Veracyte, Inc. (VCYT)

R&D Day 2020

← View all transcripts

Dec 16, 2020

morning, everyone, and thanks for joining us today for our R and D Day, which will focus on our expanding lung cancer portfolio and share the latest data for our pipeline tests into this important clinical indication. With me today from Veracyte are Bonnie Anderson, Chairman and Chief Executive Officer Julia Kennedy, our Chief Scientific Officer and Chief Medical Officer and Keith Kennedy, our Chief Financial Officer and Chief Operating Officer. We also have 3 prominent physicians joining us today. You'll hear more about them shortly. Before we begin, I'd like to remind you that various statements that we may make during this call will include forward looking statements as defined under applicable securities laws. A discussion of forward looking statements and risk factors related to the information we are sharing today can be found on Page 2 of our R and D presentation shown here. I will now turn the meeting over to Keith Kennedy, Veracyte's Chief Financial Officer and Chief Operating Officer. Thanks, Tracy, and thanks, everyone, for joining us today. We've got an exciting and full agenda planned for today. You will hear first from Bonnie describing our vision in lung cancer and why we believe we are uniquely positioned to achieve it. Lung cancer, as you will hear, is one of the areas of greatest need in medicine and we estimate has a total addressable market of approximately $40,000,000,000 After Bonnie, you will next hear Doctor. Sonali Sethi, who will discuss broadly the challenges in lung cancer care and specifically how the Cleveland Clinic is using Percepta Genomic Sequencing Classifier, our currently available test to aid in lung cancer diagnosis. Julia Kennedy of Veracyte will provide an update on our novel noninvasive nasal swab test to improve early lung cancer detection upstream and on our Percepta Genomic Atlas to inform treatment decisions downstream. Her discussion will include the latest data for each test. Doctor. Carla Lam and Doctor. Michael Bernstein will go deeper into unmet needs at specific points along the CARE pathway and discuss how our in development test can address these needs. We will then have a roundtable discussion moderated by Bonnie, which will focus on the future of lung cancer care and the role of our test followed by a Q and A session. I will now turn the meeting over to Bonnie. Good morning. Thank you, Keith, and thank you, Tracy, and thanks, everyone, for joining us today. We are so excited to share with you, along with some terrific KOLs that have agreed to join us today, what Veracyte's journey is going to look like for transforming lung cancer care for patients. And before I before we get into some of that detail, I want to give you a little bit of background on Veracyte, just bring you up to speed on who we are today. We're a global genomic diagnostics company, and we are transforming care in the clinical indications that we are addressing throughout the patient care journey. We began about 14 years ago. And interestingly, the foundation of our success and how we think about the development of clinical tests to drive the business has not changed. So I want to speak to this just briefly. Our success has been driven by our ability to hone in on and answer relevant questions in clinical care and integrate these tests into the current clinical pathway of care. So we're not asking physicians to do something different, rather use our test to answer questions when they hit a wall. Secondly, we deploy very rich and you'll see a lot more of this today, the richest scientific rigor, I believe, 1st in class in the industry. And that means not only deploying the rich science in the development of our tests, but also very robust clinical evidence so that we can inform patient care in new paradigms and have guidelines use that evidence to drive new standards of care. Our tests are designed upfront to deliver clinical utility and economic value. That was a premise under which we started Veracyte. And this backdrop has led to some of the best reimbursement successes in the industry, which as you all know, without reimbursement, it's pretty impossible to build a business. We're going to talk a lot today about our powerful science and technology platform. You're going to hear directly from our founding Chief Science and Medical Officer, Doctor. Julia Kennedy. But let me just say that we have engines that really are focused around delivering what we need to deliver high performing chest to market. And it starts with the discovery engines that begin with building extensive clinical cohorts of every piece of data you can imagine in IRB approved patient consent in clinical trials. We built extensive clinical cohorts with thousands of patients in each of our clinical indications. Secondly, it's the data engine. We use RNA whole transcriptome sequencing. And from this combination to date, we have over 20,000 patient samples in our IRB clinical cohorts and over 4,000,000,000 genomic transcripts of data. This does not include, by the way, all the data that we collect that can also be useful in the clinical stream where we don't have all the clinical data but can certainly use that clinical stream to add value as well. 2nd and in the middle is our algorithm and machine learning engine. You're going to hear a lot more about this today. This is where we take the combination of the clinical cohorts and the whole transcriptome data and determine how best to answer the clinical questions we want to answer. We have clinical genomic classifiers that inform on diagnosis. We also have tests where we are informing on the variance of that patient sample. You'll hear about both of these today in our pipeline products for lung cancer. And now we also have a clinical testing platform that is the vehicle by which we hope to take all of these tests to international patients around the world. Today, we are in 5 very substantial clinical indications. We are answering clinical questions in thyroid cancer, which was the beachhead strategy we started with, in