Emerging Bispecific Antibody Treatment Strategies in Advanced Colorectal Cancer: Dr. DeVito (2026)

[00:00:00] Manju George: Hello, everyone. Welcome to Doc Talks. I’m Dr. Manju George, the scientific director at Paltown Development Foundation, the nonprofit that supports COLONTOWN. Today it’s my pleasure to welcome, Dr. Nicholas DiVito back to COLONTOWN. He’s an Assistant professor of medicine and a board-certified medical oncologist at Duke Cancer Institute, and today he’ll be talking to us about bispecific antibodies. Dr. DeVito, welcome

[00:00:34] Dr. Nicholas DeVito: Thanks for having me back. I appreciate it. I’m glad to be here, and I will indeed be talking about emerging bispecific antibody treatment strategies in primarily advanced colorectal cancer. Here are some of my disclosures, which include companies that do make bispecific antibodies like Carbio and Astellas and Phanes that are on there.

So what I’d like to do for this talk is really get a good overview of what are bispecifics, ’cause as I’ll mention a few times throughout the talk, there are no FDA-approved bispecific therapies in colorectal cancer. And at any point, if anybody wants to ask a question or chime in, Manju, anyone else, feel free to interrupt me, and let’s try to make things more clear.

Because this, I think, is a concept where we’re describing really a whole new class of drugs. We’re not just describing a new drug target or, like a new, a pill version of a chemotherapy that used to be intravenous or something like that. This is really a specific concept. And it’s important to know, too, that bispecific antibodies are approved.

They’re FDA-approved in lots of other diseases. This includes, lymphoma, and small cell lung cancer and others. And I predict that over the next couple years, we will continue to see more and more and more of these approvals. So I think it’s a really important and timely topic, especially if you are looking for clinical trials, in particular in first line for metastatic colorectal cancer.

So depicted here is your typical panitumumab, or cetuximab that’s binding to EGFR. These are known as monoclonal antibodies. They have two Fab [fragment antigen binding] regions that bind to their target, and then they have an Fc region, which is that back region that has one end. And then it instructs the cell that’s binding to that Fc region what to do to the target that’s on the Fab region, or the antibody just serves to block that target, like in the case of EGFR and other antibodies too.

Now, you might have gathered that bispecifics instead will target two proteins instead of one protein, and there’s a number of different ways to do this. You can make it the different Fab fragments. You can have the Fc region be something that is still actionable, or you can have two Fab regions that are one target and a third target on the on the Fc region.

So you can see how this really gives you a lot of target flexibility and cell-cell proximity, both of which are really important. And why do I bring up cell-cell proximity? Because those of you that are taking any kind of deep dive into scientific literature right now have probably seen lots of technologies that are spatial technologies.

So you can tell how important not just that immune cells are there, but what they’re next to and what’s in their neighborhood. What’s the company that those immune cells keep? So a CD8 T cell, for example, that’s around a bunch of suppressive macrophages is probably not going to be effective at killing tumor cells no matter how many CD8 T cells you have, for example.

So rather than having this one monoclonal antibody, you can see here, and I’ll bring these images back throughout the presentation, that you can target again at least two things at once. So one example of this would be PD-1 and VEGF, which we’ll talk about, where the PD-1 is being blocked on a CD3 T cell, and then you’re preventing VEGF from binding to any suppressive receptors there, thus rescuing PD-1 positive CD3 T cells.

You can also bring these T cells in proximity to the tumor by putting a tumor antigen on one end and CD3 on the other, and we’ll talk about examples of that too. And then there are antibodies that really work in several different ways, such as one like amivantamab, and that binds c-Met and EGFR. So it’s binding two tumor antigens, and then the other side is competent for antibody-dependent cell-mediated cytotoxicity, meaning that macrophages and natural killer cells can come in and kill a tumor whether it expresses either of those targets.

So again, this is quite a diversity of therapeutics from just this monoclonal antibody with one other end, and it opens up a world of possibilities downstream that I think are really important. I will also note too that there are some antibodies that are kind of like this PD-1, VEGF, but even more so not really tumor-specific.

This would be like DSP107, which targets SIRPalpha/CD47, which is the “don’t eat me” signal on macrophages, and 4-1BB, which is an activation, ligand on CD8 T cells. And so when you bring activated CD8 T cells into an area of the body which has lots of “don’t eat me” signals, which is the tumor, you are converting those macrophages and drawing in activated CD8 T cells.

So it doesn’t even need to be necessarily that it’s targeted towards a specific tumor antigen or tumor-associated factor. It may be associated with a cell type that is prevalent in the tumor, regulatory T cells, suppressive macrophages, et cetera. So this really covers what are bispecifics, but why bispecifics is I think one of the next questions.

And so one of the examples that I’ve brought up here is Zolbetuximab, which is made by Astellas and ASP2138, which is made by Astellas. And Zolbetuximab is a blocking monoclonal antibody against claudin 18.2, whereas ASP2138 is a 18.2 CD3 bispecific, or what’s known as a bispecific T cell engager, BiTE, which is being developed in gastric cancer and pancreatic cancer.

Zolbi on its own had only a twenty percent disease control rate, and the overall response rate, I believe was somewhere around like nine percent. It’s not very good at all. However, ASP2138, as monotherapy in both pancreatic and gastric, so not just gastric with Zolbetuximab, had a disease control rate of forty-two percent.

