Good day, everyone. Welcome to the Leap Therapeutics DKN-01 clinical investigator conference call. Following the presentation, there'll be an opportunity for questions. Please be advised that today's call is being recorded at the company's request. At this time, I'd now like to turn the call over to Dr. Cynthia Sirard, Chief Medical Officer of Leap. Please begin. Thank you, operator. Welcome and thank you to those of you joining us today for an update on Leap Therapeutics' DKN-01 development program. I'm Cynthia Sirard, and with me today are Dr. Jaffer Ajani, a professor, Department of Gastrointestinal Medical Oncology, Division of Cancer Medicine at The University of Texas MD Anderson Cancer Center in Houston, Texas. In addition, we have Dr. Samuel Klempner, who's a member of the faculty at Massachusetts General Hospital and Harvard Medical School in Boston, Massachusetts. As well we have Douglas Onsi, the President and Chief Executive Officer at Leap. This call is being accompanied by a slide deck, so I will ask you to please turn to our forward-looking statements on slide two. I would like to remind you that any statements made during this call that are not historical are considered to be forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. Actual results may differ materially from those indicated by these statements as a result of various important factors, including those discussed in the Risk Factors section of the company's most recent annual report on Form 10-K, as well as other reports filed with the SEC. Any forward-looking statements represent our views as of today, September 17th, 2021 only. A replay of this call will be available on the company's website, www.leaptx.com, following this call. With that, please turn to slide three. Good day. Today, we are hosting a call to share initial results from our clinical study of DKN-01 in combination with BeiGene's anti-PD-1 antibody, tislelizumab, and capecitabine and oxaliplatin in first-line patients with advanced gastric and gastroesophageal junction cancer. To provide a brief overview of the data that will be presented in detail this morning, DKN-01, in combination with tislelizumab and chemotherapy, demonstrated a compelling 68.2% overall response rate as a first-line treatment for advanced gastric and gastroesophageal junction cancer with a 90% overall response rate in patients whose tumors expressed high levels of DKK1, as compared to the 56% overall response rate observed in those patients whose tumors expressed low levels of DKK1. The response rate is correlated with DKK1 expression and is independent of PD-L1 expression, as there was a 79% response rate in patients with PD-L1-low expression with a CPS score of less than five and 100% overall response rate in DKK1-high, PD-L1-low patients who would be expected to be the most difficult subpopulation for this combination therapy to treat. This new data, taken together with our understanding of DKK1 biology and our previous clinical studies, which demonstrated the activity of DKN-01 as a monotherapy and in combination with a different anti-PD-1 antibody, pembrolizumab, and with paclitaxel chemotherapy, should establish DKK1 as an important new target in gastric cancer and DKN-01 as a promising new therapy for patients with this deadly global disease. With that, please turn to slide four. Our presentation will be given by Doctors Jaffer Ajani of MD Anderson Cancer Center and Samuel Klempner of Massachusetts General Hospital, who are both investigators in this study. Dr. Jaffer Ajani will provide background on the biology of gastric cancer and the DKK1 protein and the mechanism of action of DKN-01. Dr. Samuel Klempner will review the clinical data from the study and his experience with treating patients with DKN-01. We will end the call by opening the floor to questions for Doctors Ajani and Klempner or any of us from Leap. Hello, everyone. My name is Jaffer Ajani. I'm a GI medical oncologist working at MD Anderson. I do lot of clinical research as well as I have my own laboratory. Today I'm really happy to talk to you about a molecule called DKK1, which is in the Wnt pathway. This is secreted by the tumor cells just like PD-L1. However, DKK1 is very unique in its functionality. Not only the secreted molecule DKK1 is immunosuppressive, which would be similar to PD-L1, but in addition to that, DKK1 also stimulates the stem cells, cancer stem cells, maintenance of cancer stem cells, promotes progression of cancer through different pathways, including the well-known AKT and mTOR pathway. In that respect, I think this is a molecule that is more damaging to the host-of cancer than, for example, PD-L1. It synergizes with inhibition of PD-L1. That data I'm going to share with you. On the next slide, what we can see is the prognostic value of DKK1. On the left side is the TCGA pan-cancer dataset, which is 1,000 of patients, multiple tumor types, I think 33 tumor types and 10,000 patients. We can see that the tumors that had high DKK1 had very poor prognosis. The P value is very strong, and patients' tumors that had low DKK1, their survival was clearly significantly better. On the right side, it is drawing from the same dataset, but it's customized to stomach cancer TCGA dataset, which is about 350 patients. Again, in gastric cancer, even more graphic to see that high DKK1 gastric cancer patients had much worse survival than DKK- low. This sort of makes us feel that this is an important molecule with a big prognostic impact. On the next slide, I'm showing you data from MD Anderson because we've been working on this molecule before I engaged with Leap and their clinical trial. On the left side, we were looking at normal tissue, adjacent tissue from the tumor, and also the primary tumor. You can see the expression levels are clearly much higher once the tumor is