Good day, and thank you for standing by. Welcome to the ASPEN-06 ASCO GI Data Review. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question-and-answer session. To ask a question during the session, you will need to press star 101 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press star 101 again. Please be advised that today's conference is being recorded. I will now turn the conference over to our CEO, Jason Lettmann. Please go ahead.
Great. Thanks for having us. Welcome, everybody. Today's goal is to share the updated results from our gastric study and ASPEN-06 from our phase II study, and very excited to share those today and looking forward to walking you all through it. These are our forward-looking statements on slide two. As we get into the agenda here, the goal will be for me just to do a quick reminder of evorpacept and of ASPEN-06. Following that, Alan Sandler, our CMO, will then walk through the ASPEN-06 results, and then we'll finally close with some final remarks and some Q&A. As a reminder, evorpacept is the only CD47 to deliver randomized data to date. Again, very excited about what we're seeing, and again, this follows on to our large clinical study where we've developed some really powerful results.
Again, we're advancing this across breast, across gastric, across head and neck, across multiple myeloma, and bladder cancer as well. As a reminder, on slide five, again, we are transforming cancer treatment for patients by developing evorpacept, which is a first-in-class foundational checkpoint therapy. Again, this is a very differentiated asset. We've now studied this across a number of different trials, and we've proven safety in over 700 patients to date. Evorpacept is the first and only CD47 agent to demonstrate durable improvement. And we're going to talk about that today and why we're excited about what we're seeing here with ASPEN-06. Again, this is a differentiated mechanism. We are the only CD47 in development with an inactive Fc. And what we're going to talk about today, and one of the things we're most excited about, is our clear biomarker strategy.
As I mentioned, we have multiple positive clinical studies across bladder, NHL. And importantly, we've demonstrated efficacy across both solid and heme. We're now expanding evorpacept to new indications supported by multiple pharma partnerships. And we have a strong pipeline to talk about and a strong balance sheet. And again, as we shared in our data update in San Antonio, we are now putting up data here across a number of tumor types. And in December, we shared data to show what we could do in breast cancer. And today, we're going to show updated data from ASPEN-06 and HER2-positive gastric cancer. So again, this is how evorpacept was designed. Again, a very differentiated asset, in which you can see here on slide six, is the key attributes of the molecule. Again, we are very potently blocking CD47. We have a dead Fc domain by design.
We have a low molecular weight and antibody-like PK. This is what is driving why we are seeing such a differentiated profile in the clinic, both from a safety and efficacy perspective. And as we'll talk about today, evorpacept is really designed as a combination agent. We work best in combination. Therefore, we need to both block the don't-eat-me signal as well as have a strong eat-me signal. And in this case, in gastric cancer, we need strong HER2 expression. And it's a very important attribute of how our drug was designed. And again, importantly, it's how it's a driver to what we're seeing now in the clinic. So quick reminder of the development plan. Again, ASPEN-06 is our most advanced program. We're going to focus on that today.
Quickly, we're advancing this in ASPEN-03 and ASPEN-04 with two frontline studies in head and neck, ASPEN-07 in bladder. We have a series of partnerships that have furthered our understanding of this drug and mechanism. On slide eight, one of the things we want to point to, again, as we talk about this data, is the totality of the data set. Again, we've now seen activity across six different clinical trials to date across both heme and solid malignancies. This strength of this data across so many different tumor types with different combinations is really what's driving our enthusiasm and encouragement around this program. With that, I'm going to turn it over to our CMO, Dr. Alan Sandler, who will walk through the data from our ASPEN-06 study.
Great. Well, thank you, Jason. And thanks, everyone, for taking a part in this conference. So the next slide is a nice picture, I think, of the mechanism of action, specifically evorpacept with Herceptin. So again, given the fact that the silent Fc receptor, we're able to block CD47, as you see here, but also require the partnership of Herceptin to block the eat-me signal, in essence, painting the tumor cell so that you have a more specified approach to the killing of the cancer cells with less collateral damage with normal cells. On the next slide, this is the schema for ASPEN- 06. I'll spend a couple of minutes looking at this or walking you through this slide, if you will. So again, this was a randomized phase II study designed to test the contribution of components for evorpacept in combination with TRP.