So small studies, but you can see how same manufacturer, same drug target, you get more activity with the bispecific. So we’re able to see how validated targets can then be developed even further and then slotted right in to existing clinical regimens, which is why we have a FOLFIRINOX plus ASP2138 and a FOLFOX ASP2138 plus Pembro and all these different arms that are being studied all the way in first line.

And we think that this could represent a really promising strategy that we’d like to see be biomarker driven in the future too. So again, this is something that is a pressing need, I think, to develop this in colorectal cancer because we have drug targets. We can iterate. We can develop things further. And, we know what some of those are. We know on the tumor there’s EGFR, there’s cMET, there’s CDH17, CECAM, and then there’s all these other different possible tumor antigens that we could list out here. On other cells, we know there’s PD-1, CTLA-4, VEGF, and then this possibility of using different T-cell or macrophage engagers.

So these are all things that have begun to exist or are emerging as targets in CRC, and I’m sure this isn’t a complete list and someone could go find other ones. I think too, one thing that you’ll see is that you’ll see a target that is popularized by a bispecific. Or a target that comes in as an antibody drug conjugate target, and then it expands out.

And then you can see that happening in either direction. So where there are antibody drug conjugates for lots of different targets right now, like B7-H3, there aren’t a lot of like bispecifics for that. There’s a CDH17 antibody drug conjugate, but there’s a more emerging CDH17 bispecific.

The reverse is true for a claudin 18.2, where bispecifics are already in first line, and then claudin 18.2 ADCs are in late line. So I think you’re going to see these targets play off of each other and different things kind of get tinkered with to make like better mouse traps in a way, which will make selecting for trials a challenge in some ways.

But at the same time, you can see how rapidly we’re developing in an area where there was a long period of stagnation. This slide was meant to be animated, but it was not, so I apologize for that. But I kind of just wanna go over the big picture of phases and lines of bispecifics before going into some individual data and drug targets.

I really want to advocate for the utility of these approaches in the first line. As somebody who has done first line studies with bispecifics, I think it’s of critical importance that this is considered because many of these target, targets are validated. And, I’ll pause ’cause I see a good question in the chat.

How would you distinguish between a bispecific antibody versus BiTE ? So all BiTES are bispecifics, and ADCs, which are antibody drug conjugates, could be bispecific if we wanted them to be. It just means that there’s an antibody that has a drug that’s strapped, to the back of it, and then it delivers that drug when it reaches its target.

And that drug can be a lot of different things, ranging from chemotherapy to STING agonist or something like that. So lots of different possibilities with antibody drug conjugates. But these things all overlap, I guess. And you will probably see bispecific antibody drug conjugates at some point if those aren’t already in development.

So I like that question. Thank you. Utility of these versions in the first-line, is that you can get into a maintenance phase potentially, and maybe come off of certain chemotherapies and develop a response earlier. We also know that immunotherapy in particular, this is somewhat general, but works better when given in the first-line.

I think that, that lends itself to more novel combinations and sequencing studies, not sequencing as in like molecular sequencing, like order of therapies, before we can really establish that in this disease. But I do advocate for first-line trials. They’re not a last resort. Administering more advanced therapies early on, can lead to a prolonged benefit, and then sort of the unappreciated benefit of that is that going from trial to trial, as I’ve been fortunate to have several patients do, you’re– ideally, if you’re getting a good response on a trial and you’re benefiting, you’re getting more options as time goes on.

Especially in this environment where we’re just getting this boom of options, and you could potentially be on a maintenance therapy for two years instead of one year or 1.5 years. And that’s more meaningful in a time where development is fast and furious like this and these drug targets are all over the place.

How do you pick between these first-line trials? I think the best trial is the one that’s available and is safe, meaning that you can actually physically go to the trial, and you feel comfortable with the requirements of the trial. You don’t feel like you’re taking away time from your other needs and family, et cetera.

It’s also good, and this is just kind of, again, general advice to take a look at the company’s experience with developing the drug. What are the side effects? If it’s been used in other tumor types. Where is the target normally expressed? I think is really important, particularly for these bispecific T-cell engagers.

A number of these have gotten into early preclinical studies and bombed because they simply do not have a reasonable safety profile. There’s too much normal tissue expressing it. It also informs side effects. So claudin 18.2 is expressed as a connecting tissue between gut mucosa.

And when you give a claudin 18.2 targeting drug, it causes nausea, and it’s a very like dyskinetic nausea. It’s not the same kind of nausea as like chemo nausea. it’s a different, it’s a mechanism that aligns with the target. And so I think it’s reasonable to look at these things and expect if you’re going to give PD-1 and VEGF together in a bispecific, you’re probably going to get the side effects of both together.

And these things are published because again, to reiterate this because it’s important with this concept too, is that when we’re looking at side effects, we can expect hypertension with VEGF targeting, and we can expect poor wound healing. And with PD-1, we can expect a low but not insignificant risk of colitis, pneumonitis, hypothyroidism.

And if we’re giving a bispecific, we should expect both of those. If we get, that lower than that in incidence, that’s great, and we should learn why that is. But going into a trial, I think knowing those side effects and putting them together is very reasonable. I do think that this company experience with developing the drug is really important too.

This is a hard one to ascertain, but have they been working on this for a long time? Did it come out of nowhere? Are there other companies developing this target? Are they the only ones? So lots of relevant questions to bring to your oncologist when considering these things.

