DESTINY-Breast 04 and 06: A Pathologist-Friendly Recap

Let us get started. The first part of this, program will be done by me and Dr Tolaney. Then, Dr Quintana will do the second part and then there will be a faculty discussion, and then we will move to the case discussions. Lastly, we will take the question and answer sessions.

As you all know, the first part is we are going to talk about DESTINY-Breast04/06: A Pathologist-Friendly Recap. As you all know, we are pathologists, but we need to understand the impact of the clinical trials that drives biomarker evaluations.

[00:06:25]

HER2 as a Biomarker in Breast Cancer

Let us just know about HER2. As you all know, to this audience, and this is just a recap, HER2 is a biomarker in breast cancer. We all know HER2 overexpression. HER2 gene amplification is seen in 15% to 20% of invasive breast cancer cases. HER2 expression high. HER2 expression is associated with high histologic grade, aneuploidy, p53 mutations, PI3 kinase, AKT pathway, RAS/REF/MEK, and MAP kinase pathway activation.

It is well established that HER2 is an important prognostic and predictive biomarker of response to HER2-targeted therapies.

In today's world, the targeted therapy to HER2-positive breast cancer, there are very many. Monoclonal antibodies like trastuzumab, pertuzumab, and others; and the tyrosine kinase inhibitors like lapatinib, neratinib, and tucatinib; as well as antibody drug conjugates like T-DM1 and T-DXd.

[00:07:27]

DESTINY Trials

The DESTINY trials are a series of international trials investigating trastuzumab deruxtecan, an antibody drug conjugate, as a treatment option for eligible patients with breast, gastric, lung, and additional solid cancer types.

There are 2 trials that really drives the significance of HER2 with low levels of expression in breast cancer. These are DESTINY-Breast04 and DESTINY-Breast06. Both these trials are phase III clinical trials in patients with metastatic breast cancer. Patients received the antibody drug conjugate trastuzumab deruxtecan vs physician's choice of chemotherapy.

[00:08:10]

Trastuzumab Deruxtecan HER2-Targeted ADC

Just a brief explanation about antibody drug conjugates. This is a very interesting combination, where trastuzumab, which is having the same anti–amino acid sequences like the Herceptin, it is a humanized, IgG1 monoclonal antibody. What happens is it is tagged with a tetrapeptide-based linker that links the trastuzumab with deruxtecan. Deruxtecan is the payload here, and it is an exatecan derivative. It is a topoisomerase 1 inhibitor. It is membrane permeable, has a very short systemic half-life, and it causes a bystander killing effect of all the tumor cells that is around the cell where it goes and tags.

I think, Dr Tolaney will talk about the 2 trials.

[00:09:09]

DESTINY-Breast04: Phase III Study of T-DXd vs CT for HER2-Low MBC

Dr Sara Tolaney (Susan F. Smith Center for Women's Cancers): Thank you very much. DESTINY-Breast04 was a registration study in patients who have metastatic breast cancer. Predominantly, it was designed to look at patients who have hormone receptor–positive disease that had received 1 to 2 prior lines of chemotherapy for their metastatic treatment. Again, remember we had a question at the beginning about whether or not patients were pretreated or not when they got T-DXd. These patients had all progressed if they had ER-positive disease through their hormone therapies and then had already had at least 1 line of chemotherapy in the metastatic setting.

They randomized patients to get T-DXd or treatment of physicians' choice chemotherapy. They did actually include, however, a very small subset of 58 patients who had triple-negative breast cancer as well. While the primary endpoint of the trial was designed to look at the hormone receptor–positive patients, in the ITT evaluation, it did include the triple-negative subset of patients.

[00:10:06]

DESTINY-Breast04: PFS

When you look at the outcomes, you can see an essence a doubling of progression-free survival with T-DXd compared to chemo. Clearly, the T-DXd is performing much better than standard chemotherapy. This was true not just in terms of progression-free survival, but in fact also true of overall survival.

When you look at the figure to the right that is all the patients, so the ITT population. Remember that does add in now those 58 triple-negative patients into the evaluation and you are still seeing, again, this very nice improvement in progression-free survival. This is the overall survival data in the hormone receptor–positive patients, again, showing a significant improvement in OS and really solidifying this as a new standard of care when this data was reported.

These patients had to have HER2-low disease, so 1+ or 2+ and not FISH amplified. Again, they had to have had at least 1 prior line of chemotherapy.

[00:11:05]

DESTINY-Breast04: Conclusions

These data again, do suggest that T-DXd is standard of care in someone who had HER2 low metastatic breast cancer that had received at least 1 line of chemo. Interestingly, the FDA did grant approval not just in the hormone receptor–positive patients, but also in the triple-negative patients. Even though it was a very small subset, the PFS deltas were very similar in the triple-negative subgroup relative to the hormone receptor–positive group. They felt that really the benefit was seen irrespective of subtype.

[00:11:36]

DESTINY-Breast06: Study Design

The question then was, well, if T-DXd is a standard in pretreated patients, what about moving it into patients who had not had any chemotherapy? They could have had their hormone therapies, but they could not have had any chemotherapy. That was one question, could we give it earlier?

The other question is how low can you go with the HER2 status? Could it even work in patients who had HER2-ultralow disease? That is why DESTINY-Breast06 was designed, was really to address those 2 questions.