lung cancer, in IPF and interstitial lung diseases, now in breast cancer and in lymphoma. And we believe these clinical indications create a market of over $40,000,000,000 TAM and we are on our way and a pathway to serve this market. Our strategic growth pathway is actually very clear. We are very aligned on where we are today, how we got here and what we're trying to achieve in the future. And you're going to see how the lung cancer portfolio is going to fit right into this and help build this global vision. In the early days of Veracyte, we focused on rich and robust science. We brought 3 products to market in our first three clinical indications, getting them covered by Medicare, and that fueled our initial menu expansion and growth. In 2019 2020, we focused on channel expansion, setting the company up to go global, which we didn't believe we could do without our own platform. At the end of 2019, just a short year ago, we announced a transaction to acquire the exclusive global diagnostic rights to the nCounter platform from NanoString And with that, acquired the assets for breast cancer, our Prosigna assay for early breast cancer prognosis and LymphMark, our lymphoma testing portfolio, which is undergoing FDA review today. This was a substantial strategic move. This has set us up to not only set ourselves on a pathway for global expansion through menu expansion on our own testing platform, but also has expanded our collaborations quite substantially as well. And we were super excited, which I'll talk about in a moment with the announcement of our J and J The unmet needs that we're going to talk to you about today in lung cancer are really important, really important for patient care. We still need to get patients to get screened and to have a better pathway for workup when those nodules are found. We need to make sure we have the best tools for making that diagnosis. You will hear about both of those today. And on the back end, we're not done with a diagnosis. Patients need to be determined what the best treatment pathway of care is for them. If they're early stage and can undergo a surgery for treatment, that's fantastic. But if these patients are going to need to be put on a treatment that could improve their outcomes and those are genomically driven, we want to know that information as well at the time of diagnosis. So all of this is packaged up to improve the avoidance of unnecessary invasive procedures that patients undergo today to improve the delayed diagnosis that happens often today and also improve on the insufficient and often inappropriate treatment paradigms because of the limitations around the way this all happens today. Veracyte is positioned as an industry leader. Our genomic insights to drive care at every step in the patient's journey is truly exciting. And I hope at the end of today's discussion, it's not just that you're hearing from me, but you are hearing from true experts on this being the case. I'll just remind you, today, we're going to talk about our nasal swab test aimed at aiding and helping with earlier detection of disease. You'll hear a little bit about our Percepta genomic sequencing classifier, which is on the market today and gaining traction and approving diagnosis. And you're going to hear about our Percepta Atlas aiding in the selection of the right treatments. But we're going to go beyond that. We also hope with our partnership with J and J to advance to even earlier detection of patients. Can you imagine if we can find patients through a nasal swab before they even have cancer, what the benefit of that would be. And then on the back end, future support with early stage therapeutics to help get earlier stage patients into trials to prove these drugs can actually have even more benefit than they're having today. So we are leading in a new era in lung cancer, early detection, diagnosis and treatment decisions. And I'm excited for what you're going to hear from all of our doctors on the benefits this is going to have to every patient. I want to highlight the announcement we made yesterday, and in case you missed it, on the initiation of our NOBLE trial and an expansion to our collaboration long term collaboration agreement that we have with J and J Lung Cancer Initiative, part of their world without disease. This trial is focused on enrolling over 9,000 patients in a prospective multicenter. We expect we'll be at maybe over 50 sites. And we want to study patients not only with lung nodules detected on CT, but also patients that may even be before they have nodules found on CT. This will include patients that have a history or current smoking as well as non smokers and also those that have never had other cancers so that we can build our clinical repository for future test development around those clinical factors. We will be collecting nasal swabs. We're also going to be collecting other biological samples. We want to remain on the forefront and be able to test different samples and compare and contrast which is the best to answer which clinical question. We also plan to distinguish lung cancer development and progression for 3 years until a lung cancer diagnosis is made, allowing us to identify patients that may not have cancer initially but develop cancer so that we can develop classifiers to inform that in advance. And then lastly, the biorepository that we will build from this new initiative will just expand on the thousands of patients we already have. You're going to hear from Julia about that and we're setting ourselves up to be a real lung cancer powerhouse. The potential to identify and intercept lung cancer before it develops could be a real outcome of this work. I am thrilled and honored to make a quick introduction here to the incredible doctors that are all joining us today for this presentation. And while we share our vision and discuss the unmet patient care needs, there is no better team for you to hear from than the lineup today. Doctor. Sonali Sethi from the Cleveland Clinic will start by presenting an overview of what is still a lot of unmet net needs for patients in the workup of lung cancer. She'll then talk a little bit about how our Percepta classifier works into her integrates into her practice today. Doctor. Sethi is world renowned for her expertise in interventional pulmonology, and I'm thrilled that she has agreed to join us today. 