Hi. Welcome to the TD Cowen Oncology Innovation Summit that we host annually right ahead of the big ASCO meeting. Of course, we also talk about things that are beyond just the scope of ASCO. We're really happy to kick things off this morning with the team here, with Peter, the CEO of Acrivon Therapeutics, where we're going to talk about their program, primarily the program in endometrial cancer, as well as some of the deeper pipeline assets. Maybe to start off, I will hand it to Peter to provide a high-level overview to level-set investors. I think Peter's got a few slides he's going to run through as part of that, and then we will jump into questions, which I've come up with before, but also this can be interactive. There's the portal for the investors on the line to log their own questions into the system.
You can either use that portal or you can email me directly at marc.frahm@tdsecurities.com, and I will add them to the list. With that, Peter, if you want to do the high-level overview.
Thank you so much, Marc, and thank you for inviting us to present here today. I'll give a brief overview, primarily of the lead program in phase II-B registration intent trials, ACR-368, a CHK1/2 inhibitor, but also have a few extra backup slides if needed. There are, of course, forward-looking statements. As people who know the company are familiar with, instead of looking at genetic alterations and transcriptomic alterations to link a drug with the efficacy in specific patients, we are directly looking at the activity states of the drug-regulated signaling pathways, and our goal is to match the higher-resolution effects of a drug with the disease-driving mechanisms.
The platform is called Acrivon Predictive Precision Proteomics, and it is AI-driven, and it is expanded to interpret activity states of signaling pathways in tumor cells and other disease cells such so that one can predict responders, one can identify indications prior to clinical entry, uncover resistance mechanisms, et cetera. Some of these things are shown here on this slide. This is an overview of the capabilities of the platform. This has been proven in our ongoing trials and with our internally developed phase I program, ACR-2316. We have an OncoSignature test we use for prospective responder prediction in the ongoing phase II-B trials.
In endometrial cancer, we identified endometrial cancer prior to clinical entry by screening across human intended use processed cancer tissues and types with our OncoSignature assay, could identify endometrial cancer as a tumor type with a predicted high proportion of sensitive patients, and that we have now confirmed it's a particularly sensitive tumor type to ACR-368 in the clinic. We have had a number of posters and presentations at major AACR, TRIO meetings, ASCO, et cetera, where we show how we can uncover resistance mechanisms to drugs. We identify a whole number of PD markers also unbiased for a drug and can use that in our phase I trials to assess drug target engagement and do very streamlined PD/PK efficacy correlation. All this is really developed to enable a precision medicine approach that is as de-risked as possible. What you're looking at here is our pipeline.
We have our lead asset, ACR-368, a CHK1/2 inhibitor being developed in endometrial cancer. In addition to developing it in biomarker-positive patients using our OncoSignature test, we are also having two additional arms that were added recently based on the insight that the OncoSignature biomarker proportion is particularly high in the high-unmet-need cancer type serous endometrial cancer. We'll talk a little bit more about that. Serous in and of itself is almost serving as a lineage biomarker, and we have a very impressive response rate in that particular form of endometrial cancer. ACR-2316 is our potential first-in-class dual WEE1/PKMYT1 inhibitor, which is soon going to enter into the expansion phase of our ongoing phase I trial. Oops. We have predicted lung cancer to be sensitive, which is very atypical for WEE1 and PKMYT1 inhibitors and have proven that in our ongoing phase I trial.
Both of these assets have a very differentiated, I would say, favorable safety profile. We only see transient hematological adverse events. We do not see non-hematological AEs, and especially for ACR-2316, it's limited primarily to neutropenia only. Finally, we have a potential first-in-class CDK11 inhibitor program, and we have multiple promising lead series there. There's a particular vulnerability in AML. We have complete regression in AML models across preclinical studies. In our ongoing trial in endometrial cancer with ACR-368, we had an interim analysis not too long ago, which showed that in patients with up to two prior lines of therapy, which is a particularly important thing that I'll discuss in a little bit, we see an overall response rate of 44% versus 26% in our biomarker negative arm, where we have ULDG added, ultralow-dose gemcitabine added as a sensitizer. We expect that ULDG does contribute.