This trial took hormone receptor positive patients. This trial did not allow triple negative as opposed to DESTINY-Breast04. It was restricted to the hormone receptor–positive subset and you had to have progressed on endocrine therapy, either 1 or 2 lines of endocrine treatment and not have received any chemotherapy for your metastatic hormone receptor–positive disease. Then you are randomized to get T-DXd or treatment of physician's choice chemo. They again, had patients who had HER2-low disease, but they also included a subset of patients about 150 patients that were ultralow, meaning they had some membrane staining but if they were IHC 0, they had to have some membrane staining to be allowed. This was also upon central review.

[00:12:53]

DESTINY-Breast06: Baseline Characteristics

This is just the baseline characteristics of the patient population. Again, remember these are all ER-positive patients as opposed to DESTINY-Breast 04. Patients had either received 1 or 2 lines of endocrine therapy in the metastatic setting. They are not treatment naive. Again, they have all had some form of endocrine therapy, they have just not had prior chemo.

[00:13:13]

DESTINY-Breast06: Efficacy

What you see in the primary endpoint was to look at the PFS in the HER2 low, but then also to look at the ITT population. When you look at the HER2-low population, you see a significant improvement in progression-free survival, similar to the ITT. In fact, what was really intriguing was that if you look at the ultralow patients, you see almost identical numbers in terms of the progression-free survival. Really suggesting that the benefit of T-DXd was seen both in the HER2 low as well as in the ultralow subsets.

[00:13:47]

DESTINY-Breast06: Summary

We do not have mature overall survival data yet from DESTINY-Breast06. That is still pending. Overall, I think these data suggests that T-DXd should be standard, not just in low, but also in ultralow. Again, it is now standard of care. In the US, it did get FDA approval in this population. Again, just to remember, in hormone receptor–positive patients, people have to have progressed on endocrine therapy, but then we can use this as their first chemotherapy if they are either ultralow or low.

I will pass it back to get a summary.

[00:14:24]

DESTINY-Breast04 and DESTINY-Breast06

Dr Krishnamurthy: Thank you. Basically, the DESTINY-Breast04, DESTINY-Breast06 trial results made the FDA approved T-DXd for the treatment of hormone receptor–positive HER2 low or HER2-ultralow unresectable or metastatic breast cancer.

[00:14:44]

Posttest 1

Let us do the posttest question 1. Which of the following patients with unresectable or metastatic breast cancer would qualify for treatment with T-DXd? You have the 4 choices.

Very good. You can see your responses here. Obviously, here is the answer. Patient with newly diagnosed, previously untreated hormone receptor–positive, HER2-low disease. The rationale is that T-DXd is currently indicated for hormone receptor–positive, HER2-low or HER2-ultralow unresectable or metastatic breast cancer.

[00:15:45]

HER2 Expression Today

Let us move on now. As you all understand, HER2 expression today is from 2018 when the ASCO/CAP guideline came. We all know that HER2 scoring was done as score 0, 1+, 2+, and 3+. Today, with what we understand, you can divide score 0 into 2 types. Those with absolutely no staining referred to as HER2 null, and then HER2 with some faint incomplete membrane staining in less than or equal to 10% of tumor cells referred to as HER2 ultralow. Score 1+, which is incomplete membrane staining, faint or barely perceptible in more than 10% of the tumor cells. HER2 2+, which is moderate intensity, complete membrane staining in more than 10%. Score 3+, complete membrane staining of strong intensity in more than 10 plus.

[00:16:43]

The Spectrum of HER2 Expression: Evolving HER2 Classification and Reporting in Breast Cancer

Obviously, the next part of the discussion will be understanding the spectrum of HER2 expression and the evolving HER2 classification and reporting, which will be presented by Dr Quintana.

[00:16:58]

Dr Liza Quintana (Beth Israel Deaconess Medical Center): Good morning, everyone. Thanks for coming so early, and I hope you are having a chance to explore our beautiful city here in Boston. I will be discussing The Spectrum of HER2 Expression, and the Evolving Classification and Reporting in Breast Cancer.

[00:17:16]

Pretest 2

Here is our pretest. Which of the following is recommended to optimally assess HER2 status to identify patients with metastatic breast cancer who may benefit from HER2-targeted therapy? All right. Excellent.

[00:18:01]

Pretest 3

How confident are you in identifying HER2 low/ultralow breast cancer? All right. All across the board. I agree it is quite challenging.

[00:18:25]

HER2 in Breast Cancer

HER2 in breast cancer. It is a transmembrane glycoprotein receptor with tyrosine kinase activity. When it is activated, it affects cell proliferation and survival. HER2 is present on the surface of normal cells, but increased in HER2-amplified or HER2-positive tumor cells.

[00:18:47]

HER2 Signaling in Breast Cancer

Dimerization of HER2 causes downstream signaling cascade to promote tumorigenesis, so cell proliferation, and inhibition of apoptosis.

[00:19:02]

HER2 Alterations

HER2 alterations. We see HER2 amplification or HER2 overexpression, so protein overexpression. That has long dictated HER2-targeted therapy, eligibility for patients previously HER2-negative tumors. IHC 0, 1+, or 2+ with FISH negative were ineligible for anti-HER2-targeted therapies. We have a new paradigm. HER2-low and HER2-ultralow have emerged as distinct categories responding to novel, antibody drug conjugates, like T-DXd as was discussed in the prior talk.