Of course, the T is trastuzumab. The issue is it was added to the control arm to ensure that that was not contributing to the effects of the control arm, and we could identify and isolate evorpacept's effect. The key eligibility criteria is shown on the left. It's HER2-positive gastric cancer or GE junction cancer that had progressed on or after prior HER2-directed therapy, limited to second or third line. Prior Enhertu and/or checkpoint inhibitors were allowed. However, prior CD47 or anti-SIRP alpha agents or ramucirumab were excluded. Patients were entered based on either archival or fresh biopsy results showing HER2 positivity. Primary endpoints are noted on the far right, and there are four. They are in both the ITT and fresh biopsy populations. They were based against a historical control of 30% or looking for an absolute increase of 10% over the internal control.
Secondary endpoints are listed. Importantly, again, all patients enrolled received a prior HER2-targeted therapy and were allowed to be enrolled with either fresh or archival biopsies. Next slide. This slide shows the demographics for this randomized phase II study. Simply, the two arms were well-balanced. And I'd like to spend a few seconds on the blue box here, which refers to the HER2 status overall, IHC3 or IHC2 plus ISH positive, and also describes the number of patients with fresh or ctDNA HER2 positivity. Noteworthy, fresh HER2 biopsies were obtained at a median of 1.1 months before dosing as compared to 14.1 months for patients with an archival biopsy. ctDNA was assessed for HER2 amplification and collected on cycle one, day one, just prior to dosing. And we used the Guardant360 comprehensive genomic profiling. Both fresh biopsy and ctDNA were pre-specified in the ASPEN-06 statistical analysis plan.
Fresh biopsy was defined as a primary endpoint. ctDNA analysis was defined as an exploratory endpoint. Next slide. This is the safety slide for ASPEN- 06. In summary, the evorpacept TRP was generally well tolerated with very few grade three TEAEs. Those were largely balanced across the two arms. There were 11 grade five treatment emergent adverse events, four on the ETRP arm, seven for TRP, however, only two of which were deemed to be treatment-related, one on each arm. Again, the safety profile for evorpacept was consistent with the prior experience in over 700 patients treated to date. There were no new safety signals noted. The next slide. evorpacept added substantial response activity to the TRP backbone in the ITT population, as shown on this slide. These are all confirmed response rates for both arms.
You see a response rate of 41.3% in the Evo arm and 26.6% in the TRP arm. Noteworthy, as of December 2, 2024, cutoff, and since the prior May cutoff, two additional patients achieved a confirmed response in the Evo TRP treatment arm. No additional responses were noted in the TRP treatment arm. Responses remained durable as the Evo plus TRP median DOR remained unchanged at 15.7 months as compared to 9.1 months in the control arm. 14 patients remain on treatment with the Evo plus TRP arm, including three responders ongoing treatment for over two years. This compares favorably to only nine patients remaining on treatment, the TRP arm. You'll also note about nearly one and a half years median follow-up for the two arms. Next slide, please. So this follow-up or build onto this slide now shows the PFS Kaplan-Meier curve for the ITT population.
You'll note that the medians are similar at approximately 7.5 months. However, shortly thereafter, the curves begin to separate, and when you look at the 12-month rate, you see a 38% number of patients alive without death or progression on the Evo arm as compared to 23% for the TRP arm. This, of course, is greater than a 50% relative increase in the number of patients alive and disease-free at that time. We'll move to the next slide, so the next slide shows substantial tumor shrinkage in ASPEN- 06 patients receiving the Evo TRP when compared to TRP. The waterfall plot on the left is the Evo arm as compared to the control arm on the right. Next slide. Responses with evorpacept were durable and consistent with an IO mechanism.