formed. In the middle curve, again, this is an independent cohort of patients where high DKK1 in our analysis also showed significant prognostic value. Here we are talking about 415 patients. On the right side, this is another dataset which again validates our finding of prognostic importance of DKK1. There also you can see that patients with high DKK1 fare very poorly compared to those with low DKK1. I'll show you some more MD Anderson data because what we have discovered is that DKK1 is regulated by another oncogene. Today we are not focusing on that, but this is just to expand on the understanding of how DKK1 might be operating in the cancer cells. On the left side, you see the heat map and you see four columns. On the left side, left two columns, the gene YAP1 has been knocked out and as a result of that, all the oncogenes that are shown on the right side of the right two columns with red arrows, all those oncogenes are abolished. Their expression is abolished by just getting rid of the YAP1. In the right column, which is the control, you can see number of oncogenes are overexpressed, especially in the third column. There is some heterogeneity which we expect. DKK1 is highly expressed in both patients, but more so in the third column. We have other set of data. Here again, we are also looking at the plasma of healthy individuals, plasma of patients. We also collected the supernatant from malignant ascites. There you can see that once you have stomach cancer, for example, even in the body fluid, in the plasma, the DKK1 levels are much higher than normal. I personally think this could have an implication in patients also later on. We could actually monitor therapy duration, first therapy effect, and its duration by monitoring the blood. I must say this is not validated yet. This is just very early dataset. We have also done some other mRNA level qPCR, which is on the right upper side. As I was mentioning, that YAP1 seems to regulate DKK1. When YAP1 is knocked out in two different clones, the DKK1 levels go substantially lower. Go to the next slide. This is a very small dataset, very preliminary data, just to understand whether DKK1 levels are different in different cell types in the peritoneal cavity. What we normally do is we will collect the peritoneal malignant fluid, but it is crowded with different cell types. Here we are showing that in certain clusters, DKK1 is highly expressed, and every time it is expressed, the prognosis of these patients was poor. Generally, the prognosis of all these patients with peritoneal metastases is poor, but even there, we can discriminate at least by DKK1 level, which is our focus today. On the next slide, these two graphs actually demonstrate that if your tumor has high DKK1 level, not only your prognosis is worse, but your tumor is also resistant to therapy. Now, on the left graph, I'm showing you the duration of treatment with a very popular drug called paclitaxel. Here, patients were getting paclitaxel, but patients whose tumor had low levels of DKK1, they actually received much longer treatment, mainly because paclitaxel is effective against those tumors. The tumors with DKK- high, the drug had to be discontinued early because the tumors are resistant. This is conferred by DKK1. The same sort of result is shown on the right side, except the drugs are different. This could be considered pan-resistant. This is a phenomenon we see in cancer stem cells, that they have multiple resistant phenotypes, and it is very hard to overcome that. DKK1 may be conferring resistance to the drugs that these tumors haven't even been exposed to. They sort of inherently have become resistant. The duration of treatment also is shorter for patients whose tumor overexpress DKK1. In the next slide, we again show very similar effect. This is the colorectal carcinoma cells on the left side. The experiment is little bit different here. We are knocking down DKK1. In other words, there's no way these cancer cells can express DKK1, and you can restore the resistance to 5-FU in these cells. This is like a rescue experiment we'd normally do just to be comprehensive and complete the story. The next slide I show here is regarding the effect of DKK1 on the immune system. As I mentioned earlier, that not only DKK1 supports the cancer stem cells and resistance to therapy and poor prognosis, but it also is complemented by its effect on the immune cells. Here, what we can see that the anti-tumor immune response, and that can be modulated by modulating DKK1 with the antibody. This is kind of laying the groundwork for how this antibody can actually target DKK1 and then alter the tumor microenvironment in favor of the patient, anti-tumor microenvironment or more responsive microenvironment. This is my last slide. Here, basically it shows the fact that you can magnify the effect of DKK1 by combining it with anti-PD-1. Which is sort of what led to the trial, to combine both molecules targeting cancer cells and with DKK1, and targeting immune cells with anti-PD-1. You can see that on the left side, you see the CD45 positive population, other cells that are increased by inhibition by DKN-01. On the right side is the effect on tumor volume, which is maximum with the combination. We can conclude that. First, the DKK1 expression in number of cancers, variety of cancers, in large data set is prognostic. It is from the same large data set, it is impressively prognostic for gastric cancer and from MD Anderson patient cohorts, it is completely validated. We internally validate it. We can say that it can be validated from the data I've presented to you that are not from MD Anderson. The important thing is that it does promote resistance to therapy in addition to being a poor prognostic factor. We have found at MD Anderson that it is highly regulated by other oncoproteins. In the