Immunocore Holdings plc (IMCR) Earnings Call Transcript & Summary

So welcome back to our next session. My name is Andrew Baum, I run Global Healthcare here at Citi. Delighted to introduce an old friend, David Berman, David is Head of R&D at Immunocore, before that at Bristol-Myers Squibb and has touched many of the approved immuno-oncology drugs on the market over the years. It's good to have you. Also joined by Clay, who runs IR at Immunocore. Now what I was going to do is unlike some of the majors, where many of you will be very familiar with the technology in hand because Immunocore developed and markets their own lead product, KIMMTRAK, many of you may be unfamiliar with the technology. So I was just going to hand over to David just to set a common denominator and differentiate the TCR-based approach with the bispecifics, which many of you will be familiar with, which address external antigens. So David, good to have you here, and over to you for [ the scene ] setting.

Andrew, thank you for having us. It's not only a pleasure, but I'm honored to be with you because you always have been at the forefront of immunotherapy. So anytime we can get together for a discussion, I'm very honored. So antibody -- monoclonal antibodies obviously have had a big success in oncology and other diseases and they're over 100 approved by the FDA. There is only one approved T cell receptor, and that's KIMMTRAK. And at Immunocore, we're very proud to have brought that first T cell receptor. Now there are several benefits of a T cell receptor over an antibody. An antibody can only recognize proteins on the outside of the cell, -- on the cell surface, which is about 10% of the human proteome. T cell receptors can essentially look inside the cell. They can recognize proteins that are intracellular, inside the cell, that are processed into peptides and brought to the surface by HLA. So in other words, TCR is recognized peptide HLA and use those as windows to look inside the cell. Developing a TCR as a therapeutic has been more challenging historically than antibodies because antibodies start up with a very high affinity and they're soluble. T cell receptors on the other hand, are normally membrane-bound on the T cell, and they're very low affinity. So our technology has been to convert that low-affinity TCR into a very high affinity, high specific soluble PCR, and we made that as a bispecific. So I'm just going to pause there, and I'm happy to continue, but I wanted to just give you a high level that, essentially, we are a T cell receptor bispecific company. Andrew, you're on mute. sorry.

I should have worked that out by now. So just then segueing to your first approved commercial product, KIMMTRAK or tebentafusp, and highlighting the target that you prosecute in the niche indication. And perhaps something unfamiliar with investors in terms of your Kaplan-Meier curves, the very, very material OS benefit that you saw despite some of the surrogates not being impressive.

Yes. That's great then. And Andrew, this will go back to your foresight on ipilimumab. If you remember, ipilimumab Phase III trial was in conjunction with the vaccine for a target called gp100, which is, -- gp100 is a protein that's expressed in melanocytes, which are normally in the skin and in tumors. In that trial, it was clear that the gp100 as the vaccine wasn't adding much, but it was an interesting target because it is so melanoma-specific. In our Phase I trial that was conducted a while ago, the RECIST response rate was only about 5%. And actually, as an anecdote, when I was coming to join Immunocore, I looked at that data, and I thought RECIST response rate 5%, I mean, is this really going to be a real therapy. But I ended up speaking interestingly enough, Andrew, to the same melanoma investigators I've spoken with and worked with ipilimumab 15 years ago. And they told me the remarkable story that was so similar interestingly, the parallel to ipilimumab, is that there are patients who are alive today who shouldn't be. And that, in their opinion, 40% to 50% of the patients were having long-term survival benefit, even though the RECIST response rate was 5%. And because I have worked with them back on the days with ipilimumab, I believe they trusted them. And so jump forward to the Phase III trial, as you mentioned, the survival benefit in this randomized Phase III trial where the control arm was an active therapy, this was not a placebo control arm, the therapy was -- the investigator choices were pembrolizumab, ipilimumab or chemotherapy. And most patients chose pembrolizumab. Over 80% chose KEYTRUDA. The survival benefit -- the hazard ratio was 0.51, Interestingly, so that suggests a 50% survival benefit. It's exactly what those investigators have told me. They said 40% to 50% of patients start having long-term survival benefit. Intriguingly, as you point out, the remarkable aspect is that the RECIST response rate was only 9%. And so that then triggered a long scientific pursuit to understand how do you explain that disconnect between a low RECIST response rate and a striking survival method.