[00:19:38]

ASCO/CAP Guidelines: Standardization of HER2 Testing

The ASCO/CAP. A little bit of a history of HER2 standardization and guidelines. Highly recommend if you have not already seen the amazing Maude Abbott Lecture from Saturday by Dr Kim Allison, who went into more of the history of HER2 delving more into this. This is just a brief overview.

In 1998, trastuzumab received FDA approval in the metastatic setting, along with companion diagnostic test, HercepTest. In 2007, just after trastuzumab received approval in the adjuvant setting, we had the first ASCO/CAP guidelines, which helped to set the standards for IHC and ISH testing and really address a lot of the false positive rates.

Over time, we have had updates. The most recent one was in 2023, and this was to address the HER2-low because of the DESTINY trial, that really impacted treatment of patients, who had previously not been able to qualify for HER2-targeted therapy.

In this update, there were no changes to reporting, although that has changed a little bit recently, uses the same scoring system that we had previously. We will go into that a little bit more.

[00:20:58]

Evaluation of HER2 Expression by IHC

Evaluation of HER2 expression by IHC. This is from the ASCO/CAP guidelines looking at the difference how to interpret the staining. As Dr Krishnamurthy went over the different percentages and intensities of staining that we want to see in the HER2 IHC test to categorize them as 0, 1+, 2+, or 3+. Generally, on either end of the spectrum, as with most things, it is easy to recognize. Generally, the 3+ positive, you have a dark brown slide, easy to recognize. If there is no staining at all, also easy to recognize but the challenge comes in the middle.

[00:21:38]

HER2 IHC Assays

HER2 IHC assays. There are multiple different assays. The most widely used is the Ventana PATHWAY 4B5, and the HercepTest. The 4B5 assay was the one that was used for the DESTINY04 and 06 trial eligibility. These companion diagnostic tests were really optimized for high levels of HER2 expression by IHC. They were not developed to have this sensitivity or range to detect low levels of HER2, which is what we are currently being faced with and why this has been such a hot topic for the past few years.

Previously, when we scored and reported negative, we could just lump 0 and 1+ together, which is what we did at our institution. Not everybody did that, and we did need to make that distinction, but now we do.

[00:22:20]

Current HER2 Classifications

Current HER2 classifications. HER2 low as mentioned, IHC 1+ or 2+ with ISH negativity. That is in about half of breast cancers. HER2 ultralow is 0 with some membrane standings. You will see it represented as 0+ in the NCCN guidelines and the new biomarker reporting guidelines that the CAP just came out with last week. That is seen in about 3 of 5 advanced cases that were originally classified as IHC 0.

HER2 status is no longer binary. This really requires us to refine our classification methods, an ability to detect HER2 at low levels.

[00:23:03]

HER2-Low and HER2-Ultralow

Here we have some images depicting what classifies as low and ultralow. On the left HER2 low, so 1+ or 2+ with the ISH negative. On the right, this is the HER2 ultralow or 0+. It did not have any images of HER2 null, which is no staining at all.

I have to say, it is even difficult to capture images that try to represent this because at these low levels of staining, there can be a little bit of variability across the slide. When we are interpreting these, you really want to look at the whole slide to make sure that your percentage is as accurate as possible. Again, even just getting pictures to demonstrate this can be challenging.

[00:23:49]

Variables Affecting HER2 IHC Results

There are a lot of variables that affect HER2 IHC results. Things that we are all very aware of. Different pre-analytic variables. We really want to be concerned about the cold ischemic time, the type of fixative that we are using, the time of fixation, how we are processing the tissue, storing slides and blocks. Thinking about the biopsy vs the resection. There can be tumor heterogeneity, if patients have been treated. Other analytic variables.

The antibody selection. As I mentioned, there are multiple different antibodies. The protocol, so we want to make sure we are carefully following the protocol, and also the interpretation criteria. Using the ASCO/CAP guidelines. Ideally, we will have controls on the slide if possible, but definitely want to make sure you check your controls and have controls with low levels of HER2 expression.

Then reviewer experience, so just the pathologist expertise and getting a second review for equivocal cases. I do not mean equivocal as in 2+, but anytime you are on that border of 2 different categorizations consider showing a friend.

[00:25:02]

HercepTest vs Ventana PATHWAY

The HercepTest vs the Ventana PATHWAY. Different assays. This study by Ruschoff et al showed high concordance for HER2-positive and HER2-negative breast carcinoma. Again, either under the spectrum good concordance, but a complete agreement for these semi quantitative scores was definitely lower at about 70%.

In this study, HercepTest detected HER2 expression with higher sensitivity in tumors, especially those with gene amplification. HercepTest was more sensitive for HER2-low than the Ventana 4B5 assay. More HER2-low with the HercepTest as compared to the Ventana. Although other studies have shown the opposite depending on which HercepTest you are using. The ISIMM Companion Society Meeting on Sunday, Dr Troxell went over in more detail the different assays that we use. Very interesting and helpful discussion there.