It's not known how much the effect size is. We clearly have a nice enrichment and a very nice response rate in endometrial cancer. This is endometrial cancer of all types. If you look specifically at the subgroups, we learned when we looked at all types of parameters we could assess, we learned that the main efficacy is contributed by the serous subtype. When we look at serous patients with less than or equal to two prior lines versus non-serous, and we look at all serous patients, both biomarker negative and positive, we see a confirmed overall response rate of 52% with a lower bound of the confidence interval of 33%. I'll relate that to standard of care, which is very, very insufficient already in the second line on the next slide.
You also should note that we have a disease control rate of close to 75% and a clinical benefit rate at or above 16 weeks of 65%. Very exciting data in this very aggressive form of endometrial cancer. While this form of endometrial cancer only makes up for about 8%-10% of new cases every year, it has a particularly short overall survival of about 12-24 months versus 12.7 years for all endometrial cancer. It contributes, and it's increasing up towards 50% of all endometrial cancer mortality. In Europe and U.S., where we currently are going with our global trial, we have about 20,000 deaths per year.
There are very limited effective treatment options, and the standard of care actually in second line post the new front line of immune checkpoint inhibitor, a platinum-based chemotherapy with taxane is 13% response rate, 3.6 months PFS, and these patients only have 10 months overall survival. Huge unmet need and a very significant addressable market opportunity. Again, we see here the favorable AE profile, as I've talked about before. This is data cut from all endometrial cancer, and it's very consistent with past trials. We primarily only see hematological AEs and a notable absence of GI toxicities, interstitial lung disease, and other things that can be very devastating for patients. Based on this, we have here an overview of our registration intent arms.
We have arm one with prospective OncoSignature prediction, which is ongoing, and we have now the recently added arm three and arm four, where we are adding ULDG or doing single-agent ACR-368, where we also have a pre-specified interim analysis coming up in the first part of Q4 this year. Very exciting and a very high bar. When it's a pre-specified interim analysis, it's obviously based on powering and very clearly defined terms aiming for the registration intent of each of these trials.
I think in the interest of time, I could maybe just mention that we have uncovered very strong synergy with both immune checkpoint inhibitors as well as TOPO1 isomerase inhibitors, the main payload for ADCs, which obviously opens up very attractive opportunities for switch maintenance with ICI, and that's the basis for our confirmatory trial protocol, which is designed as so-called switch maintenance, where ICI is added, 368 is added to immune checkpoint inhibitor in the maintenance phase of the new frontline therapy that was introduced by Dr. Ramez Eskander and our CMO, Mansoor Raza Mirza, a couple of years ago. It was ENGOT-EN6, it was the RUBY trial. This has basically revolutionized treatment of dMMR patients, 20%-25% of all EC patients are dMMR, they're basically cured.
That is why the serous population makes up about 50% in second line and increasing so in third line, and all patients pretty much progress to second and third line with this form of cancer. WEE1 and PKMYT1, switching quickly to the second program, are critical cell cycle checkpoints in human cancer. We have used our AP3 platform to develop a molecule aiming for superior single-agent activity with an enhanced therapeutic index compared to the existing clinical WEE1 and PKMYT1 inhibitor. Specifically, we wanted to be able to have single-agent activity in tumor types where we could get very, very strong anti-tumor efficacy. We learned that with the AP3 platform, to our surprise, that certain types of lung cancer were predicted sensitive. This has now been proven in our ongoing trials.
Here's a more recent data cut where you can see we have a very nice disease control here. When we look at the last prior line of therapy, duration of that versus our ongoing trial here in these patients who had on average more than three lines of systemic chemotherapy and immune checkpoint inhibitor therapy, we can see that patients benefit or have a very long duration of clinical benefit, either with shrinkage or with PRs. Overall, it's a very interesting tumor type for us. Again, a very differentiated safety profile. We have established two weekly oral dosing regimens already in a short amount of time and are working on a biweekly schedule with the purpose of expanding to the expansion phase in the near term. Finally, our CDK11 program.
