Good day, everyone, and welcome to the Leap Therapeutics' DKN-one Clinical Investigator Conference Call. Following the presentation, there will 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 Doctor. Cynthia Sarard, 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 Leaf Therapeutics' DK101 development program. I'm Cynthia Sarard and with me today are Doctor. Jafar Ajani, 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 Doctor.
Samuel Klimsner, who is a member of the faculty at Massachusetts General Hospital and Harvard Medical School in Boston, Massachusetts. As well, we have Douglas Anci, 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 2. 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 ks, as well as other reports filed with the SEC.
Any forward looking statements represent our views as of today, September 17, 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 3. Good day. Today, we are hosting a call to share initial results from our clinical study of DK101 in combination with BeiGene's anti PD-one 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, CK-one 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 express 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 PD L1 low expression with CPS scores less than 5 100% overall response rate in DKK1 high PDL1 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 which demonstrated the activity of DKN-one as a monotherapy and in combination with a different anti PD-one antibody pembrolizumab and with paclitaxel chemotherapy should establish DKK1 as an important new target in gastric cancer and DK101 as a promising new therapy for patients with this deadly global disease. With that, please turn to Slide 4.
Our presentation will be given by Doctors Jafar Jhoni of MD Anderson Cancer Center and Sam Klimtner of Massachusetts General Hospital, who are both investigators in this study. Doctor. Ajani will provide background on the biology of gastric cancer in the DKK1 protein and the mechanism of action of DKN-one and Doctor. Klempner will review the clinical data from the study and his experience in treating patients with DK101. We will end the call by opening the floor to questions for Doctor.
Zajani and Klempner or any of us from LEAP.
Hello, everyone. My name is Zafar Ajani. I'm a GI medical oncologist working at MD Anderson. I do a lot of clinical research as well as I have my own laboratory. And today, I'm really happy to talk to you about a molecule called dkk1, which is in the Wnt pathway.
And this is secreted by the tumor cells, just like PD L1. However, DKK1 is very unique in its functionality. So not only the secreted molecule DKK1 is immunosuppressive, which would be similar to PDL-one. 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 mTOR pathway. So in that respect, I think this is a molecule that is more damaging to the host of cancer than, for example, PD L1.
And 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 BKK1. On the left side is the PCCA pan cancer dataset, which is thousands of patients, multiple tumor types, I think 33 tumor types and 10,000 patients. And we can see that the tumors that had high DKK1 had very poor prognosis.
So the p value is very strong and patients tumors that had low DTK1, their survival was clearly significantly better. On the right side, it is drawing from the same data set, but it's customized to stomach cancer, TCGA data set, which is about 350 patients. And again, in gastric cancer, even more graphic to see that high cancer patients had much worse survival than DKK low. So 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 engage with LEAP and their clinical trial.
So on the left side, we were looking at normal tissue, adjacent tissue from the tumor and also the primary tumor. And you can see the expression levels are clearly much higher in once the tumor is formed. And in the middle curve, again, this is an independent cohort of patients where high GKK1 in our analysis also showed significant prognostic value. So here we are talking about 415 patients. And then on the right side, this is another data set, which again validates our finding of diagnostic importance of PKK1.
So there also you can see that patients with high TKK1, they 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 oncology. And today, we are not focusing on that. But this is just to expand on the understanding of how 1 might be operating in the cancer cells. So on the left side, you see the heat map and you see 4 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 oncogene, the sort of our abolished expression is abolished by just getting rid of the AP1. But in the right column, which is the control, you can see number of oncogenes are overexpressed, especially in the 3rd column. So there is some heterogeneity, which we expect. And DKK1 is highly expressed in both patients, but more so in the 3rd column.
And then we have other set of data here. Again, we are also looking at the plasma of healthy individuals, plasma of patients. And then we also collected the supernatant from malignant ascites. And there you can see that once you have stomach cancer, for example, even in the body fluids, in the plasma, the DKK1 levels are much higher than normal. And I personally think this could have an implication in patients also later on.
So we could actually monitor therapy duration, first therapy effect and its duration by monitoring the blood. But I must say this is not validated yet. It's just very early data set. We have also done some other mRNA level q PCR, which is on the right upper side. And as I was mentioning that the YAP-one seems to regulate DKK1.