2nd will be Doctor. Carla Lam, who will follow Juliet Kennedy's presentation. Doctor. Carla Lam joins us from Beth Israel Leahy Clinic, one of the premier lung cancer screening leaders in the country actually. She'll tell you a little bit about that practice. And as you noted yesterday on our press release regarding Noble and our J&J collaboration, Doctor. Kim Christ from Leahy Clinic is going to be the external PI for that study. So we're thrilled to have Leahy Clinic on board with us. And then Doctor. Michael Bernstein from Stamford, Connecticut will share his insights on how Percepta Atlas has the potential to completely transform treatment selections for patients. And of course, in the middle of all that, our own Julia Kennedy, our Chief Science and Medical Officer is going to provide an update to you on all the data and all the intricacies from a company perspective on everything she faces building these kinds of tests and why we're so excited and confident in bringing these two tests to market in 2021. And with that, I'll turn the podium over to Doctor. Sonali Sethi. Thank you, Bonnie. My name is Sonali Sethi and I am going to talk about the unmet needs in the patient journey. So a little bit about myself. I've been in practice for about 14 years. I actually started my career in New York where I was a pulmonary critical care physician for 5 years before I decided my true passion was actually interventional pulmonary and I had to reapply back into a fellowship, which I was lucky enough to get into the Cleveland Clinic. I came here and, I did my 1 year of fellowship training with every intention of going back to New York and I must have done something right because they insisted that I stay and then I joined the group here where I'm an Interventional Pulmonologist and Associate Program Director for procedural training, which is for all my fellows. Here, all of us are pulmonary critical care trained. However, we are all niche specialists at the clinic, which means we get to decide what it is that we love and what is our passion. And that's all that we do working here. And so my passion is lung nodules, lung cancer and being an interventional pulmonologist. So that is what I do. Here at the clinic, we see about 16,000 lung nodules a year. These are both incidental as well as screening nodules. We do approximately 4,000 bronchoscopies per year. What you see here on the right hand side, I actually call a Wake Castle. It's my bronchoscopy suite. This is not by any means a standard suite. It looks very high-tech. There are screens everywhere. I have to say I love it. This is where I work every day. My family is there. We do about 1200 EBIS cases, that's lung cancer staging cases a year, and 500 navigational cases, that's using navigation technology to try to get out to lung nodules. On a national level, I'm on the Board of Directors for the American Association of Bronchoscopy and Interventional Pulmonary. I'm on the Board of Directors for Women in Interventional Pulmonary. I'm the Chair for the Membership Committee at CHEST. I have the honor of being a distinguished CHEST educator, as well as being on the editorial board of many journals. So I'll start off this morning by speaking broadly about the challenges that physicians and patients face each step of the road in trying to get a diagnosis of lung cancer. So let's start with lung cancer screening. It's important for catching lung cancers really early. We know that if we catch lung cancers early, patients have a much better chance of surviving. But many patients are not getting screened for many various reasons that could be, you know, a whole talk by itself. After that, the next question is, once the lung nodule is found, which patients do we work up? As physicians, we lack objective tools in order to determine which nodules are benign or malignant. And this leads to patients that are actually undergoing procedures when their nodules are actually benign. And that's the majority of the case that you're going to get you're going to see throughout my talk. Once we determine that the lung nodule is worrisome and we need to make a diagnosis, how do we go about making that diagnosis? There are various ways to do this. There's a transthoracic needle biopsy, which is a needle that's coming from outside of the chest that your interventional radiologist or CT radiologists do to get into the lesion. There's surgical lung biopsies which come at risk of mortality. And then there's bronchoscopy, which is what I do. It's often the preferred method of getting a diagnosis because it's less invasive than other approaches and we'll go into that a little bit later. However, there is a high rate of non diagnostic results, that can lead to more invasive procedures and you'll see why, later on. Once a patient does get a diagnosis of cancer, so I do the bronchoscopy, they get a diagnosis of cancer, it's not good enough anymore to just say that they have cancer. We need genomic testing that needs to be done on these patients. We need to make sure we have enough tissue for that genomic testing so that they can go on and get the kind of treatment that they need. So these numbers show that what we are doing today is just not good enough. Looking at these numbers, there are 135,000 deaths in the U. S. Alone from lung cancer, 22% of all cancer deaths are because of lung cancer, 59% of patients if they are caught early or detected early have a 5 year survival rate, And 23% only 23% of lung cancers are detected at an early stage. So too many people are dying of lung cancer because we're not finding them, especially we're not finding them early and we're not treating this disease. We need to find more cancers early and we need to ensure that patients obtain the treatment that they need. So, this brings me back to the steps a patient