NEW YORK – The availability of AstraZeneca and Daiichi Sankyo's Enhertu (trastuzumab deruxtecan) for treating HER2-low breast cancer has shaken up long-established HER2 expression classification categories and upended how pathologists and oncologists manage their patients.
Some pathologists are questioning the ability of immunohistochemistry (IHC) testing to reliably capture varying levels of HER2 expression, while oncologists are scrambling to revisit patients' earlier test results to see if those previously deemed negative for HER2 overexpression are actually low expressers who now qualify for Enhertu.
The HER2 expression biomarker category, meanwhile, may still see further volatility as the makers of Enhertu want to see it approved in even broader patient populations — for those with very low or minuscule levels of HER2 expression. But AstraZeneca and Daiichi Sankyo don't see the current biomarker detection challenges in the HER2 space as insurmountable and believe they can be addressed by training pathologists to better interpret IHC results when tumors have low levels of HER2.
The present situation began in 2022 when the US Food and Drug Administration approved AstraZeneca and Daiichi Sankyo's HER2-directed antibody-drug conjugate (ADC) Enhertu for HER2-low unresectable or metastatic breast cancer patients who have received prior chemotherapy in the metastatic setting or experienced recurrence within six months of adjuvant chemo. Concurrent with that approval, Roche's Pathway anti-HER2 (4B5) IHC test was approved as a companion diagnostic.
Since the 1998 approval of the first HER2-targeted monoclonal antibody, Genentech's Herceptin (trastuzumab), doctors and pathologists have been using tests to identify patients eligible for these types of treatments based on if their tumors are "HER2 positive" or "HER2 negative." Historically, tumors with an IHC score of 0 or 1+ have been considered HER2 negative and not likely to respond to HER2-targeted treatment, while tumors with 3+ scores are HER2 positive and considered responders. Tumors with 2+ IHC scores have traditionally been called "equivocal" and have needed a second test like in situ hybridization (ISH) to determine whether they could respond to HER2-targeted therapy.
In the past 25 years, HER2-overexpressing breast cancer has been accepted as a subtype of the disease, and HER2 positivity and negativity is evaluated alongside estrogen and progesterone receptor status upon diagnosis. When Enhertu first came to market in 2019 for unresectable or metastatic breast cancer patients who have already received another HER2-targeted therapy, treatment eligibility was based on the traditional parameters for HER2 positivity.
And while the traditional HER2 testing parameters remain, for now, for breast cancer subtyping, for guiding other HER2-targeted drugs, and for the above Enhertu indication, the release of the DESTINY-Breast04 trial data in 2022 threw a wrench into established biomarker testing workflows with the introduction of HER2-low as a category. This new category includes tumors with an IHC 2+ score and a negative ISH result or an IHC 1+ score, denoting either weak-to-moderate complete staining for the presence of HER2 expression or faint, partial staining of the membrane in greater than 10 percent of the cancer cells. In DESTINY-Breast04, Enhertu showed efficacy in this HER2-low subgroup, demonstrating improvements in survival outcomes compared to physician's choice of chemotherapy.
With this approval, AstraZeneca and Daiichi Sankyo significantly expanded the number of breast cancer patients eligible for the HER2-directed therapy. Using the traditional definition of HER2 positive, between 10 percent to 20 percent of breast cancers were eligible for Enhertu. HER2-low patients, meanwhile, make up nearly 60 percent of breast cancer patients, according to one estimate.
Enhertu, as an ADC, is designed to bind to the HER2 protein on the surface of tumor cells and release a cytotoxic payload, a DNA topoisomerase I inhibitor. That cytotoxic payload, however, is highly membrane permeable and can kill neighboring tumor cells that may not be overexpressing HER2, using a mechanism dubbed the "bystander effect." AstraZeneca and Daiichi Sankyo announced this year that a Phase III trial has shown their drug has activity in breast cancers with ultralow HER2 expression and the firms are also exploring Enhertu's activity in HER2-negative breast tumors.
"Part of the ADC works through binding to a target, but part of the ADC also works without requirement for a target," said Paolo Tarantino, a breast medical oncologist at Dana-Farber Cancer Institute. "There is this release of chemotherapy that is prolonged, more prolonged than traditional chemo, and it does not require antigen expression. It's possible that there is activity of the drugs in HER2 0, HER2-null breast cancer due to the release of chemo irrespective of the antigen."
