Approximately 15% of breast cancers are considered to be “HER2-positive”, and the definition of HER2 positivity has been based on HER2 protein overexpression and/or gene amplification.^1,2  Since HER2 is a validated therapeutic target, a clear understanding of which patients have tumors that are truly HER2-positive, HER2-low or HER2-negative is critical in determining which patients are candidates for HER2-directed therapy. 

Traditionally, we have included all HER2 breast cancers together into a dichotomous group, with patients considered either HER2-positive or HER2-negative based on IHC and/or florescence in situ hybridization (FISH) testing of the breast cancer tumor tissue.³ In recent times, this approach has become more complicated, given the evolution of treatment options that have proven to be effective in the “HER2-negative” space, specifically within the population of patients now classified as having HER2-low breast cancer, leading to a change in the way of thinking about HER2 positivity in breast cancer.

In 2019, the FDA approved T-DXd for patients with unresectable or metastatic HER2-positive breast cancer who have received 2 or more prior anti–HER2-based regimens in the metastatic setting based on the results of the DESTINY-Breast 01 trial.⁴ With the availability and efficacy of T-DXd for patients with breast cancer, it became increasingly evident that this agent is effective not only in patients with HER2-positive breast cancer, but also in those with HER2-low disease.

This paradigm shift led the field to understand which tumors previously classified as “HER2-negative” have some level of HER2 expression, including tumors with HER2 IHC 1+ or IHC 2+ levels of expression that are not FISH amplified.⁵ These tumors are now categorized as HER2-low, and can be treated with T-DXd due to its recent approval for patients with unresectable or metastatic HER2-low breast cancer who have received a prior chemotherapy regimen in the metastatic setting or developed disease recurrence within 6 months of completing adjuvant chemotherapy based on the results of the DESTINY-Breast04 trial.⁶

The current classification of HER2 in breast cancer is based on the ASCO/CAP guidelines, written in 2018 and confirmed by the panel in 2023.^7,8 Based on these guidelines, HER2 expression in breast cancer is tested by validated IHC assay and divided into HER2-positive (IHC 3+), HER2-equivocal (IHC 2+), and HER2-negative (IHC 1+ and IHC 0), with HER2 positivity defined as IHC3+ and/or FISH amplified.

If an individual’s disease is determined to be HER2 IHC 0, that tumor has been traditionally classified as HER2-negative. The definition for HER2 IHC 0 breast cancer encompasses having tumors that have no staining observed or incomplete and faint or barely perceptible membrane staining in ≤10% of tumor cells. However, it is clear that tumors traditionally scored as HER2 IHC 1+ or 2+ have some level of HER2 expression, but not enough to lead to an IHC 3+ score.

HER2 IHC 1+ is defined as incomplete membrane staining that is faint or barely perceptible membrane staining in >10% of tumor cells. HER2 IHC 2+ is complete membrane staining that is weak to moderate and observed in >10% of tumor cells. HER2 IHC 3+ is defined as circumferential membrane staining that is complete, intense, and in >10% of tumor cells.

Per the ASCO/CAP guidelines, equivocal tumors (IHC 2+) require further gene expression testing of the same sample to look for FISH amplifications in HER2.^8-10 If a new sample is available, further confirmatory tests using IHC or ISH is required to see if the tumor is truly HER2-positive.

For patients whose tumor biopsy is IHC 3+ or IHC 2+ and ISH-amplified, it is recommended that a HER2-targeted therapy be prescribed if there is no apparent histopathologic discordance, and if such therapy is clinically appropriate.

Although the ASCO/CAP guidelines remain an important guidepost for understanding HER2 testing, the field is continuing to evolve based on clinical need.

The phase III DESTINY-Breast04 study showed a significant improvement in survival for patients with HER2-low (IHC 1+ or IHC 2+/ISH-) unresectable or metastatic breast cancer who received T-DXd compared with chemotherapy after progression on other standard therapies. Of importance for the ASCO/CAP guideline panel, patients with disease classified as HER2 IHC 0 were excluded from this study.