goes through to get a diagnosis. I want to actually focus on the middle section of this pathway. And that's diagnosing a patient once it's determined that a workup is needed and how Percepta actually improves the performance of my bronchoscopy. So when a patient comes into your clinic with a lung nodule, Doctor. Lam is actually later on going to talk to you about risk stratification and working up that lung nodule at that point in time. But what I'm going to talk about is the nodule that we've decided we need to do something about. That we've decided, we need to do something about. So this patient is sitting in my office and I like to think of them in 3 different buckets. There's the very low, intermediate or low moderate and the high. The very low is less than 5% chance that that nodule is lung cancer. So if that patient is sitting in my office, there are guidelines that I have that are very clear that help me decide, what to do for that patient. And for the most part, it's CT surveillance. And that might be that they no longer need a CAT scan. There's not enough risks for that nodule to be cancer, or some kind of CT surveillance where I'm putting them on some sort of schedule and then following them up over time to see what happens with that nodule. That bucket is fairly easy for me to be able to manage in my nodule clinic. The high range bucket is also easy for me to manage in my clinic. So this is a patient who comes in with the nodule and the chance of that nodule being transferred is greater than 65%. This patient is easy. I need to get a diagnosis. I need to get them moving on their way, whether that's them going on for surgery, whether that's chemo or radiation, I know what to do with that patient. The intermediate group, and this is the majority of the lung nodules we see, 80% of the nodules that come into into clinic fall into this group is always the one that causes a conundrum. There are so many different pathways you can go down. You could go into CT surveillance. You could go into surgical resection. Or you could do a non surgical biopsy, which is where the majority of these patients will fall into. What is a non surgical biopsy? A non surgical biopsy can either be a transthoracic needle aspiration, that's that interventional radiologist putting a needle in from the outside, or it could be a standard bronchoscopy. There are pros and cons to each of these procedures. The transthoracic needle, has a pro that they use large needles and they are able to get a yield of 92%. However, it comes at a high cost of the possibility of a lung collapsing and that could be as high as 30%. Standard bronchoscopy is safer. We have less complications. However, we have a much lower yield for being able to get a diagnosis for a nodule and there are various reasons for that lower yield. It has to do with the technology, the tools we have, where the lung nodule location is, whatever it may be, our yield is a lot lower. We have a lot of tools that are built on top of this to help us out. But the bottom line is, is we need help with our bronchoscopy, to be able to determine some of these non diagnostic nodules. So let's talk about that patient who came into my office. I sat and talked to the patient. We decided a bronchoscopy was the appropriate thing to do. They fell into that intermediate risk group. I take them into the Bronx suite. They go under general anesthesia. I try my best with every tool I have to get an answer. And I have a 40% to 60% chance of not getting an answer. And let's say that that's what just happened to my patient. I wake them up. They roll across to recovery. And now I have to have this discussion with them that I put you through this procedure. I wasn't able to get an answer, now what are we going to do? Am I going to go into image surveillance now? Are we going to be more invasive and either go to surgery or go to a CT guided biopsy? It leads to a lot of uncertainty. It doesn't lead to a great discussion that I have to have with the patient, but this is where Percepta is really helpful. It's a complementary genomic test that helps improve actually my diagnostic yield during bronchoscopy. So what is Percepta? It is a test it's a genomic sequencing classifier that actually helps to rule out or rule in the chance of a lung nodule being cancer. So it's a gene expression test that identifies genomic alterations in the lining of those airways because of a field of injury. And you're going to hear about this field of injury concept throughout the day. But what it essentially is, is it's in any patient who has smoked 100 cigarettes. That's literally only 5 packs of cigarettes in their lifetime that they've ever smoked have this field of injury. You collect cells from that field of injury with a brush. This is no different than as easy as brushing your teeth. You have your scope as you can see here in place. You put a brush out. You brush back and forth about 8 times, pull the brush out, cut it in a vial, do that with a second brush. And then what you do is you send this test off and it detects the genomic changes in this field of injury to determine the likelihood that a lung nodule is actually cancer or not cancer. So what it does is it risk stratifies. And the way to understand this test is it has a negative predictive value and a positive predictive value. And I'm not a statistician, but I think the way to simplify it is, is a negative predictive value means that you are truly negative and a positive predictive value means that you are truly positive. So if there's a patient who has a low chance of that lung nodule being cancer, what this test does is if it's low, with the 99% negative predictive value, it can go to very low. So you have with 99% assurance that that lung nodule is not lung cancer. If you are intermediate, so you're going into that bronchoscopy now and you're intermediate, you're in that 10% to 60% group, and I do Percepta on you during that time, I can go from intermediate to low with a 91% negative predictive value. That's incredible to me because I can now tell you that the chance of that nodule