Measuring the biomarker
As AstraZeneca and Daiichi Sankyo try to broaden the market for Enhertu by reframing the HER2 biomarker from a binary, positive or negative, category to a spectrum based on IHC scores, providers on the ground are worried that, using currently available test methodologies, they won't be able to reliably identify patients who will benefit from treatment. Clinical guidelines bodies and many pathologists have not fully embraced the HER2-low category amid concerns that the lower limit of HER2-low expression, like tumors with IHC 1+ scores, cannot be reliably determined using IHC.
Several studies have explored reproducibility and inter- and intra-reader concordance for HER2 IHC results, but findings have varied widely, said David Rimm, a professor of pathology in the departments of pathology and oncology at Yale University School of Medicine.
"There are studies showing that we're pretty good at [distinguishing IHC] 3+ versus 2+ [tumors], but we're not so good at 0 versus 1," Rimm said, adding that these studies have shown that pathologists tend to score low HER2 expression quite differently. "That's pretty concerning that whether or not you get the drug depends on your pathologist."
One international study from 2023 reported "moderate" inter-observer agreement for HER2-low IHC scores, but the researchers acknowledged that separating cases between IHC 1+ and 2+ cases was "problematic" in their study. When pathologists scored IHC 1+ and 2+ together as one HER2-low category, agreement was higher, at about 74 percent.
Validation studies conducted by Roche for its Pathway companion diagnostic concluded that the assay was "precise and reproducible for scoring HER2-low status," the authors wrote. The validation study, which was based on more than 1,000 samples collected within DESTINY-Breast04, found that concordance between local and central labs in determining HER2-low tumors — those with scores of IHC 1+ or IHC 2+ and ISH-negative — was 77.6 percent.
The validation study also found that there was higher agreement between individual readers when it came to determining non-HER2-low specimens — grouped together as tumors with IHC 0 and 3+ scores — than the HER2-low specimens. Pathologist agreement was 99.7 percent for IHC 0 scores and 98.9 percent for IHC 3+ scores in the study. The most scoring variability in the HER2-low category was with IHC 1+ slides. The majority of the discordance in this category, according to researchers, was due to readers scoring IHC 1+ tumors as IHC 2+.
However, the validation study did not explore the readers' ability to discern between IHC 0 and IHC 1+, a critical flaw in the study's design in Rimm's view, since research from him and others has suggested that distinguishing between the lower IHC scores is more difficult, and therefore, IHC 0 and 3+ readings shouldn't have been grouped together. In Rimm's research, the overall percent agreement for reading IHC 0 results across 18 pathologists was only 25 percent. When asked to score IHC 1+ and 2+ slides separately, agreement among pathologists was also low. For IHC 1+ readings, agreement was less than 1 percent, reflecting the fact that pathologists agreed on only one IHC1+ case out of 102 in the study. Agreement for IHC 2+ scoring was 3.6 percent.
However, when the researchers combined the 1+ and 2+ scores into one category, which reflects Enhertu's current HER2-low indication, pathologist agreement reached 27 percent. In comparison, agreement on IHC 3+ scores in the study was much higher, with half of pathologists scoring the same samples as IHC 3+. There was even higher agreement, 87 percent, when pathologists were asked to determine whether a sample was either IHC 3+ or not, reflecting the traditional HER2-positive category.
Rimm criticized the studies Enhertu's sponsors submitted to the FDA for the Pathway HER2 companion diagnostic, specifically the decision in these studies to measure precision and reproducibility between readers in two categories, IHC 1+/2+ and IHC 0/3+. He argued that the agency should have evaluated whether pathologists could use the assay to distinguish 0 from 1+, especially since many other studies have shown this to be more difficult.
One study exploring HER2 IHC scoring concordance among 16 pathologists found that the cases with the lowest agreement on HER2 IHC scores were all HER2-low or HER2-ultralow. The highest agreement, 86 percent, occurred when IHC categories were clustered into IHC 0 versus all others. Another study exploring agreement on HER2 IHC scoring on two different assays across six pathologists found a higher level of agreement in scoring HER2 IHC 0 from IHC 1+, 78.1 percent with Agilent's HercepTest and 72.2 percent with the Pathway anti-HER2 assay, but the researchers noted agreement was higher when pathologists were asked to distinguish IHC 2+ and IHC 3+, 91.9 percent for HercepTest and 86.3 percent with the Pathway anti-HER2 assay.