This is illustrated here in the swim-lane plots showing a median DOR on the left for the Evo arm at 15.7 months and 9.1 months median DOR for the control arm, and I think this gives a nice picture of the number of patients that remain on study and also the overall view of the number of patients with durable responses. On the next slide, we're now going to spend the next several slides. We're going to be talking about those patients that are known to be HER2 expression positive. Remember that the previous data that I showed are patients that were felt to be HER2 positive. However, HER2 expression is known to be highly variable in gastric cancer. This expression change can be due to the following: a loss of HER2 expression following prior HER2-targeted treatment.
Remember, those patients with archival tissue had over a year between their tissue was obtained and commencement of treatment, meaning there may have been even more than one HER2-directed therapy in the interim. HER2 expression is highly variable within the tumor known as tumor heterogeneity. Also, decreased HER2 expression following treatment with trastuzumab or other HER2-targeted agents has been observed in 16%-32% of patients with gastric cancer specifically. Therefore, confirming HER2 positivity with a fresh biopsy or via ctDNA should result in a more enriched HER2-positive population. HER2 expression was measured in this study via two well-established methods: IHC via tissue biopsy and ctDNA, as we described earlier, via liquid biopsy. Next slide. Now, this slide shows the pattern of patients starting with the ITT and ends up with the number of patients felt to be ctDNA positive.
As we look through the number of patients that had fresh tissue, about one-third of the patients on study, a little more than one-third versus archival. And then also looking at HER2 amplification via ctDNA, and it's broken down thusly. Remember that patients were eligible for enrollment with either archival or fresh HER2 biopsy. Response rate in the fresh biopsy population was a primary endpoint. ctDNA analysis suggests that about one-third of the archival biopsy patients were not actually HER2 positive and thus unlikely to respond to evorpacept. Patients with HER2 positivity based on a fresh biopsy or ctDNA analysis were more likely to derive benefit from evorpacept TRP as expected based on evorpacept's mechanism of action. And I'll spend the next several slides illustrating that point. Next slide. So evorpacept greatly improved the response rate in patients with confirmed HER2 positivity.
And what this slide shows here is a table that shows HER2 confirmed fresh biopsies between the two arms. And then on the right, we also show HER2 confirmation with fresh biopsy or ctDNA. This greatly enhanced the number of patients for evaluation with HER2 expression. Let's look at the fresh biopsy where you see now a response rate of 59% versus 23% in the control arm. Note that the impact of fresh biopsies did not impact the control arm. Median DOR is 15.7 months in the Evo arm and 14.5 months in the TRP. Now, looking on the right at the HER2 confirmation with fresh biopsy or ctDNA, that response rate is nearly 49% as compared to 24.5%. Again, no impact on the control arm in that setting.
When you have the larger patient population of fresh biopsy or ctDNA, you now note a difference in the median DOR of 15.7 versus 9.1 months. Let's go to the next slide. Now, the next slide shows progression-free survival based on investigator assessment. And this data suggests that overall response rate is translating to improved progression-free survival in patients with fresh biopsy or ctDNA positivity. On the left, HER2 confirmed with fresh biopsies N is noted here at 48. Again, here you start to see separation with the median PFS of 9.5 months in the Evo containing arm versus 7.1 months in the control. And again, I draw your attention to the 12-month rate for PFS, where you have more than a doubling of the number of patients alive at one year without death or disease progression. That hazard ratio also has improved to 0.62.
Now, let's look at the Kaplan-Meier curve on the right. This is the HER2 confirmation with fresh biopsy or ctDNA positivity, and it's now 96 patients, which is nearly 80% of the patients on study. Here, the median PFS is 7.5 versus 6.7 months. You'll note the undulation of the curve in this Kaplan-Meier, where it kind of waxes and wanes over time, but as you start to move away from that median, you see clear separation of the curve that are maintained throughout the Kaplan-Meier curve, and again, moving to the 12-month rate for PFS, 40% for the Evo arm and 21% for the control arm. And the hazard ratio, again, consistent at 0.64, so let's move to the next slide. This is, of course, a waterfall plot. However, this is a waterfall plot only with those patients with HER2 confirmation with fresh biopsy or ctDNA positivity.