And presumably, this comes down to -- I mean, the criteria for a partial response is somewhat arbitrary, right? I mean it was -- there's no particular reason that the 30% is a magic number. It's just something that is big enough to detect. But I guess it talks to that the disease control rate and the durability of response with tebentafusp is highly differentiated than can be compared with others. So therefore, even in the absence of having these more profound responses, you can still generate material OS?

Yes. So you're right. the RECIST response rate was nominally higher in the [indiscernible] control. The disease control rate, that is the stable disease [ plus BR ] was about 20% higher than in the control arm. So that we felt explained in part. But what was still remarkable is when you look at the Kaplan-Meier curves, the separation happened almost immediately. And so we thought there must be benefits in PD. And of course, Andrew, you've been along for the whole story, so you've been along for the pseudoprogression story. But when you speak to doctors, most of them will say, well, yes, pseudoprogression does happen, but it's pretty anecdotal. Nevertheless, we decided to look into it. And what's remarkable here is that -- and we can get into how we think this, but it turns out that probably 30% of patients with radiographic progressive disease as their best response are still having the survival benefit. That's the amazing pace, is the degree of pseudoprogression here is higher than I've seen with any other therapy to date. And I think that's how you get up to that hazard ratio of 0.51. Part of its RECIST response, part of it is stable disease and then you have the -- and I should add, when I saw this benefit, Andrew, in PD, I thought to myself, okay, these are probably patients with tumor size decreasing but new lesions. These small new lesions. And they're just technically called PD, but they're actually having tumor regression. So we did this really interesting analysis, which we limited the survival comparison to patients who had frank 20% or greater increase in signs. So these were not decreases in new lesion. Remarkably, we see the same hazard ratio of 0.45 even in patients whose tumors are increasing to 20% to 30% relative to baseline. So something really remarkable is going on with KIMMTRAK.

And then the assumption is that, that biological activity leading to the misdiagnosis of progression or pseudoprogression is a function of the modality, the T cell activation rather than the target. And therefore, you would anticipate seeing it within other settings.

Yes, that's a great question. And that's the exact question we had. We said, is this going to be a platform effect that every TCR target with our bispecific is going to have the same high rate of pseudoprogression with a little bit of RECIST response or is it specific to the target, that's the exact question we had. And it turns out that probably that the answer is no, it's not a platform effect. It probably is -- and it's not even a uveal melanoma effect because with our PRAME data, which we presented last year, and we can talk about the PRAME data, we had 6 uveal melanoma patients. So that is an apples-to-apples comparison. Three out of those 6 patients on PRAME who had never had KIMMTRAK, so they were essentially [indiscernible], 3 out of those 6 patients had durable PR, so a 50% response rate compared to a 4% to 9% response rate for KIMMTRAK. There was one patient who did have that pseudoprogression. So the way that I think of it is, there's elements of both to our platform with KIMMTRAK or with gp100, it was mostly a pseudoprogression and then a little bit of resist to metricate. I think with PRAME, it's the opposite. I think it's more of a RECIST endpoint that there is still some pseudoprogression. And your next question is probably what is the mechanism why -- what is this pseudoprogression, I don't really know. And we spent -- we have some initial biopsies, which we can talk about, but we spent time speaking to every type of expert academic you can think of, including even speaking to people who work in tuberculosis. We thought maybe this is a tuberculoma type, but no one really can understand what's going on. We have had some anecdotes of patients who had pseudoprogression and then they have stable disease for a very long time. And they -- surgeons have gotten and removed that lesion, and the lesion was completely nephrotic a year out from treatment or had macrophages in it, and we don't know why those lesions don't get resorbed as normally happens. Why are they still remaining as a nephrotic tumor? We don't have the answer to that.

And then in terms of the clinical implications, if -- it seems to be tissue or target-specific, actually target-specific, this pseudoprogression phenomenon. It does mean that you unable to leverage OOR as a surrogate marker as you explore another indication, which is helpful to you.

