Insight Molecular Diagnostics Inc. (IMDX) Earnings Call Transcript & Summary

Good morning, and welcome to the OncoCyte KOL call on Novel Biomarkers in Identifying Appropriate Responders to Immune Checkpoint Inhibitors. [Operator Instructions] As a reminder, this call is being recorded, and a replay will be available on the OncoCyte website following the conclusion of the event. I'd now like to turn the call over to your host, Ronnie Andrews, President and CEO of OncoCyte. Please go ahead, Ronnie.

Thank you, Tara. And good morning everyone, and thanks for taking time out of your day. We're very excited to bring to you another in our series of KOL events. Today, we have 2 amazing presenters, and we'll be happy to allow you guys to hear them. We want to give you a little introductory to the company for those that don't know us. So if we can change the slide, please? Of course, there's going to be forward-looking statements today. I won't read this to you because I'll put everyone to sleep and there's way more important things to talk about. Next slide, please? About 20 months ago, a group of us came together that had spent the last 15 years looking at large-scale genomics and the impact of genomics in oncology, specifically around being able to select patients for therapeutics, both for pharma trials and ultimately for clinical utility. What we saw was the ability to reduce the practice large-scale genomics to precision diagnostic tests that would actually map out to a specific endpoint. And those endpoints we chose were very important in terms of questions today that aren't answered or not answered appropriately or adequately by the testing that exists today. Next slide. So our road map has really been focused on 2 areas: treatment decision and then patient monitoring. And so today, OncoCyte has 2 tests that they are in the middle of launching. One has -- was launched last year in the middle of a pandemic, DetermaRx. It is a test for early-stage lung cancer patients who get surgical resection and no follow-up chemo to identify high-risk patients within that population, and we have data that adequately shows that those high-risk patients, a great percentage of them move on to a bad outcome. And so by using this test and getting that patient on an early round of double platinum chemo, you improve the life expectancy in our randomized and prospective study from 49% to 92%. And so that test is -- answers a really important question, do I give adjuvant chemotherapy for an early-stage lung cancer patients? The second question that we chose to tackle was, should I give immune therapy? And we all know the power of immune therapy and the excitement we all have around immune therapy. But the challenge with immune therapy is not everyone responds. And those who don't respond get exposed to an agent that exacerbates the immune system. And we see other comorbidities such as autoimmune diseases and other things that complicate the life of a cancer patient. So the question was, can we take some of the larger-scale genomic data and set the size down and identify a precision-diagnostic tool that allows us to identify the patients that will and will not respond to this drug. And so today's session is on our DetermaIO test, and we'll get to those data in a second. And finally, our quest continues. We announced this morning the acquisition of a company called Chronix Biomedical in Germany. They have a very unique patented assay for looking at early degradation of signal post therapeutic intervention for Determa -- or for immune therapy using DetermaIO-positive patients. And the idea is that we'll be able to, with blood only, without the use of any upfront tissue within 3 cycles, be able to identify with great positive predictive value progression of the disease. So more to come on that test in the future as we begin to put it through its clinical rigor. So today, I'd like to turn the call over to Dr. Doug Ross, our Chief Science Officer, to give you a little overview of DetermaIO before we hear our speakers. Doug?

Yes. Next slide. So thank you, Ronnie. And I thought I'd get us started by just giving the cartoon version of what we're going to be talking about today. The role of diagnostics in DetermaIO specifically in managing the response to immune checkpoint inhibitors. So just very briefly, the story of checkpoint inhibitors, which has gotten the Nobel prize over the course of the past few years is that we all have little tumors that form in our body over the course of our lifetime, but our immune system keeps them in check. That's shown over on the left. But over the course of time, perhaps the cancer begins to break through that immune surveillance, and it becomes a clinically evident cancer that has to be dealt with. And checkpoint inhibitor therapies have been invented that modulate the immune system and allow the immune system to better attack those emerging cancers. The problem is that even under the best of circumstances and the best of tumor types, these immune therapies only have about a 15% to 40% response rate, meaning 60% to 85% of those tumors are resistant. And so what DetermaIO, our test that we've developed inside of OncoCyte, is designed to do is to measure the physiology of the immune response and identify those tumors that are poised to respond to the addition of checkpoint therapy. And what we believe makes this test different from other tests is that it recognizes both what's called the hot component of the immune system, the inflamed cells in the tumor that are going to respond to the tumor, but it also recognizes the cold signal from the tumor immune environment, which is responsible for a negative signal that represses the immune response. And we believe and have data to support that the combination of those 2 is much more powerful for measuring the immune response and identifying the folks that should get the best response out of checkpoint inhibitor therapy. In the next slide, we show that we've distilled what was started as a very complex gene expression signature down to a very simple assay measuring the expression of 27 genes that can be run on real-time PCR machines. And that's important because checkpoint inhibitors are offered around the world, and those platforms, those real-time PCR platforms, are distributed worldwide. And so pharma companies who want to reach patients all around the world for their products can do that using an assay like this. So we measure these 27 genes by acute PCR. We interpret them through a proprietary algorithm. And we identify patients that are likely respond -- to respond to current immune checkpoint inhibitors versus patients that are unlikely to respond. And this test allows understanding the entire tumor microenvironment. And once again, we believe it is distinguished by really incorporating 3 different components: identifying these immune inflammatory infiltrates that are responsible for killing cancer cells; identifying tumors that have simply excluded the immune system by undergoing epithelial to mesenchymal transition; as well as identifying the repressive signal that comes from the wound response around the tumor from the cancer-associated fibro plus. So that's a brief overview of where we're going. And with that, it's my pleasure to introduce our speakers today. It truly is, as I've gotten to know Dr. David Gandara over the course of the past year or so. I've admired him at a distance for a long time that I've spent many calls with him and have appreciated his approach is rigorous approach, the patient-centric approach to driving towards advancing the way we manage patients. And more recently, was introduced to Dr. Parikh but again, have learned so much on every call that I've been with her. It's truly a pleasure to have them presenting for us. More formally, Dr. Gandara is Professor of Medicine Emeritus at the University of California School of Medicine and Director of the Thoracic Oncology program at UC Davis Comprehensive Cancer Center. He's an internationally known clinician scientist and lung cancer thought leader and has written over 700 articles, book chapters, abstracts and editorials. He is past President of IASLC and a prior Board Member of ASCO. Dr. Gandara will present an overview of successes and challenges with biomarkers to predicting in a therapy response as well as review data in non-small cell lung cancer, supporting the notion that DetermaIO has incremental utility beyond PD-L1 and tumor mutation burden and identifying responders to immune checkpoint inhibitors. Following Dr. Gandara's talk, we'll hear from Dr. Mamta Parikh, who's a medical oncologist specializing in the treatment of genitourinary malignancies and an Assistant Professor at UC Davis. Dr. Parikh is the principal investigator on numerous clinical trials, including an NCI-sponsored trial of a novel ATR inhibitor in advanced solid tumors. She was the recipient of the Joseph Sullivan Research Award in 2016 and is currently a recipient of the NCI-sponsored Paul Calabresi K12 Career Development Award. Today, Dr. Parikh will be presenting on the unmet need for better immunotherapy response prediction in bladder cancer and will be discussing new data first presented during an oral symposium at AACR just in the past couple of weeks, showing that DetermaIO has strong performance in identifying responders to atezolizumab in bladder cancer. So without further ado, Dr. Gandara, please take it way.