preclinical model, we can manipulate DKK1 as a target by DKN-01 antibody, and we can reverse some of the resistance, and we can also turn the tumor microenvironment in favor of the patient. Combining DKN-01 with PD-L1, we would clearly expect that we can get greater advantage, amplification of the effect of DKN-01. With that, I will stop. Thank you. Hello, good morning. My name is Sam Klempner. I am a GI medical oncologist at Massachusetts General Hospital in Boston, where I lead the gastric and esophageal efforts, and I have the pleasure of discussing some of the clinical data from the recent DisTinGuish trial that was presented at ESMO 2021. By way of brief background, you have already seen some of the preclinical data regarding DKK1 and DKN-01. Just to put some context to the DisTinGuish trial, DKN-01 has previously demonstrated significant activity in combination with checkpoint inhibitor in a prior trial that is most recently published. This slide is essentially showing the waterfall plot responses in the subset of patients from the prior study, and really, I think the message here is that there was a clear signal distinguishing the DKK1-high and the DKK1-low population in terms of response rate. What's shown is clearly the partial responses in the green bars are really enriched for the DKK1-high population, where the response rate was 50%. Equally important, we see that the ability of the biomarker to distinguish these populations where the response rate is low in the DKK1-low population. Which is something as a clinician, we're always looking for ways to both maximize benefit but also avoid drugs that are unlikely to benefit patients. On that background, and to further expand upon the activity, the DisTinGuish trial was designed, and this is the schema here. Essentially, we're looking at two relevant clinical situations. Part A is a frontline trial. This is really, in my mind, an important subset because we've seen there are limitations to the current checkpoint inhibitors in the frontline chemotherapy. We're looking for ways to expand the benefit and optimally select the patient. The frontline combination is standard 5-FU platinum with PD-1 agent plus DKN-01 in a cohort looking at response rate and safety. In the second line, this is really a validation and confirmation from the prior work looking at a biomarker selected population, DKK1-high, and this is a chemo-free regimen of DKN-01 and tislelizumab, very similar to the activity that I showed on the prior slide. This is just the reference for the dosing schema used in the trial. I think there's not a lot to see here beyond just saying that this is a well combinable regimen and very standard dosing in terms of chemotherapy and standard tislelizumab and DKN-01 dosing based on prior experience. This is a relatively convenient regimen for patients and not something that is significantly different in terms of schedule from what these patients are used to anyway in standard of care. CAPOX, tislelizumab, and DKN-01, everyone received this regimen on the single arm part A, and these are given in 21-day cycles. This is just a little bit more information on the study design and methods. Again, this is a single-arm phase IIA trial of DKN-01 plus the PD-1 agent tislelizumab plus standard frontline chemotherapy 5-FU and platinum, in this case, CAPOX. The primary endpoint of this phase II trial is overall response rate. Additional endpoints are shown here. Importantly, you see the interest in trying to tease apart activity by substance. The modified intent to treat is everyone who got more than one dose of DKN-01. There's analysis by DKK1 expression, which is previously established based on the prior experience using an H-score RNAscope method, which we will talk about a little bit. Finally, trying to tease out the activity, whether or not it's independent of PD-L1 expression, looking at subgroups that we know are groups that benefit and don't benefit from PD-1 agents. Here is the study population, and as you can see, this is part A, overall 25 patients. This is largely representative of a very real-world gastric and GE junction population, although you can see the breakdown in the frontline population, pretty even split between ECOG 0 and 1. You can see a slight enrichment in GE junction adenocarcinoma patients. Again, this is just reflective of the Western population and what would fully be expected in a U.S.-largely driven trial. You can see here that these patients, some of them have received prior adjuvant or neoadjuvant therapies, but in the vast majority, and consistent with our clinical practice, these are patients presenting with newly diagnosed metastatic disease. There's a few notable things to call out here, which may be of relevance, is that clinically you see that the DKK1- high and DKK1-l ow is present both in the GE junction and the gastric population. The biomarker is importantly present across the clinical spectrum of disease. Really, there may be a slight increase in patients presenting with stage IV disease in the DKK1-high population. This may reflect some of the more aggressive biology and poor prognosis that we know accompanies the DKK1-high subgroup that was previously discussed. Also important, if you look at the top-line numbers, you can see that among the 25 patients, we had DKK1 availability for 21, and 12 patients were DKK1-high and nine patients were DKK1-low. I'd say for me, this was actually quite important because when we're looking at new therapies in frontline populations, the absolute prevalence of the biomarker is quite important because it's difficult to develop drugs or conduct trials in very rare biomarker populations such as MET and EGFR and HER2. Here we see the biomarker is present in over half of the population, suggesting that if this activity is confirmed, there may be a large