And when YAP-one is knocked out in 2 different clones, DKK1 levels go substantially lower. We'll go to the next slide. And this is a very small data set, very preliminary data just to understand whether BKK1 level are different in different cell types in the peritoneal cavity. So what we normally do is we will collect the peritoneal malignant fluid, but it is crowded with different cell types. And here we are showing that in certain clusters, BKK1 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 by 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 TKK1 level, not only your prognosis is worse, but your tumor is also resistant to therapy. So on the left graph, I'm showing you the duration of treatment with a very popular drug called paclitaxel. So here patients were getting paclitaxel, but patients whose tumor had low levels of BKK1, they actually receive much longer treatment, mainly because paclitaxel is effective against those tumors.
But the tumors with DKK high, the drug has to be discontinued early because the tumors are resistant. This is contoured by PKK1. The same set of results is shown on the right side, except the drugs are different. So this could be considered like pan resistant. This is a phenomenon we see in cancer stem cells.
They have multiple resistant phenotype and it is very hard to overcome that. So DKK1 may be conferring resistance to the drugs that these tumors haven't even been exposed to their sort of inherently have become resistant. But the duration of treatment also is shorter for patients whose tumor overexpressed PK1. In the next slide, we again show very similar effect. It is colorectal carcinoma cells on the left side.
But the experiment is a little bit different here. We are knocking down the DKK1. In other words, there is no way these cancer cells can express DKK1. And you can restore the resistance to fiber Q in these cells. So 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 DJK1 on the immune system. BKK1 on the immune system. And as I mentioned earlier, that not only BKK1 supports the cancer stem cells and resistance to therapy and poor prognosis, but it also is complemented by its effect on the immune cells. And here what we can see that the anti tumor immune response and that can be modulated by modulating CKK1 with the antibody. So 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.
Microenvironment. So this is my last slide. And here, basically it shows the fact that you can magnify the effect of DKK1 by combining it with anti PD-one and which is sort of what led to the trial to combine both molecules targeting cancer cells and with PKK1 and targeting immune cells with anti PD-one. And you can see that on the left side, you see the CD45 positive population. Other cells that are increased by inhibition by PKAN1 And on the right side is the effect on tumor volume, which is maximum with the combination.
So we can conclude that first that this AK1 expression in number of cancers, variety of cancers, in large data sets prognostic than it is from the same large data set is impressively prognostic for gastric cancer and from MD Anderson patient cohorts, it is completely validated. We internally validate it. Now we can say that it's going to be validated from the data I presented to you that are not from MD Anderson. And the important thing is that it does promote resistance to therapy in addition to being a poor diagnostic factor. And we have found at MD Anderson that it is highly regulated by other oncoproteins.
But in the preclinical model, we can manipulate PKK1 as a target by PKN1 antibody. And we can reverse some of the resistance and we can also turn the tumor macular environment in favor of the patient. So combining PKN1 with PD L1, we would clearly expect that we can get greater advantage, the amplification of the effect of BGN-1. So with that, I will stop. Thank you.
Hello. Good morning. My name is Sam Klimtner. I'm 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 distinguished trial that was presented at ESMO 2021. So by way of brief background, you've already seen some of the preclinical data regarding DKK1 and DKN01.
But just to put some context to the DISTINGUISH trial, DKN01 has previously demonstrated significant activity in combination with checkpoint inhibitor in a prior trial that is most recently published. So 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 DKK1 high and the DKK1 low population in terms of response rate. And so 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%. And 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 physicians.
So 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 2 relevant clinical situations. Part A is the frontline trial. So this is really, in my mind, an important subset because we've seen there are limitations to the current checkpoint inhibitors and the frontline chemotherapy. So we're looking for ways to expand the benefits and optimally select the patients.
So the frontline combination is standard 5 FU platinum with PD-one agent plus CKAN-one in a cohort looking at response rate and safety. And then in the second line, this is really a validation and confirmation from the prior work looking at a biomarker selected population, BKK1 high, and this is a chemo free regimen of PKN-one 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 DK101 dosing based on prior experience. So 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.