Again, you see the strong depth of response in patients with HER2 positivity confirmed in this fashion. Next slide. The next slide, swim-lane plot for patients with HER2 positivity, again, confirmed with fresh biopsy or the ctDNA positivity. Again, now demonstrating the durable activity in this patient population, and I think that this, it's clear to look at this slide and note the difference from the ITT population. This looks even more dramatically different between the two arms. Median DORs are listed. We've mentioned those before as well, but noteworthy, 15.7 months in the Evo, 9.1 months in the control arm. Move to the next slide. This is a table that may be a little bit busy, but what we wanted to show you here is that the results of Evo plus TRP compare favorably to the benchmarks in the second-line plus treatment with second-line plus treatment regimens.
You have on the top half the ASPEN-06 data for fresh biopsy or ctDNA by treatment arm shown here, summarized what we've summarized earlier for you and put it in context with that of the RAINBOW study, which showed the approval for ramucirumab and paclitaxel. You see a better response rate. You see a dramatic improvement in DOR of 15.7 versus 4.4 months and also improvement in the median of 7.5 versus 4.5 months. That's the median PFS, I'm sorry. In addition, DESTINY-Gastric01 is shown here as well with Enhertu, where the response rate is 49% for the Evo arm versus 41% for Enhertu, DOR of 15.7 versus 11.3 months, and also PFS of 7.5 versus 5.6 months. Of course, I didn't mention earlier, but you have to mention that these are cross-trial comparisons.
But we just want to at least provide you with a glimpse of that and how well we believe the evorpacept containing arm is doing. Lastly, the bottom half of the study, this is fresh biopsy-only patients, which with the evorpacept TRP and then the TRP arm shown, and also Enhertu, which in DESTINY-Gastric02 was a requirement for that study. So this is a bit more of an apples to apples comparison. You'll note the response rate for the evorpacept containing arm of 59.1% versus 42% for Enhertu, DOR 15.7 versus 8.1, median PFS 9.5 months versus 5.6 months. Next slide. So I believe this is my last slide. Evorpacept demonstrates the power of engaging the innate immune system in patients with previously treated gastric and GEJ cancer. We've shown you robust and durable clinical activity, validated mechanism of action.
This agent, again, is well tolerated in this combination, and this is now the only CD47 agent to demonstrate both durable improvement in overall response rate and a well-tolerated safety profile in a prospective randomized study. Next slide.
Great. Thanks, Alan. Just to wrap up here, I think what we're also very encouraged is that this phenomenon, if you will, is not new, and again, we recently updated you all with our data in breast cancer at San Antonio last month. There, with combining with another Fc-active antibody, zanidatamab, we showed a very similar effect, so again, we've now, across two different studies, two different Fc-active antibodies and two different tumor types, shown that we can have a meaningful and durable effect. Again, I think this validates our mechanism. Importantly, we now have a clear biomarker to select the patients that are going to perform best.
Again, I think the safety here is truly different, particularly when you compare with the other CD47s with an active Fc that have been in development in the past. And then last but not least, as we highlighted here with this study, as well as in the breast study, both of these studies show good activity in patients that had progressed on prior HER2 targeted agents. So in terms of what's next for this program, very excited about what we're seeing. We're planning and actively engaging with FDA. And we are going to discuss with them plans to move this forward. And we're going to plan to share next steps and guidance with you all at some point in Q2 of this year. We're also planning to host R&D Day next month. There, the goal will really be to lay out this path to registration in combination with anticancer antibodies.
As you'll recall, the anticancer antibody combination story is one of our three large shots on goal. We think this really validates and furthers our conviction in the antibody story. Therefore, we're very much excited about sharing with you all our plans to take this forward and drive our drug to registration. So with that, I will pause and open it up to any questions on the line.
Thank you. And as a reminder to ask a question, you need to press star 11 on your telephone and wait for your name to be announced. To withdraw your question, please press star 11 again. Please stand by. We'll compile the Q&A roster. One moment for our first question. Our first question will come from the line of Michael Yee from Jefferies. Your line is open.
Hey, guys. Good afternoon. Thanks for the presentation. I guess I had two questions.