I do not think that chasing the perfect bispecific target, just like the perfect CAR T trial or anything else, is really worth it. It’s more about the one that you can get on. I think that there are plenty of options that are at least emerging or are already out there. Later line, there are earlier phase trials, meaning ones that are probably earlier in development, like second, third, fourth line.

And this includes DSP107, which is being randomized right now against Lonsurf and Avastin. These are, these are excellent options. There’s likely a role for bispecifics in many lines and settings. As I mentioned, they have monotherapy activity in several different examples, not just the one that I brought up before.

And, so there’s value in that to consider these options too, if for no other reason that they may offer disease control and getting on another trial, and that many of them are at least biomarker-driven in some ways. I think that combinations are very important to consider too. This includes anti-PD-1, which is the most logical because you’re bringing, especially with T-cell engagers, you’re bringing T-cells into the tumor.

There should be interferon gamma. When there’s interferon gamma, you get expression of PD-L1 that binds to PD-1, then you get inhibition of your CD8 T-cells through that exhaustion mechanism. So this is very logical to combine this. Not all bispecific trials do, but I suspect that you’ll see more that use PD-1 if, again, one of their arms is not PD-1, like PD-1 VEGF or something like that.

This is referring mainly to T-cell engagers. So something that I think matters though for clinical trial design and safety and what to expect going in. Okay. So I’ll start with amivantamab, which is a c-Met EGFR, bispecific we talked about a little bit before. There are several different trials called the ORIGAMI trials that are enrolling right now in different combinations with FOLFOX or FOLFIRI, looking at left-sided and right-sided, RAS wild type, metastatic colorectal cancer in the first line.

In some of their early data, they’ve shown that there’s clear deep responses to amivantamab and in a few cases. This is, this is all with monotherapy, so it lists the number of prior lines here. So that’s why they move so quickly into first-line therapy, with their ORIGAMI trials.

But this is their data essentially, for showing that you do have activity. It is more than you would expect with a panitumumab or something like that. It possibly addresses an EGFR resistance mechanism. I think it also tells you that, you’re not seeing a lot of frequent responses.

You’re seeing like mostly primarily disease control and this is why, again, that’s why I mentioned, okay, well, we really need combinations here, whether that’s with chemo or whether that’s with other drugs. And so what are our takeaways of the ORIGAMI trials

[00:18:40] Manju George: Dr. DeVito? Can I interrupt for a second? Yeah. Do you want to explain the spider plot a little bit?

[00:18:46] Dr. Nicholas DeVito: So- Yeah. Sure. For sure. I wanna mention too that at the end of this I’m gonna have all the trials listed out in like a rubric for everyone. So on the left, these are called waterfall plots And, it shows the percentage decrease in a target lesion, meaning that a tumor that they’re showing, how much it decreased by in response to amivantamab.

At the top, they’re showing prior lines of therapy. This, the spider plot shows duration of response. So what we wanna see is this down here. This green is like a complete, a true complete response or a CR versus a partial response or PR, SD, which is stable disease, or PD, which is progressive disease. So in this patient who only had one prior line of therapies–so the fewest number of lines of therapy–they had this deep complete response that went out for four years, which is impressive.

And then you can see others maybe had some benefit, some maybe stable disease, which is represented in the yellow, and then later disease progression, and then other patients who unfortunately experienced little to no benefit at all in primary disease progression. So what this waterfall plots and spider plots really tell you are depth of response and duration of response.

While the spider plot does show you like depth to some degree, like this is really, I think a lot of valuable information from having both of them together. So yeah, thank you for making me go back to that.

So amivantamab, is it just a better Pmab? The rash and AEs [adverse events] are similar.

Are there combinations other than chemo? As I mentioned before, we really don’t have a good idea mechanistically, like what are some of these resistance biomarkers? who responds? Why did that patient with a CR respond? Is it just because they had a good biology or because they were, giving it in second line, essentially?

We can’t really know. And going straight into first line with chemotherapy will blur that issue a little bit because we’re adding so many variables in. So I think it’s gonna be important, as I mentioned before, to do sequencing or like lead-in studies where we better understand the activity of these drugs and who is most benefiting and who is just not benefiting and getting all side effects, which is never what we want.

There could still be a better way to target EGFR and c-Met. Again, you can continue to see the possibilities of modifying these bispecific antibodies beyond this point, and where another company or drug developer or lab could come in and say, “Well, we’ve made a better version of amivantamab.”

So something to watch. There’s clearly activity here. Where this drug goes from an approval standpoint will be interesting. But I think, the ORIGAMI trials are certainly a reasonable option. A large topic that I’m gonna skim the surface of, in part because of a lack of really rigorous data being available, are PD-1-based bispecifics.

There are many iterations of this, and there are many different ways to create PD-1-based bispecifics. As I mentioned before, you could have two arms that are VEGF targeting. You could have PD-1 on one side. You could have one PD-1, one VEGF. And, but the whole idea is that you get these exhausted PD-1 positive CD8 T cells that are being further inhibited by VEGF, and by having that proximity of VEGF scavenging next to these T cells- You’re allowing them to infiltrate regions of the tumor that do not have a lot of oxygen and are producing all of that VEGF.

So that is the concept behind it. There are primarily most of the drugs that I’m going to talk about are now being developed in first line with chemotherapy. Most of them are randomized trials against Avastin. This is kind of what you would expect, I think, after they’ve done their late line and safety, is that they would move this in again because they’re thinking that…most of these companies are thinking that this is similar to Avastin, but maybe better. There is one drug that I will mention, a lone contender here because I know that they’re running a trial in colorectal cancer, which is, this CANTOR trial, which is a PD-1 CTLA-4 bispecific.