Based on the results of DESTINY-Breast04, if a breast tumor is HER2 IHC 1+ or IHC 2+/ISH-, the field has evolved to thinking of that tumor as being HER2-low.¹¹ 

The ASCO/CAP guidelines are traditional and have been used to classify breast cancer broadly as either HER2 positive or negative. However, the introduction of T-DXd has changed how HER2 positivity is classified. With this evolution, the HER2-low breast cancer classification was born. Of note, the guidelines were not initially built with the understanding of HER2-low breast cancer, they were built upon understanding which patients have the potential to benefit from traditional HER2-directed antibody therapies, such as trastuzumab and pertuzumab. Over time, the thinking of HER2 positivity in the field had to evolve just as HER2 test interpretations have had to change. Some challenges, however, remain with the evolution of the HER2-low classification.

It is important to keep in mind that the ASCO/CAP guidelines were not developed to be able to clearly discern HER2 IHC 0 from HER2-low or HER2-positive breast cancer. Now that these HER2 classifications are evolving, the current testing methodologies make it difficult for pathologists to easily and reproducibly make these distinctions since current HER2 testing assays are semiquantitative and were optimized to detect overexpression vs any HER2 expression.

As such, there can be poor concordance rates among pathologists when scoring the same slides, especially those classified as HER2 IHC 0 and HER2 IHC 1+, using these semiquantitative methods. In a recent study of 18 pathologists, the concordance rate was 26%.¹² These data confirm that HER2 classification by human visual scoring is not an easy task and makes it difficult to reliably determine optimal therapy for patients with HER2 IHC 1+ tumor vs HER2 IHC 0.

The specific HER2 classification has therapeutic implications, and yet, it is very challenging for our pathologists given the current scoring system. This is partly because within HER2 IHC 0 tumors, there can be some level of HER2 staining on tumor cells and it can be difficult for pathologists to differentiate a score of IHC 1+ from IHC 0.

In addition to difficulty using semiquantitative assays to differentiate between HER2 IHC 0 and HER2 IHC 1+ tumors, where and when a biopsy is obtained can also influence the results of HER2 testing. For example, there is a recent autopsy study in which samples were collected from a patient’s liver.¹³ Looking at the HER2 expression across different segments of the liver, there were areas that were sometimes classified as HER2-low and HER2 0 within the same segment. These data suggested that where the biopsy needle is placed could dramatically change the result of HER2 testing, and this is particularly important for a patient who has a HER2-low tumor vs a HER2 IHC 0 tumor. The heterogeneous expression of HER2 levels even within a specific organ makes pathologists’ task very challenging.

In addition, HER2-low expression can change over time.^14-16 When comparing the status of a primary breast cancer to its metastatic counterpart at the time of recurrence or relapse, it is important to note that a tumor can evolve from being classified as HER2 IHC 0 to HER2-low and it can go from being HER2-low to HER2 IHC 0 at time of recurrence. So, the HER2 expression level is not stable over time.

This instability of HER2 expression has led many of us to wonder whether HER2-low tumors in fact are a biologically distinct subtype of breast cancer. A way to think about this potential phenomenon is asking whether having a HER2-low tumor compared to a HER2 0 tumor can lead to a difference in prognosis.

There have been multiple studies conducted to try to assess and answer this question.¹⁷ The vast majority of the studies suggest that in fact there does not seem to be a prognostic difference between HER2 0 and HER2-low breast cancers when corrected for ER status.

It is important to be cautious when interpreting these studies, however, because most are retrospective studies where no central testing for determining HER2-low status was performed. There is a lot of variability in the determination of HER2-low status based on current pathology reads. Although not a perfect comparative study, given the large number of patients who have been tested, it seems unlikely that HER2-low breast cancer is a biologically distinct entity associated with a different prognosis.

