With the application of ADCs in metastatic HER2-low breast cancers, HER2-low expression has received increasing attention [17, 18]. However, the clinicopathological features and prognosis of HER2-low tumors remain poorly investigated, especially in Chinese patients. In this retrospective study, we identified 684 HER2-negative breast cancers to detect differences between HER2-zero and HER2-low tumors. The results showed that HER2-low tumors had a higher proportion of HR positivity and a lower proportion of histological grade 3. Moreover, low HER2 expression seemed to be a protective factor for RFI, especially in HR + patients and patients aged < 65 years.
A few studies have focused on low HER2 expression in HER2-negative breast cancer. Schettini et al. evaluated 3689 HER2-negative cases from the cBio Cancer Genomics Portal, and 59.4% of patients had low HER2 expression [12]. A similar proportion (61%) was found by Agostinetto et al., who evaluated 804 cases from TCGA [19]. Interestingly, the proportion of HER2-low patients was higher in Asian patients. In a retrospective study of 4918 HER2-negative patients from Japan, 79.1% of patients had HER2-low tumors[20]. In Chinese patients, a retrospective study of 12,467 patients reported that the proportion of HER2-low tumors was 72.6% [21], which was consistent with our results (74.9%). However, clinicopathological features and prognosis were not further explored in this study. The differences in the HER2-low proportion may be due to racial differences, disease staging, and quality control of HER2 detection.
Furthermore, we found that HER2-low tumors had a higher proportion of HR positivity than HER2-zero tumors (89.6% vs. 75.6%, P < 0.01), which was consistent with the findings of previous studies (90.2% in Japanese populations and 88.2% in cases from the cBio Cancer Genomics Portal)[12, 20]. The high proportion of HR + cells in HER2-low tumors might be explained by the upregulation of luminal-related genes and the downregulation of basal-like genes reported by Schettini et al. [12]. Other clinicopathological features varied across different studies. Horisawa et al. found that HER2-low tumors have a smaller tumor size and lower proportion of histological grade 3 [20], and similar results were found in triple-negative breast cancer (TNBC) by Jacot et al.[22]. Schettini et al. found worse T stages, N stages and histological grades in HER2-low tumors than in HER2-zero tumors [12]. We observed a lower proportion of histological grade 3 in low HER2 expression patients. The reasons for these differences are unclear, and more studies are needed.
In regard to prognosis, previous studies have shown different results. A retrospective study by Yiqun Li et al. involving 1433 patients with metastatic breast cancer reported that patients with low HER2 expression survived longer in the overall population and HR + subgroup [14]. Another study by Dehgani et al. in TNBC obtained a similar result: patients with HER2 2 + had a lower rate of recurrence and longer overall survival (OS) [23]. In addition, other studies found no statistically significant difference in OS between patients with HER2-low and HER2-zero tumors [12, 19]. Conversely, a retrospective study including 91 node-positive patients found that low HER2 expression was associated with shorter disease-specific survival (DFS) and OS, and the correlation was more significant in HR + patients [24]. In another study of 5907 patients, moderate HER2 expression (HER2 2+) was also considered an adverse factor for DFS [13]. The different results of previous studies may be caused by several reasons. First, the inclusion criteria varied in different studies; some focused on TNBC, and some focused on early-stage or advanced breast cancer. Second, as an important prognostic factor, therapeutic regimens were not mentioned in most studies. Third and most importantly, breast cancer with low HER2 expression may be a highly heterogeneous disease, and more efforts are needed to define HER2 levels.
Age is an important factor affecting the prognosis of breast cancer but is poorly investigated in HER2-low patients. In our study, there was no significant difference in the proportion of HER2 statuses between patients aged < 65 years and ≥ 65 years. However, the survival analysis results showed that in patients aged < 65 years, those with HER2-low tumors have a longer RFI and BCSS. This finding may have implications for future research.
Our study has several limitations. First, although we have relatively complete clinicopathological and follow-up data in our database, this was a single-center retrospective study. Second, different criteria for HER2 evaluation were used due to the updating of the ASCO guidelines. Third, some patients with HER2 2 + did not undergo ISH detection and were not included in this analysis. However, we provided data from Chinese patients with HER2-low early breast cancer and performed analyses stratified by HR status and age.