Again, well-known drug targets. These are the targets of pembrolizumab and ipilimumab. We know that these– there are many drugs out there that, already go after these targets. So they have a phase III that’s already underway in lung, and then they have this phase II. There is a restriction that no liver metastases are allowed.

This is a trial, again, just to be aware of in the spirit of mentioning all bispecifics. But there aren’t a lot of PD-1 CTLA-4 bispecifics in development. There are also some PD-1 IL-2 bispecifics, which include drugs that stimulate, anti-tumor immunity and T cell proliferation, that are much more toxic when you give them systemically.

So lots of other iterations off of PD-1 that I think you’ll see, and which one is better than the other one is impossible to know until these trials start to read out, and then it will be difficult to compare them to each other. But I think over these next few years, you’ll see how we really gather this information.

And it’s also why I point back to what I mentioned before, which is the best trial is the one that is available to us and feasible. These bispecifics are being used in the first line because a lot of– there’s a lot of reason to think that they are going to be better than the standard of care.

Again, if they’re not targeting VEGF most of them are including Avastin. So these are some of the studies that are PD1 VEGF. So this is pumatimig, ivonescimab, and then this Pfizer drug that is just a bunch of numbers. And, these are the examples for those trials.

So HARMONY, ROSETTA, and SYMBIOTIC, and then I’ll have those at the end of the presentation in a rubric as well. And so you can see how this is essentially attempting to use first-line therapy and replace Avastin. This is the same song, a different key every single time with each one of these.

So I think there’s no need to belabor the point with that, but I think all of these trials have a reasonable chance of being more successful than standard of care. We just need to learn more. So another interesting PD-1 bispecific that I think is worth attention, because they had some monotherapy data that I think was very compelling, was the use of a PD-1 TGF-beta R2 bispecific.

TGF-beta is a lecture in and of itself that I will not go into a lot of detail on. It is a suppressive cytokine that is released in the microenvironment by various cells. It has effects on tumor cells, and it has effects on immune cells. It’s used normally in the body for bone morphogenesis and all these other things, and of course, tumors co-opt it.

Inhibition of TGF-beta has been very elusive. It can actually cause outgrowth of metastatic lesions, by shifting tumors towards what’s called an epithelial state, so they grow more quickly. So the proximity, how you inhibit TGF-beta, what cells you inhibit on is very, very important clearly, and this was not appreciated until it got all the way to the point of clinical trials.

‘Cause there have been several TGF-beta drugs that have looked great in mice and then have just flopped in clinical trials, either for lack of efficacy or in a couple cases, overt harm. And so I think it’s really important to tread cautiously with TGF-beta inhibition. And I think that they’ve done their due diligence by creating a novel drug that does not inhibit TGF-beta when it is not bound to PD-1.

So essentially what it is doing is only blocking TGF-beta when it’s bound to activated CD8 T cells. So you’re not getting systemic TGF-beta inhibition that caused all of the toxicity and problems that we were worried about before. Instead, you’re doing it in this very localized environment and saying, “I’m just gonna protect my tumor-killing T cells that are going to an environment that has too much of this suppressive cytokine,” similar to VEGF, and that’s one of the reasons why I put them next to each other.

They’ve shown that the drug is safe. They’ve shown some monotherapy activity that could be pretty compelling, and they’re already also moving to a phase III, but this is in combination with Avastin. As I mentioned before, if you don’t have a VEGF arm, you’re usually combining it with Avastin in most of these trials. So a compelling mechanism of action and an interesting drug that I think is unique.

This is the bulk of the first-line trials I think that I’m gonna talk about so far. Let’s see. I’ll take a second for Q&A. High-dose vitamin C, I do not advocate for, the main thing that needs to come out of that is a randomized trial that would show some sort of benefit.

That would not be a bispecific though. Bispecific refers to the agent itself, which has to be a dual targeting antibody.

Is TGF-be-beta relevant to consensus molecular subtype four? I agree. It certainly could be. It could certainly be relevant to other treatments as well. All of these trials that I’m mentioning are for microsatellite stable colon cancers.

Though some do accept MSI high, particularly in the late line after immune checkpoint blockade. So if you are a patient with microsatellite unstable colon cancer, I think that these bispecifics may be a viable next line therapy after dual checkpoint blockade, which is different than a bispecific antibody.

So dual checkpoint blockades an- so anti-PD-1 and CTLA-4.

And I think that enrollment in first line trials can be anywhere really. With a lot of these randomized trials, we don’t have a lot of them at Duke because we don’t enroll very well to randomize trials. Everyone wants to get the drug, and we understand that.

That’s why you’re coming to an academic center. And I think in the community, they enroll a little bit better, and ones with larger networks than us enroll a little bit better in phase three trials, which is fine. I mean, we all are contributing in different ways. But there are a number of centers ranging from MD Anderson to Sarah Cannon, that are running these phase three studies and they’re randomized in the first line.

So I think the number of trials is really a benefit, ’cause it does give access to more people to at least potentially get drug. So great questions. Appreciate everyone-

[00:30:25] Manju George: Dr. DeVito, Do you want to elaborate on that part a little bit? The community clinics versus tertiary centers about the randomization.

