Targeting CDH6 in Platinum-Resistant Ovarian Cancer: Emerging ADCs, Biomarkers, and Sequencing

This transcript was automatically generated from the audio recording and may contain inaccuracies, including errors or typographical mistakes.

Targeting CDH6 in Platinum-Resistant Ovarian Cancer: Emerging ADCs, Biomarkers, and Sequencing

Dr. Elizabeth Lee (Dana-Farber Cancer Institute): The context for this podcast is that we are in a really exciting time for therapeutics for platinum‑resistant ovarian cancer. We are also excited about the many different types of therapies, including targeted therapies that we have at our disposal. A lot of which are still investigational and on trials, but we are just so excited about the activity that we are seeing.

In that regard, CDH6 is emerging as a target. Right now, it is primarily for a class of therapeutics called antibody‑drug conjugates, which is a very rapidly growing class of therapeutics for platinum‑resistant ovarian cancer. There are a number of different investigational drugs in various levels of development, but those are targeting CDH6.

Dr. Rimel: 100% agree. It is interesting that cadherins are being targeted in this way. The biology of CDH6 is very interesting. Cadherins act primarily as tumor suppressors, especially E‑cadherin, because cadherins are transmembrane proteins that basically require calcium and they stick cells together. They also are really important in cell aggregation and especially in the metastatic effect of cellular movement that we think happens, especially in ovarian cancer, which has a unique presentation with a lot of small‑volume disease in plaque‑like areas or sticking on to peritoneum, which may actually be one of the roles of the cadherin.

CDH6 is frequently expressed in these high‑grade serous ovarian cancers, and it is really interesting that from this basic science, these cancer cells move and spread the way that they do, we came to this opportunity of having this as a target for these antibody‑drug conjugates. Structurally, are we using biomarkers CDH6 expression either by IHC or other mechanism to select for ovarian cancer patients for ADCs?

Dr. Lee: Right now, we are not. This is based on the trials that are using these CDH6‑targeting ADCs. They are not selecting at this point. We do need to understand on a more granular level, how expression might be changing the clinical benefit of these drugs in these patients. The worry, of course, is that if there is someone who has very low levels of CDH6, the ADC may not be quite as effective compared to those who might have higher levels of CDH6 expression. This is also all in the setting of we do not right now have a validated assay that we can reach for clinically in order to test for CDH6. This currently remains an investigational assay.

The whole thing with ADCs is really relying on the expression level. We are really understanding a lot from the HER2 experience with trastuzumab deruxtecan, as well as the folate receptor alpha experience with mirvetuximab soravtansine, that there may be nuances to how expression can impact the efficacy of the drug. We need to get a bit more of that data in the setting of the CDH6‑type drugs.

There have been some preliminary data reported out related to CDH6 expression that we anticipate in ovarian cancer, the most common being high‑grade serous ovarian cancer. The CDH6 testing, based on IHC, has its limitations again, as we understand from our HER2 and folate receptor alpha experience. Even if something scored as low expression or "1+", does that really mean there is not going to be enough for a drug? I do not know that we can answer that directly. I think it also depends on the drug that is targeting CDH6. There may be differing levels of potency of the drugs. There may be different payload offloading and tumor microenvironment interactions that may be just as important as the surface expression of CDH6. It ends up being a fairly loaded question.

What are your thoughts on this, Dr. Rimel?

Dr. Rimel: CDH6, we have data that it is expressed more commonly in high‑grade serous and endometrioid subtypes. So, you are right. This is an opportunity for us to think about are there other subtypes of ovarian cancer that might benefit. It is exciting to see that we have phase I and II data that demonstrated some exciting responses.

We are in a time, as you mentioned, learning from our prior experience with HER2 and folate receptor‑targeting ADCs that we are just beginning to open the door to understanding how much expression is required to get a good response, how valuable is it for the linker to be dissolved in different spaces within the cell? How impactful is that payload? If the payload is similar between two drugs, do we choose based on a target and expression when we are trying to think about how we might deliver these? Then the drug‑antibody ratio, we do not talk about very much, but this is something that I am curious about, especially in low‑expressing situations. If the drug‑antibody ratio is very high and we have lots of drug per antibody, will there be sufficient activity in a drug like that in a low‑expressing phenotype?

These are all questions that we just do not have answers at this moment. We have the REJOICE‑Ovarian01 study for R‑DXd which is ongoing. The original phase I and II data was very exciting and got us to this point where we can start looking at the phase III in prior‑treated patients with platinum‑resistant ovarian cancer.

However, we are starting to see some toxicities. Dr. Lee, you have a lot of experience with this. Would you like to comment on the toxicities that we see from the CDH6 ADCs?