[00:26:05]

Challenges With HER2 Assays

Challenges with the HER2 assay in the study by Fernandez et al. They looked at the CAP survey and found that, 0 vs 1+, 19% of cases had less than 70% of concordance. Not great. They also surveyed 18 experienced pathologists to see how they scored 0 vs 1+ and 2+ vs 3+. Zero vs 1 had only a 26% concordance, 2 vs 3 had a 58% concordance. Again, either end of the spectrum we tend to do pretty well, as you can see, on the right, but when we are really parsing out the different levels, it is a bit more challenging.

Of note, these pathologists were not made aware that there was going to be a distinction between 0 and 1+ in the results. Something that they had mentioned that they would potentially take more caution and review the slide a little bit more carefully, had they known.

[00:27:10]

HER2 IHC Rescores

HER2 IHC Rescores. This is study presented at San Antonio Breast Conference by Dr Krishnamurthy in December, showing that interobserver concordance was moderate, whereas intraobserver concordance was quite high. They looked at 400 metastatic breast cancer tissue samples that originally scored as 0. Some of those got rescored as either ultralow and about a third as low.

[00:27:40]

HER2 Status Can Change Over Time

We know that HER2 status can change over time. This study looked at 457 patients and they found HER2-low vs 0, so about 38% discordance from the primary to the metastasis. It was almost even so 0 to low and low to 0. Whereas this study by Tarantino and colleagues found that the metastatic setting was a bit more enriched for the HER2-low.

[00:28:12]

HER2 Retesting

HER2 retesting. When do we want to retest? The ASCO/CAP guidelines provide us some suggestions for that, as you can see here. If you have got equivocal or discordant results, if it is a really small or non-represented biopsy, if there were issues with fixation or staining. One thing to consider is that we can see heterogeneity between samples.

As a cytopathologist, I diagnose a lot of metastatic breast carcinoma, and we have found that in our samples we fix everything in CytoLyt and then do cell blocks that are fixed in formalin and paraffin-embedded. We have found that there is less staining of HER2 in cytology samples as compared to either the primary or concurrent samples.

This is an example here of a patient who had HER2-negative primary breast carcinoma, with widely metastatic disease. She had a femoral head biopsy that was HER2 1+ and a lymph node biopsy, that was HER2 null. Both of these samples were formalin-fixed paraffin-embedded. It is showing that spatial and temporal heterogeneity that you can see in HER2 expression.

[00:29:24]

ASCO/CAP: Distinguishing 1+ From 0

Distinguishing 0 from 1+. Want to look at the whole slide and use the scoring criteria. High-power objective lens, because the staining can be quite faint. Show a friend if you are close to different thresholds, because this really impacts treatment. We want to use controls with a range of expression. We want to pay attention to pre-analytic conditions of samples. Again, challenge with metastases, as I mentioned. One issue with cytology. Other issues can be getting specimens decalcified, if it is a bone met.

[00:30:04]

Global Analysis of HER2-Low Diagnosis in BC

This study looked at a global analysis of HER2-low diagnosis in breast cancer. There was a set of 77 pathologists from 14 different countries who scored HER2. Then they were trained by experts and then asked to rescore and they had a significant improvement on concordance after the training just highlighting how challenging this can be.

Here are some of the challenges and recommendations they found if you are underscoring or overs scoring, the pictures on the right; at the top are examples of cases that were underscored, and on the bottom, those that were overscored.

[00:30:44]

Augmented Interpretation of HER2, ER, and PgR in BC

Perhaps AI or augmented intelligence can help us with interpretation of HER2 ER and PR in breast cancers. This study used an AI-powered analyzer that can help achieve consistent results. They found that the agreement with HER2 scoring really improved with AI assistance. I know there are some presentations this week that are also looking at AI, Dr Krishnamurthy, and her team has one.

[00:31:17]

HER2 Reporting

New reporting. This just came out last week. Probably many of us found out about it at the Maude Abbott Lecture. We are going to have to parse out 0+ or ultralow now for HER2. Definitely recommend checking out the new CAP guidelines, and including these scores in your testing.

[00:31:41]

 
Posttest 2

Posttest. Which of the following is recommended to optimally assess HER2 status to identify patients with metastatic breast cancer who may benefit from HER2-targeted therapy?

All right, excellent.

[00:32:19]    

Posttest 2: Rationale

Okay. For this, the ASCO/CAP guidelines recommend using a high-powered objective lens 40x to confirm staining. You really want to exclude areas of granular staining, cytoplasmic staining, really pay attention to pre-analytic conditions and use this high-powered objective lens. There are no FDA approved AI tools currently to help us evaluate but, hopefully that will be coming.

[00:32:53]

Poll 4

Which of these patients with breast cancer would be eligible for HER2-targeted therapy?

Again, it is difficult to capture these images when reviewing the slides.

[00:33:28]

Conclusions

In conclusion, HER2 spectrum is no longer binary. HER2-low and ultralow do require refined diagnostic accuracy. There is variability in results due to factors including interpretation and the assay used, and different pre-analytic and analytic factors. Ideally, we would have improved diagnostic tools. We will see what is to come and we want to report per the newly published CAP guidelines that came out last week, and perhaps future directions include AI or artificial augmented intelligence.