population of patients who may benefit from this approach. Here are the tumor characteristics in terms of the biomarker details from the population that's presented. PD-L1 patients, nearly everybody had PD-L1 expression available. As you can see, this largely reflects what we see in practice. I think the CheckMate 649 data suggesting that 60% of patients are CPS 5 or higher is somewhat hard to explain and not what we see in clinical practice. In my experience, the rates of CPS greater than five or greater than 10 is probably in the 10%-15% of what I see in a high-volume center where we see more than about 400 patients a year. You can see that about a quarter of patients were CPS- negative, three-quarters were CPS less than five, and consistent with what I was saying, there was only two patients here who were CPS greater than 10. TMB- high is a population we think may have a greater chance of benefiting from checkpoint inhibitor. The vast majority of patients are mutation burden low here, and high mutation burden is outside of MSI is very uncommon in gastric and GE junction cancer. This is just reflecting what we really see in the clinic. There are no microsatellite high patients among patients with available data. This is a population that you're seeing that, one, reflects the real world, and two, is not enriched for a population where you would say, "Okay, these are patients who are very likely to benefit from just checkpoint inhibitor and chemotherapy anyway, and therefore it will be somewhat difficult to assess the relative contribution of DKK1." I think here we're seeing a little bit the opposite, where this is a population that's perhaps biased towards patients who are unlikely to benefit from a checkpoint inhibitor in addition to chemotherapy. This is a very straightforward slide, just looking at patient disposition on the trial. Again, with available follow-up, we see the mean duration of treatment is five months, and the longest duration is over 10 months at this time, and 16 patients remain on therapy. You can see, like what you would envision for a consort diagram, the most common reasons for study discontinuation are progression, and that is unfortunately just the nature of this population. This is just highlighting the ability to do the biomarker and just representative images from biomarker testing. This is DKK1 expression using the in situ hybridization RNAscope assay complemented with digital pathology. This has now been, I believe, published by the company about this method. Essentially, tumor specimens readily obtained in paraffin are stained for DKK1 expression and quantified using a digital image analysis algorithm. An H-score is calculated. An H-score is a well-validated method in pathology to provide a general appreciation of the degree of expression of a given biomarker. Here we see, comparing a high versus a low, I think it's quite clear to the lay eye that there's clear differences between DKK1-high and DKK1-low. To me, I think this is when we look at biomarkers, we want to think, is it both analytically valid and clinically feasible? I think a reproducible biomarker testing is important, and I do believe that this RNA ISH assay is a reproducible assay. Now we're getting into the meat of the study. This is best overall response by DKK1 expression. Of course, this is a shape of a waterfall plot that we would all love to see in any of our trials, where essentially everyone has had some degree of tumor shrinkage here, and you see the dotted line is the cutoff for determination of response by RECIST. One thing that's immediately clear is the depth of response and the enrichment for responses among the DKK1-high population, which is the green bars. There are still some responses in DKK1-low population. We'll get into this a little bit more. This is colored by GE junction, adeno, and gastric, and this largely just reflects the trial population. To put some numbers to that waterfall plot, it's really clarified in this slide. Here, what we're showing is the evaluable DKK1-high GE junction in gastric patients, and everybody had a partial response. As you can see, there's one non-evaluable patient in the DKK1-high group, which consists of a total of 10 patients, and the partial response rate is 90%, and everyone had a response among the evaluable patients. Interestingly, among the DKK1-low population, the response rate is still 55% with four of the five people still on therapy. Just to provide some context, in the PD-L1-lower population from the now published CheckMate 649, the response rates are in the 40s, largely. This is the spider plot demonstrating the durability of responses among evaluable patients here. What you see is, again, consistent with the waterfall plot. The vast majority of these lines are below the zero, suggesting some degree of tumor shrinkage. There's a clear enrichment of the green, which are the DKK1-high patients, both in terms of depth of response but also durability. Now you see this does, of course, not have significantly long follow-up. At the available follow-up, you see that there's a large proportion of patients on therapy at three and six months. These are sort of early landmark time points that we're always looking at in phase II type trials where the primary endpoint is response rate. This is a swimmer plot looking at the differences between durability and DKK1 expression. On the top, you see the DKK1-high population. On the bottom, you see the DKK1-low. Then the far left is the unknown population. This is marked by both complete and partial responses, and this is a pretty standard way of showing both durability and percentage of response rate patients across the trial population. Again, I think this is a very encouraging signal and something that would get anybody quite interested in the potential biomarker here. This