ALX Oncology Holdings Inc. (ALXO) Earnings Call Transcript & Summary

Welcome, everyone, to the 31st Annual Healthcare Conference. I'm Tiago Fauth, Biotech Analyst here at Credit Suisse. We're joined today by ALX Oncology. Sophia Randolph, CMO of the company. We're going to do a presentation with 5 to 10 minutes Q&A at the end of the presentation. So Sophia taking away.

Thank you. So as we just said, I'm Sophia Randolph, I'm Chief Medical Officer here at ALX, and happy to give you an update on our program. Here are disclaimers. So here at ALX, we're working on advancing a very differentiated pipeline. And our lead molecule is evorpacept, which is a CD47 innate system checkpoint inhibitor. Evorpacept is designed to really unleash the full immune response against the patient's cancer. So it exclusively inhibits CD47 as a myeloid checkpoint inhibitor, but does not target CD47 as a cancer-specific antigen. And because of that, we can activate the innate immune system against the patient's cancer, but we spare healthy normal cells that may also be expressing CD47, and I'll explain that a little bit more as we get into the data. So evorpacept, we have demonstrated activity both in solid as well as liquid tumors. We have 3 ongoing randomized Phase II studies in solid tumors, as well as 2 studies that we've just begun looking at evorpacept in combination with antibody drug conjugates. And then we have our hematology studies that are ongoing. We continue to build our pipeline, where [ we initiating ] a new IND at the beginning of the year with a molecule called ALTA-002, which is a SIRPalpha TLR9 agonist. And in addition, working on some ADC platform in the preclinical space. And currently, we have a strong financial position with cash equivalents of about $293 million as of the end of September, a $100 million loan facility available to us. And together, these things have extended our cash run rate through mid-2025. So CD47 is a target that is expressed very broadly on healthy tissues as you can see in the black histograms here. And cancer cells take advantage of that. And they upregulate overexpressed CD47 as a marker of self as a way to evade the immune system. So for macrophage, which is the business end of the innate immune system for it to phagocytose cancer cell, it needs 2 signals. The first is a prophagocytic signal that comes from an anti-cancer therapeutics, such as Herceptin or Rituximab. And that comes from the engagement of that antibody into the Fc gamma receptor on the macrophage. But those anti-cancer therapeutics must overcome the CD47 checkpoint in order to be effective, and that's this don't eat me signal shown in red. So other CD47 blocking agents try to accomplish just by using CD47 as a tumor-associated antigen. So most of the other CD47 agents out there will try to provide those 2 signals on the same molecule. So they'll block the CD47 SIRPalpha interaction on one end, and then they have an active Fc on the other end that can engage the Fc gamma receptor on the macrophage. So they're blocking that checkpoint and they're providing the prophagocytic signal in the same molecule, which is great if you're a cancer cell expressing CD47, but the problem with this, as I mentioned is, CD47 is broadly expressed. So it's expressed on red blood cells, platelets, neutrophils. And so you end up activating that macrophage against healthy normal cells, and this can result in cytopenias, anemia, thrombocytopenia and neutropenia in the clinic. In addition, when combined with an anti-cancer therapeutics, such as anti-cancer targeted antibody, you can see that theoretically, you've got competition for that Fc-gamma receptor between the CD47 blocker and the cancer therapeutic. And so together, these dose limitations, the competition for the Fc makes us less optimal as we believe as an approach to anti-cancer therapy. So evorpacept takes a different approach. So what we are doing here is only blocking the CD47 SIRPalpha interaction, so we're targeting that myeloid checkpoint. The Fc gamma portion of evorpacept is dead, it is inactive Fc domain. So it cannot provide that second prophagocytic signal. So because of that, normal healthy tissues are spared. So you're not targeting -- you're binding to red blood cells, but you're not targeting them for destruction, similarly, platelets and neutrophils. But you do need to get that prophagocytic signal from somewhere. And so our drug has been designed to be used in combination with an anti-cancer therapeutic that provides it. So for example, here, we'll talk about gastric cancer, Herceptin will bind its active domain, will bind to the Fc gamma receptor on the macrophage, it will target HER2 on the cancer cell. And in the presence of evorpacept, which is inhibiting that myeloid checkpoint, now you have those 2 signals present and you've maximized this innate immune response, but in a cancer-specific way, sparing your healthy normal tissues. So evorpacept is shown here on the left, it is a fusion protein, it has 2 high affinity CD47 binding domains that bind with picomolar affinity, and it has an inactive Fc domain. It is about half the size of an antibody. And so between the 2 CD47 blocking modes and the decreased size, you basically about 10 mgs per kg of evorpacept is equivalent really to 20 mgs per kg of a traditional CD47 antibody. The presence of that Fc domain does ensure a good half-life of 30 days. And so we have now these 3 characteristics where we can potently block CD47, we're sparing healthy normal tissues, so we do not have dose-dependent cytopenias in the clinic. And then because of that, we have a very simple and flexible dosing. We don't have to do priming, loading maintenance-type schedules. So as you can imagine, much like a T cell checkpoint inhibitor, evorpacept, as a myeloid checkpoint inhibitor is designed to improve the activity of many different cancer therapeutics. And this results in a very broad potential utility of this drug. So in