The first question was just thinking about your message around the effects seen with fresh biopsy and/or ctDNA, which I guess is really slide 20 and 21 and 22. And I think maybe the market is trying to figure out if things are so much better, why is the PFS, I guess, on slide 21 not hugely different from a median perspective? And maybe you might say, well, it's a hazard ratio. But then I think when you look at Kaplan-Meier curves, you can see that a lot of the events are in the first 12 months. So despite a hazard ratio, it's really driven by sort of maybe a proportion at the tail. So could you maybe just triangulate your confidence in the effect if the median isn't woefully different between the two arms? And my second question is just what the next step would be.
If you're so confident in the results, do you fully expect to finance and be able to run a phase III? Thank you.
Yep. Great. Thanks, Mike. Appreciate the question. I think on the data on the patients that are, again, confirmed HER2 positive with either positive fresh biopsy or ctDNA, I think what's really exciting for us is to look at the entire curve, right? And median PFS is a point estimate on the curve. But oftentimes, when evaluating an IO drug, that can be misleading, right? We've seen that over the years with many IO drugs, including Pembro. So what we're excited about, as you mentioned, is the hazard ratios here.
Again, looking at HER2 expression two different ways to see a hazard ratio of 0.62 in 48 patients and then 0.64 in 96 patients, which represents over 75% of the study, really speaks to the strength of this data. So certainly, we see a tail here. I think from our perspective, we're excited about the tail. Again, that's what you hope to see with an IO mechanism, and we see it translating. Alan, do you want to add to that?
Yeah. Hi, Mike. I just want to add, obviously, everything I agree with Jason's comments. And just one other. And I tried to mention it during the presentation as well. But if you're able to look at that slide or have it in your memory bank, the curves sort of undulate, particularly the curve for the Evo arm, and they actually are separated early on in the curve.
They just sort of kiss right at around that 50% mark. And then they separate again. And that's just the natural undulations that you see on the Kaplan-Meier curve. And with a hazard ratio of 0.64, I think we're very confident with the contribution of Evo to this regimen.
Yeah. And then I think on your next question, Mike, just in terms of our plans going forward, I think our immediate next step is to discuss with the FDA how to proceed. I think we're going to certainly introduce the notion of an accelerated approval here, given the strength of our data, both versus historical benchmarks, as Alan walked through, as well as versus our control. And I think the other thing that we're looking forward to discussing with them is this biomarker-selected population. I think, as you'll recall, the design here is a phase II, III design.
So we do have agreement from FDA in prior discussions around running a pivotal versus RamPac. I think it's smart and wise to go and revisit that because here, I think we now are able to define a pivotal study that is different and informed by the biomarker. And again, that's what you want in oncology. We think we know exactly where this drug works best. And we're looking forward to talking to the agency about that. Okay.
Thank you.
Thank you. One moment for our next question. Our next question comes from the line of Chris Raymond from Piper Sandler. Your line is open.
Hey. Thanks. And maybe just this might be a little bit of a cart before the horse question because I know you still have to talk to the FDA.
But just on the go-forward path and thinking about the trade-off between using fresh biopsy and ctDNA, obviously, ctDNA is much easier to implement than fresh biopsy. But fresh biopsy is sort of the gold standard, right? And I think you give off something on accuracy with ctDNA. How do you guys think about using one versus the other? Is it too early really to make a determination? And how do you feel about the enrollment dynamics if you decide to go the fresh biopsy route?
Yeah. No, thanks, Chris. I think it's a great question. I think one of the questions that we had coming out of our July update was, how else could you look at HER2 expression, right?
So I think first and foremost, what we're most intrigued with is when you take a totally different approach to measuring expression, i.e., with a liquid biopsy, how does that data look? And so for us to see what we're seeing across, again, two different measures of HER2 expression, to see them line up exceeded our expectations going into looking at a blood-based measure of HER2 expression. I think importantly, we want to get FDA's feedback. I think as everyone likely recognizes, ctDNA has come a long way over the years. It's used regularly in studies and in the commercial setting. And again, what we're talking about here and what we're excited about is an OR, right? So it's a fresh OR ctDNA. And again, to see that in roughly three quarters of the study here is a really powerful signal.