Dr. Lee: Yes, sure. The CDH6‑directed ADCs, and actually many of our other ADCs, are now increasingly using topoisomerase I inhibitor payloads. We are beginning to see more of a class effect based on the payload type and the payload mechanism.

We have our topo I inhibitor payload ADCs, and this includes drugs like R‑DXd, as you mentioned, which has a deruxtecan payload. There is also CUSP06, which is using another topoisomerase I inhibitor payload. SEM‑505 is yet another one.

What we are seeing is primarily cytopenias, so neutropenia, thrombocytopenia, and anemia. They are severe drops in neutrophils, platelets, red blood cells.

That speaks to how we think about folks who might be suitable for these types of agents. We have to think very carefully about mitigation strategies, for example, G‑CSF to support neutrophils or use of a thrombopoietin mimetic in order to support platelets if needed. Just being very vigilant for our patients.

Right now, these drugs are being studied in the trial setting where it is as homogeneous a patient population as we can get, typically on the healthier end. They are selected to have met certain criteria for their blood counts or their functional status, etc, but in the event in the future, we end up having one or more of these agents brought into the real world where we might not be applying directly per the trial‑based patient population criteria, we might not be applying these drugs to just those folks. We might be using these drugs in patients who are a bit more heavily pre‑treated. For example, maybe not three prior lines of therapy, but four or five or six prior lines of therapy. They might be slightly older. They might be slightly less functional or sicker. These types of cytopenias that are high grade might start rising higher in terms of the frequency we are seeing, and we just have to be so vigilant for our patients. Especially with the DXd experience, the trastuzumab deruxtecan experience with interstitial lung disease, that is something that I was paying very close attention to when the R‑DXd data was being reported out.

From ESMO in 2025, the phase II dose optimization part of the REJOICE‑Ovarian01 study did comment on a few cases of ILD. About 4%. Perhaps less so than the T‑DXd reported experience, but it is not zero. Especially for folks, as we are thinking about maybe using multiple different types of ADCs in any one patient's treatment journey, each of them potentially with some pneumonitis risk. It could be that ILD in the real world with these drugs may be higher than what is reported in the trials. ILD pneumonitis might be a treatment‑related side effect that right now seems to be at a low frequency. We will have to see.

Primarily, it is the cytopenias, some GI toxicity, nausea, diarrhea, and then cumulative fatigue related to multiple doses of the drug.

What do you think about what you have been seeing with the CDH6 ADCs?

Dr. Rimel: You are making an incredibly important point. The phase I and II trials demonstrated an overall response rate of 44%‑50% in heavily pre‑treated platinum‑resistant ovarian cancer patients. They landed on the dose of 5.6 mg per kilo for the phase III. My experience has been when I am explaining the potential benefits of participating in a clinical trial in this setting, or any of these ADCs where the response rate looks really exciting but there are toxicities that may not match the patient that I have in front of me, as you mentioned, because they are heavily pre‑treated in a way that they might not have been eligible for the clinical trial. Being extra careful and vigilant about these deep and prolonged cytopenias is really important, and also making sure that we counsel our patients about the rates of diarrhea. That is just another one of the GI toxicities we can see and nausea. Making sure those medications go home with them before the first dose is delivered to minimize some of the callbacks and needs for IV fluids and things like that.

You are right. As we look at these pre‑treated patients, their bone marrow flexibility or tolerance to these kinds of medications is likely to be limited. We do need to be more careful and ready to jump in, as you mentioned, with G‑CSF or thrombopoietin agonists; that is a really important point.

I have seen no ILD in my practice, but we have been very carefully ensuring that these patients, of course, on trial are getting chest CTs because we know that early evaluation and recognition of ILD, even when it is asymptomatic and seen on the CT alone, is grounds for close monitoring, follow‑up, and sometimes delay of drug. It is important to point out that these kinds of things make these ADCs different than our traditional chemotherapy, where we maybe did not always scan the chest with every disease assessment. It is important to just note that there are things that make these drugs a little bit different.

I am curious about how we will think about sequencing this. As you mentioned, this is one of several different topoisomerase I inhibitors that is used as the payload. I am curious what you think when you have a patient maybe that was treated with a HER2 ADC with a topoisomerase backbone. Are you imagining that you would consider a CDH6 with a topoisomerase backbone ADC for that patient if they progressed on that line of therapy?

Dr. Lee: These are hard questions, and I really wish we had much more biomarker correlative translational data to inform on making these kinds of clinical decisions. I do not really know that we understand well enough the mechanisms of resistance to topoisomerase I inhibitor ADCs in order to be able to adequately sequence because there are just so many factors.