That is a topic that is discussed in COLONTOWN. So if you can expand a little bit-  that would be awesome.

[00:30:40] Dr. Nicholas DeVito: Sure … Yeah, yeah. So, most, not all, but most phase one studies and phase two studies are not randomized. I’m actually looking at an exception to that right now, which is DSP107.

There are some phase two studies that are randomized. Phase three studies, the reason why we… To take a big step back, why are we doing a phase three study? We are trying to change the standard of care. We feel that it’s safe in phase one, and many companies, especially with bispecifics with known targets, are doing these combined phase one, phase two.

They show some safety data, they show a little bit of efficacy data, and they say, “Boom, we’re targeting VEGF. We’re just adding PD-1 arm. We can replace Avastin. We can go for a first line trial.” So they have to randomize. Statistically, that is the only way to get an outcome that truly means that there is benefit with this drug and it’s just not a spurious recruitment value or the fact that you got lucky with your phase two or something like that.

And so most of those studies are done at very, very large centers, so large sprawling academic centers. This is Sloan Kettering, Dana-Farber, MD Anderson. They do lots of these phase three studies through their large network. Duke operates a little bit differently where we do our phase one and phase two at the main campus.

We enroll so poorly to phase three since again, no one wants to be randomized at our main institution. We tend to have a community network that has all of these phase three studies. And this of course is not hard and fast. We have plenty of randomized studies throughout, but we are more likely to take on these studies that are earlier in development, handle potential side effects and difficult explanations of how the drugs work and whereas our community sites are better at handling randomization and a more surefire product that’s a known entity.

And I think that’s one of the reasons how– that’s how we balance the best of both worlds, and I think others can balance it lots of different ways. But the overall point is that these trials can be conducted in the community, and that is not trivial. Like many bispecific antibodies before, in early development, had to be conducted at academic centers.

This includes tarlatamab, which still has to be given inpatient. This is a bispecific that’s approved in small cell lung cancer And, that’s because of some of the side effects we’ll talk about later. So I think that there is both necessary and it is also a convenience that these studies are able to be done at community centers.

But unless we made them that safe, unless these drugs were that safe, then how would we ever get them approved in a cancer as common as colorectal cancer? So there’s a lot to take away from how these drugs are being development and the availability of these different trials.

Great. Thank you for those questions.

So briefly I’ll mention this “don’t eat me” signal in tumors, which I’ve had up here for a little while. And it, essentially the “don’t eat me” signal has been well described in oncology. There are other “don’t eat me” signals other than SIRPalpha CD47.

This drug, DSP70, or I’m sorry, DSP107 blocks CD47 signaling and draws in these activated CD8 T cells. And we currently have this trial being randomized against fruquintinib in the fourth line, after three lines of previous therapy. It’s a very strict criteria that they’re trying to develop this in the fourth line.

We think this is an important agent to give patients access to given the low response rates to fruquintinib in general. So it’s one of the reasons we’ve taken on this trial. They have some interesting results in their early phase studies. So we look forward to continuing this study and hopefully getting more patients on, and then hopefully this would be a drug that could be moved into earlier lines of therapy.

It is important to note that it has to be combined with anti-PD-1. They really did not see a lot of responses or disease control when they didn’t give it with atezolizumab. There’s one critical aspect when talking about combinations here, okay. I’m not going to spend a lot of time on this.

I just think this is an interesting story in bispecific development, ’cause this drug was very toxic when they first tried to develop it. It was CEA, CD3 bispecific, and one of the actual reasons for that, they discovered, was this high occurrence of anti- antibody drug antibodies. So that’s when you create an antibody to the drug that you’re being given.

That is not great for that to happen. So this was a very difficult drug to develop at the time.

“Are you seeing responses in these with peritoneal mets? ” I wish I could answer that question more specifically, but I think we don’t have enough data. I don’t think we have patients parsed out well enough.

But I think your point is very well taken. I think metastatic site matters is a common phrase that, that I use, and I’ve picked up from others too, and I think that this is something that needs to be reported in all of these trials. That would be an interesting topic of a review of these drugs as like PD-1, VEGFs, and others start to move on.

Most of us will tend to focus on liver metastasis due to their commonality, and companies wanna say that we have the immunotherapy that treats liver metastasis, but this is no less, — extremely important. So yeah, it’s a great question.

So yeah, everyone here I think is familiar with CEA, and, that’s what this drug, cibisatumab targets, in combination with CD3, which is again, a bispecific T-cell engager.

And so after the anti-antibody drug antibodies, the solution was to combine it with another bispecific, which is an FAP41BB ligand bispecific. So, the T-cell activation marker 41BB, and then, FAP is expressed on these cells called cancer-associated fibroblasts, which are just wound healing cells gone wrong, basically.

So the idea would be that it would enhance these CD8 T-cells coming in. And then the problem is you still have to deal with these ADAs [anti-antibody drug antibodies]. And so their solution to it was to give obinutuzumab, which is a drug that’s used in CLL and depletes B-cells. So they’re saying like, “Well, just eliminate your ability to make antibodies then and give this.”

So I think it was an interesting solution to do all of this at once. They did get responses though in a subset of patients, and this was primarily in this obinutuzumab prep arm, and when they gave it together in like a Q3-week fashion. There were some responses, that I think is something to take away, and I think this is a possible target.