addition to being able to combine with checkpoint inhibitors, targeted monoclonal antibodies such as EGFR, CD20, HER2, for sure. We can also combine because of the good safety profile with antibody drug conjugates, which may have their own cytopenia type toxicities. And then as well, combining with chemotherapy, which itself is known to provide a prophagocytic signal in cancer cells. So together, based on all of these different types of antibodies, classes of antibodies, evorpacept is really a combination agent that can combine with agents that represent almost half of U.S. cancer drug cells. So it has a potential for very broad utility. Our development plan, you can see reflects that potential broad utility. So up on top, we're combining with checkpoint inhibitors, KEYTRUDA in head and neck cancer. And we're in randomized Phase II studies there in collaboration with Merck, we have fast track approval. We are combining with anti-cancer targeted antibodies such as Herceptin in HER2-positive gastric cancer, and we have a randomized Phase II study ongoing there in collaboration with Lilly also with Fast Track approval. And then combining with, for example, antibody drug conjugates, we have 2 new studies, one in urothelial carcinoma with PADCEV, another in collaboration with the ejfv group looking at evorpacept in combination with in HER2-positive breast cancer. And then with Zanidatamab, which is an anti-HER2 bispecific drug, a naked antibody in collaboration with Zymeworks. And then on the hematology side, combining with small molecules such as azacitidine and venetoclax in our heme program with MDS and AML. We also have some IST studies that are ongoing and planned. Again, combinations with checkpoint inhibitors as well as anti-cancer targeted agents in ovarian metastatic colorectal and non-Hodgkin's lymphoma. So based on the design of the molecule, as I described, it was projected that evorpacept would have a very good safety profile, because it does spare those healthy tissues that are expressing CD47. And this has translated into the clinic. So we have a very consistent safety profile, we've treated over 300 patients to-date with different combinations, including combinations with anti-cancer agent, anti-cancer antibodies, checkpoint inhibitors, small molecules, very few grade 3 and above toxicities and the most common toxicity we see across the program is low-grade rash and fatigue. In terms of the demonstrated activity of proof-of-principle in the clinic, I'm going to focus to talk on gastric cancer as well as head and neck. On the gastric cancer side, when we combine with Herceptin and on top of the global backbone of ramucirumab or Cyramza paclitaxel, you can see under our objective response rates, a significant increase in ORR of 72% compared to the benchmark of 28% that's seen with RAINBOW and looking at Cyramza paclitaxel by -- as a doublet by itself. On the head and neck side, when we look at objective response rate from the KEYNOTE series, KEYNOTE-048, which was a pivotal study with pembro and chemo in first-line head and neck cancer. And in the second-line setting, when you look at KEYNOTE-040, earlier pivotal study, there -- and we'll talk a little bit about it. But there, when you talk about immuno-oncology agents, response rate and median PFS have never been shown to predict the clinical benefit for which pembrolizumab was ultimately approved. But what does predict are some of the later time points, such as OS at 12 months, and of course, median OS as the gold standard. So when we look across our first-in-human studies in small numbers of patients, first-line and second-line head and neck, we are also seeing an improvement in those OS rates at 12 months and median OS within this data set, we'll go through that in more detail. So first, I'm going to focus on gastric cancer. Again, evorpacept is inhibiting that myeloid checkpoint between SIRPalpha and CD47 and the prophagocytic signal is coming from Herceptin's active Fc domain that engages with the Fc gamma receptor on the macrophage against that HER2-positive cancer cell. Just to level set, these are the benchmarks on top, you can see RAINBOW, which was the study looking at ramucirumab and paclitaxel in a broad gastric population with an objective response rate of 28%, median overall survival of about 9.6 months. On the very bottom, you can see in DESTINY-01, this was the other pivotal study in the third-line population of HER2-positive gastric, so in HER2 or trastuzumab deruxtecan. In this study, we saw an objective response rate of 41% in HER2-positive third-line gastric and a median overall survival of about 12.5 months. In between, also in the second-line setting are 2 single-arm Phase II studies, tras/ram/pac, a single institution study done at Yonsei in Korea, where we saw an objective response rate of 52%, median OS of about 13 months. And then DESTINY-02, a single-arm study looking at in HER2 in Europe and the U.S. with a 38% objective response rate and OS not published yet. So these are kind of the comparators and some of the relevant studies that are in this space. In our first-in-human study, we looked at, first, a combination of evorpacept in combination with Herceptin or trastuzumab in the second-line setting and then looked at it in combination with that same doublet, but now on top of the global standard-of-care, Cyramza and paclitaxel. It's important to note that many of these patients had actually were more than second-line in the doublet, many of them, about half had received 2 prior lines of anti-HER2 directed therapies, including prior checkpoint inhibitor therapy as well. Both groups had aggressive disease as noted by their visceral metastases. Shown here in the waterfall plot is the 4 drug regimens, so evorpacept on top of trastuzumab, Cyramza paclitaxel. And again, you can see just a really nice objective response rate of 72%, that compares well to both RAINBOW as well as DESTINY-01 and a median overall