So kPOQ, tislelizumab and DKN-one, everyone received this regimen on the single arm Part A, and these are given in 21 day cycle. This is just a little bit more information on the study design and methods. So again, this is a single arm Phase IIa trial of VKN-one plus the PD-one agent tislelizumab plus standard frontline chemotherapy 5FUN platinum, in this case, kPOQ. The primary endpoint of this Phase II trial is overall response rate. Additional endpoints are shown here.
And importantly, you see the interest in trying to tease apart activity by substance. So the modified intent to treat is everyone who got more than one dose of VKN-one. And then there's analysis by DKK1 expression, which is previously established based on the prior experience using an hscorerna scope method, which we will talk about a little bit. And then 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-one 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 G junction population. So you can see the breakdown in the frontline population, pretty even split between ECOG01. You can see slight enrichment in GE junction, adenocarcinoma patients. Again, this is just reflective of the Western population and what would fully be expected in the U.
S. Largely driven trial. You can see here that these patients have 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 BKK1 high and BKK1 low is present both in the GE junction and the gastric population. So the biomarker is importantly present across the clinical spectrum of disease.
And really, there may be a slight increase in patients presenting with stage 4 disease in the DKK1 high population. And 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 have DKK1 availability for 21, and 12 patients were DKK1 high and 9 patients were DKK1 low. And 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. And here we see the biomarker as present in over half of the population suggesting that 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 presented. PD-twenty two patients, so nearly everybody had PD L1 expression available. And as you can see, this largely reflects what we see in practice. I think the CheckMate 649 data suggesting that 60% of patients are CPS5 or higher is somewhat hard to explain and not what we see in clinical practice.
In my experience, the rates of CPS greater than 5% or greater than 10% is probably in the 10% to 15% of what I see in a high volume center where we see more than about 400 patients
a year.
So you can see that about a quarter of patients were CPS negative, 3 quarters were CPS less than 5 and consistent with what I was saying, there is only 2 patients here who are CPS greater than 10. TMB high is a population we think may have a greater chance of benefiting from checkpoint inhibitor. Again, the vast majority of patients are mutation burden low here and high mutation burden is outside of MSI high, very uncommon in gastric and GE junction cancer. So again, this is just reflecting what we really see in the clinic. And there are no microsatellite high patients among patients with available data.
So really, this is a population that you're seeing that, 1, reflects the real world and 2, 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 VK101. I think here we're seeing a little bit bit the opposite where this is a population that's perhaps biased towards patients who are unlikely to benefit from checkpoint inhibitor in addition to 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 5 months and the longest duration is over 10 months at this time and 16 patients remain on therapy. And 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, 1, the ability to do the biomarker and just representative images from biomarker testing. This is ZKK1 expression using the In situ hybridization RNA scope assay, complemented with digital pathology. And this has now been, I believe, published by the company about this method. So essentially, tumor specimens readily obtained in paraffin are seen for DKK1 expression and quantified using a digital image analysis algorithm. And then an H score is calculated.
And H score is a well validated method in pathology to sort of comparing a high versus a low. I think it's quite clear, 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. And really to me, I think this is when we look at biomarkers, we want to think, is it both analytically valid and clinically feasible? And I think a reproducible biomarker testing is important.
And I do believe that this RNA ish assay is assay is a reproducible assay. Now we're getting into the meat of the study. This is best overall response by DKK1 expression. And of course, this is a shape of a waterfall philosophy 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.
So one thing that's immediately clear is the depth of response and the enrichment for responses among the DKK1i high population, which is the green bars. But there are still some responses in DKK1 low population, then we'll get into this a little bit more. And then this is colored by a GE junction, adenoid gastric, and this largely just reflects the trial population. Let's put some numbers to that waterfall plot. It's really clarified in this slide.
And so here, what we're showing is the evaluable DKK1 high GE junction and gastric patients, and everybody had a partial response. And so as you can see, there's 1 non evaluable patient in the BKK1 high group, which consists of a total of 10 patients, and the partial response rate is 90%, and everyone had a response among the valuable patients. Interestingly, among the DKK1 low population, the response rate is still 55%, with 4 of the 5 people still on therapy. And 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.