"[The companies] grouped the 3+, which we're really good at, with the 0, [with the rationale that] those are the patients that aren't qualified for the drug and the 1+ and 2+ patients are qualified," he said. "[The FDA] didn't require the company to do what the pathologists actually have to do when they sit there in front of the slide."
To make matters more challenging for pathologists, the American Society of Clinical Oncology and the College of American Pathologists haven't fully incorporated HER2-low into their joint guidelines as a new category. In 2023, nearly a year after Enhertu's US approval in HER2-low breast cancer, ASCO and CAP updated their recommendations for HER2 testing in breast cancer to acknowledge that there was a new indication for Enhertu as an option for breast cancers that don't overexpress HER2. However, they also concluded that there were limited data on the efficacy of HER2-targeted ADCs in IHC 0 tumors, and therefore it was premature to change the traditional HER2-positive/negative categories and add a new HER2-low or ultralow category.
Given Enhertu's approval for HER2-low breast cancer, however, the organizations recommended that the best practice for measuring the biomarker is for pathologists to report whether a patient's tumor is positive or negative for HER2 overexpression using the traditional categories, but then include the semiquantitative IHC score and a footnote explaining how eligibility for Enhertu can be determined using IHC 1+ and 2+ scores.
The groups also recommended best practices for pathologists to distinguish IHC 1+ from IHC 0, which includes examining HER2 IHC-stained slides at high-power magnification of at least 40X; getting a second review from another pathologist when results are close to the 1+/0 threshold; validating tests using controls with a range of HER2 expression; and paying careful attention to the pre-analytic conditions during the processing of breast cancer tissue samples.
For their part, AstraZeneca, Daiichi Sankyo, and their companion diagnostic partner Roche have mounted a large-scale education and training effort around gauging HER2-low based on the best practices outlined by ASCO and CAP. "We understand [the challenge] because for the last 20 to 25 years, everybody was used to reading HER2 in one way," said Shalini Singh, chief medical partner of oncology at Roche Diagnostics. "Everybody knows HER2, but this is a little bit of a shift [in the scoring method], and that's why the training is there. With [the training], I'm pretty confident that it can be done."
The companies have taken several approaches to their training efforts, said Dan Switzer, head of Daiichi Sankyo's US oncology business division, using national commercial and medical field teams to conduct one-on-one engagements with oncologists; employing third-party training programs for pathologists; offering company-sponsored educational programs at larger oncology centers; and disseminating dozens of videos, resource guides, and case study walk-throughs of HER2-low cases by oncologists and pathologists through educational websites.
The goal of these programs is twofold: to raise awareness for the new category so pathologists know to more precisely score HER2 IHC results and to train pathologists on how to recognize HER2-low results.
Switzer said that one of the educational websites, HER2know.com, has reached more than 25,000 pathologists and that the companies are partnering with Roche to train pathologists on HER2-low scoring at the 25 largest labs in the US. The Roche Pathway HER2 assay is used by about 80 percent of the US lab market, he estimated.
"Those [25 high-volume] labs represent almost half of the market share in the country," Switzer said. "We're pretty fortunate the testing landscape is very concentrated so we can get to the high-volume labs relatively easily."
The two main pillars of the companies' HER2-low awareness efforts, according to Switzer, are encouraging oncologists to ask pathologists to reevaluate patient samples, even older samples previously deemed HER2 negative, in light of the HER2-low category; and encouraging pathologists to get second opinions if they're unsure about an IHC result.
Finally, the companies are also continuing to generate evidence supporting the HER2-low and ultralow categories that may be used to inform future clinical guidelines, said Omar Perez, head of medical diagnostics and US medical affairs for oncology at AstraZeneca. In December, Perez said that AstraZeneca researchers plan to report additional data on HER2-ultralow scoring.
However, before the field has a chance to become comfortable gauging HER2-low, the testing landscape may get even more complicated if AstraZeneca and Daiichi Sankyo gain FDA approval for Enhertu in the HER2-ultralow breast cancer setting. The FDA is expected to decide whether to approve Enhertu for treating HER2-ultralow advanced or metastatic breast cancer patients in the first quarter of 2025. The ultralow group would expand Enhertu's eligibility to tumors with an IHC 0 result with faint, incomplete membrane staining for HER2 expression in less than 10 percent of tumor cells, also defined as an IHC score greater than 0 and less than 1.