Thank you, Doug. Can we have the next slide, please? So as you can see from the topic here, my task for today, which I would argue, is one of the greatest unmet needs in all of oncology, is to develop better predictive biomarkers for checkpoint immunotherapy efficacy. And I'm going to use advanced non-small cell lung cancer as a model here because it is a poster child for immunotherapy and also for those unmet need in many ways. My disclosures for this presentation are that I am a consultant and collaborator with OncoCyte, and I thank them for the opportunity to participate today. And also, I am a consultant for several other biomarker companies, including Foundation Medicine and Guardant Health, as well as a large number of pharmaceutical companies. Next slide, please? This is a quite complex treatment algorithm for Stage 4 non-small cell lung cancer based on biomarker-driven therapeutic strategies. If you can advance once. In this algorithm, which was in a presentation at our International Lung Cancer Congress last year by Solange Peters, the first thing that's done in a new patient with advanced non-small cell lung cancer is the tumor is tested by tissue and/or by liquid biopsy by next-generation sequencing for oncogene drivers as well as tumor mutational burden and PD-L1 immunohistochemistry. Advance once. Those patients that are found to have 1 of the 8 guideline recommended oncogenes are then treated selectively with targeted therapies, TKIs directed against these oncogenes. Everyone else on this slide, and this amounts to at least 75% of the patients in the western world, will receive immunotherapy as the first-line treatment in some way. But how that happens is dependent on these biomarkers. PD-L1 high, and the cut point here for high is 50% or greater; or TMB, tumor mutational burden, high or low. And you can see the various categories of treatment, either monotherapy with checkpoint inhibitors or combinations of these inhibitors with platinum chemotherapy or combinations of immunotherapy themselves, such as nivolumab plus ipilimumab. If you are, what I would call, a splitter oncologist who divides up your patients for personalized medicine, then you try to use this algorithm. But the point of my presentation will be that it is imperfect. On the other hand, because it is imperfect, many oncologists, particularly those in the community setting, will treat all of the patients the same way, with checkpoint immunotherapy plus platinum chemotherapy regardless of these markers, as long as that patient doesn't have an oncogene. So this is the dilemma. Next slide. As Doug has already introduced, developing biomarkers for checkpoint immunotherapy is incredibly complex by comparison to discrete biomarkers such as mutations, which we use for the oncogenes. And this is an article by [ Blank It All ] from science a few years ago, which I have updated. But it shows the complexity of tumor neoantigenicity, tumor microenvironment, host environment and tumor immune suppression innovation, all the points that Doug made earlier. And I will focus on the tests which have FDA approval relevance that is PD-L1 TMB as well as some that are in development. We don't know these are not standard of care yet. But these are markers of either immune suppression, which prevent the efficacy of checkpoint inhibitors or perhaps predicted biomarkers for efficacy. That last one is AradA1 (sic) [ Arad1A ], the former, STK11 and KEAP1 mutations. Advance once. And then I will give the lung cancer information regarding this new test, DetermaIO, the test that Doug just described, as an example of a way to move forward. Next slide, please? And I developed this slide to show you this complexity and why it is different from oncogenes. Advance once. So the first point to be made is that biomarkers for checkpoint immunotherapy, by and large, are not discrete variables like driver mutations, which are present or absent. Instead, PD-L1 and TMB in particular, are dynamic, meaning they can change over time, and they are expressed as continuous variables across a context-specific range. And what I mean by that is that range is quite different from lung cancer to breast cancer to hematologic malignancies. Advance once. And as continuous variables, they can be analyzed by different algorithms. They can be analyzed across a continuous range or as a binary variable, by quartiles or the highest 10%, et cetera. Advance once. This is an example of our own work developing a blood TMB assay analyzed as a continuous variable and picking what we feel is the most appropriate cut point. Advance once. And in comparison, a very nice study by Sam Stein [ it all ] using the Memorial Sloan Kettering IMPACT platform, analyzing TMB across multiple tumor types as a binary variable, showing if you pick the highest 10% or 20% for a given tumor type, then TMB is highly predictive across tumor types. Advance. For biomarkers such as TMB, we can also analyze them by research tools such as Roche Sequencing or an even more sophisticated way of looking at neoantigenicity that is new antigen load or now by commercially available next-generation sequencing platforms. Advance. And we can also now do this in tissue and blood. Advance. So importantly here, because DetermaIO is being developed as a pan-tumor biomarker, for PD-L1 and TMB, the issue is what is the context? What's the tumor type? And for lung cancer, this could be squamous or non-squamous non-small cell lung cancer or small cell lung cancer. What's the regimen? Is it a checkpoint inhibitor by itself? Is it in combination with chemotherapy? Or is it an IO-IO combination? Advance. And part of the hypothesis for the work that I will describe to you is that chemotherapy is likely agnostic to these immune biomarkers, and it will dilute out some of the predictive value. Next slide. There are data in the literature supporting several potential checkpoint immunotherapy markers for the efficacy of these drugs, and I show 4 examples here: the first to the left for PD-L1 high; next, tumor mutational burden; next, a gene expression signature combined with TMB, a composite biomarker; and then lastly, tumor infiltrating lymphocytes by themselves. Advance. For PD-L1 briefly because I know many of you are quite familiar with this biomarker. It's in the FDA approval for lung cancer in many regimens as well as other tumor types. This is one of the initial presentations on PD-L1 expression across tumor types, across drug categories, across multiple PD-L1 assays. Advance. And what it showed was that PD-L1 positive cancers had higher response rates to these drugs regardless of the tumor type, regardless of the assay for PD-L1 and regardless of the tumor type. But as you can see at the bottom, we don't even assess PD-L1 the same way in different tumors. For example, in non-small cell lung cancer, we use TPS. This is the tumor content. And the example there is for a high-expressing tumor. But in many other tumor types, we use a composite biomarker of TC plus IC, the immune infiltrating lymphocytes. And there is an example for -- for example, where the TC is 0, but the IC is high. So quite a bit of variability here. Next slide. But the bottom line is in the clinic. In non-small cell lung cancer, about 30% of patients will have this very high level of 50% or greater. As shown on the left, in the KEYNOTE-024 study using pembrolizumab as monotherapy, it is quite an effective biomarker. As shown on the right, in combination with chemotherapy, PD-L1 is still an effective biomarker. But if you can advance. But it's not a complete biomarker. As you can see here, those patients with PD-L1 0, meaning no staining, whatsoever, not a single cell staining for PD-L1, there is still improved survival in this combination with chemotherapy. Next slide. And this is true in our own study as well. This is the HOPE trial of atezolizumab versus docetaxel. If you can advance. So using the SP142 assay, which is one of the only PD-L1 assays that test for both PC and IC, you can see it as positive for survival, even if both of these are negative. If you can advance again. Mimicking other studies, you can see even those patients with TC3 or IC3 have the best benefit by far. So PD-L1 is a useful biomarker, but I don't think we will be using it the way we do now in 3 to 4 years. We'll be doing something else. Next slide. And that brings us to tumor mutational burden. Doug has also already given a very nice