being cancer is very low. It might intermediate and that's fine. I'm going to stay on the same path I always had planned for this patient coming out of that bronchoscopy. But it can also go to high with a 65% positive predictive value. So in those cases where I wasn't sure if I was going to do CT surveillance or not and I went from intermediate to high, I now know I'm not sitting on that lung nodule. I'm worried about this patient having cancer. I need to go on to that next step. I'm a lot more comfortable about making that decision of going to that operating room or going and having, a needle with the risk of collapsing the lung. If a patient has a high chance of having lung cancer and they go into bronchoscopy and I wasn't able to get a diagnosis and I do Percepta, it could take me to a very high chance of this nodule being cancer with a 91% positive predictive value. That is very valuable information for me because if a patient is a lung nodule and they're not a surgical candidate and they're high risk and they've got comorbidities and they're on oxygen, their only option at this point is radiation. And radiation oncologists are really hesitant to want to radiate a nodule if they don't know it's cancer because there are risks, quite frankly, associated with radiation. However, if I present to them at tumor board that the chance of this nodule being cancer is very high, they are a lot more comfortable moving forward in radiating that patient who has the potential of having cancer. Again, the sweet spot for us though really is this intermediate risk group of using the test in that group. So I just want to briefly do a case. This is a typical patient. This is a 44 year old male smoker who came into my lung nodule clinic. They were found to have this 12 millimeter nodule on the right upper lobe. I know you guys don't look at lung nodules, but this is an ugly looking lung nodule. It's worrisome to me. They were able to get a pet. There was some avidity on it. And, the pretest risk of this lung nodule being cancer actually came out at 27.7%. Based on that, we decided to go to bronchoscopy. What you see here is an image of my tool that's sitting right in the middle of the lesion. That's what that white streak is that you see. And during the bronchoscopy, I only got inflammatory cells. I was still really worried about this nodule being cancer with the smoking history. And I did a Percepta on it. And I went from an intermediate risk chance of this nodule being cancer to a low risk. So a nodule that I was worried about that I probably would have sent on now, either to surgery or an IR guided biopsy, I was very comfortable talking to the patient that's saying in combination with getting inflammatory cells and the fact that you your perceptive downgraded, I'm now comfortable going into CT surveillance, which is exactly what we did. And a year later here, you can see the lung nodule is now calcified in the center. I don't know if you can appreciate that, but that's a sign of a benign nodule. And the Percepta test saved this patient from getting an avoidable needle biopsy. I want to talk about our Percepta use at the Cleveland Clinic and how we use it. It actually starts at our lung nodule clinic itself. So as a group, we have all standardized our notes and we all risk stratify patients exactly same. And if we decide that a patient needs to go to bronchoscopy, what we do is we put in a bronchoscopy request. But in that request, there is a hard stop that asks, is this patient Percepta eligible? And you have to answer that question before going on and filling out the rest of it. And the reason we do this is we were missing cases initially when we started doing this years ago. And we found that if we put that hard stop, we were no longer missing cases because when you do the case a couple weeks later or days later, you forget. And that this way no patient would ever get lost. The Percepta is performed the same. It's performed at the beginning of the case. We have on-site cytology that sits right outside the room. They read my slides for me right away. If I don't get an answer, it's non diagnostic. I have the Percepta test run. I don't unfortunately have data or recent data because of COVID having happened, but I have looked at our data on first generation use of Percepta here at the Cleveland Clinic, and we were able to downgrade 34% of our patients from intermediate to low risk. So with that, I'd like to thank you and I'm going to look forward to answering any questions you might have later on in our Q and A session. Thank you, Doctor. Sethi for that highly engaging presentation. I'm pleased to be here to give you an update on our lung cancer portfolio and our focus on developing genomic tests that provide doctors and their patients with answers to their diagnostic dilemmas. As Bonnie laid out in her presentation, there are multiple points in the journey for patients who have a lung nodule where there can be ambiguity, and we've positioned our 3 lung products at pivotal points in this pathway where we feel they will help solve the diagnostic dilemmas. We just heard from Doctor. Sethi. She set the clinical stage for this journey and we heard her describe an overview of the types of patients she encounters in the middle part of this continuum and how Percepta GSC brings more value to the bronchoscopy procedure. So I'm here to share with you our ongoing R and D efforts to develop our nasal swab test, which is positioned at the beginning of the diagnostic pathway as well as our R and D work on Percepta Genomic Atlas, which is positioned toward the latter part of the patient journey. So our goal is to create a genomic test with a nasal swab classifier that we believe will transform physicians' ability to determine which patients with lung nodules are at high risk for cancer and therefore need a work up and which are at low risk and can avoid a work up altogether and simply be monitored with CT scans. So we're doing this from swabs taken from the nose. Why the nose? Well, there's a large body of literature demonstrating that the field of injury