We are very pleased to have the CEO of Allogene here with us, Doctor David Chang. Allogene, despite the volatility of the markets, is obviously continuing to execute on the, I call it phase III, but pivotal phase II randomized study in a consolidation approach to first line DLBCL. They have been working on that study, but also had some recent announcements on the last earnings. Maybe it would be a good place to start if you could tell us about the study and particularly the developments on the quarter and how we should interpret that, because the market was a little bit concerned about the delay. You say that, no, we got plenty of people enrolling and coming onto the study. Tell us about the execution of the study and the recent developments.
Yeah. First of all, Mike, thanks for having me on your panel. The study that you are referring to is our ALPHA3 study of cema-cel in the front line consolidation setting. One thing, just as a reminder, the approach that we are making in that study is applying the right product, cema-cel, where we have shown very compelling efficacy data as well as safety data coming from our phase I study into the right patient population. These are patients who have completed front line R-CHOP treatment, but still remain MRD positive. There is the selection of the patient based on MRD testing and then also at the right time, which is treating them before their disease manifests in a way that jeopardizes the safety of the patient or patient's well-being.
In terms of the strategy of the ALPHA3 program, this is very well established and recognized by not only the hematologist and investigators involved in our study, but also by the investor community and also is very well liked by the FDA in terms of design. One thing about that study, as we rolled out in the United States, is that study design in itself is somewhat unique. There has been no such design as a frontline consolidation. It's usually adding onto the frontline on top of the R-CHOP. Investigators have been telling us this may require a little bit of education, patients, and other sub-investigators alike. We sort of heard that numerous times, but as we were operationalizing the study, we definitely come to appreciate that it does take a little bit of time.
Part of the communication that we announced at our last earnings call is pushing out our first milestone by about two quarters to the first half of 2026. The main reason for us to do that is that initial ramp up that took more time than that was needed. That was mainly coming from clinical sites, even after they complete all the regulatory paperwork to be qualified as activated in terms of actual contribution of the patient screening took a little bit of time. This all had to do with the site resource issue more than anything else. The second part is the design in itself. There is a flow of the patients. The patient gets identified for the MRD testing first, and they have to wait till they complete the R-CHOP regimen.
It can be up to about 18 weeks, and then they undergo MRD testing before they can be randomized. The patient flow has been likewise, it takes a little bit of time.
Specifically, you've given some numbers, right? You gave some numbers on the call as to how many people you have visibility on or had signed consent to get tested, right? What were those numbers and how does that play into how many could get enrolled?
We have said that right now, this was at the time of the earnings call, which was a couple of weeks ago, about more than 250 patients have consented to undergo MRD testing. That number is growing pretty rapidly. I think that is a really good indication for the demand of the study, which investigators have been telling, and it is actually materializing and we are seeing it from the patient consenting. Now, how does that eventually translate into patients being randomized to the ALPHA3 study? They have to wait a little bit, they have to undergo MRD testing, and then after that they get randomized to the ALPHA3 study.
It does take a little time and we expect in terms of getting to enough patients to make the lymphodepletion selection, which is our first important milestone in our first half of 2026, we feel very confident that we can get there.
Let's walk through some of that math. You said 200 and something, how many?
More than 250.
More than 250 people have consented to get MRD testing. That is obviously a funnel because those people would then, what percent would then be MRD positive after R-CHOP? I think the number could be what? The R-CHOP and then they have to actually not have gotten a CR. Of those people, walk me through some of that math.
Yeah. Of the people who have consented for the MRD testing, some of them would not be eligible to undergo MRD testing. Those are patients who did not get good enough response from the R-CHOP regimen. These are so-called primary refractory large B-cell lymphoma, and they should really go to the next line of therapy.
What percent of people would be eligible to get MRD testing?
We currently estimate the majority. We haven't given the exact number till we figure out more actual numbers from our clinical study, but we expect a large majority of patients will undergo MRD testing. Based on all the data that's available from the retrospective sample analysis, we expect one out of five patients who undergo MRD testing to be MRD positive. Those are the patients who will be eligible to go on to the ALPHA3 study.
Now are the people that consented, are these people that have not yet begun the R-CHOP? Are these people who are in the middle of the R-CHOP? Where are they in the stage of that where they consented on this form?
Most patients would have already started R-CHOP treatment. In terms of consenting, it can happen while they are undergoing frontline, first cycle of R-CHOP to when they complete the R-CHOP. It is essentially about the 18- week period, which is the length of the six cycles of R-CHOP.