"If approved [in HER2-ultralow], it will require a higher recognition of the IHC 0 to be truly 0," Perez said. "That will necessitate the separation of what is IHC 0 versus IHC 0 with some membrane staining. That category will [need] to be split and recognized by pathologists."
Exploratory data from the HER2-ultralow subgroup who received Enhertu in the DESTINY-Breast06 trial was presented earlier this year and demonstrated similar efficacy to the HER2-low subgroup, which had a median progression-free survival of 13.2 months on Enhertu.
If approved, Switzer noted that Daiichi Sankyo plans to take a similar education approach for the ultralow category at launch. The firm plans to emphasize using the companion diagnostic, which Roche is also developing, to diagnose patients and distinguish between IHC 0 and IHC 1 tumors. "We feel we take a lot of responsibility and accountability to help pathology and oncology adopt these new practices as quickly as possible," Switzer said, acknowledging that the companies have heard from pathologists that distinguishing HER2-ultralows from IHC 0s "can be really challenging." The closer a sample is to "hugging that 0," the more important it is to get second opinions, he stressed.
In Perez's view, scientific advances in the HER2-directed ADC space are happening quickly, and it may take some time for guidelines bodies to update testing recommendations to reflect the ultralow category if Enhertu is approved in that setting. "It will require the separation of the IHC 0 category into two new categories, and given the fast pace of science, it does take some time to deploy [those changes], but hopefully with our partners, we'll be able to do that very quickly," Perez said.
Rimm is more skeptical about the ability to reliably measure HER2-ultralow as many studies of HER2 IHC scoring have shown the highest discordance between scores of IHC 0 and 1+. Even the validation study of Roche's Pathway anti-HER2 (4B5) IHC test approved with Enhertu's approval in HER2-low breast cancer showed that distinguishing between 0 and 1+ was more challenging than other scores. In that study, nearly 20 percent of samples were scored HER2-low in local labs but later determined to be IHC 0 by central testing, versus only around 3 percent of samples deemed HER2-low locally and later determined to be HER2-positive centrally.
"Pathologists are doing their best, we're not trying to make mistakes here," Rimm said. "We're just asked to do the impossible."
The testing uncertainty has spilled over into the oncologists' workflow and challenging how they determine which patients to give Enhertu. ASCO and CAP recommended considering HER2 IHC results from multiple tissue samples, including previously collected samples, multiple primary tumor samples, or metastatic tissue samples. The organizations noted there may be heterogeneity in HER2 expression levels within these samples.
Dana-Farber's Tarantino noted that there is a lot of confusion around HER2-low as pathologists adapt to reading these low IHC scores. Amid this confusion and in light of the latest guidelines, oncologists are now looking for HER2 expression in any patient biopsy, new or old.
"The way patients were enrolled in the trials was that they were required to have one biopsy to test as HER2-low, but in our case, our patients often have several biopsies in the past, some of which are HER2-low, some of which are IHC 0, " Tarantino said. "There is a lot of discordance in this, and so usually we accept any biopsy that has proven to be HER2-low to use [Enhertu], which is a very broad usage."
At Dana-Farber, he noted that their labs are beginning to report whether samples are IHC 0 with some membrane staining in anticipation of the potential HER2-ultralow approval for Enhertu. Most labs are not doing this yet, however, in the absence of Enhertu's approval in this setting or guidelines supporting it, he noted.
Given the rapidly changing parameters for patients' eligibility for Enhertu, Tarantino said he's taking an "open-minded" approach to prescribing the drug. "If we find any sign of HER2 expression somewhere, we use the drug [Enhertu]," he said. "If we don't find it, we try to do a new biopsy to see if we can find it."
What's next for HER2 breast cancer management?
While AstraZeneca and Daiichi Sankyo used IHC to stratify patients in Enhertu trials, the companies are keeping an eye on advances in digital pathology and even new assays that may be used to refine patient selection.
"For the most patients to benefit from this targeted agent right now, we wanted to leverage what's currently available in practice and used, and it is well established in clinical practice that IHC, particularly in breast cancer, is widely utilized," Perez said. "That's not to say that other modalities can't be introduced; however, with any new technology, there is an adoption period."
Perez highlighted various tools that AstraZeneca's translational research teams have explored for refining the measurement of HER2 expression, such as a deep learning-based quantitative continuous scoring (QCS) method for analyzing digital whole-slide images that are IHC stained. A retrospective analysis by AstraZeneca and Daiichi Sankyo researchers of breast cancer samples from a Phase I trial of Enhertu with the QCS method found it was more precise than IHC for selecting the patients who responded to treatment with Enhertu.