survival as of -- I think this was at SITC last year of 17.1 months, so a really nice improvement. This is the basis for our randomized Phase II study, which is ongoing now. So here in the Phase II portion, we'll be looking at the contribution of evorpacept to that trastuzumab, ramucirumab and paclitaxel backbone versus the 3 drug regimen. And once we show that contribution in a randomized way, then we'll move on to a planned Phase III, looking at the 4 drug regimen versus what is the global standard Cyramza and paclitaxel. So as you can see, if we can show proof-of-concept in the Phase II portion of the study, that bodes well for the Phase III, because all we're doing is taking away trastuzumab from the control arm. Tras, Cyramza paclitaxel is not an approved regimen in HER2-positive gastric, which is why the Phase III uses Cyramza paclitaxel as the comparator. For ASPEN-03 and ASPEN-04, which is our studies in head and neck, evorpacept is still working as a myeloid checkpoint inhibitor. But here SIRPalpha is expressed on dendritic cells as another key player in the innate immune system. So when that myeloid checkpoint on the dendritic cell is interrupted or inhibited dendritic cells become activated and are better at cross-priming T cells, cytotoxic T cells. So on the right hand panel, when those activated T cells are now in the presence of a T cell checkpoint inhibitors, such as KEYTRUDA, you can now theoretically unleash that adaptive immune response against the cancer cell. So it's a way of bridging the innate immune response through the dendritic cell with the adaptive immune response through the cytotoxic T cell. Okay, so as we talked about before, the 2 main pivotal studies are KEYNOTE-048 in the first-line setting and KEYNOTE-040 in the second-line setting. And the main thing I just want to call out is that, again, this point that for immuno-oncology agents, markers surrogates of benefit are a little more tricky. So you can see in KEYNOTE-048 whether the test arm with pembrolizumab versus the extreme regimen on the control arm, both had overall response rates of 36%. And in fact, the PFS on the test arm was worse than on the control arm. It's really not until you get to OS rate at 12 months and of course, median OS that you start to see this improvement from the pembrolizumab containing arms. A similar story is in the second-line setting, although there is some difference in the objective response rate, small between the control and the test arm. But again, PFS goes the wrong direction. And then when you get to OS at 12 months, a median OS, this is when you start to see the benefit of the pembrolizumab component. Visually, you can see this on the Kaplan-Meier curves, it's the same point. It takes time for those 2 curves to split apart, the pembrolizumab containing in the blue and the control arm in the red. So if you look at time points early in the course of disease, you can be fooled, you may not see the clinical benefit. So we've looked at in our ASPEN-01 study, 2 small cohorts, of 10 patient cohort in a 13 patient cohort. Looking at patients who are checkpoint naive in the second-line and in the first-line setting, our drug in combination with KEYTRUDA in the second-line setting and in the first-line setting in combination with KEYTRUDA chemotherapy. And what we saw is actually -- it's very similar to almost to the KEYNOTE experience. So again, with smaller numbers. But here in the chart at the bottom left, you can see the first-line head and neck population, where we combine evorpacept plus KEYTRUDA chemotherapy an OR rate of 38%, not too far different from the KEYNOTE series and a median PFS, a little bit higher. It is really not until you get to the median OS and OS rate at 12 months, that you start to see a difference with historical controls. Similarly in the second-line setting in the absence of chemotherapy you do see a slight increase in the OR rate, again, small number of patients, but it's that median OS and OS rate at 12 months, where you potentially are seeing some increase in activity. So this became the basis for 2 randomized Phase II studies, where with the much larger and more homogeneous population, we can look at in a randomized way evorpacept plus KEYTRUDA versus KEYTRUDA alone and a second separate randomized Phase II study, evorpacept plus KEYTRUDA chemo versus KEYTRUDA chemo. Primary endpoints 12 month OS rate as a co-primary with objective response rate. So with that our milestones over the next few years, here we are in 2022 were for our gastric Phase II study, we initiated and in our AML study, we will be presenting our AML dose escalation, that's a study where we've looked at evorpacept in combination with venetoclax and azacitidine and we'll be presenting that at ASH this year. We will be starting our ASPEN-07, our Urothelial carcinoma study evorpacept in combination with enfortumab as an ADC. We have ongoing collaboration with Zymeworks and as we mentioned, our investigator-sponsored studies. Most notably, in 2023, we do anticipate that the Phase II portion of our gastric study will read out. And in addition, the dose optimization portion of our MDS program, where we're looking at ALX -- excuse me, evorpacept in combination with azacitidine. We'll be presenting that data over 2023 as well. And then the 2 randomized Phase II studies with head and neck cancer, ASPEN-03 and ASPEN-04, we anticipate those reading out in 2024. And then as we mentioned earlier, we'll be filing that IND in 2023 for our SIRPalpha TLR9 agonist. So lastly, in terms of the financial information. Similar to the first slide, so we have approximately $545 million in net proceeds raised to-date. We have a loan facility for $100 million that is available to us and the cash and cash equivalents and investments give us as of September 30, approximately $293 million, which will take us through with an expected cash runway into mid-2025. So I think that's it, unless there are any questions.