So I'd say more to come after conversations with FDA, but certainly excited about sharing this.
Thanks.
Thank you. One moment for our next question. And our next question will come from the line of Sam Slutsky from LifeSci Capital. Your line is open.
Hey. Good evening, everyone. Thanks for the questions. Just two for me, I guess. One, could you just elaborate a little more on what you think happened between the interim analysis in terms of the updates on response rate and DOR versus kind of the second half of the study? And then I guess also just anything to say on overall survival and how that's trending in the study?
Yeah. Sure. Thanks, Sam. Both good questions. So we've given a lot of thought to interim versus post. Obviously, heard the feedback from investors in July that there were some questions around that.
I think at the time of the July update, we spoke to the rate of HER2 fresh biopsy, if you will, between pre and post interim. I think since then, we've prepared more data, so on this slide that we have up now, we asked the same question but looked at ctDNA, so what you can see here on the chart is the ctDNA analysis of pre-interim on the left and post-interim on the right for the Evo TRP arm. What we had hypothesized back in July is that there was a higher rate of HER2 negative patients post-interim. It makes sense for a lot of reasons, and it makes sense in terms of what we were hearing from the sites, and in fact, the ctDNA analysis supports that.
If you see on the left, we had about a 20% rate of ctDNA negative versus an almost doubling of that in the post-interim population of closer to 40%, and as we've pointed out, for patients that are truly HER2 negative, our drug will have limited to no activity by design, right? We need the eat-me signal, and we need to block the don't eat me signal, and again, we think that's translating into the difference in those two populations. The second point I want to make on the bottom right of this slide is that regardless of how you cut the data on fresh biopsy, whether it's interim, post-interim, or ITT, the ORR and signal is strong with very little deviation on the control arm, so you can see the fresh ORR at interim was about 67% versus 25% on control. The post-interim population was 54% versus 20%.
And then if you look at the full population, that's the 59% versus 23% number. So again, I think what's encouraging for us pre, post, full, regardless of how you cut it, that signal is strong. And again, I think this is a strong driver as to what we saw in terms of the differences post-interim in the ITT population. On your second question on OS, that, of course, is event-driven. We're looking at events. I think as that data matures and as we get to the level of events that we'll need to call it mature, then we'll give an update. It's, of course, difficult to predict. But that's something we're going to continue to monitor as events come in.
Thank you. One moment for our next question. Our next question will come from the line of Lu Tang from UBS. Your line is open.
Oh, hi. Good evening.
And thanks for taking time to walk us through the data tonight. I have one main question. So how convincing are the 12-month PFS rates in telling the potential progression-free benefit given the MDOR greatly exceeded 12 months in ITT and was about 9 months in the TRP control arm? So those who respond to eTRP are then supposed to respond longer. And given they're 40% of the ITT, we'd assume this population will hold the 12-month PFS rate high. So I thought this is more of a derivative effect of the DOR versus looking at PFS rates at later time points or the overall survival rate to be a more independent metric.
Yep. Great. Thanks, Tanguy. We always kind of knew for the detailed question, a good question for sure.
Again, I think what I'm hearing you ask is around how is this going to flow through the data in terms of the correlation between DOR and 12-month rate? Is that correct?
Yes. Yeah. I thought 12-month PFS rate is more like a derivative effect from the DOR given your treatment arm has a DOR exceeded 12 months, but the control arm was under 12 months for the ITT population.
Yeah. I think that is an important question. And again, I think there would potentially be flux in those numbers if you were looking at an earlier data set, i.e., a less mature data set. I think what gives us conviction that these numbers are real is the median follow-up, right, which is 17 months on this study. So therefore, I think we have a lot of conviction in these being stable, if you will.
But I'll let Alan add to that as well. Yeah.
Hi. So yeah, a very interesting question. The only thing that I would add is, of course, something that you're very well aware of, that PFS includes patients with stable disease, patients that have not progressed, as opposed to DOR, which are responders only. So it's sort of a subset of what you might find in the PFS that may have been your derivative comment that you made. So I think both of those provide evidence for the durability of response and also the prevention of disease recurrence. And I think that both of them providing, again, confidence in what Evo is bringing to the table here.