introduction on the effect of neoantigens here and how this translates then into measuring tumor mutational burden, which we can now do by exome sequencing as well as commercially available next-generation sequencing platforms in both tissue and blood. If you can advance. This shows the Foundation 1 assay in comparison to whole exome sequencing or neoantigen load, and it is quite comparable. So this is a good measure of TMB. Advance. This commercially available assay and tissue correlates quite well with the efficacy of checkpoint monotherapy quite well by comparison to whole exome sequencing. And advance. And perhaps most important, and this has been surprising since day one, but it is highly consistent. PD-L1 and TMB largely are non-overlapping. They measure largely different populations of patients. Next slide. Shown here are 3 studies, which show that a high-tissue TMB is associated with checkpoint inhibitor monotherapy. If you can advance. On the left, the CheckMate 026 trial, whole exome sequencing, showing, although this was a completely negative trial for efficacy of nivolumab, that if you look at the high TMB versus low, there was predictive value to the biomarker. And then below, you can see that a combination and an index of PD-L1 plus TMB high was most predictive. Advance. On the right, these are our own data using the Foundation 1 assay, again, positive for the efficacy of atezolizumab. And you can see in the bottom left, agnostic to the effects of chemotherapy. And then advance once more. This is the binary evaluation by Sam Stein, showing, again, using this as a binary variable predictive value and again, no association with the efficacy of chemotherapy. Next slide. TMB high by using the Foundation assay at a cut point of 10 or greater mutations per megabase was approved by the FDA as a predictive biomarker for pembrolizumab monotherapy based on a trial KEYNOTE-158 that assessed it across multiple tumor types. You can see the response rates total and then also for the multiple different tumor types, TMB high or low. Now although this gained FDA approval, it was based totally on response rate, and there was no clear association with progression-free survival or overall survival, so much left that would be needed. Next slide. Now all of those studies were in tissue. We developed in collaboration with Foundation Medicine and Roche/Genentech the ability to measure TMB in blood. I won't go through the computational algorithm, but it is very similar to that of the Foundation test in tissue. We had a test set of the [ popular ] study and a validation set with HOPE, and we -- and that assessed by continuous variable. If you can advance once. And we picked 16 mutations per megabase is the most appropriate cut point. You can see there was very little overlap with PD-L1. And in the table in red, those patients that had high levels of PD-L1 plus high levels of TMB did the best for progression-free survival and overall survival. But this was a retrospective study, so prospective studies are ongoing. Advance. Well, let's turn then to our algorithm. Advance. What happens when you add chemotherapy to one of these checkpoint inhibitors? Next slide. Oh, and I'll also comment right after that about emerging other biomarkers such as some of the genomic correlate. Next slide. This is my own analysis of several studies done to assess tumor mutational burden of checkpoint monotherapy in certain trials for combinations and others. These are 4 studies, all Merck studies using pembrolizumab, the first 2 pembrolizumab monotherapy and the last 2 with chemotherapy. As you can see, and this is consistent across multiple trials now that TMB high correlates with therapeutic efficacy when the drugs are given as monotherapy but not in combination. And in the bottom left, you can see one of the trials of monotherapy where there's a nice split in the curves; and on the right, no split. In other words, not an effective predictive biomarker with chemotherapy. If you can advance. And then more recent trials of immunotherapy combinations: nivolumab plus ipilimumab or durvalumab plus tremelimumab. And here are the results, including our own study in lung map are mixed. Next slide. Then returning to some of these genomic biomarkers at least in early studies. The mutations in STK11 and KEAP1 were felt to be predictors of poor outcome to checkpoint immunotherapy. This is the trial by colitis, which we participated in, showing that if you combine these markers, that is STK11, KEAP1 wild type plus high TMB, you get the best separation of the curves for progression-free survival and overall survival. So very encouraging that we might put these together in a combination index. Next slide. But other studies, and this is an analysis of the MYSTIC trial or durvalumab plus tremelimumab against chemotherapy. The authors felt in their analysis that these genes, STK11 and KEAP1, were prognostic. As you can see from the Kaplan-Meier curves, not predictive, but they identified a new biomarker Arad1A, which they felt had predictive value. There have been several studies, including our own ongoing looking at this things. Next slide. Well, this slide, I think, shows why we need better biomarkers. This is an interesting analysis from a couple of years ago, showing how many trials there are of all the various drugs that's shown on the left and what sorts of combinations are being done with radiation or chemotherapy or checkpoint or other checkpoint inhibitors. So at this time, 2 years ago, there were 2,250 clinical trials ongoing. We're acquiring about 400,000 patients. 750 trials just in non-small cell lung cancer, less than 5% of these trials, having a predictive biomarker involved. So this, to me, is much like throwing more spaghetti against the wall just to see if it sticks. It has been estimated that there are now at least 3,500 trials ongoing. Next slide. So with that unmet need then, there is -- here is just a sample of the clinical data that are being acquired with DetermaIO. And of course, because I'm a lung cancer specialist, I'm showing you the data in non-small cell lung cancer. This is the pilot retrospective analysis in non-small cell lung cancer. There are other studies ongoing as well right now. But this is 71 patients who received checkpoint inhibitor therapy, pembrolizumab or nivolumab. Retrospective analysis, analyzing by DetermaIO, the primary endpoint progression-free survival. Next slide. And the data are quite impressive. Shown on the right -- or, I'm sorry, shown on the left, are the percentage of cases positive and negative for DetermaIO versus PD-L1 immunohistochemistry versus TMB. So you can see for DetermaIO it is about 50%. That's actually quite good for a predictive biomarker. You don't want it skewed one way or the other too greatly. And then the 3 Kaplan-Meier curves, you can see that DetermaIO positive or negative, highly predictive in this retrospective analysis and better than the efficacy per predictive value of PD-L1 or TMB. Next slide. Furthermore, and these are in the silico data from the TCGA, DetermaIO seems not to predict for the efficacy of chemotherapy. So very important in the development of a biomarker in this space that it'd be predictive but not prognostic. These data were presented at the SITC meeting in 2019. Next slide. And this just emphasizes here at the levels of predictive value for this test. Next slide. And much as I showed you before on combining PD-L1 or TMB, combining the DetermaIO score with PD-L1, you can see in the upper Kaplan-Meier curve does not seem to add to the benefit of DetermaIO. However, again, this is a small retrospective study. It seems the DetermaIO plus high TMB may add to predictive value. Next slide. So I'll close with this slide. And for many of you in the audience, you will get a chuckle out of this. This is not from a medical journal but rather from the economist, June 2007, when I was interviewed about development of predictive biomarkers and another class of drugs. But it emphasizes the so-called untargeted use of targeted therapy because I do think that these drugs are targeted therapy if we can find the right biomarker for this unmet need. And you can see my quotation here. None of the recent trials were based on selecting patients whose tumors were producing [ targets ]. So with that, I'll close. Thank you very much. And I will now introduce Dr. Parikh, who will take us into bladder cancer. Mamta?

