In the DBCG cohort, HER2-low BC was associated with lower histological grade, higher ER-positivity rate, more frequent LN involvement, and marginally larger tumor size compared to HER2 0 BC. This is consistent with other studies, where lower histological grade, higher ER expression, and higher LN positivity rates have also been reported in HER2-low BC compared with HER2 0 (13, 23, 24). Of particular note a large investigation of 1,136,016 patients with invasive BC from the US National Cancer Database found that HER2-low BC was associated with a higher proportion of ER-positive BC and higher LN stage (34). Because the absolute differences in clinical and pathological characteristics are relatively small, the small p-values are primarily driven by the high number of patients in the cohort.
The multivariable survival analysis did not show that HER2-low status was associated with a statistically significant difference in the primary clinical outcome TR. Of the secondary clinical outcomes showed that HER2-low status was an independent prognostic factor for OS (HR 0.85, 0.77;0.94) from 0-2.5 years. No interaction with ER status or histological subtype was observed in the survival analysis.
Retrospective studies have found a variable prognostic significance of HER2-low BC ranging mostly between no effect and a modest positive effect on survival (8–11, 13, 14, 16, 23, 25, 26). A large retrospective cohort study of 1,136,016 patients from the US National Cancer Database reported a marginal improved survival in 5-year OS for HER2-low BC mirroring the DBCG cohorts results (34). A meta-analysis of published studies found that HER2-low status was associated with significantly improved OS (HR 0.84, 0.77;0.92, p < 0.001) and DFS (HR 0.86, 0.79;0.92, p < 0.001) in the early setting, but not for the metastatic setting (5). An European Society for Medical Oncology (ESMO) expert consensus statement found that there was insufficient evidence to support a prognostic value of HER2 expression within HER2-negative BC due to the inconsistency of the studies results and small size of most studies (6).
One fundamental issue with this retrospective survey and others is the interlaboratory and intraobserver variance of HER2 detection in current clinical practice. The differences in clinicopathological characteristics between HER2 0 and HER2-low BC are only modest in absolute percentages. Furthermore, the current HER2 IHC testing guidelines are designed to distinguish between HER2 overexpression and normal expression, but not between lower levels of HER2 expression. Of all BC cases in the national DBCG register, 56% were classified as HER2-low; in other studies, the amount of HER2-low BC has been around 40–50% of clinical HER2-negative BC (7, 15, 35–38), whereas others again have reported frequencies up to 70% (9, 10). Part of this variance can be attributed to interobserver and interlaboratory variance especially within the HER2 0/HER2 1 + categories (39–48). A pre-print study from our group found high inter-laboratory variation of HER2-low BC in our cohort ranging from 46.3–71.8% during among laboratories during individual years (49). This could partially explain the discrepant results about the prognosis of HER2-low BC reported from other studies, and further highlights the need for standardization and an increased focus on HER2-low BC, including alternative assays (50, 51). Other reasons for the discrepancy of the results in the survival studies of HER2-low could be the use of different endpoints, follow-up intervals and variance in the makeup of th and the above-described variance in the makeup of the HER2-low and HER2 0 subgroups make any conclusion drawn from retrospective studies fraught with e makeup of the HER2-low and HER2 0 subgroups makes any conclusion drawn from retrospective studies fraught with a degree of uncertainty. This has been noted previously along with the need for more precise assays (34, 52).
A strength of this study is that the dataset is collected on a national basis with a close-to-complete degree of coverage and with widespread standardization of the guidelines for management, diagnostic workup, and pathological examination. One limitation of this study, which is comparable with other studies investigating the prognosis of HER2-low BC, is the retrospective design and the lack of central review. The ASCO/CAP guidelines were also updated during the study period (see Table 1), which changed the HER2 scoring scale, possibly accounting for some variance in the HER2-low category.