For example, let us say I have a platinum-resistant ovarian cancer patient who is HER2 IHC 1+, and at the time she needs treatment, all I can reach for is maybe a T‑DXd‑based regimen off‑label or on a trial, and she progresses. Do I know if it is because of the low level of expression? Because from DESTINY‑PanTumor02, we do already see differences in terms of objective response rate and PFS differences based on levels of HER2 expression. Or is it based on something that is mutating in the target HER2 itself? Is it a truncated p95 form? I do not know. Is it actually more payload‑directed, and it is a DNA repair mechanism of upregulation to bypass the double‑stranded DNA breaks? Maybe, but we do not know.

If I have someone who has maybe progressed on T‑DXd, would I feel very comfortable with going to R‑DXd? It is the same exact payload, slightly different in terms of the target, but same payload. I do not know that I would feel comfortable, but I do not have hard data to support that. Would I feel comfortable going from, let us say, T‑DXd over to one of these other drugs that might have a 'newer generation' type payload? If it is exatecan‑based, belotecan‑based, what have you, but a different payload than T‑DXd? Maybe. Do I have data to support that? Also no. This is just really hard. We have been using ADCs for less time in the ovarian cancer space compared to other cancers, and so, with the limitations of doing cross‑tumor‑type comparisons, cross‑trial comparisons, just leaning more on the data that is available, for example, in breast, the indications are that we might not be able to sequence topo I inhibitor after topo I inhibitor.

I do not know what is your read on the data.

Dr. Rimel: Yes, I agree with you. We do not have this data for ovarian cancer, because we have not had as many drugs or as many opportunities to use ADCs in our population. However, the data from breast cancer and some of the basic science data, looking at cell line responses to multiple lines of ADC treatment, we are switching target versus switching payload does suggest that, at least in people and cell lines retrospectively assessed, it appears that the payload may be an important feature of resistance. In those cases, you are right; I probably would feel less comfortable switching from the same backbone to the same backbone again, even if the target changes.

That being said, for things that do not share the same backbone, for ovarian cancer, we have mirvetuximab soravtansine, which is a different backbone; that is a maytansinoid. I would probably feel more excited about treating with the CDH6 ADC with a topoisomerase I backbone. I would probably be okay if the patient had progressed in the line before on a mirvetuximab regimen. There are opportunities for us to treat with ADC, but keeping an eye on what the backbone is, it is going to be important. It is imperative for us that we keep our eye on the data as it emerges, because I do think it will help us choose the right drug for the right patient at the right time, which is everybody's goal.

 When you are thinking about how you are treating your platinum-resistant ovarian cancer patients, are there other agents in this space that you are thinking about?

Dr. Lee: Yes. We have the luxury of being able to think about multiple agents. Rolled into that question is how we are looking at multiple different targets. Certainly, folate receptor alpha, as you mentioned, is important from a clinical standpoint, a standard‑of‑care standpoint, in assessing whether a patient is appropriate for mirvetuximab either alone or potentially with bevacizumab. Also looking at HER2 testing to see whether they might be eligible or appropriate for trastuzumab deruxtecan. CDH6 is there in the platinum-resistant ovarian cancer setting as well. We have so many other targets, B7‑H4, B7‑H3, TROP2, etc. It is getting to be a crowded area of development, which is great because for patients, having more opportunities and more potential therapies is only good. The question then just falls onto us as providers as to trying to figure out and parse out the data of which is the most appropriate

The question of expression of these different proteins is also really important to think about across someone's lines of therapy. We do not really have good data, even for folate receptor alpha, which has been 'around the longest,' of how it might or might not be changing over time, how it might or might not be changing in response to folate receptor alpha‑directed therapy. We do not have that data for CDH6.

The biomarker data that was presented from the phase I of R‑DXd was looking at a static point in time, not following it over time, which I acknowledge is very hard to do both on trials and clinically, to subject patients to repeated biopsies, to evaluate expression. That is very hard and a big ask of our patients. That means that we are conducting our clinical care in almost a vacuum of understanding what happens to the protein. We do not really have a whole lot of data either on co‑expression of target antigens. Some are starting to come out based on TMAs, etc, but these are all retrospective. However, as we have more and more targets coming out, we really need to understand these panels of targets for any one particular patient in order to inform on what might be the most appropriate. It gets to be a very complicated area, an area where I wish we had more data, but it is exciting as these trials are being done and hopefully as their translational work is also being reported out that we will start to fill in some of those gaps.

Dr. Rimel: Yes. You bring up a really important point about biopsy and that snapshot in time. Most of the tissues that we have for our ovarian cancer patients come from their cytoreductive surgeries, which are at the beginning of treatment, either before chemotherapy, after neoadjuvant treatment, at the time of interval cytoreduction. That gives us at least some idea of what the tumor might look like after exposure to some platinum‑based regimen.