But they did not correlate with a biomarker. It was really not clear which patient population was benefiting the most here, and what should be changed to keep this to, to be a more tolerable regimen, more feasible one that hopefully wouldn’t involve depleting everyone’s B-cells.

That’s a great question, Anthony. They didn’t address that about a more sensitive CEA marker, or again, any kind of biomarker here. You can see their CEA levels are like somewhat clustered in their responders. They seem to have higher ones, but the numbers are so small here that there’s no way you could say that, “Oh, these ones in the middle look just like this.”

It’s not a differentiator. And in some cases, it may be a prognostic biomarker, not a predictive one, which is also a challenge. But I think what we’re going to have to do, in short, not to jump ahead about biomarkers, but we’re gonna have to have composite biomarkers where it’s the CEA and it’s something else– CMS subtype, the immune phenotype, all of these different things are gonna have to all come into play so we can actually predict who’s responding.

‘Cause who would you expect to respond to an FAP41BB drug and CEA, bispecific T-cell engager, CMS4 high CEA? That’s who you’d expect because you expect lots of cancer-associated fibroblasts, and then you have the high CEA. But that’s not what they see. So there’s clearly something else here that is going on.

The FAP correlation was not there either. So yeah, I like that question a lot. Another question that was mentioned was the VEGF R2 anti-PD-1 antibody trial. This is similar to what I would say for all of the anti-PD-1 VEGF drugs, is that these are very reasonable options. I think we know these safety profiles here.

I will not spend a lot of time on this. There are some companies that make a CAR T-cell and an ADC and bispecifics. The ADC company is called MediLink. Sorry I didn’t put that on the slide. I think these are interesting, trials because cadherins are normally supposed to be tucked away in between cells, and then on tumor cells, they’re all over the place ’cause they’re disordered tissue.

So they could potentially be preferential tumor targets. And cadherin 17 is expressed in like screaming high levels on liver metastasis. There are some cadherins too that have been described in ovarian cancer that are expressed in peritoneal metastasis, and I think that this class of drugs that targets cadherins, you’re going to see repurposed for different cancers.

And it potentially being something, to go back to the other question that was mentioned that’s based on site of disease and not just disease histology, which could be very interesting. But it makes sense. They’re adhesion proteins that dictate tissue organization. A very interesting early phase trials here, primarily in late line colorectal cancer, biomarker independent. They’re not even testing for CDH 17 as clearance.

I mentioned this a little bit before about some of the expected side effects. I really think it’s important to try to look for what others are doing or have done. These drugs are being developed heavily in China, as well as the US. There’s folks that publish their experience with them, even if it’s a small amount of experience, and I think it’s worth trying to look for those.

This concept of cytokine release syndrome, or CRS, is one that we’ve overcome to some degree with drug design, it seems or administration. Apparently sub-Q is better than intravenous. It smooths out the PK [pharmacokinetic] curve and makes it less likely that you get T cell aggregation and this big immune response [like CRS].

Cytokine release syndrome is often described as the worst flu that you could possibly imagine. Fever, muscle aches. The dangerous thing is low blood pressure, which can send you to the ICU. We don’t want that, obviously. We don’t really see that with a lot of the bispecifics that I’ve listed here.

Um, so Cibisatamab, the CEA bispecific certainly had that. These are more of an issue with T-cell engagers than the– to that distinction that was mentioned earlier, than PD-1, VEGF drugs or even DSP107, where we comfortably give these outpatient and it’s not a problem. If companies are worried about this, they’ll often start with an inpatient protocol.

It is my opinion that if CRS is not observed on the first two infusions, that they should drop that hospital requirement. Most companies have been pretty good about that. Again, designing their drugs or reformulating their approach to prevent this as a problem. But it is something that should be known, with all bispecifics as a possibility.

We do not know if giving it with chemotherapy will alter these kinds of things or other immunotherapies. As you add combinations, I think you have to revisit these questions every single time and say, “Okay, but now would it cause CRS with this?” ‘Cause CRS is a well-defined immunological entity, but is a difficult and dangerous clinical entity.

It is treated with a drug called tocilizumab, which blocks a protein called IL-6. And again, it happens because all these T cells are getting close together, and then a macrophage produces a bunch of IL-6 and again, can drop your blood pressure and make you extremely sick. So not something we want, something to review and look for in the literature as you’re reviewing these trials. You can see that these drugs have activity though.

So these are some early overall response rates from different trials, and I think amivantamab with its forty-nine percent overall response rate but eighty-eight percent disease control rate still represents benefit and no possible signals that we would really worry about.

That’s all very reassuring. And, so I think that there’s a lot to continue to learn about these drugs, and this is just a snippet of some trial results that have come out recently. What’s more relevant to this audience are actively recruiting trials, which these are most of the ones that I could find, and I will send this slide deck to you, Manju, and you can distribute to everyone.

But a list of all the different trials, their mechanisms, there are probably just a dozen more that I didn’t find on here. But one of the reasons I put this together too was to say, there are lots of options, here.

[Q&A] Ivo being combined with a KRAS inhibitor. I did see that trial. That’s a really interesting one. I’m certainly not against it, KRAS inhibition leads to a more favorable immune microenvironment. I wonder how long that window lasts. There was a trial when we were looking at targeted therapies, that was done by Mike Atkins at Georgetown. This was in melanoma, and he took patients with BRAF mutant metastatic melanoma and either gave them BRAF MEK inhibition first and then anti-PD-1, or anti-PD-1 first and then on progression, BRAF MEK inhibition. And the overall survival difference at two years was 70% in the PD-1 first arm and 50% in the TKI first arm.