Thank you, Dr. Gandara. So thank you for the opportunity to speak. I will disclose that I am a consultant for Janssen, CGEN and OncoCyte and have received research funding from Karyopharm. So today, I'm going to discuss the role of immunotherapy in bladder cancer and the application of the DetermaIO testing to a particular study in advanced bladder cancer. Next slide. So just like in other tumors, bladder cancer treatment exists across the spectrum based on the stage of disease. If you go to the next slide, you can see -- next slide, yes, thank you. You can see that the treatment has largely consisted of surgery and chemotherapy for quite some time. But immunotherapy has become integrated into the treatment of bladder cancer at various stages, and I'll go through those. So next slide. The first of those indications is BCG refractory non-muscle invasive bladder cancer. So bladder cancer has been known to be responsive to immunotherapy for quite some time because BCG has been used as a treatment in this non-muscle invasive setting. But when patients become refractory to this treatment, they have the option of undergoing a cystectomy, which is a very morbid procedure or surveillance, which involves cystoscopies. So this trial enrolled patients that were BCG refractory and were treated with pembrolizumab for 2 years. And the primary endpoint of this KEYNOTE-057 trial was the complete response rate at 3 months. Next slide. So the study did meet its endpoint. The complete response was 41%. About half of those patients maintained their complete responses for over a year. And in January of 2020, pembrolizumab was approved in this indication. Interestingly, it's hard to see here, but PD-L1 status was evaluated in this study. And it's albeit small numbers, but PD-L1 positive patients actually appear to have a slightly lower response rate compared to those that were PD-L1 negative as defined in the study. Next slide. So now if we turn to advanced metastatic urothelial carcinoma. For a very long time, we didn't have any treatment options besides platinum-based chemotherapy. But in the last 5 years, there's really been an evolution in bladder cancer treatment in this advanced setting that has incorporated immune checkpoint inhibitors as well as subsequent therapy. Next slide. So the first of those patient populations we'll discuss are those patients with metastatic disease that are ineligible to receive cisplatin, which is otherwise the cornerstone for our treatment. So 2 immune checkpoint inhibitors have been studied in cisplatin-ineligible patients. These 2 trials that are shown are Phase II open-label studies of atezolizumab and pembrolizumab. Both of them had objective response rates that were higher than historical controls, which averaged between 10% to 15%. PD-L1 status did seem to correlate to an improved overall survival and a better response rate. But interestingly, follow-up studies that were a Phase III randomized studies looking at patients in the same population compared to chemotherapy showed that patients that were PD-L1 negative as defined by these studies actually had decreased overall survival if they were treated with immune checkpoint inhibitors. So atezolizumab and pembrolizumab both have altered their labels that in the cisplatin-ineligible patient population, patients must have PD-L1 positive staining by IHC. Next slide. But by and large, treatment for metastatic bladder cancer does involve chemotherapy that's platinum-based. And a significant portion of patients that receive platinum-based chemotherapy respond, which leaves us with the question of what to do with those patients after their response. So the JAVELIN Bladder 100 study addressed this question. It took patients that had responded to platinum-based chemotherapy and randomized them to receive avelumab plus best supportive care or best supported care loan and looked at overall survival in these patients. So in the next slide, you can see that overall survival in the total patient population was significantly better with avelumab. It was 21.4 months versus 14.3 months with best supportive care alone. If you go to the next slide, though, you can see that in what this study defined as PD-L1 positive, which was greater than 25% PD-L1 positive staining by IHC, there was a wider separation of curves in the patients that -- where PD-L1 positive did even better with avelumab when you compare to those that were PD-L1 negative. Next slide. Now we turn to patients that are platinum refractory that have metastatic bladder cancer. And many immune checkpoint inhibitors have been studied in Phase II studies in this area. And all of them have shown objective response rates that are better than historical controls, which range, again, around 10%. You can see that there are some modest improvements in objective response rate when we look at PD-L1 positivity, which is defined slightly differently for each of these immune checkpoint inhibitors. But of note here, there were 2 randomized Phase III studies that were conducted: one looking at atezolizumab versus chemotherapy. This is the IMvigor211 study; and the other one looking at durvalumab versus chemotherapy, and this is the Phase III DANUBE study. Both of those studies did not show a significant overall survival benefit for the immune checkpoint inhibitors versus chemotherapy. And so recently, Roche and AstraZeneca have voluntarily withdrawn their approval for atezolizumab and durvalumab specifically in these indications. Pembrolizumab is the one study that they did have a Phase III study that was positive in the syndication. Next slide, please? So I'll give you my opinion of why a better biomarker could matter for patients with bladder cancer as well as for oncologists treating bladder cancer. First of all, as you may have noticed from my presentation, there's a lack of harmony in how we define PD-L1 positivity depending on the immune checkpoint inhibitor. Overall, this discourages use of the biomarker for providers because it becomes very confusing. And so that discourages the use of PD-L1 status. For patients with non-muscle invasive bladder cancer, they have a noninvasive cancer that they could just undergo surveillance for. And so receiving 2 years of systemic therapy is really a change in how their disease is treated. It would be more compelling if we could identify patients that are most likely to benefit that would help us with those conversations with patients. Among patients that are cisplatin ineligible, there may be patients that we're missing here because of this revised label where patients need to have PD-L1 positive disease. There may be patients that are PD-L1 negative that may actually benefit from treatment if we have a better biomarker that could tease those patients out. For patients that have responded to platinum and are being considered for treatment with avelumab, again, the question would be, can we better identify who would benefit from avelumab versus those patients that may be equally well off from having a treatment holiday because none of these immune checkpoint inhibitors are benign either. For platinum refractory advanced bladder cancer, the question becomes whether we can predict who will benefit or whether we need to lose the indication altogether for patients in this patient population in terms of receiving immune checkpoint inhibitors. And that's an important thought because immune checkpoint inhibitors in bladder cancer are still the things that are most likely to respond in a durable response for patients. So it's important not to miss these patients that could benefit. Next slide. So I'll turn to discussing the DetermaIO assay and just recap some of the things Doug already discussed. So this is a gene expression signature run as an algorithm on whole transcriptome RNA-seq data. It's also translated to the RT-PCR assay. It measures 3 distinct components of the tumor immune microenvironment, includes immune infiltrates, fibroblast, the extracellular matrix, epithelial mesenchymal transition. So it really gets a snapshot of the physiology of the tumor in that micro environment. Next slide. So the DetermaIO assay was applied to the IMvigor210 study. Now I'll just remind you that the IMvigor210 study was a Phase II study of atezolizumab in patients that either were Cisplatin ineligible or were platinum refractory. So in this patient population across all of the cohorts, 41% of the patients sampled here were positive by the DetermaIO assay. And DetermaIO positive patients had a longer overall survival. You can see it was 15.4 for those that were positive versus 7.9 for those who are negative. The complete response or the objective response rate was also higher with the DetermaIO assay. The objective response rate was 32%. And the prespecified endpoint in the IMvigor210 trial was to achieve an objective response rate greater than 10%, which the Determa IO assay does quite well on. Next slide. So the DetermaIO assay was also compared to other clinical factors. And so when you look across a bivariate analysis looking at performance status, gender, site of metastases, race and tobacco use, these were all independent factors from the DetermaIO assay. Next slide. Similarly, we -- similarly, bivariate analysis was conducted to look at different biomarkers that had been studied in the IMvigor210 210 data in the past. So comparing to CD8 T-cell effector signatures; TCGA subtypes, so basal versus luminal subtypes; looking at PD-L1 status; and tumor mutational burden, again, the DetermaIO assay was independent of those factors. Next slide. So now turning specifically to the second cohort in this IMvigor210 study, these were the patients that progressed on platinum treatment. So either they were platinum pretreated. And when we look at patients that were platinum pretreated, again, about 40% of them were positive by the DetermaIO assay. And the DetermaIO-positive patients had a median overall survival of 12.9 months compared to 7.2 months. And I'll just point out provocatively that the IMvigor211 study, which was the subsequent Phase III study that compared atezolizumab to chemotherapy in this particular patient population showed an overall survival of 11.1 months for atezolizumab versus 10.6 months for chemotherapy, and that was not statistically significant, and that was what led to the revision in the approval for atezolizumab. Next slide. So just to conclude, the DetermaIO assay in this study met its prespecified endpoint of identifying patients that would respond to atezolizumab and showed an overall survival extension. Bivariate analysis also demonstrated that there was independence from other biomarkers and gene signatures as well as clinical factors. And this represents the third tissue type where DetermaIO has shown a significant association with response to immune checkpoint inhibitors. It's been studied in triple-negative breast cancer. And as David pointed out, in lung cancer. And these studies included agents that were PD-L1 and PD-1 antibodies. So taken together, these data began to establish DetermaIO as a pan-cancer predictor of response to immune checkpoint inhibitor therapy. Thank you.