He added that AstraZeneca has ongoing collaborations evaluating digital pathology methods for optimizing HER2-low and HER2-ultralow detection and that the company hopes to report data from this research at upcoming meetings.
Rimm is also advancing a new, more sensitive assay for measuring HER2 expression.The assay was designed specifically to measure low HER2 expression in a quantitative manner using mass spectrometry methods to standardize and calibrate immunofluorescence-based measurement. It uses quantitative immunofluorescence methods to highlight the tumor cells expressing the HER2 protein. Rather than trying to count cells or visually inspect the sample under a microscope like in IHC, pathologists only need to indicate the boundary of tumor tissues and normal tissue, then the test uses mass spectrometry to measure HER2 expression in the tumor sample.
In the validation study for Rimm's assay, his team determined the low HER2 expression range, between 2 and 20 attomols per square millimeter, and used this threshold to measure HER2 expression of 364 breast cancer cases at Yale Cancer Center. They found that 67 percent of cases fell within that low HER2 range.
Rimm's lab has partnered with Danaher subsidiaries Leica Biosystems and Cepheid to conduct clinical trials of the assay and standardize its use. Rimm currently offers the assay as a laboratory-developed test out of the CLIA lab at Yale but can only conduct about 50 tests per month. He hopes the partnership with Danaher will enable more testing with the assay to meet increasing demand.
Rimm's lab is just one of many exploring alternative methods for measuring HER2 expression. Researchers reported this year that a liquid biopsy circulating tumor DNA assay had 89.6 percent sensitivity for determining HER2-ultralow breast cancer samples. Researchers in the Netherlands evaluated a HER2 PET imaging method to identify HER2-low expressers or those who had whole-body HER2 heterogeneity. Apis Assay Technologies reported results from a study of its Breast Cancer Subtyping Kit to detect mRNA expression of standard biomarkers (ERBB2, estrogen receptor, progesterone receptor, and Ki67) as a method to identify patients with low HER2 expression.
"There are methods of HER2 quantification that are going to happen, whether it's my method or some other method," Rimm said. "I agree that IHC is a lot easier, but if it doesn't work or if people ultimately prove that being quantitative is more important, then quantitative [testing] will happen. It probably won't happen at every lab, but rather it will occur in the big central labs that have the ability to take on expensive machinery and take on expensive validation processes in order to give a better result."
The future of the HER2 biomarker and HER2-targeted treatment will also undoubtedly be affected by the indications AstraZeneca and Daiichi Sankyo advance Enhertu into next. The firms have begun studying the drug in hormone receptor-positive, HER2-negative unresectable or metastatic breast cancer patients in the Phase III DESTINY-Breast15 trial, which has two cohorts of HER2-low and HER2-negative, or IHC 0, breast cancer patients.
Tarantino suggested that with the current data on Enhertu showing activity in each of these HER2 expression categories, it may be the case that almost all breast cancer patients could benefit from this drug. "Probably you can see a benefit with this drug irrespective of HER2 expression," he said. "To some extent, the HER2 expression can further boost the activity of [Enhertu], but even in the absence of it, you can have activity."
The companies have not disclosed when they expect to report the first data from the DESTINY-Breast15 trial and the HER2 IHC 0 cohort, but the trial is recruiting patients worldwide and has an estimated completion date in 2027. "We also know that even IHC 0 patients have minuscule levels of HER2 expression, so it truly is not binary, and that's the rationale of going after the IHC 0s," Switzer explained. "Potentially, this drug is so sensitive to the HER2 protein that all it takes are these tiny amounts of HER2 for the drug to attack the cancer cells with its payload."
Tarantino was also part of research that aimed to characterize whether HER2-low was a distinct subtype of breast cancer with different clinical features from HER2-negative disease. After exploring clinical pathologic features, genomic profiles, and the prognosis of HER2-low versus HER2-negative breast cancer patients, his team found no major differences to define a new subtype.
"As soon as there is a new drug for this new category called HER2-low breast cancer, everybody starts thinking that HER2-low is something distinct or a subtype of breast cancer, " Tarantino said. "What was apparent is that it's basically a continuum of HER2 expression and without striking differences among those that express low HER2 or zero HER2. This is something that helps us to consider treatment with this drug in HER2 0 patients."