is, again, just doing some formal comparison between the DKK1-high and DKK1-- sorry, and DKK1 expression by response. Here, if you compare the partial response rate versus the stable disease, you see that there's a clear enrichment in responders among the DKK1-high population. We look at biomarkers both from a predictive and prognostic standpoint. Are you identifying a biology that is aggressive? That's more of a prognostic biomarker. How likely is this patient to do independent of the therapy? A predictive biomarker is, do you have the ability to identify a population that's more likely to respond to your given drug? I think we've seen some data on prognosis already. Now we're suggesting that this is also a predictive biomarker, identifying a group of patients who are more likely to respond to the drug. I think an important and immediate question is, what is the relationship between DKK1 and PD-L1? Which, of course, is an imperfect biomarker, but a biomarker nonetheless, and certainly used clinically throughout gastric and esophageal cancers. Here we're seeing best overall response by PD-L1 and DKK1 expression. What you see is the DKK1-high is again colored in green, and the low is colored in blue. Above the bars, sorry, you see the CPS, the pluses are the DKK1-high, the minuses are the DKK1-low, the green is CPS- high using a five cutoff, which is what was used in CheckMate 649, and the blue is a CPS less than five. I think visually what stands out here is you see blue bars with green pluses. These are DKK1-high, PD-L1- low patients that are responding. This is a population of patients where you would not expect significant benefits from CAPOX and PD-1 because of the PD-L1- negativity or less than five. Again, that's a population that had an unremarkable hazard ratio in the CheckMate 649 trial. Here we're seeing independent of PD-L1, if you're DKK1-high, you are, for the vast majority of these patients, responding. This is a little bit another way of showing some of the same data. This is just trying to confirm and convince us from this available data set that DKK1-high patients are responding regardless of PD-L1 status. Just to break this down, their response rate was 79% in patients with PD-L1- low expression. CPS less than five and 100% in the DKK1-high, PD-L1-1 ow. Again, for reference, the response rate in PD-L1 less than five from the CheckMate 649 population is about 50%-55%. Here we're seeing that even in the not all DKK1-high population, there may still be some enhancements of and perhaps additive or synergistic activity even in the DKK1-lower population. Clearly, it's significantly enriched in the DKK1-high population here, where the response rate was 100%. You can see, although the numbers are relatively small, the response rate in the PD-L1-high population still exceeds what was seen in CheckMate 649. This is a significantly smaller numbers, but certainly an encouraging signal. Here is again a spider plot showing durable responses independent of PD-L1 expression. Here CPS is coded as green for CPS greater than or equal to five and blue for patients who are low. What you see is, again, responses independent of PD-L1 expression. You don't see clear clustering between the PD-L1-high and low population. Just to formally compare this, there is no correlation between DKK1 and PD-L1 expression, and this is consistent with what was seen in prior data with DKN-01 and pembrolizumab. I think that the message here is that really DKK1 is not marking a population of patients who are more likely to respond to checkpoint inhibitor, which would raise some concern about the biomarker. In fact, this is the opposite. Here it's very independent of the PD-L1, and so it's certainly independent, in terms of its ability to predict and identify patients. Obviously, anytime we're adding a drug on top of a standard of care regimen, we certainly need to understand the toxicity profile. I think this is largely consistent with what you would expect from chemotherapy and immunotherapy combinations across many of the large data sets. The most common DKN-01 related events are fatigue, nausea, diarrhea. Again, these are consistent with what was seen in the prior experience with checkpoint inhibition. There are some cytopenias which may or may not be more related to the chemotherapy in my own opinion. There are rare grades greater than or equal to three. Again, diarrhea, there was one, two pulmonary emboli on the trial. Again, that may or may not be related to the disease underlying more likely. This slide here is really just a busy slide showing some context, which I've tried to provide throughout speaking through this data. It's just reference to give you some context of the encouraging response rates seen in this DKK1/DKN-01 trial. To summarize, I think what I've tried to show is DKN-01 and tislelizumab and CAPOX is a well-tolerated regimen. Again, the 5-FU platinum and PD-1 are something we have a lot of experience with. Adding DKN-01 does not seem to enhance the toxicity above and beyond what we would expect from chemotherapy and immunotherapy. This is based on my own experience, both with this triplet combination on the trial as well as a large experience with chemo and immunotherapy on multiple other trials. The response rate is very encouraging. Sometimes this is, I believe, underappreciated because these are very symptomatic patients when they're presenting to the clinic because they have tumors that are interfering with their ability to swallow and eat, and they have significant pain and weight loss. A biologically active combination is very important because response rate is very tightly associated with improvements in quality of life and performance status because now these patients are able to eat, et cetera. I can speak from personal experience but also general experience