And that was data that I took very seriously, given that there is an abundance of data that resistance to MAP kinase pathway inhibition, your RAS, RAF inhibitors and even EGFR and MEK is associated with an immune microenvironment that’s not responsive to PD-1. So this goes back to the sequencing issue before, would we wanna give it all together?

Would we wanna give PD-1 first and pulse dose the KRAS inhibitor? I don’t really like the idea in colon cancer of giving a KRAS inhibitor without inhibiting EGFR, would be the other thing that I would say. So like to me, an ideal trial would be something that really targeted all these mechanisms and maybe the KRAS and the EGFR therapy is like an induction therapy and it’s not keeping that tonic inhibition on to where you develop resistance, but trying to get instead tumor clearance.

But again, that’s these similar kind of sequencing studies need to be done rather than just throwing the kitchen sink at everybody every time. And so we can learn scientifically, could you get a better benefit and potentially give less drug and which would be valuable.

So yeah, no, great question. I definitely think there’s rationale to work at that further, buy it definitely needs to happen in a well-thought-out hypothesis-generating fashion. So that kind of is exactly what I wanted to point to. So biomarkers are not panning out so far, which is a real problem.

We really need to know who responds and who does not respond. And so we’re gonna need some sort of like monotherapy lead in, adjuvant studies, things where we get direct mechanism of action, and try to determine why we’re getting response in some patients and not in others.

And then we can maybe design these combinations in a more thoughtful way and not over or under-treat people, preferably as early on as possible. One thing that is frequently brought up when I talk about bispecifics with folks that are– This is kind of the naysaying side of things, is like, “Oh, we’re just iterating.”

Yeah, and it’s the same thing. You’d say the same thing with antibody drug conjugates. You’re just iterating. You’re just taking drug targets. You’re making it slightly better. It’s a slightly better mousetrap, just the additive result. It’s not synergistic. I would argue against this to some degree, especially with bispecific T-cell engagers and some of the durable responses that we have seen.

The jury is out though, and I think every bispecific and every new drug deserves this level of scrutiny because you are just involving two known targets, and it’s not a guarantee that it’s going to be better than the sum of its parts, just because you’re spatially co-localizing those two populations or scavenging a suppressive molecule in the proximity of an activated cell.

It may work just slightly better than both of those together. On the other side, you do not see a lot of responses to CD47 by itself. But when you combine it with these other drug targets like claudin 18.2 and 41BB, suddenly you start to see responses. And so there is some evidence here that we may be getting synergistic gains where you would not have a monotherapy result, and others where that’s not the case.

So I think evaluating this on an individual level, ..

[In reply to Q&A] Well, I guess what I would say about ADCs, in general too is that they’re great at validating drug targets, but it does seem like the responses are fairly brief. Those might– This is why I would kind of flip this criticism onto ADCs and say “Are we just making better chemo?

We’re just making more targeted chemo that’s like a little bit more tolerable. It’s just better irinotecan than some of the other formulations that we’ve been using. I’m not super-duper excited about ADCs, and I’ve also seen targets that looked very promising just, completely fall flat.

So I think that goes to the biomarker issue too. It’s like ideally, if you’re gonna go on an ADC trial, there’s some sort of biomarker that’s directing, you to go down that path. I do think that you’re going to start to see this get like infinitely more complex. There are many ways to do this. You can make pentamers with IgM that go into mucosal areas and can technically have ten different targets.

I mean, the sky’s the limit with this. You can do nanobodies, which are very small and can detach in certain environments. You can have masking that is hypoxia– low oxygen state or acidic state activated. So there’s a lot of future design that could come from this, and you could see it again being iterative or totally transformative, and that it wasn’t a target before and now it is.

So this is going to get, I think, more complex and we’re always going to need to learn more. I did– To the point of ADCs, I did wanna mention that there are ways to target immune cells with ADCs. There’s like a B7 H3 ADC that’s made by GSK and a few others that targets suppressive myeloid cells and kills them.

There’s also delivery of STING agonist that’s been used for dendritic cells to activate them in a microenvironment. Lots of cool ideas that can be combined with bispecifics and it’d be all part of the same molecule. I mentioned these bispecifics like Ami, I mentioned combos. I will hold off on one piece of exciting data, even though it came out recently, is some of the GCC CAR T data.

We are hopefully, going to open that study here at Duke in the near future. But I have also heard of companies developing these GCC bispecifics, which is, again, to go to the point I mentioned at the very beginning, you’re gonna see a CAR T target, an ADC target, a bispecific target, and ones that span the whole gamut and go in between them.

It’s very hard to know which one is the best. So it’s good to look for trials, especially in the first line. Access your resources. I will provide this PowerPoint, and I thank everybody for their time

Thank you so much, Dr. DeVito. This was very engaging and you covered like a whole vast area really well.

I want you to give a little bit, a couple of sentences about your lab and the work that you do so that people know about the other kind of stuff that you do.

Yeah. Well, Annie asked a much more important question, which is if I still have the sword in my office, and yes, I do have my foam sword in my office from before.

I’m happy to talk about the lab, thank you for bringing that up, and some of the translational work that we do. So from bench to bedside and back is really that in the lab we utilize a colorectal cancer model that’s implanted, orthotopically using a colonoscope[in mice]. It metastasizes to the liver and to the lymph nodes, and to the peritoneum, and to the lung.