Well, so thank you very much, Dr. Parikh and Dr. Gandara, for a really beautiful overview of the role starting with Dr. Gandara, PD-L1 in tumor mutation burden as biomarkers and yet also some of the challenges with them and focusing on lung cancer and bladder cancer. As my role is responsible for clinical development of the DetermaIO product, and we are very cognizant that we're bringing this to market in the context of existing biomarkers, PD-L1 IHC and tumor mutation burden, and acknowledge and want to develop them bringing the best information to patients. So tumor mutation burden clearly carries a biologic signal that's important to the immune system, and measuring it is clearly important for some tumors, but not all tumors. And clearly, there's different cut points for different tumors. Same with PD-L1 IHC. It carries some information, but as Dr. Gandara pointed out, clearly does not carry all the information about the immune system. And so this slide is an overview of the clinical development that we're doing, progressing through non-small cell lung cancer, where Dr. Gandara showed you the study that we've done in collaboration with Dr. Vidal at the West Clinic and showing that Determa outperformed PD-L1 IHC and perhaps added information to tumor mutation burden. A key question, as Dr. Gandara pointed out, is, can you identify patients PD-L1 negative or positive PD-L1 or tumor mutation burden negative or positive depending upon the cut point that you use? Who are responsive, independent of those biomarkers? And can you add information to those biomarkers. And so we've been very fortunate with the British Columbia Cancer Agency to put together a cohort that has over 200 patients, all treated with monotherapy. So this will give us a very clear signal separating from adjuvant chemotherapy. We hope to complete that study in -- towards the end of Q2 this year. We've also published in our methods paper, a small cohort out of Korea that really showed a very dramatic relationship between DetermaIO and outcomes in lung cancer. Similarly, in neoadjuvant treatment of triple-negative breast cancer, which we did not get a chance to talk about much today. With Pusztai and Dr. Ueno at MD Anderson and Yale, respectively. We've completed a 55-patient Phase I/II trial, showing that Determa outperformed, once again, PD-L1 IHC. And it was this study that got us -- asked us to a randomized clinical study that is looking at neoadjuvant treatment of triple-negative breast cancer around 250 patients. This is with our collaborators over in Italy at the Fondazione Michelangelo in Milan. And so we expect those results to be presented by our collaborators over there probably at ESMO in the fall. And Dr. Parikh gave a very nice overview of the work that was presented at AACR recently, moving into bladder cancer and really beginning the story of DetermaIO emerging as a pan-cancer biomarker. And the really, quite frankly, remarkable thing is that you would expect -- that you would expect DetermaIO -- you would expect the measurement of the immune system to be tumor-type agnostic. There's no reason the leukocytes in the bladder cancer would be different from leukocytes in breast cancer versus a lung cancer. And in fact, that's what we found. So remarkably, as we move from triple-negative breast cancer to lung cancer to bladder cancer and to renal cancer, an abstract has been submitted to ASCO, we've not changed the algorithm nor remarkably have we changed the threshold, defining those who are likely responders versus those that are likely nonresponders. And we've shown an approach to looking at the classifier function, independent of predicting checkpoint inhibitor response, and shown that really this classifier function is very robust and apparently, tumor-agnostic, at least in the solid tumor space. So as Dr. Parikh reviewed, it's been shown to be a successful predictor of response in all indications tested to date. It's -- we've -- all those studies have included agents that are targeting both PD-L1 and PD-1, 4 different drugs to date. And it regularly, every time we've looked now in bivariate analysis or multivariate analysis, has outperformed existing biomarkers, including PD-L1 IHC in tumor mutation burden as well as in the study that Dr. Parikh described exploratory analysis of many different gene expression patterns. And so obviously, we're excited to advance this, focusing on the specific clinical indications in lung cancer and triple-negative breast cancer but building the evidence base for this being a pan-cancer test. So with that, we'll move to the Q&A, and I'll send it back to Tara to manage this. Thank you. Thank you very much for your attention.