Given all these caveats, several questions remain. To many treating physicians, it is still not clear what approach should be taken in the clinic to clearly define who actually has HER2-low breast cancer. Algorithms have been proposed to define HER2-low breast cancer, which has been defined as a tumor that is IHC 1+ or IHC 2+ and one that is not FISH amplified.¹⁸ However, because this status can change over time, another question arises: how do we know what time point to use to make a clinical decision about whether or not a patient should receive T-DXd? For instance, if a patient’s primary disease is classified as being HER2-low at the time of diagnosis, but the most recent biopsy is classified as being HER2 IHC 0, should this patient receive T-DXd or not?

A dataset that can help with answering some of these questions, at least to some extent, is from the DESTINY-Breast04 trial.¹⁹ When you look back at the tissue that was utilized in DESTINY-Breast04 to determine tumor HER2 expression level, and you look at the median progression-free survival (PFS) results based on the tumor location, specimen type, specimen collection date, and tissue sample characteristics used, it is clear that these factors did not make a difference in PFS outcome. Regardless of these features, the median PFS favored T-DXd in patients with HER2-low (IHC 1+ or IHC 2+/ISH-) unresectable or metastatic breast cancer.

It did not matter if the tumor was tested using the primary tissue sample or whether the sample was collected many years prior to the patient going on to the trial. There was benefit with T-DXd regardless of whether HER2 was tested using the most recent tumor tissue, fresh, or archived tissue. In general, a relative PFS benefit of T-DXd compared to chemotherapy in DESTINY-Breast04 was observed across all different tissue sources utilized.

My interpretation of this dataset is that, even if the patient had a HER2-low primary cancer and yet has a HER2 0 metastatic sample, I am willing to consider T-DXd for that patient, because in DESTINY-Breast04, patients who had primary tumor samples collected years before enrollment on the trial that were classified as HER2-low benefitted from T-DXd.

With the consideration that it is not clear which of the patients had disease that evolved to be HER2 0 prior to receiving T-DXd, it is impossible to clearly understand the true difference in efficacy compared to a patient whose disease retained its HER2-low status. Nonetheless, this dataset provides a signal of benefit to suggest that it would be reasonable to take any HER2-low status at any time point to make a decision regarding the utilization of T-DXd.

At present, an important question is: how low do we go for HER2 expression to know if a patient will benefit from T-DXd? Some very provocative data that have been published on this topic are from the phase II DAISY trial, which suggested that, in patients with HER2 IHC 0 tumors, treatment with T-DXd resulted in an objective response rate of 30%.²⁰ Even though this study was conducted in a small cohort of patients with HER2 IHC 0 tumors (n = 37), a median PFS of 4.2 months was observed with a robust response rate, suggesting that there are patients who do seem to benefit from T-DXd even if they are HER2 0. This study requires further evaluation and validation to better understand exactly how low we can go with HER2 expression.

The ongoing randomized phase III DESTINY-Breast06 trial will help to shed more light on how low we can go with HER2 expression for a patient to derive benefit from T-DXd (NCT04494425). DESTINY-Breast06 includes patients with HER2-low, hormone receptor (HR)-positive advanced/metastatic breast cancer, but also includes a cohort of patients who are HER2-ultralow (IHC >0 and <1+). Patients without any prior exposure to chemotherapy in the advanced/metastatic setting will be randomized to receive either T-DXd or physician’s choice of chemotherapy (capecitabine, paclitaxel, or nab-paclitaxel). Randomization will occur after disease progression within 6 months of initiating first-line metastatic treatment with an endocrine therapy-based treatment or after disease progression on at least 2 prior lines of endocrine-based therapy with or without a targeted therapy for metastatic disease. The primary endpoint of the study is PFS in the HER2 IHC 1+/2+ cohort by central review.

Not only will this study help us to understand if T-DXd can be utilized earlier in patients with HR-positive metastatic disease, but it will also help to determine whether T-DXd has any benefit in the cohort of patients with HER2-ultralow breast cancer.

The paradigm about HER2 expression level in breast cancer has evolved to include the HER2-low status, and this will continue to evolve with time.⁵ Right now, we think of cancers as being either HER2 IHC 0, HER2-low or HER2-positive. This paradigm could evolve even further very soon to include a HER2-ultralow status, and maybe even towards a more quantitative assay in the future.