And so we can track it, we can look at the immune microenvironment, we can look at how tumors change as they go from one site to another, and we can also use therapeutics in that model to develop a better understanding of response and resistance as particular as it relates to metastatic site.

And in particular, how this affects the myeloid population. So most immunotherapies focus on T cells, but the majority of immune cells in tumors are myeloid cells, which are macrophages, neutrophils, et cetera. So that is the primary focus of my lab is how these myeloid cells interact with tumor cells when a metastatic site is developed.

And then we also bring this to human samples by doing spatial transcriptomics and proteomics, so we can proximate these cells and see which genes they’re expressing, and better develop therapeutics, and then, eventually these things go into to clinical trials or at least we hope, or we conduct clinical trials that we learn from.

So, I thought I talked about on my previous COLONTOWN talk, we’ve now finished the first line trial of botensilimab and balstilimab in stage IV colorectal cancer, microsatellite stable, without, liver, bone, or brain metastasis. We recruited fifteen patients in fourteen months. And so, really a rapidly enrolling trial.

But we also can learn why some patients responded, why some patients had stable disease, and why some patients had progressive disease by looking at circulating tumor DNA markers, but also again, complex transcriptional markers too, that would give us insights into biology that we can then bring back into the mouse model.

[00:55:11] Manju George: Okay. Okay. Thank you so much. So I think you covered a bit of this, I mean, a lot of this about patients looking for clinical trials. You said that if you’re newly diagnosed, then look for first-line trials, but there are also, second-line, third-line, fourth-line trials of many of these agents. So any other thing that you want to tell patients?

[00:55:33] Dr. Nicholas DeVito: I think it’s really that we just really need to rethink what standard of care is. Standard of care is, A, not good enough, and B, therefore a clinical trial. And really that thinking of trial first. The reason we didn’t go on trial is ’cause we couldn’t wait or we missed an opportunity or there wasn’t a slot, because there’s so much development going on right now, and the agents are not just another chemo.

We’re not doing FOLFOX versus FOLFOXIRI and setting number 19 like we’ve done for so long. These are promising agents that have a lot of potential. And so that continual search, always being engaged for clinical trials, is why COLONTOWN exists. It’s why Colorectal Cancer Alliance exists.

It’s why all these organizations exist, because they are the window to a better treatment future for colorectal cancer. So I think that is first and foremost is that we need to be thinking trials first, moving the field forward, but getting the chance to have much better responses on a trial than, again, some kind of stereotypical, salvage therapy, I’m a guinea pig, all of these kinds of things. It’s no, this is nothing short of a miracle that these drugs have been developed and are going into people. And it’s great that we can actually get to that point where we are seeing real long-term responses with some of these drugs, like botensilimab, like vilastobart, some of these other bispecific drugs too, not just second gen CTLA-4s.

So, really important to emphasize, trials first and always. Many of my patients have standing second opinions or they’re seeing an academic doc in between their community. Even if we miss on one trial, maybe we’ll get the next one.

[00:57:34] Manju George: Yeah, and thank you. And then you mentioned that you had patients who went from one trial to another. Can you give some examples of, what would be a sample scenario of that type?

[00:57:44] Dr. Nicholas DeVito: Yes. Yeah, well, the famous one is Bianca Harvey, who everyone has heard of probably, ’cause she went from a FOLFIRI pMab, first-line trial, and then on to vilastoabart and atezo, and has had since a complete pathological response.

So she was only treated on clinical trials. Her dad would not have had it any other way. It was like, “You’re going … you’re in your 30s, you have colon cancer that can’t be resected. We’re gonna get the most cutting-edge therapy that we can.” And being enrolled in a first-line trial at an academic center makes it so much more likely that you’re going to get on a second-line trial.

Because you’re already plugged in with us. We’re watching for you, we know you. We’re trying to think, “Man, if this doesn’t work for them, we’re gonna get something else next.” And we’ve had other examples of that too, where even when we’re taking on a trial, we will say, “This first-line trial will not exclude patients from this second-line trial.”

So we try our best, as hard as that is, to line up a trial portfolio where you can go from one to the next to the next. It’s also why we test a lot of these biomarkers early, and that’s something to always advocate for too.

[00:59:00] Manju George: Okay, okay. And then again, going back to what you said earlier. So if you’re a newly diagnosed patient, then look for first-line trials which have, standard of care plus something novel.

Yeah. And then, you rightly mentioned, people in COLONTOWN, they’re also concerned about being randomized. But going on that trial you have a 50/50 chance of getting the new therapy on top of the standard of care. So in case you missed it, you have the opportunity to continue in that trial, or you can look for another trial, or you can think of going second line, when you progress on that, to a different trial, right?

[00:59:41] Dr. Nicholas DeVito: That’s right. Yeah. Yep. And I think one of the key points there too is, the … again, the best trial is the one that we can get you on. ‘Cause, like, first-line trials are really important. It’s also really important to get started on therapy in those settings. So it’s always a balance, but thinking of it early and often is the key there.

[01:00:04] Manju George: Okay, okay. Thank you so much. Thank you. It’s 3:00 and the time passed so quickly. Thanks everyone for attending, and thank you so much for taking the time to be with us and explain such a complicated topic in such a nice fashion.

[01:00:18] Dr. Nicholas DeVito: Thanks, everyone. Appreciate it. Have a good one.