[Operator Instructions]

Great. Tara, would you like me to start now?

Yes. You can go ahead.

Great. Okay. I think this one's for Dr. Gandara. This was from Bruce Jackson at The Benchmark Company. His question is, one, if PD-L1 testing goes away, what do you believe are the most likely replacements for PD-L1?

Well, that's, of course, a great question, and the answer is unknown. My own impression is that PD-L1 may not go away, but it may be further refined and that it will be part of a composite algorithm. The -- just recently, as a matter of fact, 2 months ago, the Swanton Group from the U.K. presented on their concept of a composite biomarker. And interestingly, in their data, PD-L1 persisted along with TMB and some other markers there. So there are a lot of people working on this. But I can tell you and my opinion, the DetermaIO data have, despite the fact that it's early, it has certain advantages, it's agnostic to chemotherapy, that's incredibly important. And the data so far suggest that it is across various tumor types, even though it's a positive negative result. In other words, not requiring some binary analysis. So in short, I'm not sure if PD-L1 will go away. But these data, at least in the lung cancer data set suggest that DetermaIO plus TMB could be a good combination. I'll also add that the TMB that we're doing today in the next-generation of sequencing assays is undergoing refinement as well. So again, a year from now or 2 years from now, it may also be a different beast.

Great. Thank you very much. The other question comes in from Steve Mah at Piper Sandler for Dr. Gandara, I was wondering if you could comment on gene expression analysis methodologies such as DetermaIO and TMB versus some of the new approaches in development that are using -- which are using spatial biological approaches with multiple protein marker expressions to stratify immune checkpoint inhibitor response?

Well, I'm not sure if I know the data for the assays that you're talking about. There, of course, is a lot that goes into developing a biomarker in terms of analytics and then clinical validation. So at this point, there are over 1,000 publications on potential biomarkers for immunotherapy, over 1,000. So we have more trials than we do publications interestingly. But I'm not familiar with those data. I don't think I can comment on them specifically. Even in relationship to the sort of testing that we know about, there are a lot of complexities. And even a test as simple as PD-L1, our organization in lung cancer, the IASLC, did a blinded analysis -- analytic analysis of the 4 most common PD-L1 assays, which is quite interesting. 3 lined up very clearly. One was an outlier. So even for PD-L1, there is diversity in terms of how we apply these in the clinic.

Got it. Okay. The other question has come in from Joe Conway over at Needham. And can you elaborate further on Dr. Ross' content -- this could be for either doctor, can you elaborate further on Dr. Ross' comment about not needing to change the algorithm or the threshold across cancers. In other words, what would be the difference in process between a bladder and lung cancer tissue sample?

This question is for both of us. I'll let Dr. Parikh go first.

I'm not sure exactly what the question is. The samples for bladder cancer come usually from surgical specimens or needle biopsies when patients have advanced bladder cancer. And my understanding is that's very similar for lung cancer as well. So I don't think there's too much variability there. When you get into things like small cell lung cancer, I think it becomes more challenging because getting tissue is more difficult. So maybe Dr. Gandara could speak to some of that.

Well, as I understand the question, it may be about what's different about this test from other tests. And to me, and I think Doug has educated me here about the fact that this test is like -- I wouldn't say a Jekyll and Hyde, let's say, it's 2 sides of the coin. If this biomarker is based on an assessment of the genes that reflect the hot portion of the tumor or one tumor versus the other, but also the cold. So again, how this is weighted, how this is done, of course, is proprietary. But I think that's the difference. In other words, it's a more complete assessment of the immunophenotype from what I understand. And you may want to add to that.

Yes. No, I appreciate that. And at the risk that I may have confused folks when I said that, just let me explain what I meant by that. And that is I come at this, I'm a pathologist by training. And so I've spent my career thinking about classification of cancers and classification of the immune response. And so there's really 2 pieces for a diagnostic assay to be a robust assay. It has to be a really good classifier. It has to divide things into 2, 3 groups really, really well and really reproducibly. And then once you've identified a group that you think has a clinical impact when it's there, you have to generate the clinical data that supports the clinical activity. And so the point that I was making in talking about this test, we're not changing the algorithm and we're not changing the threshold as we move between different tissue types, really speaks to the power of the test as a classifier. It's a really good classifier. It started off as a 2,000 gene expression pattern, which was first distilled down to 1,000 genes. And now we've distilled it down to the 27 genes that we need, running either as a -- on RNA-Seq data from whole-transcriptome RNA-Seq data or on real-time PCR. That specification function doesn't need to change when you go from tissue type to tissue type. That's not true of tumor mutation burden. It appears that different thresholds for tumor mutation burden are important for different tumor types. And as Dr. Gandara elegantly showed, there's many, many different ways of scoring PD-L1 IHC that's not yet been standardized. And so it's quite remarkable that this algorithm does not seem to be modified as you go from tumor type to tumor type. And so that's what I was trying to find out.

Great. Okay. Thanks very much. The other question we had was, as treating oncologists, what role do you see DetermaIO playing in enabling treatment selection, especially relative to other biomarkers like PD-L1. That's for both of the KOLs.

I'm sorry, Bob. I was looking in the chat to see what the next question was. So I totally missed this one.

It's okay. I'm sorry. So the question here was, what role do you see DetermaIO playing in enabling treatment selection, especially relative to other biomarkers like PD-L1?

Yes. So again, I think I went through some of this in terms of bladder cancer at various stages. I think it is a little -- it's -- so far, the data suggests that it is a more powerful biomarker than PD-L1 alone or tumor mutational burden alone. And so I think it gives us a little bit more confidence in terms of using it as a marker. And I've gone through the various stages of bladder cancer, where I think it could be of value. But I think it really helps, especially in non -- in patients that don't have any measurable disease, these kinds of biomarkers can be very powerful to help guide discussion with patients in terms of whether they want to put themselves through a systemic therapy that otherwise the alternative would be to just undergo surveillance.