I will not have time to go through the details, but it's actually the CDK with the broadest controls in cell cycle regulation. It controls global transcription of cell cycle-related genes and pre-mRNA processing and splicing of mitotic gene, and is predicted to be a very key driver in aggressive AML models. We have verified that we have a number of equally promising backup series to the recently declared development candidate, ACR-6840, and have a catalyst of IND in first half of 2027 to be able to really find the molecule with the best properties. We see here why that is the case. We have complete regression with multiple of these series across aggressive AML models, as shown here with a very favorable safety profile. It's a very potent molecule, and we are very excited about this program. I think I'll end here and take questions, Marc.
Great. Thanks, Peter, for the overview there. Maybe just follow up with some of your earlier comments on 368. You want to put the response rate that you have seen, which is pretty impressive in endometrial, in the context of the outcomes that, particularly in serous and what is the unmet need there, and how rigorously is that known, particularly for the post-checkpoint setting?
[To Arnold], it's a great question. There's been very few, almost no subgroup analysis of drug efficacy in the serous form of endometrial cancer. We did identify one, it was presented at ASCO some years ago, and it was based on KEYNOTE-775, which was the second-line therapy introduced by Vicky Makker, based on pembrolizumab and nivolumab, which was done before the new frontline of immune checkpoint inhibitor and platinum-based chemo was introduced a year or so later. It was the opening session on ASCO, both of these two new frontline therapies. In that subgroup analysis, there was a control arm with physician's choice of 115 patients that were serous. In those serous patients, we saw that there was a response rate of only 13% and a 3.6 months PFS, and again, they also had a survival of 10 months. This is a very aggressive cancer.
Consistent with the nature of these serous cancers, they are pretty much obligate P53 mutated, which means they have a dependency on the S and G2/M checkpoint, which is where CHK1/CHK2 acts. We can say in hindsight that it makes sense we have such a very high activity in that form of endometrial cancer. There's a huge unmet need. Seeing 52% is obviously more than 3x higher response rate, and I showed you that the lower bound already in our interim data cut is at 33%, which is also severalfold above the 13% we're having now. We think that there's a clear second-line opportunity. Now with the emerging ADCs, we allow or we actually mandate patients with HER2 3+, which is about 10%. They must have had trastuzumab deruxtecan or HER2-directed therapy prior to our drug treatment if they are serous.
With the synergy we see with ADCs, we think that the resistance to ADCs and triple-negative tumors is actually the G2/M checkpoint kicking in. We think that once patients progress with ADCs, they will, if anything, be more sensitive. We have already had patients with prior ADCs, obviously, that are sensitive to our drugs. Yes, we are aiming for second line. Worse comes to worse, that ADCs penetrate in second line, which a lot of ADCs are being developed. There's basically based on our KOL event and other data, it's very clear that pretty much all patients transition to third-line therapy, where we certainly have a very low bar, probably lower than the 13%. Right now we are aiming for second-line therapy.
Yeah.
Second and third line.
Thanks. Very helpful. Maybe, again, you touched on the different kind of cohorts that have pivotal intent. Do you want to walk through the important differences there between arm one, arm three, arm four, and maybe a little bit about how the bar might be subtly different for those arms?
Yeah. Arm one is the one we originally started with, which was based on prospective response prediction. It's based on a pretreatment tumor biopsy, OncoSignature assay read. If you're biomarker positive, we predict sensitivity to single agent, and if you're biomarker negative, that you're not sensitive to single agent or as sensitive, and that's why we have an added ULDG, ultra-low dose of gemcitabine as a sensitizer. That's only 1% of standard dose of gemcitabine, and it's based on a wealth of preclinical studies and uncovered with the AP3 platform. For that arm, which is all subgroups and up to three prior lines of therapy, we are seeing 44% overall response rate. The disease control rate is over 80%, as we have previously shared. The OncoSignature predicts control addiction to the CHK1/2 target, but it doesn't predict whether you shrink 30%.