[00:34:06]

Faculty Discussion

Dr Krishnamurthy: Thank you, Liza for that overview of the changing landscape of HER2 expression. Let us have a little discussion now. Liza, you want to summarize for people the key pathology related issues of HER2 testing in breast cancer?

Dr Quintana: I think, it is just challenging to interpret. I think everyone has slightly different thresholds. Things that I have found helpful is there are websites online that have HER2 that has been scored by experts. You can really try to compare your sample with others.

Then as I mentioned, just some of the analytic variables. As a cytopathologist, again, if I have the opportunity to stain a metastatic sample, I try to pick a sample, like a surgical pathology biopsy vs a cytology sample knowing that the HER2 expression can be lower and at least our samples that are alcohol-fixed.

Dr Krishnamurthy: Very good. Systematically if you think of it and so there are 3 things that we all have to remember, pre-analytic factors. The CAP clearly gives you the guideline that the cold ischemic time should be less than 1 hour. Putting the tissue as quickly as possible in formalin. I know it is a challenge for surgical resections compared to biopsies.

Then fixation in formalin between 6 and 72 hours. That is another thing that we all have to pay attention. Even in my practice, I think, the surgical resection specimen, fixing it in formalin in time becomes a challenge, particularly when you do lot of intraoperative evaluations like how we do in our practice.

Pre-analytic is extremely important. When it comes to bone biopsies, I think we all do not like very harsh decalcification of the bone specimens, which is going to decrease the staining. Formic acid-based flow, decalcification, or EDTA-based decalcification is what is preferred for bone specimens.

When a patient comes with a metastatic breast cancer, the current ASCO/CAP guidelines are that we do biopsy the metastatic tumor and do the standard biomarker testing.

With respect to analytic factors, I think as you all know, the DESTINY04 and 06 used the 4B5 antibody. I think for our testing, that is what is recommended either the companion diagnostic kit. If that is not done, at least have the 4B5 antibody validated as a lab developed test in your lab. That is one thing.

The second is, the last that Liza mentioned, interpretation. You all know the CAP guidelines. The 2023 update clearly said that there are certain best practice recommendations to identify HER2-low, ultralow well in your practice. What are those? We all know that we have to evaluate the tissue at 40x magnification.

I also feel that, you do end up 20x, but particularly you have to go under 40x to really determine whether it is truly ultralow or low. Then we always remind the pathologist to be very careful to avoid pitfalls.

What are the pitfalls? Cytoplasmic staining with a little accentuation near the membrane is not really membrane staining. That is one thing to remember. Also edge artifacts or staining near necrosis really are artifactual staining and not real. I think identifying true incomplete membrane staining is the key. When in doubt, it is good to take your opinion from your colleagues. I think that is all is available right now.

We all know all the issues of identification for a pathologist, but right now immunohistochemistry is the standard and we all have to live with it until we learn more. I think there is DESTINY-Breast15 trial that will tell us whether patients with even 0 is going to respond. I think we have to wait for that. Until then, I think we all have to live with this and so we have to do our best to identify these patients who can be selected for therapy.

I want to ask Dr Tolaney some questions as to, what does she expect from the pathologist and what she desires she gets from the pathologist and any tips and advice for all of us?

Dr Tolaney: Thank you. I will say thank you to our pathology colleagues because we would not be able to figure out what to do without you all. I will say we have been calling on you a lot these days. I think our group a few months ago after DESTINY-Breast06 had been reported had switched the reporting structure so that our pathologists were reporting null, ultralow, 1+, 2+, 3+ sustaining. That was particularly helpful because we needed to understand who had ultralow. Obviously, prior to this, we were not reporting that.

The challenge though is that we have patients who have older biopsies that were reported prior to this new system. We are going back to our pathologist a lot and saying, "Hey, we have this patient who has had every biopsy be IHC 0, but can you tell me are they ultralow?" They have to then go back and be willing to do that on an older specimen and let us know if it actually has any HER2 staining or not. It is causing a lot of work we realize because these are specimens that have already been read and now need to get reread.

We are also biopsying patients more. I do talk to the pathologist and they say, "Hey, it is really null, there is no staining." We will re-biopsy and see if now maybe they are ultralow. I think as was pointed out, patients have heterogeneous disease. They also have disease that is different across different sites in their body. We realize that there is a little bit of imprecision with all of this because we are only biopsying 1 site. Obviously, we are not really testing all the locations.

I think you are right that things will shift if DESTINY-Breast15 is positive because then all of this becomes not an issue any longer. Although I will say I think in the future I do wonder if quantitative assays are going to be needed because I do not think the benefits are probably quite as large in someone who has very minimal staining compared to someone who has much stronger staining. Maybe in the future we are going to have to switch to AI technologies and really get a quantitative look because we are going to have so many drug choices and we are going to need to figure out what makes sense. For now, we appreciate you all being willing to go back and re-look at slides.

Dr Krishnamurthy: Thank you. Also I think, even in this meeting you may have heard what are the alternate ways of evaluating HER2 at the lower end of the spectrum? There is getting the RNA levels of the HER2 expression in the tumor cells. There was a very good platform presentation on that.

In addition, there is mass spectrometry-based techniques to evaluate protein. There are several other, I think, evolving research technologies for evaluating HER2 at the lower end of the spectrum. All of those are very research methodologies and not ready for prime time, not ready for standard of care.