Just one thing to add on that, Tanguy. I think importantly is this is now a December data cut, right? So we've added seven months since May.
I think a lot of the questions we heard from investors is, how will the data mature? And what we've shared is we've added a couple of responses on the treatment arm. The DOR has continued to be stable, which is great. And so I think as we're stacking months here, if you will, to a more mature data set, it's been continuing to hold. So hopefully, that answers your question.
Yeah. Yeah. That's helpful. Thanks, Jason and Alan. I have a quick follow-up. So I noticed the median follow-up was actually slightly longer in the eTRP arm versus the TRP control arm. However, there are considerably higher percent of patients censored in all three assessment groups in ITT, fresh biopsies, or those with confirmed HER2 expression. So what happened to those patients being censored?
Yeah. So great question. So the median follow-up is 17.5 versus 16.8 months.
So pretty close. I think on the censoring question, what you hope to see, of course, on the treatment arm is less events. And I think that's what we're seeing. Again, that's what we hope to see. But I don't know if Alan, do you want to add to that?
Yeah. I assume that's what you're looking at the censored events on the Kaplan-Meier curves, those are just patients who haven't had an event yet and should be greater on the Evo arm.
Okay. Okay. Sounds good. Yeah. Thanks. That's all my questions for today. Thanks.
Thank you. Thanks, Dave.
Thank you. One moment for our next question. Our next question will come from the line of Brad Canino from Stifel. Your line is open.
Hi. Good evening. Thanks for the questions.
Maybe on the patient selection strategy with the IHC selection of fresh biopsy or ctDNA, it looks like you're getting about 75% of your ITT falling into that category. Do you think that's representative of the broader population of HER2-positive gastric? And then what would you expect that proportion to be for HER2-positive breast if you apply the same patient selection criteria?
Yeah. I think it's a great question. So we'll take that in two parts. I mean, I think on the gastric side, there's a great paper from Dr. Shitara, who is actually our PI, who is presenting today. It's a paper in Nature Medicine from 2024, which looks at this. And in that paper, he showed that about 16%-32% of the patients had lost expression.
So again, when you line that up to what we're seeing with all the caveats, Brad, of it being a smaller study, blah, blah, blah, I think we're about what you'd expect in gastric. Again, this is a known phenomenon in gastric. And again, it speaks to the importance of fresh. And frankly, it's why with other studies like Daiichi's study with Enhertu, they required fresh. So we think what we're seeing is reflective of what's happening in the real world. On the breast side, I think we feel like this is less of an issue, if you will. But again, it's very important for us to also make sure we have confirmed HER2 expression. And that's why in the study that we reported with zanidatamab in December in breast, we were very keen to look at this patient population with centrally confirmed HER2 positivity.
There, as you know, yet again, is where we saw the strongest signal. We think the read-through to breast, if you will, is very strong. I think it's good learning for us, if you will, in terms of how we take this forward and how we design future studies.
Okay. Then the experience you're seeing with the late curve separation, it kind of looks like around month 10. How do you like to explain that with the mechanism at play here of enhanced phagocytosis? Thank you.
Yeah. I mean, again, I think this is also consistent with our broader data set. It's always tempting, albeit dangerous in single-arm studies, to try to conclude that you're seeing a tail. But we've seen signals that we are seeing similar across multiple studies. Here, I think when you think about driving ADCP, again, that should be a long-term effect.
And just like any immune modulating agent, we would expect to see a tail. And I think that's what we're seeing. And again, randomized data always speaks. And to see it in a randomized setting is super powerful.
Thanks again.
Thanks, Brad.
Thank you. As a reminder, that's star 101 for questions. One moment for our next question. Our next question will come from the line of RK from H.C. Wainwright. Your line is open.
Thank you. Thanks, Jason, for doing this. Just trying to look through some of the responses that you have. Do you know if the distribution between second-line and third-line patients between the arms is about similar, especially between the CRs and the PRs?