And so what you see is, again, consistent with the waterfall plot, the vast majority of these lines are below the 0, suggesting some degree of tumor shrinkage. And there's a clear enrichment of the green, which are the DKK1 high patients, both in terms of depth of response but also durability. And now you see this is, of course, not significantly long follow-up, but at the available follow-up, you see that there is a large proportion of patients on therapy at 3 6 months. And 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 and largely. This is a swimmer plot looking at the differences between durability and DKK1 expression.
And so on the top, you see the DKK1 high population. And on the bottom, you see the DKK1 low, and then the farthest is the unknown population. And 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. And 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 expression by response.
So 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. So 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?
And then a predictive
biomarker is, do you have the ability to identify a population that's more likely to respond to your given drug. And so I think we've seen some data on prognosis already. And 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 how does the 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. So here we're seeing best overall response by PD L1 and DKK1 expression.
And what you see is the DKK1 high is again colored in green and the low is colored in blue. And then above the bars, sorry, you see the CPS, the pluses are the DKK1 high and the minuses are the DKK1 low and the green is CPS high using a 5 cut off, which is what was used in CheckMate 649 and the blue is a CPS less than so on. 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. And so this is a population of patients where you would not expect significant benefits from kPOXX and PD-one because of the PD L1 negativity or less than 5.
Again, that's a population that had a unremarkable hazard ratio in the CheckMate 6.9 trial. But 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. So this is just trying to confirm and convince us from this available data set that EKK1 high patients are responding regardless of PD L1 status. And so just to break this down, the response rate was 79% in patients with PD L1 low expression, CPS less than 5 and 100% in the DKK1 high PD L1 low.
Again, for reference, the response rate in PD L1 less than 5 from the CheckMate 649 population is about 50% to 55%. And so here, we're seeing that even in the even in the BKK1 lower population. But clearly, it's significantly enriched in the BKK1 high population here where the response rate was 100%. And you can see, although the numbers are relatively small, the response rate in the PD L1 high population still exceeds what was seen in CheckMate649. Again, this is a smaller significantly smaller numbers, but certainly an encouraging signal.
And here is again a swimmer spider plot showing durable responses independent of PD L1 expression. And here, CPS is coded as green for CPS greater than or equal to 5 and blue for patients who are low. And 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. And just to formally compare this, there is no correlation between PKK1 and PD L1 expression.
And this is consistent with what we've seen in prior data with DK101 and pembrolizumab. And I think the message here is that really, a population of patients who are more likely to respond to checkpoint inhibitor, which would raise some concern about the biomarker. And 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 the standard of care regimen, we certainly need to understand the toxicity profile. And 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 VKN-one related events, fatigue, nausea, diarrhea. Again, these are consistent with what was seen in the prior experience with checkpoint inhibition. And there are some cytopenias, which may or may not be more related to the chemotherapy in my own opinion.
And then there are rare greater than or equal to 3, again, diarrhea, and there was 2 pulmonary amyloid on the trial. But 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. But it's just reference to give you some context of the encouraging response rates seen in this CKK1, DKAN-one trial. So to summarize, I think what I've tried to show is DKNO1 and ciculizumab in kpox is a well tolerated regimen.
Again, the 5FQ platinum and PD-one are something we have a lot of experience with. And adding DK101 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. I would say that response rate is a very important endpoint consider in gastric and esophageal cancers in particular.
And 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. And so 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, etcetera. So I can speak from personal experience, but also general experience with biologically active regimen and patient outcomes. Certainly, the response outcomes compare favorably to current standard of care. And again, this is an unselected population.
And I think this is something that's important because when you look at KEYNOTE-five ninety and CheckMate 649, there really is no benefit to the CPS less than 5 or CPS less than 10 populations, respectively in those trials. So 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 benefits. It does seem in this data set that the efficacy is driven largely by the enhanced response rate, which is very high in the BKK1 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. And 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, 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. But 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. So that's always attractive to us clinically when we're looking to select patients because it's the majority of our current biomarkers like MSI is only 3% to 4% of our patients, PD L1 high, so CPS greater than 5% or 10%, again, it's 10% to 15% in my own experience. And HER2 is maybe another 15%. So biomarkers that have a high prevalence are something that are attractive, I think, to many of us as clinicians. But I'd like to thank you for your time, and I'll be happy to take further questions.