I think we have gone through this before, but most people do get a durable CR, so that most people would be eligible to at least be considered for the MRD. Is that fair?
That would be a fair statement.
Yes. I'm going to throw out a number, like I think 80% do get a durable CR so that we could do some testing. Is that fair?
I think you are in the right ballpark figure.
Then of the 80%, if they all do get the testing, then one out of five could be MRD negative, excuse me, MRD positive.
MRD positive, yes.
Therefore they would choose to now move into the study. Is that a fair assessment?
You know, one thing I've learned from you is I'm not going to argue the math with you.
I'm okay in math. I'm decent in math. To be fair too, because we want to make another adjustment, which is once they get there, they do still have to then agree to now go under the protocol to actually get the CAR T.
Yes.
It's just not like, okay, we got a test, you did the test, you're not consented to the test, then you got to consent to the drug. Is that a fair statement?
Yeah. The way that you have described funnel is exactly how we think about it. We start with as many patients as possible. The funnel, the angle of the funnel, we will determine as we do the study. In terms of the actual number that could end up in the study, we still believe the majority of patients who consent to the MRD testing can potentially end up in our study.
Right. I'll put it this way. If they do become MRD positive, not negative, they still have minimal resistance, that most, as your number, will end up choosing to say, yeah, I would want to get this drug.
Correct.
Okay.
Especially that's the part that we had to do a lot of educational part, because that concept, just imagine from the patient perspective, when they start R-CHOP, most of the conversation is, you have a pretty aggressive lymphoma, but the good story is R-CHOP is a very effective treatment. After six cycles, your chance of getting a cure is very good. Now from there, just coming in at the end of the R-CHOP regimen, oh, by the way, I'd like you to undergo MRD testing. And if you are MRD positive, you may have to prolong your treatment. That's not the easiest conversation for physicians to have with a patient. That's why part of the exercise and implementation that we have done is try to have that conversation early on during the treatment to prepare the mindset of the patient.
That mindset is part of where they were educated about this as to why they were consenting for this test. Is that fair? Is to see if they would be eligible for a treatment?
Correct.
Right. It's not like, oh, we're taking this test and we're just collecting some data. It was you're taking this test because you would be considered for this therapy?
Because if you are MRD positive, you have an option.
Okay. All right. And then again, so we can do the math and therefore they would choose to get this off- the- shelf therapy.
Correct.
This was an earnings call a few weeks ago. At the earnings call, it was just an announcement that we're pushing out the timing of the interim, which we'll get to in a second. That interim would be pushed into first half 2026 from sort of mid. Now, have you actually tested people that are MRD positive? Have they actually gotten the drug and been randomized?
We haven't provided that level of detail, but there are definitely a number of patients that have been randomized to the study already and received cema-cel treatment.
Okay. All right. Good to know. All right. Therefore, that should be actually picking up now because you obviously said there was just an initial delay and now the sites have been ramping up. Based on what you're seeing, can you say here, sitting here in June, that things are picking up and you definitely feel very good and on track and feel a lot better than you did the prior earnings call where you guys said this was going to be on track? Actually, you guys said it was going to be a mid- readout. That is supposed to be a catalyst. I kind of got slipped up there on the next one.
Yeah. I mean, any slippage in timeline is something that we did not want to do. We were watching the study with hawk eyes to see when the hockey puck effect of the enrollment would occur. By the time we got to the earnings call, we felt that it would be better just to push out the timeline to be very transparent with investors about how the study is performing.
Sounds like you could be conservative on it. We'll see.
We already pushed out the timeline once. I don't want to do that again. Let's just keep it as a first half.
Let's put it this way. It is public disclosure in the 10-K that the interim is based on around 36 patients. Is that fair?
As I said, I'm not going to argue about the numbers or math with Mike.
It's in the 10-K. Now, what is the interim? What is it looking at? What is the, actually, is it a futility? What are we testing in this?
Yeah. So interim, there is a futility component. But as you know, most futility boundaries are set pretty low, unless something is fundamentally wrong with what you're doing.
Usually causing harm is one thing.
Yeah. So yes, we are looking at safety. We are looking at, I mean, especially keep in mind that the comparison in this study is observation. We know that CAR T has some safety issues, although that was not so prominent in our phase one study. That is one thing, just for full transparency, that we will be looking for any imbalance in safety. The second thing that we will be looking for is we are enrolling patients who are MRD positive. How many of those patients convert into MRD negativity? The data that is out there is pretty good that if you are MRD negative, your chance of having a recurrence is very low. We believe that surrogate biomarker of MRD conversion is a pretty good indicator about how the study is going. That is how we are going to look at the initial interim analysis.