For standard of care, it is still the HER2 immunohistochemistry. All laboratories I think are encouraged to use the 4B5 antibody. When you use the controls on the slide, it is good to use, not just the high 3+ and a 0, and a 2+, but also to include 1+ in your controls when you are doing the staining.

Other new technologies like the AI-based decision support tools, there are several available right now. Interestingly, AI tools do help in bringing the consistency and improving the interobserver reading of HER2. Almost all the AI tools show the data.

Remember, the AI tool is not a magic. What happens is you do improve the interobserver concordance, but it does not mean that it becomes perfect with an AI tool because obviously you are training the tool. It is us training the tool and then when we use it, it is good, it makes it much better. All the data from the AI tools clearly tells you that if you used it vs if you did not use it, there is a significant improvement.

The issue with AI tools is that there are several available currently. There is no idea about which one is better, how to use it, and you know that the issue, the challenge of digital pathology is that not all laboratories have digital pathology. I think that is an evolving space right now.

Stay tuned to that. I think there will be more decision support tools for us. Right now we all have to follow the criteria of ASCO/CAP, the 2023 update. Very recently now there is the CAP biomarker template, some modifications to capture the 0 ultralow in your report. I think, when we do the CAP template now, they clearly say that you have to categorize your 0 into the null, which is like no staining and the 0 with some membrane staining in less than or equal to 10% of tumor cells. In fact, the 0+ is also being used for that category.

Okay, so we are doing well with time.

With that, now we move to case discussions. We have 2 cases for you and Dr Tolaney will go over the cases.

[00:44:31]

Case 1: Patient With 2-Yr History of Palpable Breast Mass

Dr Tolaney: Thank you. We will try to put this into context. This is a 72-year-old woman who had presented with an enlarged breast mass and she had staging scans when this was found and did unfortunately have multiple sites that looked like metastatic disease, including several enlarged lymph nodes, and what appeared to be bone metastases.

The breast mass was biopsied and came back as an invasive ductal carcinoma that was hormone receptor–positive and HER2 IHC 1+. Given that she had metastatic hormone receptor–positive disease, she was started on endocrine therapy with a CDK4/6 inhibitor and was on this for a little over a year.

However, her disease then did progress and she developed new liver metastases. We had looked for potential mutations and did see that she had an ESR1 mutation. She also was found to not have a PI3 kinase mutation. It was decided to treat her with fulvestrant and everolimus. However, that did not work very long and after 4 months had disease progression.

In essence, this is a patient who has had 2 lines of endocrine therapy for metastatic disease that is known to be HER2-low. She had progressed again on the endocrine treatments with worsening lung, bone, and liver lesions and had normal liver function testing.

[00:45:57]

Poll 5

Now the question is what are you going to do in terms of therapy? Here you can see the options include:

1. A couple standard chemotherapy agents such as capecitabine; or
2. Eribulin;
3. Another choice being sacituzumab govitecan, a TROP2-directed ADC; or
4. Trastuzumab deruxtecan.

I think people got the theme here that this patient again did have HER2-low disease. Given now data from DESTINY-Breast06, you can use T-DXd as your first-line of chemotherapy in someone who has progressed on endocrine treatment.

[00:46:47]

Case 2: Patient With T2N1 ER+/HER2 1+ Breast Cancer

I will say that it does not mean we always do. Sometimes we will use oral capecitabine or capecitabine because patients do not lose their hair. It is a pill. With T-DXd people are coming in for IV infusions. They do have about a 30% to 40% rate of hair loss, and there is about a 10% to 15% risk of interstitial lung disease. It is a discussion with the patient, even though T-DXd is clearly much more efficacious than standard chemo. These are patients who have multiple therapies ahead of them.

We do not really know that if you sequence one before the other that it will impact survival. We are still waiting for survival data from DESTINY-Breast06 to know if you really have to give it first vs second.

I think most of the time we will give it first, but again, we do take patient preferences obviously into account here because some patients are very concerned about the risk of hair loss or just find it too difficult to come in for infusions or maybe they have pulmonary contraindications. Maybe they have a history of prior pneumonitis, and they are not even a candidate for T-DXd. All of that stuff has to be taken into account.

Here is a another case. This is a 54-year-old woman who had a prior history of a breast cancer that did have nodal positivity. Her original breast primary was ER-positive and HER2-low, so IHC 1+. She had gotten adjuvant systemic chemotherapy and an adjuvant endocrine treatment. While she was on her endocrine therapy, she unfortunately developed right upper quadrant pain and was found to have liver metastases.

Her liver was biopsied and did come back having some hormone receptor staining. So ER 20%, PR-negative, and HER2 was IHC 1+. Now again, she has metastatic disease that arose on an aromatase inhibitor. She was treated with fulvestrant and a CDK4/6 inhibitor. She was on that not for a very long period of time, unfortunately. Just after 5 months of treatment did develop a disease progression.

[00:48:50]

Poll 6

What would be your next appropriate step for therapy? Here, we have listed some choices.