It's a good question. I think when you look at the demographics overall, the majority of the population was second line, as you know.
In the third line, we had about 22% of our patients in the Evo arm were third line versus 31% in control. So arguably for that metric, the deck was slightly stacked against us, if you will. I don't have the answer off the top of my head in terms of the breakdown across those two lines. I'll put Alan on the spot here to see if he does. But we may need to circle back with you on that.
Yeah. Sure. It's a good question. I don't have the specific numbers. But not surprisingly, the second-line patients did perform a little bit better than the third line. We can get back to you with precise numbers.
Okay. And then based on the data that you're seeing here and you're looking at other drugs, what seems to be appropriate combinations for phase three?
I know you will tell us a little bit more later, but I'm just trying to figure out what you're thinking as of now.
Yeah. I think it's a great question, and that's an area we've been focused on, frankly, for the last, I'd say, 18 months, RK. Because what this data, again, does is demonstrate what we can do with an anticancer antibody. How we are approaching CD47 is very novel. It is novel mechanistically than a conventional approach with an active FC. And what we know now is with an FC-active antibody, we feel this drug works, and that unlocks many, many indications, right? We've talked about breast. Again, we showed what we can do in breast already. So we think advancing there would make sense, but we also think this is applicable outside of HER2.
There's a big population of HER2 patients that progress on standard HER2 therapy. So of course, that's a natural place to go. But I think when you think about other Fc-active antibodies like cetuximab, for example, how would we combine there? And I think mechanistically, what's beautiful about this drug is it should work. And so for us, again, gastric is important. But what this really does is de-risk our entire development path with anticancer antibodies. And that's what we're going to share with you all in February at the R&D day and really lay out the plan to execute on that.
Thank you. Thanks, Alan. Thanks, Jason, for taking this.
Thanks very much.
Thank you. One moment for our next question. And our next question will come from Li Watsek from Cantor Fitzgerald. Your line is open. Hi, team.
Thanks for taking my questions.
I wonder if you can comment on the PFS from the control arm and what might have contributed to the better PFS versus the RAINBOW study, and the second question is just thinking about different outcomes from the regulatory discussion in Q2. What would be the best case and base case scenarios for you and ASCO possibility?
Yeah, so on the first one, I think it's a great question. I think it wasn't lost on us when we looked at our control arm that our control was, frankly, performing better than some of the benchmarks. Again, I think as Alan often reminds us, that's what you see in clinical trials is improvement in care, which of course is great for patients. I think one note to make here is that this study is a first in many ways.
But one of the things from a broad perspective in gastric is this is the first randomized study that we're aware of in the second line post-pembro being approved in the front line. So we do think it's reasonable that because of improving care in the front line, that's impacting what we see here, which again is why you run controlled studies and why we're excited both how we stack up with our control arm as well as with the benchmarks in the space. So I think that was your first question. On your second question, I think we do feel like AA and accelerated approval is a reasonable ask at this point. Of course, we're not going to play FDA or try to handicap that in any way.
But I think for an AA, and Alan can certainly talk about this too, we hope to see an improvement versus control and an improvement versus the big studies that set standard of care. And we feel like we've met that. Alan, do you want to add?
Yeah. Just the key elements for accelerated approval are non-overlapping confidence intervals typically based on a single arm study. But if you look at Evo as we compare to control, particularly with RAMPAC, I think we've shown that in this particular study. And then also evidence of durability, typically looking for DOR in excess of six months, which we've shown as well. So we think that there's enthusiasm with the data that we've shown here today that we're excited about an upcoming visit with the FDA for the discussions.
All right. Thank you.
Great. So just to wrap up, thanks again, everyone.
Again, very excited about this data set, excited about our next steps, and very much appreciate the engagement. This is a novel mechanism, and certainly, we're excited about seeing how this is translating in the clinic and yet again another positive study, so again, appreciate the interest in the program and look forward to future updates. Thanks.
Thank you for your participation in today's conference. This does conclude the program. You may now disconnect. Everyone, have a great day.