I forgot 36 patients. Sorry, I got to do a little math. It's three arms. One is with the lymphodepletion, one is without, and then one's observation.
Correct.
Observation, I don't think people are going to go to MRD negativity by magic luck.
Correct.
With lymphodepletion, there could be a very good chance that they're going to go to MRD negativity based on your hypothesis of how the drug is working, which is getting rid of that last residual disease. The arm, I couldn't say with or without lymphodepletion, but the other arm that's also getting the drug, they should also get it too to convert. You think that with lymphodepletion, it could be even better. You have not said whether or not getting lymphodepletion is a good thing or a bad thing.
Here, the lymphodepletion.
Honestly, I think Wall Street is a little bit confused too. Is it good or bad? I don't know.
Okay. First of all, there's no confusion here. CAR T therapy is always done together with a lymphodepleting chemotherapy. What we are doing in the study is testing two different regimens of lymphodepletion.
Fair. Yes.
Mainly because we are treating patients with a low volume. Minimal residual disease is essentially the patient population that we are treating. If you look at our phase I data, in the patients with the relapse recurrence, these are patients with evidence of disease, what we have shown in the phase I data is complete response rate of about 60%. If you look at within those patients, according to whether they had low volume versus high volume disease, low volume disease patients, the response rate goes up to 100%. It's a small data set.
CR rate.
Yes.
The CR rate was 100%. Can't remember how many numbers. There's a handful.
Yeah. Correct.
Yeah. They all had, and not only did they all have CR, but it was durable.
Correct.
Therefore, based on basic science principle, the lower the bulkiness of the disease, the greater the T cells are going to be able to clear that. When you got MRD, we're talking about very small levels of disease.
Correct. Back to doing the math and just becoming a little bit of math discussion. The first interim analysis, when we do it, on each arm, we would like to have about 12 patients per arm. In the control arm, exactly as you said, there should be almost no one converting into MRD negativity spontaneously. In the treatment arm, the range of MRD conversion can be either around 50%- 100%. That is the sort of unknown that we like to see as the data matures.
Based on the fact that you're adding, I can't remember the name, but the number of the drug, but you're adding the anti-CD19.
So that's ALLO-647.
That's with the Flu/Cy, right?
Yeah. The lymphodepletion regimen that we are comparing is Flu/Cy, which is a standard lymphodepletion that's used in the autologous setting. The second lymphodepletion is Flu/Cy with ALLO-647, which is anti-CD52 antibody.
Why would that, that's just to help again, further reduce the rejection of the cells? That should may or may not be necessary. That's your point, right?
Exactly. Especially when the disease volume is so low. That's something that we have not tested before.
Here's what would happen when we get to this. I got to be honest, because we're trying to figure out for investors who are looking at this, okay, should we buy this and get excited about this event? On one hand, obviously it's reasonable that the drug study continues. They're going to either pick with or without 547, and we're going to continue. We got to get there. We got to make sure Allogene executes and gets there on time. We got to do that first. Second of all, it seems reasonably obvious that they're just going to pick one of those arms and keep the observation arm going. Good news. Thumbs up. You didn't fail the futility. On the other hand, there's safety things that could come up. That would only be the only other reason why things have a problem.
I mean, I got, this is not your case, but there's another CAR T company named Cargo, which ran into a big surprise because the efficacy was no good, but there were safety problems. That was a big disappointment there. Given the fact that we're dealing with fairly healthy patients, is, other than CRS or other things, what are the types of safety things that would pop up that should not be an issue here? Theoretically, this is the stuff that you have to check. That's why we're doing the interim.
Yeah. Exactly.
What type of safety stuff?
These are patients who.
It should be reduced because it's low volume.
Precisely.
You're not going to get the type of inflammatory reaction CRS that you would see in third line DLBCL.
If you look at our relapse refractory data set, when you look at the cytokine release syndrome or the neurotoxicity, we had no grade three events.
No grade three events.
Yeah. There were some grade one or two, which are milder ones. There was also a grade three infection rate that was much favorable than what Autolus CAR T has shown. The exact number is in the patient cohort that's relevant to the study. We had about 18%. In the Autolus CAR T programs, the rate was somewhere in the 20% range. The infection most likely is coming from the lymphodepletion more than anything else. Those are the kind of safety issues that we are dealing with.