Yes, that's exactly right. So with KIMMTRAK in particular, that's been a problem. We have had some success with circulating tumor DNA, and we found an incredibly strong correlation with ctDNA reductions and survival benefit. It's probably the strongest. It's almost a linear correlation, the greater than depth of ctDNA reduction to longer survival. But for KIMMTRAK, in particular, RECIST response rates that cannot really be used. And this was particularly important because we wanted to expand beyond uveal melanoma to cutaneous melanoma. And so we looked at all of the same surrogate endpoints in cutaneous melanoma, and it actually mimics the picture exactly as what we see in uveal, including the response rate that survival. So that gave us enough confidence to initiate a cutaneous melanoma trial, but it does require some leap of faith because it is not a RECIST endpoint.

Yes. And where was I going to go there? Well, maybe we should just move from the melanoma landscape, which you have a unique approval. Oh, I know I was going to ask. I was going to ask about the, the once-weekly scheduling, which, obviously, in uveal melanoma setting. It's -- there's not much else that you can use for these poor patients with really quite nasty malignancy. If you start entering including non-uveal melanoma there starts to be more treatment options and therefore becomes more of a hurdle. So what are the challenges in reformulating in order to come to a either a subcu or a less frequently dosed IV formulation?

Yes, that is an important question. And certainly, it's something that we're keeping our eye on. The once-weekly regimen was just empirically chosen and there was -- they actually didn't test less frequent regimens. We did test more frequent, and it didn't appear to add any benefit. But once-weekly delivered the speed survival benefit. Intriguingly, the half-life of the drug of KIMMTRAK in the blood is only 8 hours. And that was another -- there have been so many dogmas that have been overturned with this therapy. It happened something with a plasma life of 8 hours, dosed once weekly still give you a survival benefit. And it turns out, we believe what's happening is that the residence time, and the target is actually several days, not 8 hours. And so once-weekly seems like it's a good regimen. We have data from the cutaneous melanoma that when you reach a certain plateau, you can move to less frequent dosing and you still retain the benefit. So there may for our other programs or if we move that into our first-line setting, we certainly could explore less frequent dosing. We have also in -- so there's nothing that dictates that less frequent dosing would have worked, I think, especially when you're combining it with another active regimen. The patients aren't rapidly progressing, that might be acceptable. And then with our PRAME program, we have recently announced that we have a half-life extended version, which should help with patient convenience. So technology can be amenable to half-life extension.

The half-life extension is due to -- is from PEGylation? Or...

We've looked at everything. And what we found is that for our platform and FC is probably the best modality.

So maybe shifting gears and instead of focusing on PRAME and other tumor-associated antigen. Could you talk to your development program in that? I know you've mentioned it. But you've, I think, presented some early data at ESMO. Do you want to just remind those investors who may be not so familiar with what you've demonstrated so far and when we expect the next data?

Yes. So for those more of a generalist investor who haven't followed the immuno-oncology field as closely as Andrew has, you're going to start hearing about targets that you've never heard about before because most of the targets you heard about more our self-service targets for 2 EPFR, et cetera. So now you're going to start to hear about a new class of targets that are only found inside the cell, and no one's ever developed therapies against them because no one's ever had a way to look inside the cell until we pioneered the T cell receptor. One class of targets that are really interesting are called cancer testes antigens. And these are targets that are normally only expressed in the testes, which is an immune privileged site. And then they are aberrantly expressed in cancers. They offer a very good therapeutic window. It turns out in this large group of family of proteins, PRAME is the largest of the opportunities because it's not only expressed in melanoma, but it's expressed in lung cancer, ovarian endometrial and a whole boost in other tumors. So PRAME became of interest for us to determine -- to show whether our platform can work outside of uveal melanoma. And so I'll just give you the high level of the data, we can certainly go into it. But the high-level data is, interestingly, PRAME is better tolerated. Our F-16 molecule is better tolerated than KIMMTRAK because there's no rash. Remember, I said KIMMTRAK with gp100 skin, so you get cytokine release syndrome and ranch. With PRAME, there's virtually no rash, the CRS is milder, and the RECIST response rates are higher. We saw a 50% response rate in uveal melanoma. We saw 2 out of 6 PRs in cutaneous melanoma. All patients progress on epilumab and an anti-PD-1, and 2 out of 4 patients with ovarian cancer who were platinum prior had confirm PR. So we do believe that this platform can work outside of uveal melanoma, and we think PRAME is going to be the place where we can improve this.