It appears that the HER2 expression spectrum will continue to evolve with time; we are awaiting data from DESTINY-Breast06, which may help us understand if T-DXd is effective in patients with HER2-ultralow tumors. This will mean that patients with tumors that have traditionally been categorized as HER2 IHC 0, but in fact have some level of HER2 expression that lies somewhere between a IHC 0 and IHC 1+, may have a targeted treatment option available to them. If the DESTINY-Breast06 trial data show benefit in this subgroup of patients, it could lead to a further change in the definition of HER2 expression in breast cancer. My hope is that in the future, we will be able to develop more quantitative assays for HER2 that could lead to more of a continuous understanding of HER2 expression levels.

There are multiple ways in which we can have more accurate and reproducible assays for the detection of HER2 levels in HER2-low breast cancer.^24-26 For instance, via the utilization of digital pathology to take away some of the challenges that come with subjective human assessment. Another method is to develop more quantitative assays, using techniques such as quantitative immunofluorescence. Such an approach will allow for more quantitative reads. Finally, techniques such as RT-qPCR utilizing messenger RNA may help to improve our understanding about HER2 expression.

All of these different modalities are currently under investigation and their association with benefits to ADCs is being evaluated. The availability of more standardized testing modalities can help with making decisions regarding the selection of ADCs in the future and can assist with selecting patients who will benefit from each of the different ADCs. Hopefully, the development of newer and more accurate testing methods will give us ways to have more reproducible test results, and circumvent the challenges associated with discordant reads of the same slides from pathologist to pathologist.

A promising approach is to select patients with HER2 IHC 0 staining by the traditional methodology for further testing. This group of patients may in fact have quantifiable HER2 expression if a sensitive quantitative assay were available. In support of this notion, we have some data that were recently presented looking at a quantitative assay that utilizes a deep learning-based image analysis approach.²⁷ Using this artificial intelligence–based algorithm, the investigators looked at membrane staining for patients previously classified as HER2-negative using the traditional method in a quantitative fashion and then associated that level of staining in a quantitative way with efficacy from T-DXd using a cellular optical density cutoff to generate a HER2 quantitative continuous score (QCS).

A higher objective response rate was seen among patients with a higher QCS compared with patients who had a lower QCS (53% vs 24%). Similarly, the median PFS was higher among those with a higher QCS compared with those with a lower QCS (14.5 months vs 8.6 months). This method was useful in better selecting patients who benefitted from T-DXd compared with the more traditional IHC methodology.

Obviously, only a small number of patients were included in this study (n = 65). Even though the results need to be further validated, these data suggest that the presence of a stronger target expression using a sensitive quantitative measure can help to better select those patients who are more likely to have even more robust benefit from the available ADCs such as T-DXd.

Another recently presented promising approach is the assay that uses a reverse-phase protein array technology, which revealed that among patients who traditionally were classified as having HER2 IHC 0, approximately 30% of those with ER-negative tumors and 40% of those with ER-positive tumors came back with quantifiable HER2 expression levels.²⁸ These results further suggest that patients with disease traditionally classified as having HER IHC 0, in the absence of a more sensitive and quantitative assay, may actually be deprived of deriving benefit from T-DXd.

Studies by David Rimm and colleagues are looking at an automated quantitative assay (AQUA). The AQUA method assesses HER2 expression using immunofluorescence staining.²⁵ This technology has integrated immunofluorescence staining into a more quantitative HER2 score termed the AQUA score.

Using the AQUA HER2 scoring approach, 67% of patients with disease previously classified as HER2 IHC 0 showed HER2 expression. Hopefully, the AQUA approach will help us to be able to more accurately select patients who will likely benefit from HER2-directed ADCs. In the future, if we have multiple available ADCs, more sensitive and quantitative HER2 scoring approaches such as the AQUA technology may help us to figure out optimal sequencing approaches for ADCs based on the true expression level of the target of interest. Of importance, studies to validate the AQUA technology are currently ongoing.