with biologically active regimens and patient outcomes. Certainly, the response outcomes compare favorably to current standard of care. Again, this is an unselected population, and I think this is something that's important because when you look at KEYNOTE-590 and CheckMate 649, there really is no benefit to the CPS less than five or CPS less than 10 populations respectively in those trials. There's really a large unmet need to address this PD-L1- negative population or low, but also identify additional biomarkers of how we can really select patients for therapies that are more likely to provide them with benefit. It does seem, in this data set that the efficacy is driven largely by the enhanced response rate, which is very high in the DKK1-high population, where everyone that's evaluable has had a partial response. This doesn't appear to be associated with PD-L1 expression, which is relevant. Early on, the duration of responses, of course, are not yet mature, and we expect this to come later on. I also think that this, and this is my own opinion, is important in the field. I think the field of gastric and esophageal cancers is really moving towards biomarker selection. We have a pie of biomarkers when patients come to see us, and we're testing for PD-L1, mismatch repair, HER2. Ultimately, we may test for FGFR2. The paradigm clinically is really to slice up the populations into biomarker selected groups. I think one of the things that I would emphasize is that at least from this data set, the biomarker appears to be present in a large portion of the population, in this case, 57% of the frontline population that we're seeing here. That's always attractive to us clinically, when we're looking to select patients, because the majority of our current biomarkers like MSI is only 3%-4% of our patients. PD-L1-high, CPS greater than five or 10 again, is 10%-15% in my own experience. HER2 is maybe another 15%. Biomarkers that have a high prevalence are something that are attractive, I think to many of us as clinicians. I'd like to thank you for your time, and I'll be happy to take further questions. Thank you, Dr. Ajani and Dr. Klempner, for your participation in today's program, and thank you all for your time and attention today. We'd now like to open the call for questions. Operator? Thank you. Our first question comes from Joel Beatty with Baird. Your line is open. Hi, good morning. Congrats on the data, and thanks for this presentation this morning. The first question is for the physicians. I'm curious, do you use nivolumab in all of your patients in this patient population or just patients who are PD-L1-high? Related to that, how do you anticipate using DKK1? Is there an opportunity in the patient to diagnose this process to be able to assess all this upfront before that first-line therapy is started? Yeah, I can answer it first. This is Jaffer Ajani. We don't use nivolumab or pembrolizumab if the CPS score is low, like less than five. For pembro, it has to be less than 10. We don't use it. I use it only for those patients with CPS of five or higher. Yeah, I'm sorry. Yeah. I was also going to say it depends on how the phase III trial is designed with DKN-01. It may be focused on, maybe Cindy can elaborate on this. If it's focused on DKK1 expression irrespective of CPS, I think that's the way the trial will go irrespective of CPS. I totally agree with Dr. Ajani. We do not offer checkpoint inhibition in the first line to our patients who have the same CPS stratification that Jaffer was mentioning, less than five and less than 10. I'll also just say one thing. In terms of feasibility of selecting on a biomarker like DKK1, I think we saw some data from this trial called FIGHT-101 in FGFR2 patients that actually, sometimes you can just give it one cycle of chemotherapy while you're waiting for a biomarker result, and then stratify for that in the trial design. It makes waiting for biomarker, if you need to wait, much more palatable to patients and investigators, and it didn't seem to compromise the outcomes whatsoever. I think, yes, there's definitely opportunity and feasibility to do something like a DKK1. I do think in situ hybridization-based biomarkers may become more common as well. I like the strategy. Got it. Thanks for that perspective. Maybe 1 follow-up question. In trying to assess how this data from this trial looks, I think 1 of the comparisons is with the nivolumab data. I noticed there's different data that's on the label versus what was presented this year at ASCO. I believe in "The Lancet" publication, with the ASCO data seeming to be a little bit higher than what's on the label as far as ORR. I guess I'm curious, either for the physicians or the company, would you be able to kind of help assess what that difference is and what might be the more relevant comparison for the data on Leap's study that we're looking at today? Sure. Thanks for the question. This is Doug. I'll answer. Which is that we presented both. We acknowledge that the FDA label includes all of the patients that were enrolled in the study, that the supplement to The Lancet article had 370 fewer patients, that represents the ORR and the subgroup reported on the supplement of The Lancet. We provide both of those for context because people can find them when they search. In general, we're inclined at the company to look at the FDA label as representing the full results of the study and the full N of patients enrolled. Got it. Okay. Yeah, that's helpful. I noticed there's somewhere around 20%- 25% of the patient population seems to have dropped off between the label and the ASCO data that was presented. Thank you very much for taking the questions. Thank you. Our next question comes from Joe Catanzaro with Piper Sandler. Your line is open. Great. Thanks. Appreciate you guys taking my questions and