And the origins of Immunocore, you had a sister company and you still do [indiscernible], the 2 companies are very separate. The initial target, which the sister company looked at was NY-ESO, and a lot of money has been invested by GSK until very recently when the whole program was shut down with very little to show for it outside solid tumors. Are you surprised, given you're -- basically using the equivalent of a nuclear missile that in a solid tumor, in contrast to what you're seeing in uveal melanoma that you're not seeing that level of activity. Do you think it's something unique about NY-ESO expression or cell therapy dynamics or quality of the cells? Or do you think it's something which is target-specific? And I'm just curious, as you contrast because essentially your -- it's 2 different ways to approach intracellular targets, and so it's just interesting that we're getting such divergent outcomes. Obviously, some of the data for other antigens looks a bit more promising. But I'm just curious on the observations you have.

Yes. I think it's an open question a lot of people have, which is preferable T cell that's engineered with new TCR to a target X, whether it's NY-ESO or PRAME or a off-the-shelf TCR by specific. So we certainly obviously have our preference, which is off-the-shelf. I mean there's some inherent benefits just in the platform, right? We don't have to give a chemotherapy conditioning. We're not autologous lower cost of goods. But there's some operational benefits to an off-the-shelf price specific. But I think scientifically, where the big question is, do you get the durability of response with the TCR engineered T cell in solid tumor trade? With CAR-Ts and EM and with antibody bispecifics and EM, clearly, you get similar strong activity. In solid tumors, it seems to be a different...

You think exhaustion and even with TGFB did not have rest of it, it's just -- it's not enough.

Yes, because you get the T cell persistence, right? They're showing you get the TCRs to persist, but they just have not yet shown durability of benefit. And so we think -- I think what we're thinking, Andrew, is that the repeat dosing with the bispecific brings in fresh T cells almost every week, and so you don't have to worry about exhaustion. Because if you have a T cell that gets exhausted, you give a new drug the next week and you get new fresh T cells.

And presumably thinking of especially in uveal melanoma, but you may have some data in PRAME. For those patients who have had a response and then subsequently progressed, have you had the opportunity to rebiopsy and look at whether these cells no longer -- the presentation of -- is no longer there or just other explanation?

Yes. So we had a presentation at CITC last year. We don't have that data on PRAME yet because that's I think earlier about our contracted uveal. We had about 18 patients who progress radiographic. And remember, we don't know what that really means, and we got biopsies at that initial progression. And then we followed up the patients for survival, and it turns out that about half of the patients were ended up having long-term survival, and about 1/3 of the patients died pretty rapidly. And so then we asked the question, what's going on in the tumor between those 2 groups of patients? The ones who died rapidly have launched HLA on the circuits, the cell. Whereas the ones who live long at HLA and the surface, and so it's exactly what you would expect. In terms of they retain expression of the gp100, they just seem to have less HLA expression. And that's probably would have the same mechanism of resistance for checkpoints, of course, which just activate T cells. The TCR on the T cells also require HLA.

Yes. mix, you might think about using an NK approach, right? If you're -- so okay. So I've got a question from one of the viewers. You mentioned the extended half-life version of PRAME. Could you talk about what dosing you're using?

Yes. So that's a program that will go into the clinic next year. We -- they have life in mouse for the tool molecule is about 7 days. And so we think just by extrapolating, we could have maybe every 3 weeks or 4 week dosing with that half life. Of course, we need to get into clinic to really understand. Definitely, I think it could go less than week, but...

And in terms of the -- I mentioned sort of thinking aloud about NK-based therapies to address the secondary resistance. But thinking about the more here and now, is there any rationale for combining with PD-1 inhibitors? And presumably, you've looked at that.