A. Elacestrant is an oral SERD;

B. Olaparib, a PARP inhibitor;

C. Datopotamab deruxtecan, a TROP2-directed ADC;

D. Sacituzumab govitecan, also a TROP2-directed ADC; or

E. T-DXd.

Everyone is also sticking with the T-DXd theme. Again here she did have HER2-low disease and so you could think about using T-DXd here again. We do use TROP2-directed ADCs as well. The preference generally is to use if they have any HER2 expression, T-DXd first, and then consider a TROP2-directed ADC in the future.

I think the challenge is sequencing 1 ADC after the other is something where it does not always work because if you could imagine you have just given someone T-DXd that has a deruxtecan payload, a Topo 1 payload, and if they developed drug resistance due to having payload resistance, meaning really resistance to the chemotherapy component of the ADC, one would think if you are going to give them a similar or the same payload with the subsequent ADC, the chances of it working may be less. That does seem to be the case. We do see that progression-free survival is generally shorter with the second ADC, when given in sequence, but not always.

We do have some patients who, for example, may not respond to the first ADC and then have a home run response to the second. We are still trying to understand ADC sequencing a little bit better and better understand resistance mechanisms.

Here the first ADC that you could give should be T-DXd if they have any HER2 staining.

[00:50:46]

Posttest 3

We will turn to our posttest question. How confident are you now in identifying HER2-low and ultralow disease?

[00:51:18]

Question and Answers

Dr Krishnamurthy: Very good. Okay. We are coming to the last part of this program. We will take some audience questions now, but I just want to remind you all to complete the evaluations to get your credits. First, we will take some question and-answers now. Any questions here?

Dr Tolaney: There is one of the audience.

Dr Krishnamurthy: Yes.

Speaker: I have a question about [inaudible]. Thank you. Avoid the edge entirely. How are you managing that in your practice?

Dr Krishnamurthy: Basically, we are not using AI tools for standard of care. It is still the manual evaluation. If there is staining only at the edge, you have to discount it. You cannot use that in your scoring. It has to be in the bulk of the tissue to give your final scores. Any other questions?

Speaker: I was just wondering if you have any studies which have compared your HER2 ultralow in the true cuts[?] and the resections. Is there any variation in the 2 that has been reported?

Dr Krishnamurthy: I think there are some studies, but I think, Liza showed something, where between core and the surgical resection, there can be a shift. There is heterogeneity. At the lower end of the spectrum, there is more heterogeneity. That is the reason why I think your results on the core may not be the same on surgical excisions.

I think for HER2-low, there are studies that clearly tells you that what is 0 can be 1+. What is 1+ can become 0. Exactly for ultralow, I think, there is no reported studies so far, but it is expected because of heterogeneity. That is the reason why we recommend that if a patient has stone 0, null in the metastatic tumor, you go back and see any tissue that is HER2-low or ultralow to make them eligible.

To answer your question, exactly. How often, how much is the shift between that for ultralow? I do not think we have that data, but I think for HER2-low, it is clearly established that there will be a shift between the 2.

There are some questions from the online audience. What does AI stand for, artificial intelligence or augmented interpretation? That is very interesting. I think the American Cancer, our medical societies, recommend to use the word augmented intelligence, not artificial intelligence, because the rationale is that all these tools in conjunction with human judgment is supposed to produce better results than the tool alone. What the American Medical Society has said that we should use the term augmented intelligence, not artificial intelligence.

There is another question about billing codes, AI tools for billing codes. Interestingly, I work a lot for CAP for the digital and computational pathology efforts of CAP. Interestingly, that is the key question in the field now that there is really, the CAP, we clearly tell people that there is the category 3 codes for digital pathology utilization. They are the utilization codes that has been set up just to encourage people to show how much they are using digital pathology.

Remember, it is category 3 codes. Then for it to be transformed to category one, it is going to take time. Right now we are just gathering data with category 3 codes as to who is using digital pathology. There is no money tied to these codes. The CMS is going to decide as to how much money can be tied to these codes.

Also remember the category 3 codes are utilization codes to capture the technical cost of getting your tissue scanned. The professional interpretation of a digital pathology image vs a microscopic image is a discussion in itself. To answer that question, right now there is no money to using any of these digital pathology tools.

Speaker: May I ask a related question?

Dr Krishnamurthy: Yes.

Speaker 6: [Inaudible].

Dr Krishnamurthy: I am not sure I can answer that because that is little bit, people are not clear about that. Even in my own practice, say, for instance you repeated the ER/PR HER2 of a patient, the reimbursement codes for that is not very clear. I think unfortunately I cannot answer that question as to what are people doing for billing. I do not think you get paid for that when you re-review a case that was already signed out years back, I do not think we get paid again. But I have to check. I am not so sure. Any other question? Yes.

Speaker: Thanks for a lovely presentation. I noticed in the DESTINY-Breast06 the 2 survival curves seem to cross after 24 months on that graph. Is there any significance to that?

Dr Tolaney: No, because the follow up time is short. That it is actually less than that so that is a problem. The other issue is there is crossover. We already know a little over 20% of patients on the control arm have crossed over to get T-DXd with subsequent therapy. I think it will be interesting to interpret overall survival because they did go on to get subsequent ADC. We are not sure to be honest if there will be a statistically significant OS difference when it is mature enough. I would not worry about the crossing because it is just maturity of data.

Dr Krishnamurthy: Any other questions from the audience here? Yes.