congrats on these data. First a couple here, one for Dr. Klempner. Knowing that follow-up is perhaps limited at the moment, I'm wondering if you see anything within the data set as it currently stands, whether it's safety or efficacy, that perhaps portends well for longer-term outcomes like durability of response and progression-free survival. As a follow-up to that, maybe one for the company is what they see as next development steps, knowing that likely needs a little bit more mature data, but maybe they could speculate on the exact biomarker strategy moving forward, whether it's a selected population like DKK1-high and PD-L1-low, or enrolling irrespective of CPS status. Thanks. I have a follow-up as well. Sure. Yeah. The data is the data, so we're limited by follow-up, as you mentioned. The encouraging things, the shape of the waterfall plot and the depth is something that we do tend to have some loose associations with translation to longer-term outcomes. Deeper responses may be more durable, although that is not a completely hard and true fact. Similar kind of phenomenon in KEYNOTE-811, where you look at the shape of the waterfall plot, and the FDA approved that combination based on response rate alone so far. The other thing that I think is notable is that among the responders, essentially what seven of nine are still on therapy. The majority of people who responded continue to remain on trial. The proportion of responders on trial is high. That may be something to look forward to as a marker of maybe an encouraging progression free survival. I think the data is what is presented, and we can future cast only so well based on what we have. The high rate of patients who remain on trial and the depth were things that were noted to me. Dr. Ajani, did you want to add anything before I kind of respond on biomarker strategy a bit? Yeah, sure. I think as Sam mentioned, we are sort of limited by the total number of patients in this study and also the duration of follow-up. I think those issues have to be considered in judging these data. I have patients who are on for more than nine months. One of them is still getting therapy after almost one year, and their quality of life is very good. As all of you know, what we have to do when we start with three-drug or four-drug combinations is we have to drop oxaliplatin going forward, and often we will drop the second cytotoxic, which is a fluoropyrimidine. The patients stay on a targeted therapy or immunotherapy. I think if you consider that, it is very likely that some of the patients will continue combination of DKK1 and trastuzumab for a long time. The toxicity profile will be probably more intense in the first three, four months, but then we can titrate it out, and patients should do very well, as some of my patients are doing. With respect to your question, obviously we at the company feel the same way, that we feel very encouraged and enthusiastic about the overall response rates seen to date. Now our mission is to continue to follow these patients and see what the duration of response will look like, the median progression-free survival. I think durability is very important to us. We've seen that be a real strength of the drug in other studies due to its tolerability profile and as Dr. Klempner mentioned, the depth of response and the kind of curves of the response. We are still at an initial presentation of the data here and continue to follow these patients. With respect to the biomarker strategy, as a company, it was our initial development thesis for the drug, is that we would be able to find a way to identify patients whose tumors are expressing high DKK1 in indications where that high DKK1 was associated with worse outcomes and be able to bring them a targeted therapy. On a philosophical basis, we support and believe that the right strategy would be to identify a DKK1-high patient population focus development at least initially in that patient population where you would expect to have the greatest treatment effect, right? The worst outcome for your control group and the strongest outcome for your DKN-01-based treatment group. Look forward to having that discussion with our partners at BeiGene as the data matures to really make that final assessment and strategy. Okay, great. That's helpful. I could just squeeze in a quick follow-up. Would you happen to be able to provide the details of the DKK1 status of the two patients who reported pulmonary embolisms? Relatedly, maybe if the KOLs could speak to their experience in this population around such AEs. It looks like pulmonary embolisms were noted in both CheckMate 649 and KEYNOTE-062. If you could help there, that would be great. Thanks. Why don't I let maybe Dr. Klempner speak about the pulmonary embolisms, and then we can. I think it's noted, as we said earlier, that the non-evaluable patient who had the pulmonary embolism was DKK1-high, also PD-L1 greater than 5%. In terms of the frequency, maybe Dr. Klempner or Dr. Ajani can comment on it or its customary in us. Yeah, I think I was a participant in the prior trial where there was no chemotherapy component to the DKK1 and PD-1 combinations. Broadly, I think all of us in oncology have a fair experience with pulmonary emboli and cancer-related venous thromboembolic phenomenon. I think these are events that happen essentially across all therapies. We see them at relatively similar frequencies, whether you're looking at chemo-only trials, chemo PD-1 trials, chemo HER2 trials. I'm not aware of any enrichment or association between necessarily therapies that we use commonly and increased DVT risk. Certainly with anti-angiogenic drugs maybe, but those are more true with maybe ramucirumab and some of the VEGF TKIs, which is quite different than DKK1, in my opinion. Then, I can comment broadly on the toxicity profile, and