Yes. So we have some interesting preclinical data, which is we took TILs from a cancer patient, and we segregated them into PD-1 high and low. And then we use our impact in a tumor cell line that have high PD-L1 and low PD-L1, and it turns out the only place that an anti-PD-1 helps is if you have PD-1 high T cells and PD-L1 high tumor. If one is high and the other are low or both are low, adding a PD-1 doesn't really help at all. So we've wondered whether for KIMMTRAK at least in cutaneous melanoma with the anti-PD-1 health. And so this -- the Phase III trial, we just started actually a randomized Phase II/III trial in patients who progressed on PD-1, CTLA-4 and BRAF and it's randomized with the survival endpoint has 3 arms. One of them is pembrolizumab plus KIMMTRAK, one of them was KIMMTRAK, and one of them is a control arm. So there's that scientific reason to combine, then there's a strategic reason with PRAME, we want to move into early lines so we need to combine with PD-1. But I think, Andrew, one of the dogmas that we've uncovered with this or we've turned around is one of the dogmas has always been, if you give a T cell activating agent, it will come in releasing the term gamut, induce PD-L1. And you always have to add a PD-1 on to have activity. And our answer is no. We see a monotherapy survival benefit with KIMMTRAK with no PD-1. And even PRAME in the 6 patients who progressed on prior PD-1 and CTLA-4, even PRAME monotherapy had RECIST response. So you don't need a PD-1. Your question is still valid, would a PD-1 make it even better? And I think that's an open question.

And in terms of that data, which indication was it where you have activity? There was presumably more than one indication where you have a...

Yes. So for PRAME, we had durable PRs in cutaneous melanoma after checkpoints in uveal melanoma and in the ovarian carcinoma. So I look at this as the 3 different types of tumors. One is epithelial, no matter that's platinum retreated. One is a hot tumor that's failed checkpoints, and uveal a cold tumor where checkpoints don't work.

And in terms of broadening the addressable market through MHC restriction. And also a separate question, the heterogeneity of PRAME expression or -- because presumably, not all cells are going to express may get differential levels and whether that guides outcomes or whether there is some local bystander effect because you're creating an adverse environment.

Yes. So 2 questions and great questions. The first is expansion beyond the HLA-02 restrictions about 40% to 50% of the west depending on what tumor type we're looking at. One of the ways we're doing that is by going over -- by going after other HLA. And we announced a PRAME program against HLA-A24, which has increased the opportunity by 30%. It's about 60% of patients in Japan are HLA-A24 positive. So that's one way we're doing it. The other way we're doing it is by going after non-HLA polymarket targets. We announced that collaboration with Gadeta recently, which is using a gamma delta TCR that recognizes. So it's a TCR-based bispecific, but it doesn't recognize the classical polymorphic HLA, it recognizes non-polymorphic. So there would be a universal of growth. With regard to your second question, it was another fascinating insight. If you look at gp100 expression, we use H4, but you can consider also a percent of tumor positive. And you plot that for uveal melanoma, you get a range. Some patients have very heterogeneous expression, very low expression. Some patients are 100% positive for gp100. And what you would imagine is all the survival benefit would be in the [ CP interaction ] would be at the top. And you wouldn't have any benefit in patients with very heterogeneous, that's your point. Remarkably, what we see is we see survival benefit in [indiscernible] almost the same where the patient is 100% positive for gp100 or whether they have very heterogeneous low expression for gp100. However, what we do see is we see an enrichment for RECIST responses in patients whose tumors are more homogenous for gp100. When we move to MAGE and PRAME, we're seeing the exact same story pan-out which is RECIST responses tend to be enriched at the higher homogeneous expression, as you said. And at the lower ones, we believe we see ctDNA reductions, but we need to see -- we're still boiling up for survival. The hypothesis is exactly what you said, and we have a couple of hypotheses. But one of them is we are getting cell-independent adjacent killing either by epitope spread or by cytokines and perforin and granzyme, which are released into the milleau and can have an effect on adjacent cells. That's currently our hypothesis about how we have a survival benefit even if a minority of the tumors are positive with the target.

And I should have asked earlier, but obviously traditionally one looks in the adjuvant -- in the metastatic setting and often last line, as we think about the potential of TCR in the adjuvant setting, where you don't have tumor microenvironments and all the challenges that, that poses, how do you -- what are your plans for entering that space, either neoadjuvant or adjuvant to?