We're seeing 44% response rate, or call it about 40%, and we are having a very high disease control rate. Now, with the emerging therapies, that bar is probably going to change a little bit. Again, we are having up to three prior lines of therapy. The enrollment has been slower because of a pretreatment tumor biopsy than we wanted, and that trial is only running in the U.S. The recent insight that the biomarkers and the OncoSignature fraction is much higher, positivity fraction is much higher in serous endometrial cancer, consistent with them being up against P53 mutated, has made us very excited because we can now do an all-comer serous where we are treating all patients that are serous, treating that as a lineage biomarker with either ACR-368 plus ultra-low-dose gemcitabine sensitization over the ACR-368 single agent, and that's arms three and four respectively.
The enrollment is going incredibly fast there. We are having an interim pre-specified interim analysis, which is the highest bar you can do. It's a very, very experienced biostatistician we're working with that has been part of Dr. Mirza's phase III approved trials from the past couple of decades, and that is obviously with the intent to have a registrational intent for those. Based on the data we are seeing so far, we actually are also repowering the patients. We are saying we enroll up to 90 patients. We actually, with the data we have, we wouldn't need that many patients, obviously, but for now, we are keeping that.
Yes.
We are expanding that in, sorry, in Europe, which you cannot really do easily with an OncoSignature, and that obviously also further accelerates the enrollment of both arm three and four because we can go in all comer serous there in these two regions and are doing so.
Yeah. You mentioned the enrollment has been going much faster with the serous cohorts. Most of that experience has been arm three since that's been open for a few extra months. Just any reason to think arm four will be on a different enrollment pace than arm three has been, either faster or slower?
Both of them, I think, are going to be similar. There's such a strong enthusiasm from KOLs. I encourage everyone to listen to our KOL event that happened during ASCO in, I think it was end of February or March, and from Copenhagen, where we also had our KOL event. We had a late-breaking oral presentation by Panagiotis Konstantinopoulos from Dana-Farber of our serous data. The KOL enthusiasm is extremely strong because of the unmet need, for good reasons, and with that, the enrollment is going very fast. I'd say they're both going equally fast. Obviously, if we have comparable efficacy in arms three and four at the time of interim read, we would always aim for going for arm four, the single agent, because it's a very, very clean path for single-arm accelerated approval intent with single agent.
Yes. Can you walk through some of the mechanics of that interim? What exactly triggers it? How many patients should investors expect across arm three and four ? Do those arms end up being pooled for the analysis, or are they independently analyzed just at the same time?
They are independently analyzed, and the purpose is to do it without a direct comparison of them. We have an opportunity to glance at both arms at that time. Typically, what you see with a powering like this and a pre-specified interim analysis, you do it at about a little over a third towards 50% of patients enrolled, you can anticipate that that's a number there. It's in part event-driven when we do it. There's some flexibility, but it's a pre-specified interim analysis written into the SAP. As I said, the read of that will dictate which arm we choose. We'll probably choose the best one to move towards the registrational intent, towards the registration.
In terms of disclosure plan around that, do you expect to reveal the data, just what the decision is? Just what level of detail should investors be expecting?
What I said, that it's a little more than a third of patients up to the 90 patients enrolled, more towards maybe between a third and half, and that gives you power based on the very, very significantly higher response rate we are observing and the clinical benefit rate, which is 65% over 16 weeks. That beats obviously the standard of care by margins, and this gives an opportunity for that earlier read. If it just holds up, there's obviously room for a lot of degradation in the data and the response rate and still be very, very superior to the standard of care.
You'd expect to disclose whatever the response rate is, and I guess the emerging durability, obviously, that will be pretty limited in follow-up.
There'll be some follow-up.
Yeah
on quite a few of the patients we are aiming for. Yes, we'll not have full follow-up on all patients. Overall response rate is the primary endpoint. Duration of response is the primary or key secondary endpoint, as is usually the case for accelerated approval intent.
Okay. What are the range of next steps that that interim is going to drive? I think, obviously, one would be hopefully not the futility, but then there's also expand to larger within that, there's opening phase III. Just can you walk through what that decision tree?