I think Dr. Ajani mentioned as well. These patients are getting chemotherapy, immunotherapy, and another immunomodulatory drug in DKK1. We have a lot of experience and expectations with the toxicity profile from that combination. In my own experience, I personally have not seen an enhancement of the side effect profile with the addition of DKK1. I think the cytopenias are largely chemotherapy related. There's a known AE profile from the PD-1 agent, and there's some experience with DKK1 toxicities from monotherapy and from prior PD-1 combinations, and I think that's been similar in this trial from my own experience. I'll leave it at that and let Jaffer add whatever he wants. This patient population and pancreatic cancer, for example, they have hypercoagulable state, and thromboembolic phenomenon are not uncommon. I get calls from radiologists in a new patient. I see a patient in the clinic. We order an imaging study, you get a call in the evening saying there is a pulmonary embolism. This is untreated situation. I think having two episodes here of PE in a 25-patient population would be baseline. I personally didn't think that DKN-01 is contributing to PE in this study. I think if there were seven or 10 cases, we would be worried about that toxicity. I think this is expected in this population, untreated, as well as it increases with on-treatment. We have published data just with chemotherapy. There is an increase. Okay, great. Thanks so much for taking my questions. Thank you. Our next question comes from [Steven Seedhouse] with Raymond James. Your line is open. Yeah. Hi. Thank you for taking our questions. Congrats on the data. Just had a quick question on potential durability. In terms of the response trend in DKK1-high patients, so the two patients with the longest follow-up, around eight months, looks like they had a tumor reduction of about 50% and holding steady at that. What do you think is the chance of those patients progressing to a complete responder status? Other patients, they also appear to be sort of steady. Maybe also for Drs. Ajani and Klempner, how feasible and clinically significant is it for patients to remain in sort of this 50% tumor reduction state and be there for longer term? Thank you. I can begin with that. Longer duration of response is really very important. Of course, deeper the response, you can consider what is called consolidation. In other words, someone can start with big geography, tumor geography, and they have dramatic response, and then that response prevails. Then there are strategies to get rid of the active tumor. This has been published from our institution and others. This particular strategy can become very important if you have treatment or if you have a tumor very sensitive or if a treatment very active. I think that is one very good option available. Now, the question about 50% reduction and prolonged duration of response. I think that is also very important because I tell patients and families that the next treatment, because they are presenting with incurable condition. Most of these patients will require another treatment. Longer we can wait to use the second treatment, better it's going to be. Suppose this second-line treatment we don't have to use for another year, the portfolio will be much better. I think in that regard, it's great to have 50% reduction and a very long duration of response. Yeah, I agree. I have very little to add to what Jaffer was saying. I would say, also, just to put some context, the rates of complete response in metastatic gastric and esophageal cancer are very low. Maybe 5%-10% at most with chemotherapy and maybe, and Jaffer knows this better than me, maybe in the HER2 positive population, we get little higher rates of complete responders, but really the expectation of anyone having a complete response is very low. I think it's hard to tease too much more out from the spider plot that you were referencing. Largely, I completely agree with Jaffer. People who have a 50% reduction, we're very happy with that, and often it does translate to better quality of life and symptom control and ability to tolerate future therapies, and get future therapies. You got to remember that across the board, a lot of these patients actually don't even go on to get second-line therapies in the community. We get really our best shot in the beginning and to see an active regimen is really what we hope to see. Okay. Thank you very much. Then just maybe a quick follow-up, based on what Dr. Ajani had said. seven out of nine DKK1-high responders are still in the study. Just want to clarify, are those patients, they're no longer getting chemotherapy regimen? They're just getting DKN-01 and tislelizumab? Thank you. I don't know about all of the patients, but I have currently two patients. One is just getting DKN-01 and tislelizumab. The other one is also getting capecitabine with immunotherapy drugs. They are doing extremely well. Functionally, they are almost close to normal. Sorry, Cyndy. Go ahead. No, go ahead, Sam. I was just going to say, we've taken the exact same approach, slowly backing off the chemotherapy as standard chemotherapy side effects accumulate, primarily neuropathy and the oxaliplatin. We've also done capecitabine, tislelizumab, and DKN-01 as sort of the transition, and our patient continues on trial as well, feeling well. Yeah, I was just going to add that the protocol does permit them to stay on any combination of the three, provided they remain on the DKN-01. I think consistent with what both Sam and Jaffer have noted today, many sites are dropping a cytotoxic therapeutic, so the oxaliplatin being dropped first, followed by the Xeloda as necessary. Okay, great. Thank you for taking our questions. As we are showing no further questions, thank you again for dialing in today's call. Have a good day.