Yes, I'll just add to that a perspective on the United States versus the rest of the world. The United States is unique in having a very large proportion of community oncologists who are out in practice, large towns, small towns, but they, right now, can decide entirely how they manage their patients. They have -- don't have the same restrictions for either ability to test or reimbursement for drugs that their colleagues in the rest of the world have. And what will change their mind is having a test that you can apply to every patient who comes in the door. In other words, a community oncologist, if they have 50 patients sitting in their waiting room and they can order one test for immunotherapy that is easily accessible, has a good turnaround time, which maybe Doug can comment on, it's good for this test. Then they will embrace it. Now there has to be some landmark associated with it, either a landmark publication or FDA approval or CMS approval. In other words, something that gives it the gold seal to use. But I think that's what's different. And I go back to the hot and the cold, being able to have a test like this, I mentioned about the PD-L1, even PD-L1 0, not one cell staining, overall survival is still improved in a small proportion of patients. So you want something better than that. You want to find out who is best going to respond? And then who you should move on to another sort of therapy, which avoids the toxicities associated with immunotherapy.

Yes. Just a quick comment. Thank you, Dr. Gandara. Yes, this -- we are launching this later in the year as on the CLIA-LDT path, and we will be taking it to CMS for reimbursement. But because of the success at distilling this down to 27 gene real-time PCR assay that can run on platforms that are distributed worldwide, we can bring it to the market all over the world so patients all over the world would have access to it. And quite frankly, pharma would have access to the information to drive adoption of their drugs worldwide. So I think that's a very key and important part of our strategy here, to get this to patients and really inform better use of these drugs that are both very beneficial in a subset of patients but also very toxic in a subset.

Understood. Okay. Great. The other question comes from Mark Massaro of BTIG and this is for Dr. Gandara. Basically, in your opinion, how much additional data needs to be generated for DetermaIO to be more readily used in the clinic? I don't know if you could talk about whether they need to be retrospective or prospective data, sample sizes and basically whether other cancers can be used to predict IO response across solid tumor types?

Okay. Well, thank you. So that's the question of the day, of course. I will say, I've mentioned already landmark publication or approval by FDA or CMS. But the other thing I'll mention, because I know our audience is forward thinking, is that our criteria for drug development and biomarker development is not the same today as it was 10 years ago. A few years ago, there was quite a distinction between a drug or a biomarker being experimental versus being used in the treatment of patients, in other words, clinical application. These days, that distinction is blurred, and the time line is blurred. In other words, what was research yesterday is in clinical practice today. And if anything, that is more prominent in biomarker development than it is in drug development, where typically a drug has to be FDA approved in some indication, even if an oncologist wants to use it off-label. But for a biomarker, if it is CMS approved, if it's incorporated into a commercial test. There will be a group of oncologists, and I would say this is most prominent in the United States, who are what I would call early adopters, and they will embrace it. They will get experience with it. And as I mentioned earlier, if it can be applied across all cancer types, that is a huge advantage. So again, for a drug, you're going to say, I want a Phase III randomized controlled trial, typically. But even that's gone away. We now have a lot of targeted drugs approved based on expanded Phase I studies. So everything we thought was dogma is changing, and this is really true for biomarkers as well.

Very helpful. Okay. Great. Just while we're on this topic of DetermaIO, we have Padma Sundar from the company on. Maybe, Padma, you want to talk to the commercialization path for DetermaIO.

Yes, happy to do that. Yes, I completely -- we take Dr. Gandara's advice and completely concur with him. So we have 2 paths to commercializing the test. The first, as we said, we are intending to launch this test later this year. That's the clinical path, and we are assembling a dossier to take to CMS for coverage. So that's the clinical path. Simultaneously, we are targeting a number of pharma, including the usual suspect big pharma, certainly, but also several small pharma that are developing therapeutics in combination often with IPI to see if they would be interested in including biomarker in their clinical trials given the evidence we have in multiple tumor types. So again, 2 paths, continue to pursue pharma, big pharma as well as the smaller pharma; as well as the clinical path of trying to secure CMS coverage for the biomarker.

Great. And Doug, maybe that's just another follow-on question we've gotten from the audience here in terms of what type of the biopharma companies have you targeted and how did you address some of that? And then also, what's been the response to date, and that's something maybe you and Ronnie want to talk about?

Yes. No, I appreciate the question. So there's obviously different classes of pharmaceutical companies out there. There are folks that have their products on the market. They're being sold to patients sometimes in a biomarker-dependent fashion, but at least in the case of lung cancer, patients are getting these drugs almost irregardless of their biomarker status. And so we're talking to those folks. We're having conversations with large pharma who has their product on the market. It's clear to me, and I think clear to them that our product can inform the management of those patients, perhaps the inclusion or not inclusion of adjuvant cytotoxic chemotherapy. But those are big companies and changing in a biomarker strategy is slow. And so active conversations, but hopefully, we'll have more to report in another time. On the other hand, there are biopharma and pharma who have not established that presence in the market and a companion diagnostic or a complementary diagnostic that can show higher response rates based upon targeting the appropriate patient population, we're very attractive to them, obviously. And so we're having those conversations as well. And there are indications for which, like bladder cancer, where clearly the drugs are struggling to make it to market and where we've shown data now that, in fact, DetermaIO can identify patients who are responsive to the drugs and even those that are PD-L1 negative, which is sort of the first stop here for most companies. Even folks that are PD-L1 negative are responsive to the drugs. Last but not least, there are small biopharma who have understood that our DetermaIO is really characterizing the entire phenotypic landscape of the immune response and are very attracted to it as a potential diagnostic for agents -- secondary agents that are meant to complement checkpoint inhibitors. Those folks are -- we're having very active conversations with and we'll be telling you more about that as the year progresses. Ronnie, did you want to chime in on that?

I think you said about all we can. I mean, obviously, many of these pharma companies, as most of you know, are very particular about what we can and cannot say and whose name we can use and who we can't use. But I think it's safe to say, Doug, that these emerging biopharma companies that have therapeutics in the IL-12, IL-15, and similar, they're very interested because these are complementary and/are unique to certain solid tumors where we don't see an ICI perform very well. And so there's a lot of opportunity there, especially given that we can identify these negative signatures that are actually responders. So a lot more to come folks, and we wish we could talk more about it, but there is a lot of activity going on. So...