Basically, we are result and data-driven as always 100%. The data from the interim read will dictate a choice of which arm we pursue towards registrational intent. On the tail end of that, we'll be able to try to initiate our confirmatory phase III trial, which is currently designed as a switch maintenance. With the insight of the serous sensitivity, which we had based on the interim analysis at the beginning of this year, we are aiming to do a hierarchical with serous first and then followed by a pMMR all-comer. It's written as a particular plan where it's called, I've got the name for it, inverse normal combination test, so it allows you to not get alpha penalty with the two primary endpoints.
That's how we are thinking about, and that we'll aim to also partner with towards the end of the year after we have the interim read, confirming that it holds up with the prospective enrollment.
Okay. When you mentioned partner there, that's just for that trial conduct just to get drug supply, or would you be anticipating a broader partnership?
You obviously are open to both. These are very big trials with many hundred patients, but it will be minimally aiming for drug supply. You can also imagine that we would entertain any interest in partnering discussions.
I know we're running up on time, I do want to get to.
Yeah
2316 as well. You quickly touched on it, but maybe the activity you've seen so far, you can run through how that differs from what-- There have been a number of WEE1 programs before?
Yeah
At least one other PKMYT1 program. Just how does the activity you're seeing so far differ from what any of those programs have been able to establish?
Thank you for that question, Marc. There are three main differentiators with ACR-2316 from other WEE1 PKMYT1 inhibitors. The AP3 platform enabled us to develop a compound with superior single agent activity, not only by hitting the PKMYT1 resistance mechanism to WEE1 inhibition, but also by, in the cells, ensure that there was activation of PLK1, which is essential for mitotic catastrophe. That translated to complete regression across tumor mouse models. In head-to-head studies against all the benchmarks, none of which at highest tolerate dose were able to result in complete regression. This was the first observation. Using our AP3 platform for indication finding, as we also did for ACR-368, the team predicted that lung cancer would be a particular sensitive tumor type, and I was personally very skeptical given the lack of activity of previous WEE1 inhibitor in that tumor types.
We have now confirmed that with pretty much all except one. There's only one PD in the lung cancer patients we've treated so far. The durability seems to be very, very good. There's an ongoing clinical benefit, either tumor shrinkage in most of them or PRs. That is completely differentiated also, and something we are obviously thinking of at least including in part in our expansion phase. There's a huge unmet need in second and third line, obviously, in various lung cancers. Finally, the safety profile. We don't have GI tox, we don't have broad heme AEs. They're transient, and it's primarily only neutropenia. These are the three primary differentiating features. It's moved very fast from the very first lead till we entered all dosing. It went in 15 months, and in four or five months, we had two weekly oral dosing regimens established.
We are very excited about that program. Both ACR-368 and 2316 are obviously fantastic combination agents also, developing for single agent first gives us opportunities to have fewer patient numbers to watch approvals. As Timothy A. Yap, our lead investigator for 2316, has often mentioned, these are perfect combination agents. They have non-overlapping AEs with immune checkpoint inhibitors. Both of them have very significant synergy there. With topoisomerase inhibitors, we have such strong synergy that you can imagine lowering the ADC doses so that you avoid some of the non-hematological AEs that can be very severe.
Yeah. Okay. Maybe can you frame, there is an update expected in the second half of the year. You want to frame just how investors think about patient numbers, follow-up, and then this is a broad trial, so the representation of different tumor types.
You can imagine that in addition to some representation of lung cancer, we are also looking into other, using our AP3 platform and molecular markers, we are looking into other ways to more tumor agnostically define a population to be sensitive to ACR-2316. The numbers, it is an expansion phase, so the numbers will still be somewhat limited, but with the encouraging activity we've seen already, we hope to have the expansion phase well underway here mid-year. The numbers will be based on the speed of enrollment and there are many competing trials both in lung cancer and beyond. What we can get in six months with durability will probably be maybe eight to 10 patients. We haven't guided on that, with durability and more patients on the drug.
Okay. All right. All very helpful. Thanks for joining us, Peter, and thanks for everybody joining on the line as well.
Thank you so much. Thanks, Marc. Have a great day.