlation of ER status and crosstalk low rate, along with previous studies demonstrating that endocrine therapy resistance may be associated with the crosstalk between ER signaling and growth factor signaling (Massarweh and Schiff 2006) (Eggemann et al. 2015), inspired us to further explore the relationship between HER2 IHC0 and HER2-low BC at different ER levels. However, no such regularity was observed in the present study.
Patients may experience a dynamic change in the HER2 IHC score during the process of tumor progression, a type of HER2 heterogeneity that raises scientists’ interest and may be associated with resistance to anti-HER2 antibody-drug conjugates (Miglietta et al. 2021; Almstedt et al. 2023; Ocaña, Amir, and Pandiella 2020). We explored and confirmed the instability of HER2 IHC status and described the HER2 IHC status from the primary site to the corresponding metastatic site. 52 paired cases were analyzed in our study, with 28.84% of cases showing a discordance in the HER2 IHC score between the primary and paired metastatic sites. As studies have demonstrated, the administration of endocrine therapy, chemotherapy, and radiotherapy may upregulate HER2 expression (Seol et al. 2012; Tarantino, Gandini, et al. 2022).
Our study has the following limitations. First, as a retrospective study, some clinicopathological information was missing, and treatments of some patients could not be obtained, which may have influenced the prognosis. Second, our sample size was insufficient and we enrolled only the Chinese population at a single institution. Third, the median DFS and OS were not reached in our study, and a longer follow-up time is needed to better compare the survival differences between the HER2 IHC0 and HER2-low groups in early BC. More metastatic patients with longer follow-up periods should be observed. Fourth, due to the retrospective nature of the study, we failed to obtain all relapsed biopsies, which may have influenced the analysis of HER2 IHC evolution. Finally, the molecular characteristics were not described in this study, which are important to better understand the landscape of the HER2-low population, especially in the precision treatment era. Further improvements are required for the molecular stratification of ERBB2 expression.