Bob, if I can add to this conversation. I'll just mention as a practicing oncologist, and I'd be interested in having Dr. Parikh chime in as well, that this is a high-prevalence biomarker. And those lung cancer data, it was about 50%. I think Mamta in the study -- one of the studies, she described 40% or 45%. So this is a biomarker, which has a lot of traction because of the high prevalence. If a company develops a new biomarker, and it's only positive in 5% of the target population, pharma is not very attracted to that. That restricts their market. The oncologist is not very hot on that either because it's a deselection factor. If it's 50%, that's exactly where you want it. So again, that would be something in favor of being able to adopt this test into clinical practice. And Mamta, did you have comments?

No, I agree completely. In the study that I presented, 40% of the patients were DetermaIO positive and that, that is attractive because, like you mentioned, we would like to see this be positive. These are drugs that practitioners want to use. So finding the right patients for it would be good, and it will be good if they're a good significant portion of the patient population.

I'll just close the loop on this by saying in one of the studies I described, the so-called MYSTIC study from AstraZeneca, they incorporated TMB. And what they showed is that not using the typical level of a TMB, but driving it up to very high level, like, TMB 26 mutations per megabase or 20 mutations per megabase, then the study was very, very positive. That was 8% of the total population. So in other words, that's not what you want out of a predictive biomarker. You don't want to deselect 90% of the population with your biomarker.

Great. Got it. Yes. And maybe this is a follow-on. I think during the presentation, it was mentioned that a relatively high percentage of patients, and this was from Steve Mah at Piper, perhaps as high as 75% get immune checkpoint inhibitors as a first-line therapy almost regardless of their PD-L1 expression. But given that high number of nonresponding patients, could you comment on the side effects of these drugs and quantify the unnecessary cost to the health care system because of the use of ICIs basically in first-line therapy.

Mamta, you want to start with this one?

I can. I mean I think the first-line therapy is kind of specific to lung cancer, but we are using immune checkpoint inhibitors in various solid tumors, across various stages of disease. So it is used very frequently, and it is -- these are not benign drugs. They can be perceived that way but the immune checkpoint inhibitor adverse effects can be very significant. They can require hospitalization. They can require high-dose corticosteroids, which come with trade-offs as well. And some of -- in rare cases, they're irreversible. I think the -- and the tough part is that immune checkpoint inhibitor-related adverse effects can happen at any time during the course of therapy as well. So it's something that could happen at any time. So trying to protect patients who would be unlikely to benefit from these drugs from receiving them is something that we should be considering. So yes, that's my take on that.

And I'll just add that because these drugs are activating the immune system, in some patients, this is like developing an autoimmune disease where all of a sudden your immune system, your T-cells start reacting against your normal lung or your normal liver or one of your endocrine glands. So the toxicities can be quite diverse. We've had a patient at UC Davis who developed a severe myocarditis. So again, immune reaction on one of the checkpoint inhibitors against the heart. Patient went into heart failure and could not be saved. So again, as Mamta said, in most patients, these are fairly well tolerated. But there is also financial toxicity. These are really expensive drugs, governments -- and again, I point out the U.S. as maybe an exception here compared, let's say, to Western Europe or they also want to know that the patients who get the drugs were the ones who are going to benefit because not only is there toxicity of the drugs in terms of these side effects, but also the costs are extraordinarily high. So you have to factor that into the equation. My own impression is 3 to 4 years from now, that will be much more important because it will be a requirement.

Got it. And another question has come in, maybe this one is for the management team, because you have both DetermaIO and now your ability to start to move into the monitoring market, how does that confer some sort of unique advantage to OncoCyte in the marketplace? That's for Ronnie or another member of the team.

Yes. Let me start. But Doug, you can certainly pile on. Our vision has always been to -- when we all decided to come back and do this and attempt this one more time in the area of lung cancer and now across more solid tumors, the vision of OncoCyte was always to make sure that we create a precision diagnostics to get the right patient the right drug at the right moment in the disease cycle where they had the best outcome. And then to monitor and follow that to make sure that the therapeutic efficacy was sustained. And that's the piece, I think, that we're trying to accomplish with the CNI assay. And there'll be other technologies that we bring in to the fray, if you will, in the monitoring world because we know that you can't use one type of test to answer all the critical questions in a blood-based monitoring environment. You have cost that have to be considered for payers for repeating testing. You have the sensitivity and specificity of these technologies, which, in some senses or just a long-term disease-free patient, is it recurring or not, having a surrogate that's a low-cost supplement to something that would identify the need for a larger scale genome panel from the blood to identify what has changed in that patient that's disease-free. Those are just different questions that 1 test like an MRD test really just can't answer. We love MRD, Doug and I have worked in that world for a long time. MRD is a hammer in the questions we believe we're trying to answer, not nails. And so CNI is a blood-only, doesn't require tissue. It's -- look, we've got work to do. There's only 4 papers published. But the papers that are published are consistent across the positive predictive value at cycle 3. So we're very -- we feel very I guess, bullish, if you will, on our ability to play this out and prove it out. But it is only one piece of information that a Dr. Parikh or a Dr. Gandara would use to identify, "All right, now what do I do with that patient? Okay. The drug is not working, what do I do next?" And I think someone mentioned it earlier, having spent a lot of time in Europe and build a number of businesses in Europe, we know that in single-payer systems in Europe, there are way more app used in monitoring test in the therapy than we are in the United States. And so as I think about our monitoring experience and where we want to go first, one of the reasons we are excited about the Chronix acquisition is they're in Germany. They already have deep connections into the German medical and reimbursement systems. And so our goal is obviously to prove the thesis out there with some trials that are already getting started. And then assuming the data plays out, we'll continue to look at how we transfer that to the U.S. in the second half of this year for pharma trials, and we'll see where that goes. But it's truly to complement, DetermaIO, being able to select patients that are -- have the potential to respond, but we know that these tumors are smartest -- all get out, and we don't want to the tumor to escape the immune system. If it does, we want to see it when it happens. So that's sort of the theory and the strategy, Bob.

Great. Okay. Good. For Dr. Gandara and Dr. Parikh, I really appreciate your participation. And for the OncoCyte team, I'll turn it over to Ronnie now to wrap up.

Yes. I wanted to say thank you to everyone for your time this morning. I wanted to say a special thank you to Dr. Parikh and Dr. Gandara. We, as a little company kind of breaking through in a really important world, you're only as good as the people you get to work with. And these are 2 incredible professionals who care about their patients and are willing to look at the horizon of new opportunities to practice better medicine. And we just are very blessed for both of you to be alongside us, you give us great wisdom, great insight and you've helped craft what we hope is going to be a great company doing good things for patients. So thank you, too, for joining us. And again, thanks, everyone, for the time today. We will be -- this was recorded, so it will be up on our website soon. If you have further questions, you can certainly reach out to either Dr. Ross or myself or through our -- through LifeSci, and we'll be happy to get you answers. So again, thanks, everyone, for your time, and we look forward to our next KOL event in the coming months.