and smoking related damage extends from the bronchial airway all the way up to the nose. And we believe the nasal swab offers a remarkable way to get biologically close to the cancer on a non invasive sample and it demonstrates superior sensitivity to the liquid biopsies for early stage disease. Here you can see the preliminary feasibility data that we shared at CHEST last year. So we used machine learning to create a single class ifier with 2 cutoffs. The test identifies high risk patients with high specificity so that patients don't undergo unnecessary procedures for false positives and that same classifier using a different cutoff identifies low risk Remember, we Remember, we're trying to reduce ambiguity at those points in the pathway, not create more ambiguity. And high sensitivity and specificity along with NPV and PPV are the avenues to achieving that. Now this was great data to demonstrate proof of concept, but in reality that's when the real work started. This classifier from 2019 was developed on a single cohort of patients where only high quality samples were used in the training of the models. So why do we choose to do it this way? Because had we not seen this level performance in a pristine setting such as this one, we wouldn't have had any hope of developing a test that would work out in the real world. Well, here's the real world and what we need to contend with. We needed to expand the size of the training set in order to capture the spectrum of patients that embodied all these clinical and technical features. In machine learning language, we call this generalizability. There's no point developing a classifier trained in this one pristine setting that then doesn't apply anywhere else. So we have to deal with a whole spectrum of patients that have a range of history from a lighter smoking history to a heavier smoking history. And that smoking history is against a backdrop of both benign and malignant patients, patients with cancer. We needed to represent small nodules, big nodules and everything in between. We needed to represent high quality samples, low quality samples. So over the past year, we've more than doubled the size of our training cohort and built many, many models using a variety of modeling methods, feature engineering and tools in our machine learning toolbox, which I will elaborate now. So most people would agree that developing a test that classifies lung cancer risk from a non invasive nasal swab is extremely challenging from a biological and technical perspective and they'd be right. You can imagine how much biological and technical noise would be present in such a system. And we found that when we attack these really complex biological questions, there's really no better tool than a combination of RNA whole transcriptome sequencing and machine learning. They're each really powerful in their own right, but combining them is truly powerful and takes it to a whole new level. By measuring all the transcripts expressed in the tissue, we're not limiting ourselves in any way to known genes. There might be genes that are not even in the literature that we don't know about that could inform them on lung cancer biology. And if we didn't look at everything, we'd miss those. By not placing limitations on these features, the machine learning models are given free reign to select those genomic signals that will do the job that we're asking it and training it to do. But with any really powerful tools, you need to be able to know how to use them. So we found that the more complex the biology, the less a one size fits all approach works and we have to make allowances for that. We built special cassettes or indexes is what we call them to deal with these unique complex biological and technical challenges such as traces of blood from the nasal swab. Not all the samples have touches of blood, but some do and we have to deal with those. There's differences in cell type composition that occur when the airway is swabbed, different places swabbed slightly differently. We built special machine learned classifiers that separate the effects of smoking on gene expression that are common to all patients who smoke benign and cancer alike and another different classifier that distinguishes those patients whose smoking has led to molecular damage that puts them at high risk for cancer. So ironically, smoking is both the signal and a noise and we have to be smart about how we deal with each of These are all complexities we deal with as sub models in the classifier. So we call this an algorithm of algorithms and we've used it successfully with our other products. So in the real world of developing clinical diagnostics, machine learning models don't just appear overnight. They're usually a result of an evolution of progress as different types of data are added into the pipeline. Each of these columns is a separate training exercise with its own set of performance data. For example, after chest, which is all the way on the left, our first order of business was to accrue more nasal cohorts, which we did. We added 2 cohorts and developed a set of models. We also included all the poor quality RNA samples that we had left out in the proof of concept and we had to develop what are called cell type indexes to deal with that technical variability. Then we had to update the clinical database with cleaned clinical data that comes in over time. And then we created different models to explore the question, should we weigh some of the cohorts more heavily than the others during the learning process, the training process? The point here is that we build many models. I'm just showing you a few here. The point here is that each of these training exercises adds to the canon of knowledge and make the model smarter and smarter over time. Here's this last one is very interesting approach we recently took, which was to create 2 full fledged machine learning classifiers, each with their own subclassifiers, where one sits on top of the other. If you take home anything from this slide, it's performance of these models, as you see here, is consistently