Asking for a subsequent biopsy to look at protein expression in later lines of therapy has been, I agree, very difficult to sell. Although more recently I have found in my clinic, when I am talking to patients about how these drugs work and what my options are, and I say, "We have done tests for HER2 and folate receptor and CPS and all the things I have clear approved testing for," I often will have patients say, "Is that true still?" When I say, "I do not know; it is possible it has changed," I have been surprised at how many of my patients are willing to undergo a biopsy if they have accessible tumor to answer the question about is it possible that another drug may be available to them that changes may have happened over time?

I know you are aware of that very interesting paper that looked at a series of ovarian cancer patients who had been assessed for HER2 expression from their cytoreductive surgeries, and then had a paired biopsy that was done during a later line of therapy. It was not a lot of patients, only 16 patients. However, in that, just looking at a single protein marker and whether or not it changed over time, there was drift from the initial cytoreductive sample to the subsequent secondary biopsy.

I do think that there may be a role for us to be considering offering patients access to a second biopsy to see if a drug may be appropriate. There may be opportunities for us to consider for some of these ADCs, if we decide or learn that expression is really important in how we think about this, to re‑biopsy looking for better expression of something if it was a little bit there before maybe just to push us into an area where we are more confident of a response, or at least to be able to get it on label. That is a potential situation.

CDH6 ADCs are fortunate that they appear to be pan‑expressed, but I do not know if that data includes after multiple lines of therapy or if all of that is derived from primary tumors, either just neoadjuvant or before chemotherapy treatment. It will be interesting to see if subsequent data can help us find out if that expression changes over time and, if so, does it change the efficacy of what we think is likely a pretty exciting drug?

Dr. Lee: These are really pertinent points. I would love to get your thoughts on: do you repeat tumor DNA testing, NGS, or RNA‑seq to see if that might impact whether you go from one type of ADC or targeted therapy to another?

Dr. Rimel: I have not seen measurable change in next‑generation sequencing of DNA in my patients that have had sequencing at primary, and then after subsequent lines of therapy. I had patients that participated in a clinical trial where they were allowed to have a second next‑generation sequencing test. I did not have patients paying for off‑label testing, but we did not see a lot of differences in the DNA. The study was done a while ago. We did not have RNA‑seq at that time. I imagine that there are probably changes that are more visible in the RNA and protein than they are in the DNA.

I have been using a little bit of circulating tumor DNA for patients that are getting immunotherapy‑based regimens because that is where it is on‑label, and using that to look for minimal residual disease or to follow. I am excited to see because some of the companies are saying that they may offer resistance identification mechanisms in the coming years, and that does sound really exciting. I do not know if that will really help me, though, in ADCs, because protein expression seems to be what I am looking for and outside of tumor biopsies and IHC, I do not know if the RNA correlate will be strong enough. What do you think?

Dr. Lee: Hard to say. My hope is that we are able to eventually get down to a point where we have better diagnostics that we can continue tracking through a treatment journey for a patient, rather than at those static points in time. As you are alluding to, the jury is still out on really how to most effectively use 'liquid biopsy' mechanisms if we are unable, or a patient is just not safe or feasible to have a repeat tissue‑based biopsy and tissue‑based assays done.

Dr. Rimel: As a clinician, I find that the more of these tests and answers that I have, the more I need to develop a strategy of reporting out in the patient's chart of a matrix, basically, of these are the tests I have done, these are the answers, this is the time point, the temporal space in which that data lives. These are the subsequent tests I have done. This is a protein test. This is an RNA test. This was a DNA test. The time that it was done and then what the results were for each of the drugs that I have given, what the responses were.

I have actually started working on coming up with a template to try and help me understand for each patient, because there are so many questions that help me try to think about what the right drug is for that person that I see a future where I am going to need help. We are going to need help figuring out who should have what and when and how.

I hate to bring up the spectre of AI, but this might be a time where we could at least have that data in front of mind and in front of us as we are thinking about the patient and her wishes and needs. What kinds of toxicities are intolerable to her? Where does she want to receive her treatment. How do you do it?

Dr. Lee: I do not have a great way. It is getting very convoluted, as you are alluding to. I think the heart of the matter is that the platform, the type of test, the timing, the exact alteration, the exact expression level is now becoming very important.

Dr. Rimel: Yes, it is really important. This has been really exciting, and I am excited to see the CDH6 ADCs move into this exciting space and be able to offer them to more people if the phase III trial is as positive as we hope it is.

Dr. Lee: I agree; it is a really exciting time. The early‑phase data clearly, R‑DXd being a little bit further along, the data from R‑DXd is very, very encouraging. Same for CUSP06, a little bit further along as well. They had some data that was presented back in 2025, very similar in terms of an objective response rate. These are really encouraging numbers when we think about platinum resistance and comparison to our historical chemotherapy.

Dr. Rimel: It has been great talking to you today, Dr. Lee. Thank you for your time.

Dr. Lee: Same. It was a pleasure.