Just a question on the mechanism of action of evorpacept. Would you say it's most active in combination with antibodies or also in combination with different chemotherapies?

Yes, it's a good question. There are 2 different MOAs. So it's hard to know. What we can say is, I think in the clinic where we have mostly demonstrated proof-of-principles in combination with anti-cancer antibodies. And part of that is just because that's the most mature data that we have to-date. So combinations with rituximab, with Herceptin. But our data coming out of the head and neck, it's small numbers, but it's very encouraging. And so we'll have to wait until the randomized Phase II studies to get a better read on the combination with the checkpoint inhibitors. In terms of the small molecules, so those are the heme side of the world, so looking at azacitidine with venetoclax. There again, we have Phase I data, which is a very heterogeneous population in our clinical studies. So really, the dose optimization for our MDS study will be the first time that we actually have a larger, more homogeneous population of newly diagnosed patients, and that will read out next year.

Our drug seems -- sorry, will start it over again. In principle, your antibody or your construct, sorry, seems to potentiate existing cancer antibodies, right?

So different buckets. I think the most data that we have is in combination with anti-cancer antibodies. We'll be developing the data in combination with checkpoints that will read out in 2024. And in combination with small molecules will read out in 2023.

Just a couple of follow-ups based on the expected data flows. So specifically for ASPEN-06, is there a pre-specified bar on the Phase II that would make you confident moving that for the Phase III? What do you hope to see from that readout in 2023?

So from the Phase II portion of that study, where we're looking at the 4 drug regimen versus the 3 drug regimen. So evorpacept plus tras/ram/pac versus tras/ram/pac. What we'd like to be able to see there is a difference of at least approximately 10% between the 2 arms to convince a study evorpacept is truly adding to that backbone. And then in addition, the study is powered to look at the 4 drug regimen versus historical control of ram/pac. So there are sort of 2 things that we want to be able to see, an improvement over historical control ram/pac, an improvement over the 3 drug backbone within the randomization.

Got it. Perfect. And then a follow-up on hematological malignancies, and that seems to be less of a part of the story these days. But either way, you have been able to achieve quite high exposure in AML patients. And in some survey work that we've done, there was actually some excitement mechanistically in AML, specifically not just in MDS. So I'm curious what do you expect to find out in ASPEN-02, the main form potential for AML? And for ASPEN-02 specifically, given magrolimab's lead, were there still room for potential differentiation in further development?

So for ASPEN-02, so that's our MDS study. So coming out of our MDS dose optimization study, where we'll be looking at 40 patients at the 2 highest doses confirming safety as it's a Phase Ib study. But also, it will help inform the doses that we're evaluating to optimize an ASPEN-05, which is the AML study. So we have put ASPEN-05 on pause until we get that dose optimization, it's just a more efficient use of funds. And -- so that's what we want to see coming out of the ASPEN-02. In terms of how it compares to magrolimab, I think one of the key differentiators is, of course, safety between our drug and magrolimab. Magrolimab certainly has dose-dependent cytotoxicities, but it's in a population that can be transfused. But from a development standpoint, it allows, I think, greater opportunity for evorpacept -- for our drug -- for evorpacept, because you can combine with other cytotoxic agents. So for us to be able to combine, for example, with ven/aza, whether it be an AML or an -- there is certainly is an effort to look at magrolimab ven/aza in MDS as well. But there, where you have the cytotoxic toxicities of the backbone, we would have a natural advantage. In terms of the activity, we'll just have to see. Hopefully, with an improved safety profile, you can dose higher, you can dose longer. But we'll have to see how that translates into benefit.

Okay. Perfect. I think we're out of time. So again, I want to thank you again for joining us in the healthcare conference. Thanks so much for your time. I appreciate it.

Thank you.