Yes. It's I think it's clear for all bispecifics that the earlier you go, the better the benefit that was seen for BLINCYTO many years ago that the minimal residual disease, the better the benefit. We saw it with KIMMTRAK in our Phase III trial. The survival benefit in patients with very smaller tumor burden was -- has a ratio around 0.3 to 0.4 remarkably, Think about that. So it's clear -- to me at least, I think your hypothesis is clear, which is if you took our platform into the adjuvant setting, I think the benefit would be really high because as you said, there's no tumor burden that T cells aren't exhausted. So with gp100, our plan is to do some form of adjuvant trial in uveal, patients with adjuvant uveal. I think with PRAME, which is where we want to invest very heavily because of the exciting data, the goal is to move it into earlier stages as rapidly as possible. It's just early to go the longer the trial, so we need to do things in a staged fashion. But I do think this platform should work even better in the adjuvant setting.

And so maybe in the last 10 minutes, and by the way, if there are any additional questions, I'm happy to ask some anonymously on your behalf, [email protected], [email protected]. I was going to shift gears, and this is not irreversibly, but I do want to talk about it, given Clay is wrapping up and you did share some data that. Could you talk about your attempts to target the GAG protein? And how you think about prosecuting the potential opportunities within HIV as a way to address the reservoir and the interest from current HIV players? And there's not a lot of them, there's probably about 3 that matter. So what stage are we at? Is it just data generation, and it's too early and the -- or is there already tentative signs of interest from some of the existing HIV players?

Yes. There's some -- so thank you. So for the -- T cells recognize both virally-infected cells and cancer cells, and so it was natural to apply our technology for functional cure. We have a...

David?

Oncology, which is to redirect any T cell to kill these viral reservoirs. We're particularly excited about our HIV data that was shown at COI yesterday for a couple of reasons. The first is that we went up to 15 micrograms. And 15 micrograms with KIMMTRAK, with PRAME, with MAGE, we always saw cytokine induction. We always saw cytokine release syndrome, clinic points. So patients get fever, tachycardia. With 15 micrograms of our HIV program, we see IL-6 induction, but we see no clinical signs of CRS, no clinical symptoms. And we think this makes complete sense because with tumors, with our bispecifics, you have these very large tumor burdens and tumor antigen sinks. With HIV, it's very rare. It's one in 1 million cells literally in the blood are -- have the target. So that's number one. It was really interesting to see that we see the IL-6 induction, which we see across every program, but we don't see clinical cytokine release syndrome. We also -- with our oncology programs, when you give 15 micrograms, you see lymphocyte in the blood go down initially and then go up, and we think that's redistribution into the tumor. With that exact same dose with HIV, we don't see any T cells going out of blood, which once again makes sense because the target is in the blood there. It's not in a solid tumor. So we met our predefined threshold to move to the MAD portion of the study. This was in a single dose, single dose. We met the predefined, which is a safe dose where we are biologically active. We now move to repeat dosing. In the repeat dosing phase, we're going to be looking at antiviral activity. So we're going to look at the viral reservoir by purifying CD4 T cells and looking at RNA and DNA. But most importantly, we're going to be interrupting an entire retroviral therapy. The participants in HIV will be on the background of ART. We'll interrupt the ART, and we'll stop treating with our therapy, and we'll look for viral rebound. The goal there being -- and we should be at active doses even in this first cohort. So the goal there is, do we see delay or can we eliminate even rebound? This is the major question. I think we have in the past, had discussions on both HBV and HIV with companies that are more experts than us. Right now, we believe we have enough ability to demonstrate proper confidence before we move to any partnering stage. But probably, at some point, that would make sense because HIV and HBV are much larger trials, and there's certainly others with more expertise at large base infectious disease trials than us.

And in terms of exploring things like bNAbs as well, which have also been implicated in reservoir or addressing the reservoir, you're not going to be able to do that independently. You're going to have to go to one of the -- or you could go through an academic center, but it seems more likely you're going to be JNJ one of the companies active in the space?

Yes. And that's one of the nice things, by the way, about our platform. We had the same discussions in HBV, what if molecule modality X trends out to be the next big thing. The nice thing about our platform is that it can be combined with almost anything. It can be combined in a cancer with chemotherapies, with targeted therapies, with checkpoints. And in infectious disease, it can be combined with literally any of the other modalities out there. So we do believe as anything that reduces the burden of the disease and holds the virus and check from mutating, I think our platform can synergize with that.