Thank you, Doctor. Ajani and Doctor. Klumna, for your participation in today's program, and thank you all for your time and attention today. We'd now like to
And 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? And then related to that, how do you anticipate using DKK1?
And is there an opportunity in the patient to diagnosis process to be able to assess all these upfront before that first line therapy has started?
Yes, I can answer it first. So this is Zafarajani. We don't use nivolumab or pembrolizumab in if the CPS score is low, like less than 5 and for pembro, it has to be less than 10, then we don't use it. So I use it only for those patients with CPS of 5 or higher. I would
I'm sorry.
Yes. I was also going to say, it depends on how the Phase III trial is designed with DKN1. It may be focused on maybe Cindy can elaborate on this. If it's focused on DKK1 expression irrespective of CPS, then I think that's the way the trial will go irrespective of CPS.
I totally agree with Doctor. Ajani. We do not offer checkpoint inhibition in the first line to our patients who have the same CPS stratification that Jaffar was mentioning, less than 5 and less than 10. I'll also just say one thing. In terms of feasibility of selecting on a biomarker like PKK1, I think we saw some data from this trial called FIGHT101 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.
And 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. So I think, yes, there's definitely opportunity and feasibility to do something like a DKK1. And I do think, in situ hybridization based biomarkers may become more common as well. I like the strategy.
Got it. Thanks for that perspective. And then maybe one follow-up question. In trying to assess how this data from this trial looks, I think one of the comparisons is with the nivolumab data. And I noticed there's different data that's on the label versus what was presented this year at ASCO and I believe in the Lancet publication with ASCO data seems to be a little bit higher than what's on the label as far as ORR.
So I guess in interest either for the physicians or the company, would you be able to kind of help us assess what that difference is and what would be the more relevant comparison for the data on lead station 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 and that the supplement to the Lancet article had 370 fewer patients and then that represents the ORR and the subgroup reported on the supplement of the Lancet.
And so we provide both of those for context because people can find them when they search. But I think 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 end of patients enrolled.
Got it. Okay. That's helpful. I noticed that somewhere around 20% to 25% of the patient population seems to have dropped off between the label and the ASCO data that was presented. So 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. I appreciate you guys taking my questions and congrats on these data. Maybe first, a couple here. One for Doctor. Pemberm, 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?
And then, 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 selected population like DKK1 high and PD L1 low or enrolling irrespective of CPS status? Thanks. And I have a follow-up as well.
Sure.
Yes. I mean, the data is the data. So we're limited by follow-up, as you mentioned. I mean, the encouraging thing, 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. But similar kind of phenomenon in KEYNOTE-eight eleven where you look at the shape of the waterfall plot and the FDA approved that combination based on response rate alone so far.
But the other thing that I think is notable is that among the responders, essentially, what, 7 of 9 are still on therapy. So the majority of people who responded continue to remain on trial. So the proportion of responders on trial is high, and that may be something to look forward to as a marker of maybe an encouraging progression free survival. So I think the data is what is presented, and we can future cast only so well based on what we have. But the high rate of patients who remain on trial and the depth were things that were noted to me.
Doctor. Ajani, did you want to add anything before I kind of respond on biomarker strategy a bit?
Yes, 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. So, I think those issues have to be considered in judging these data. But there are patients I have patients who are on for more than 9 months. One of them is still getting therapy after almost 1 year.
And their quality of life is very good. So as all of you know, what we have to do when we start with 3 drug or 4 drug combinations is we have to drop oxaliplatin going forward. And often we will drop the second cytotoxic, which is the fluopyrimidine. And then patients stay on a targeted therapy or immunotherapy. So I think if you consider that, it is very likely that some of the patients will continue combination of PKM-one and tislelizumab for a long time.
So the toxicity profile will be probably more intense in the 1st 3, 4 months, but then we can titrate it out and patients should do very well as some of my patients are doing.