Now, that's still, although these patients are definitely healthier than third line DLBCL, how do you think about infection risk? Because technically you're putting people at a risk of infection. Although, yes, obviously we're supposed to get a therapeutic benefit too, but the FDA wouldn't want to, people don't want to see things like that without evidence of efficacy.
I mean, first of all, you're giving Flu/Cy. So there is always infection risk. But infection, as long as that's within the range that we have seen in the relapse refractory setting, I don't think that is an issue. Yeah, I would agree that if there's no efficacy associated with the infection risk, that's a different story. All the data would indicate we should have a very good conversion rate. I mean, frankly, this is one of the reasons that a lot of investigators involved in the study who know the phase I data so well said, "You guys, this study is so de-risked. I mean, you have to do something wrong to fail the study.
That's what I'm saying. Like what would go wrong? It's a safety problem. Something would happen. That would derail a situation. Again, this feels fairly de-risked. Now, if we march forward, then what is, because I think I can predict that one of those arms is going to go forward versus observation, how much more time after that, those 36, do we need to continue to enroll to get to the efficacy endpoint?
The efficacy endpoint you're talking about is the next interim analysis, which is looking at the clinical endpoint. We will be doing another interim analysis. This is a statistically powered interim analysis that will look at the event-free survival, which is the primary endpoint. We will provide at the time of the first interim analysis, based on all the projections that we can gather from the enrollment, about when that will occur, rather than giving you numbers that are so many sort of unvalidated.
What's the target? The target is that the enrollment for the interim is 36. What is the number of patients needed for the actual primary?
We haven't disclosed about what those numbers are. We are intending to enroll about 240 patients to the study. Event-free survival, if you look at it, usually you'll be looking at 40%-50% having the event before doing the primary analysis. That is a norm at which most statisticians will think about when to do the primary analysis. Interim analysis will happen when those events are either about 40%-50%.
You plan to enroll 240. Certainly an interim could happen on 50% of the patients. There is an interim. You would plan to spend alpha on that because you believe that it should hit on that, right? That's the thinking?
We are spending some alpha on that.
Yeah. Yeah. Okay. All right. Another question that would come up too, and this is important in the last five minutes or so, is if this takes place and if it's positive, and we get to all that, that CBER, Vinay Prasad, and obviously just CAR T in general, there had been provocative comments by CBER, although his account has been deactivated and said that's in the past, about MRD and about CAR T. That was in myeloma, but these are comments that worry investors because they believe that certainly in a healthier population near too, because that was also the concern was that CARTITUDE-4 was in healthy. In a healthier population, how should we think about MRD? This is actually going reverse. We're trying to get people to go to MRD activity, but this is the idea of MRD.
In the absence of significant safety database, how should we think about that with this type of FDA?
Yeah. I mean, from that regard, we feel very comfortable about the study design and endpoint that we are using. Just to make it clear, ALPHA3 study is a randomized control study. That is the gold standard at which you do any kind of study. For CAR T study, that is pretty unusual. Number two, we are using very well-defined patient population who are at high risk. We use MRD assay to risk stratify and select the patients. MRD is not a clinical endpoint that we are using. The clinical endpoint that we are using is event-free survival, which is the same endpoint that has been used in the second- line study of Yescarta and Breyanzi that led to the approval of those drugs. There is a regulatory precedent of using event-free survival for drug approval.
What is the ultimate?
Also, in our study, the MRD conversion, that's not the primary endpoint. We use that to inform internal decision making, but it's not how we're going to make an argument with the agency about the benefit of cema-cel .
Right. The primary endpoint is not conversion of MRD positivity to negativity. The primary endpoint is event-free survival.
Correct.
You obviously believe that when you convert someone to MRD negativity, that that person obviously will have a longer event-free survival.
Exactly.
Okay. That is the primary endpoint. By the way, you believe that if you do not convert to MRD negativity, which the observation arm definitely is not going to have a lot of those, that they are going to fall off based on the Kaplan-Meier curve of some literature that they should fall off within six to 12 months. Because if they go from, if they are MRD positive, the data suggests that those people do convert pretty quickly. It is just that nobody tests them.
Yeah. That's a growing data. I mean, that's the initial data set that our diagnostic partner, Foresight , has published. They also recently updated using different samples, basically verifying what they have said before. MRD positivity at the end of R-CHOP treatment indicates that your chance of having a disease recurrence is very, very high.
Okay. So then this is, let me ask the next question then. The next question is, what is the clinical, I haven't memorized, what is the clinical benefit of someone who is MRD negative? So if you can get them MRD negative, what do you think their event-free survival should be versus the person, I think you just said, if you do not, then you're going to progress pretty quickly, like within six months. So what do you think the EFS curve would be for the drug arm? Pretty much flat or what?