atorabovethe90% mark for specificity for classifying samples as high risk and consistently atorabove the 90% mark for sensitivity for classifying samples as low risk. So we're seeing consistent behavior and this gives us confidence that we're on the right track. And that's not the only thing we look at. We also look at the percentage of truly malignant patients that are classified as high risk and truly benign patients that are classified as low risk. And what we see here is that about 40% to 60% of the truly malignant patients are classified as high risk and about 40% to 50% of the truly benign patients are classified as low risk with high sensitivity. So the results of this is we now by moving patients into the high risk categories, we are now able to reduce this middle no man's land of indeterminate risk, this area which gives the pulmonologist so much problems with ambiguity. And so ideally, this nasal swab test will help pull more patients into the work upstream through screening. And you're going to hear next from Doctor. Carla Lam on how such a nasal swab test would help her assess patient risk of lung cancer at the Leahy Clinic. So from here, what's next? Well, we need to continue to build our models. We haven't chosen a classifier yet. We need to lock it, do all the analytical verification studies to ensure the test is robust. Once we're satisfied with those results, we'll unblind the independent validation test set, which we have already. It's sitting in a lockbox. And we'll ultimately unblind that when we feel that we're ready. And we will then formally transfer the test to our CLIA lab and commercialize the test and ultimately publish the data in a peer reviewed journal. We remain on track to launch this test in the second half of next year. So you already heard from Doctor. Sethi how our Percepta GSC is helping to reduce unnecessary procedures and diagnosis by improving the performance of bronchoscopy. Now I'll turn to our work to develop the Percepta Genomic Atlas, which will inform treatment decisions once patients have been diagnosed with cancer. Our Percepta Atlas will provide comprehensive genomic profiling using the same tiny biopsy sample, the original biopsy used to diagnose the cancer, either whether it's through a bronchoscopy or transthoracic needle, as for example that we heard Doctor. Sethi talk about. The biopsy sample becomes quite important in lung cancer molecular profiling, which you'll be hearing from Doctor. Bernstein in a few minutes. Now this clinical question requires a much different type of genomic tool. Here we're not doing classification per se. We're measuring genomic variance, mutations that guide treatment decisions. And so here, we turn to a more targeted gene approach. We measure changes in very specific genes that are tied to therapy and that are included in the guidelines. So once again, we're going back to the concept of feasibility. Here we show some proof of concept data. We had to first see whether we could detect mutations in reference samples using our assay that's on the left, and we showed extremely high concordance between our assay and the reference method. On the right, we wanted to see if our assay would pick up a variety of genomic alterations and tissue biopsies. And here we see a whole variety of types of changes that we're able to pick up CMVs, insertions, deletions, copy number variations, fusions, translocations. And that is all well and good. It's great that we can pick these up in reference controls and in the tissues where we would expect to detect them, but can we pick them up much earlier in the process? The answer is yes. We can pick them up. Here's a set of bronchoscopy samples taken from a number of patients with these variants. And we've taken multiple samplings from these patients. And we're very excited to report it. We were very delighted to see this data because not only are we highly concordant with the reference method over 95%, But much to our delight, we're picking up the mutations and repeated samplings from the same patient. So you'll be hearing from Doctor. Bernstein Bernstein on how these types of results can be used to help guide treatment decisions in patients. So in addition to adding to the biorepository with nasal swabs in the early stages of the patient journey with our recently announced NOBLE trial. Now with the addition of Percepta Genomic Atlas, we'll be able to enrich the biorepository collection of samples along the entire stage of diagnosis cannot be overstated. We'll have a biorepository of clinical and whole transcriptome data on early as well as late stage cancers, thereby helping to advance the future of precision medicine and facilitating earlier diagnosis so patients can be treated earlier and to save lives. So as we move towards our launch of Percepta Genomic Atlas in the second half of next year, We'll be continuing our analytical work on our assay, more kicking the tires. We'll continue to build consortium of investigators who can help us build this repository and we'll be launching the test and we're on track to do so in the second half of twenty twenty one. And we're really looking forward to keeping you updated with our progress. Thanks very much. I would really like to thank the team for inviting me to speak today on something that I feel incredibly passionate about. My name is Doctor. Carla Lam. I'm the Director of Interventional Pulmonary Medicine here at Leahy Hospital Medical Center. And I really want to spend the next few minutes really talking about the lung nodule, identify it, what do we do about it and how can we improve upon our process. A little bit of background about myself and my career over the last few decades. I've really dedicated myself to be part of one of the largest lung screening programs outside of the NLST in the country. As the Director of Interventional Pulmonary Medicine, it really goes glove in hand management and assessment of patients who have abnormal CT scans, namely pulmonary nodules. I'm the co founder of the Multidisciplinary Thoracic Oncology Center Clinic. I'm a current