Speaker: I would like to ask Dr Tolaney a question regarding the bystander effect. How can you assess the impact of bystanders effect in terms of the scoring? Because sometimes it can be diffused all over the tumor, sometimes it can be just as little spot. How do you think that this relates or not to toxicity effects, to activity of the drug? How can you feel that we, as pathologists, can give this information in order to optimize the selection of these drugs?

Dr Tolaney: It is an interesting question. I will say that I have had patients who have had heterogeneous tumors where one area has been biopsied, for example, that has HER2 staining and another does not. You see response in both areas. Whether that is due to bystander effect or whether that is due to the fact that T-DXd still works in HER2 null tumors, I cannot really answer.

We had done a really interesting study where we looked at HER2 heterogeneity in the preoperative setting. We gave patients T-DM1, which does not function by bystander effect. We biopsied 2 sites in a primary breast tumor. If there was heterogeneity, none of those patients achieved PCR. Whereas if there was no heterogeneity, there was a 60% PCR rate. It was very clear that heterogeneity impacted response to preoperative therapy.

We actually pitched this to AstraZeneca and said we would like to actually repeat this and do it with T-DXd because we would bet that you would see PCRs in patients who have upfront heterogeneous tumors, which is different than what we saw in T-DM1 but that study never moved forward.

One could retrospectively now look at this. They are going to report soon DESTINY-Breast 11 which was a preoperative trial with T-DXd. It had T-DXd alone or T-DXd followed by THP or it had an AC-THP arm. It would be actually quite interesting to look at heterogeneity and look at PCR rates in that trial.

Dr Krishnamurthy: There is another way of thinking for the effect. Even in this meeting there has been a lot of discussion about that. What we are seeing as HER2 null or 0 does not necessarily mean that there is not very low level of protein expression on the tumor cells. Perhaps that is the reason why it responds even if there is 1 cell that is showing some incomplete membrane staining that we are calling ultralow, why is it that the tumor is responding?

One could be heterogeneity. Another is that the way we are testing, it is not that sensitive to capture the tumor cells with very low levels of protein expression. That means there is still some level of very low expression that can be captured with other more sensitive methods.

There is one question from the online audience, which I think is important message. If concurrent metastatic primary tumors have different HER2 results, do treatment follows higher HER2 reading?

The answer is that HER2-low or ultralow, in any one of the biopsies, whether it is core biopsy, resection, or metastatic tumor will make a patient eligible. Because even in DESTINY04 and 06, they took the most recent biopsy, but if that was not available, they took the archival core or archival resection.

For routine practice, we, of course, do the metastatic tumor, but if any one of the 3 specimens shows low or ultralow, then I think the oncologist selects the patient for therapy. Any other question?

Speaker: Can I ask one more? Thank you so much for your talk. I really appreciate Dr Quintana's point about cytology spaceman because I am practicing cytology too. Do I have any comments or suggestion regarding this thing like how we report for cytology basement in those HER2-ultralow samples? Thank you.

Dr Quintana: How we do what? Sorry, I missed the very last.

Speaker: How do you report? Any guideline or comment suggestion?

Dr Quintana: I report using the ASCO/CAP guidelines. We still use the same guidelines, and again, if I have the option for which specimen I can select, which actually happened yesterday. One of our oncologists reached out to us because they wanted testing on metastatic disease and the patient had a cytology sample and a surgical pathology sample, and made sure to go with the surg path sample knowing that the cytology, at least in our hands with alcohol fixation and then subsequent formalin fixation, the HER2 results can be lower.

If it is 3+ positive, it is usually okay in cytology. When we are looking for low and ultralow, that slight difference in a decrease of staining can make a big difference.

I think, there is another discussion at the ISIMM Meeting, where they went over some of the overall guidelines for implementing tests and they also discussed a little bit of cytology. You can try to optimize your assay for cytology samples, if possible. There is a nice article that came out last year by Dr Goldsmith that also goes over cytology. Definitely recommend looking into that. Sometimes cytology is all that you have and I think it can be important to relay to the clinicians that maybe, if we are on that borderline or we are looking at null, maybe consider another sample or something like that. It is definitely a challenge.

Dr Krishnamurthy: I just want to ask Dr Tolaney, in both DESTINY04, 06 cytology specimens were not included. FNA specimens were not included. I think this topic is coming up using cytology specimens. One of the main issues with cytology specimens is the pre-analytic factors are not that regulated as it is in surgical specimens. That is the reason why in clinical trials, they do not like to take FNA specimens.

Having said that, in routine practice, we do evaluate effusions. We do take FNA specimens when core is not available. Core is the optimal specimen, but when core is not available, you have to dip into your cytology specimens, knowing fully well that the pre-analytics are not that regulated. There will be a lot of underestimation of ultralow and low in cytology specimens, but I think we all follow the same ASCO/CAP guidelines with the cytology specimens.

There is a lot of work to be done with respect to pre-standardizing pre-analytic with cytology specimen. That is the issue. For now, like what Liza said, it is following the ASCO/CAP guidelines on cell blocks, cell blocks of effusion, cell blocks of FNA specimens. Smears cannot be used for HER2 protein expression analysis. But would you treat based on cytology specimens' results alone?

Dr Tolaney: If there was any HER2 staining, we will take it and we will certainly try T-DXd in that setting.