And then 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. And 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 do its tolerability profile and as Doctor. Klevener mentioned, the depth of response and the kind of curves of the response. But 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. And so on a philosophical basis, we support and believe that the right strategy would be to identify a DKK1 high patient population, focused 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 DK101 based treatment group and 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. And I could just squeeze in a quick follow-up. Would you happen to be able to provide the details of the EKK1 status of the 2 patients who reported pulmonary embolisms? And, 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-sixty 2. So if you could help there, that would be great. Thanks.
Why don't I let maybe Doctor. Klemmer 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 PDL1 greater than 5? But in terms of the frequency, maybe Doctor. Kleiman or Doctor. Ajani can comment on it or its customer in us.
Yes. I mean, I think I was a participant in the prior trial where there was no chemotherapy component to the DKN-one and PD-one combinations. And broadly, I think all of us in oncology have a fair experience with pulmonary emboli and cancer related venous thromboembolic phenomenon. And 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-one trials, chemo HER2 trials.
I'm not aware of any enrichment or association between necessarily therapies that we use commonly and increased EVT risk. I mean, certainly with anti angiogenic drugs maybe, but those are more true with like maybe rambucirumab and some of the VEGF TKI, which is quite different than PKNO1 in my opinion. And then, I can comment broadly on the toxicity profile, and I think Doctor. Ajani mentioned as well. These patients are getting chemotherapy, immunotherapy and another immunomodulatory drug in DK101.
And so we have a lot of experience and expectations with the toxicity profile from that combination. And in my own experience, I personally have not seen an enhancement of the side effect profile with the addition of the DK101. I think the cytopenias are largely chemotherapy related. There's a known AE profile from the PD-one agent. And there's some experience with TCAN-one toxicities from monotherapy and from prior PD-one combinations.
And I think that's been similar in this trial from my own experience. And I'll leave it at that and let Jafar add whatever he wants.
So 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. So I see a patient in the clinic. If you order an imaging study and then you get a call in the evening saying there is a pulmonary embolism. This is untreated situation.
So I think having 2 episodes here, a PE in a 25 patient population would be baseline. So I personally didn't think that DKN1 is contributing to PE in this study. I think if there were like 7 or 10 cases, then we would be worried about that toxicity. But this is 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 Timur Ivankov with Raymond James. Your line is open.
Yes. Hi. Thank you for taking our questions. Congrats on the data. So just had a quick question on potential durability.
So in terms of the response trend in DKK1 high patients, so the 2 patients with the longest follow-up around 8 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? And other patients, they also appear to be sort of steady. And then maybe also for doctors Ajani and Klentner, 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. So 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. And this has been published from our institution and others. So there is 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. So 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. So most of these patients will require another treatment. So longer we can wait to use the second treatment, better it's going to be. So suppose this second line treatment we don't have to use for another year, the portfolio will be much better. So I think in that regard, it's great to have 50% reduction and a very long duration of response.
Yes, I agree. I have very little to add to what Jafar was saying. I would say, also just to put some context, the rates of complete response in metastatic gastric and esophageal cancer are very, very low. I mean, maybe 5% to 10% at most with chemotherapy and maybe in and Jaffra 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, very low.
I think it's hard to tease too much more out from the spider plot that you were referencing. But largely, I completely agree with Joffrey. 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. Because 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.
And so we got 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. And then just maybe a quick follow-up based on what Doctor. Ajani had said. So 7 out of 9, 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 KAN-one and tislelizumab? Thank you.
I don't know about all of the patients, but I have currently 2 patients. 1 is just getting DKN1 and tislelizumab. The other one is also getting capecitabine immunotherapy drugs. And they are doing extremely well. I mean, they are doing functionally almost close to normal.
Yes. We've taken the exact sorry, Cindy, go ahead.
No, no. Go ahead,
Tim. 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 in the oxaliplatin. So we've also done capecitabine, tislelizumab and DKN-one as sort of the transition, and our patient continues on trial as well, feeling well.
Yes. And I was just going to add that the protocol does permit them to stay on any combination of the 3 provided they remain on DKN-one. And I think consistent with both Sam and Jaffray have noted today, many sites are dropping the cytotoxic chemotherapeutics with Doxali being dropped first, followed by the ZILOTA as necessary.