The modeling that we are doing is that once they are converted into MRD negativity, they will remain recurrence-free.
For how long?
That we will have to see. I mean, ultimately, if you look at the existing data, they will say for the next two to three years, your chance of disease recurrence is high.
How do I know that? Just by looking at people who were on R-CHOP and who were MRD negative? What is the evidence?
That's a one line of evidence. I mean, that's published data. Another line of.
If you get R-CHOP and you're MRD negative, your EFS is like a few years.
It's like flat. I mean, you don't have a recurrence.
Right. Okay. So that's one line of evidence.
The second line of evidence is coming from retrospective analysis of people who are treated with CAR T therapy in relapse recurrent setting. The ones who achieve MRD negativity, they are the ones who have a prolonged durability of the response that goes beyond.
That's right. If you look at Yescarta or others, right? This is what we're talking about.
These are published data.
Okay. The people who get a CR with MRD negativity, they have significantly better EFS than people who are CR without MRD negativity? Let's just start with that.
I mean, I can tell you the available data would indicate patients with CR, but MRD positive, they will relapse quickly.
Interesting.
The reverse is also true if you're an MRD negative.
I think that's because most people on Yescarta, if they do get the CR, isn't most of them MRD negative? What percent is that? You would know. 70%, 60%? It's got to be most because people get a CR with Yescarta, they get past the first three months. We know that they can have a long duration.
Yeah. So if you look at.
It's most.
Not the Yescarta. If you look at across all the approved autologous CD19 CAR T, the initial CR rate is between 50%-60%.
Right.
However, over a period of next six months, some of them do progress.
Coming out at 35%. 55% coming out. I say most, I'd say half. And that's your delineation between being MRD negative and MRD positive.
Also, the retrospective analysis would in a way predict what CRs are, insufficient CRs with a CAR T therapy.
We just went to ASCO, and it's interesting because one of the, well, it was the oral plenary at ASCO. The oral plenary on Sunday was in breast cancer. They talked about testing for ESR positive molecular breast cancer just by ctDNA. Think about that. That's kind of interesting. No radiographic tumor, just looking at molecular testing of ctDNA, kind of like MRD positivity. If you do, we switch you to an amazing new SERD, which would treat you rather than observation. I'll check that out. That had a huge seven-month PFS, hit the primary endpoint, hazard ratio 0.44. Isn't that kind of analogous to it?
Their only question in the discussion, realizing that we all agree it was a positive study, and they said, "We'll consider how that could be used if approved," that it's almost like we got to test for PFS too. Here, the question is, "Oh, if you waited for the person to progress and there were an observation after six months, what happens if I give them Allogene after that or Yescarta? How would that PFS look?" That's an interesting one. We're going to spend time thinking about that another time.
Precisely. I think that's a secondary question after we conduct the ALPHA3 trial. If the data is positive, there will be questions that are coming up on how to best utilize the MRD information.
If they did progress, isn't the standard of care to then get Yescarta in the second line?
That is.
Correct.
Is that what people are going to ultimately do in your study? Because if they do progress in their observation, they do not cross over to Allogene, they would go and get probably Yescarta?
Yeah. If you are in the observation arm, and it also applies to people who are in the treatment arm, if they have a progression event, they can go into any of the available treatment. Obviously, if you are transplant eligible and if you can get access to CAR T, there will be an option. If not, there's a second line chemotherapy regimen that I used, including so-called bone marrow transplantation.
That would be the way to sort of look at PFS too if that was the case, but that would be getting another CAR T.
Yeah. And frankly, in the breast cancer, where the follow-up is much longer, you can talk about PFS too. But in the lymphoma setting, it will be more like overall survival.
Overall survival. Okay. First of all, thank you very much again, David, for spending time. I thought we did a great job walking through the dynamics of that and why your result should be positive, both at the interim, well, fertility and then the interim, and why it would be transformative if we can get people to MRD negativity. So thank you very much.
I mean, I would just say ALPHA3 is really innovative design. The field is moving to MRD-based stratification. I do believe ultimately MRD will be used as an endpoint, but how long it will take? It will take some time.
Usually it starts with a trial that ran the results and showed you that it happened. That's how you changed the standard of care.
Yeah. It's at the very beginning. We are at the forefront of the beginning.
Thank you very much, David. Appreciate it.
Thank you. Thanks for the insightful questions.
Thanks.