So I'm just going to see check latest on the questions to make sure I'm not ignoring any that I should have done already. No, we've already asked that. Okay. So I think we've already catched them. Is there anything that you think I should have asked about? I know you've taken additional TCR or it's pre-IND, but about to go into Clay, I'm happy if you want to say some few words about that. Or more broadly, is there anything on the existing suite of drugs or forthcoming news flow that you wanted to highlight that we haven't already counted?

Yes. I mean, just -- I'll touch on those 2 quick things, Andrew. So on the PRAME, which we're very excited about, our focus now is on getting sites activated and getting patients on because we're asking a lot of questions. We're looking for monotherapy efficacy signals, combinations with the different agents. So we're going to need a lot of sites because we have a lot of patients. So this year is an execution here for PRAME. And the idea is that by the first half of next year, we will share the efficacy follow-up on these patients. We believe we'll have a robust data set. Our strategy is not to give small patient updates, but rather give a comprehensive data set that will be more interpretable. The other point I wanted to highlight is the initial programs we took and most of the other TCR companies are the low-hanging fruits. The targets that academic sites have worked on for a number of years and gp100. And that makes sense because this is a new field, and you want to go in with a derisked target. I think where we as a company are doing now are unique targets, pushing the boundaries of TCRs. And we're excited to show as an initial example, T01, which is our next target heading to the clinic later this year. And the reason we're excited about TO1 is has many of the same features about PRAME. It's a negative prognostic marker. And so we think it's in both in the cancer progression. And so the tumor is less likely to lose expression of the T01. Secondly, it's expressing colorectal cancer, MSS and MSI high. And MSS has been a tumor that has been immunotherapy insensitive. So it reminds us a little of uveal melanoma. And finally, in about 1/4 of CRC patients, T01 is homogeneously expressed. Once again, gives us hope that we can see RECIST responses with T01. Of course, it's -- colorectal is a very large opportunity. So we're taking all of the learnings about how we do clinical trials, and we're applying it to [indiscernible] for later this year.

And 2 questions just for your speaking spring to mind. Firstly, for the existing -- for tebentafusp but also for PRAME, what's the duration of the infusion? How long does it take to infuse?

Yes. I mean it is -- for KIMMTRAK, it's a 15-minute infusion.

It's fast.

Yes, it's pretty fast.

Okay. And then the second question is on -- is there any need for prior administration of steroids?

Yes, that's the interesting thing. So with KIMMTRAK, the answer was no, we didn't give any. The patients just come in and get the KIMMTRAK. This is very different than what other bispecifics have been doing, antibody bispecifics where they premedicate with steroids. Even without steroids in KIMMTRAK, we did see a lot of CRS about 60% to 80%. But half of it was Grade 1, and half of it was Grade 2, less than 1% Grade 3. And part of the reason for that difference is that the antigens that we're targeting the antigen burden is so much lower for TCR than it is for, say, a CD19 or BCMA. With PRAME and our other programs, we are exploring the use of steroids to see whether we can make it an even easier regimen. But there's nothing inherent about our platform that requires the use of steroids.

And then the other thing is competition. Who else, I mean, do you have blocking IP? Who else? I seem to remember from [indiscernible] was prosecuting a TCR-based approach.

Yes. So there -- so we have the IP and pioneered the natural soluble TCR receptor, which is the alpha beta chain into bispecific. There is one company that has a different version of a soluble TCR bispecific, and so they're taking that into the clinic. But they haven't -- I mean they haven't reported dosing any patients. We've treated almost 700 patients with about 6 programs.

And sorry for making you promote a competitor, but who is that, David?

It's iMATUS.

Okay. All right. So this is the latest pivot.

Yes. But TCRs in general, have a much higher technological antibodies because they're normally membrane bound. They're low affinities. We have to find ways to make them into a soluble therapeutics. So there's a little more of a technological variant.

We've reached time. Exhausted questions from the audience, but it's always a pleasure, David, to see you again. And Clay, thanks for joining. And thank you to our listeners for participating. Say operator, if you could put David and Clay and myself in the side room, that would be great.

Thank you, Andrew. Appreciate it.