board and steering member of both our research and steering committee for Rescue Lung, Rescue Life Society For Lung Screening. I have co authored with our 2 major national societies, the early policy statement on implementing in a responsible way low dose computed tomography, CT scan screening in clinical practice. I've been 21 years in practice. Time flies, I guess, when you're having fun. And I would say that on average, I will see anywhere from 15 to 20 patients with any form of lung nodule or nodules either through an incidental avenue or by way of lung screening. And so, it's a very active part of my practice, which I care very much about. This is a really broad overview just to highlight what we've been doing at my program. I don't work alone. We have a really robust team who feel equally passionate. But I'm highlighting here that these colored markers from states and different countries throughout the globe, we have shared tirelessly with those to help others start responsible lung screening programs to provide them our toolkit and how we've implemented successfully a lung screening program and continue more than ever to advocate that for our patients, for all patients everywhere, in fact. So what I really want to take you on a journey with in the next few minutes with me is I'm going to focus on the very beginning. This is the door of opportunity, if you will, where if we are given a CT scan and we identify the lung nodule, it is truly just the beginning. And to do justice to the evaluation of that patient, it's not enough to not note that there's a nodule. We have to be responsible in what we do from that point on. We often will use the term the diagnostic funnel. And what that means is this gateway allows us the amazing opportunity to where by doing things right from the very beginning, getting the right patient to the right clinician with the right intervention is really the premier level of care that we would all strive to in our specialty. So let's take a journey and see how do we how does that happen in the real world. I think it's also important for you to have a background of what's the volume, what is the scope of this issue. And more than ever, we're seeing the annual rate of incidentally not screening nodules, but incidentally found nodules through CAT scans increased by 95% between 2,006 2012. That's amazing. It's like a tidal wave, if you will. And now with lung cancer screening, over 10,000,000 people today in the U. S. Alone qualify for annual low dose CT scan screening. It's mind boggling actually. What a great opportunity. And I think it's also important, while this is a large volume, it's important to know that the majority of these patients who have lung nodules will have the 9 disease. For example, 95% of these patients who go through lung screening, 95% of those nodules will not be cancer. And those who have their nodule discovered by incidental CT scan findings, 75% of those will be benign. So you can see there are 2 extremes here. It's assuring that we identify the nodule that is cancer by diagnosing and getting them through the workout the work algorithm quickly and effectively, minimally invasively. And by the same nature, the patient who does not have a lung cancer, we follow them responsibly when necessary and we reassure the patient to really be patient facing and helping the patient alleviate anxiety by discerning in the most efficient way possible what the nodule is or is not. This is also the data. So it's not just my passion for this, it's because it's supported by real facts. Lung cancer screening saves lives. It continues to be proven over and over and over again. I'm going to highlight 2 key studies, 2 key randomized trials. And what this brings to our attention is that more than ever all nodules, all lung nodules are in the spotlight. So how do we manage every single nodule is really it really is on us to do better do best. And so I think from here on, when we look at the National Lung Screening Trial, we highlight that 20% mortality benefit in patients who were screened and were found to have lung nodules that ultimately were cancer. 20% reduction in mortality in the Nelson trial, European randomized trial proved this an even higher rate of mortality benefit, 24%. I'm also going to highlight, it's really important that we do screening in the right patients for the right reasons. And we know that by the data, we're refining who gets screened based on their demographic and risk factors. And by doing so, we further have potential to reduce mortality at even a greater scale. But we have to find a better way to make screening more effective and help patients trust these evaluations by refining the science. So I'm going to walk you into my daily routine. So you see this word cloud. This is my word cloud in my own brain when I'm thinking about a patient who has a lung nodule. So let me just kind of pull apart here what this word cloud represents. So when I see a patient with a nodule, I'll first look at their scan, I'll go into the room and meet the patient, then I begin a whole series of questions to get to know the patient. So I put the nodule in context. I'm asking the patient, have they traveled? What do they do for work? What are their environmental exposures? Are they a smoker? Are they a former smoker? Do they have a family history of lung cancer? Then I'll look at the scan. What does the nodule look like? What location is it in? How many are there, what shape, size, density, all those things go into play. And then I'm starting to stratify using my own physician assessment by inherently doing that. Are they high risk, intermediate or low risk? I do that every day multiple times throughout the day. What else do I do? Well, there's also validated risk calculator models, where you can actually plug in some demographic features, radiographic features and come up with a pretest probability of malignancy. It's helpful, but it's not good enough. And you can see here in this bar graph, they do not perform reliably well. And so we're still basing decision making with limited prediction tools, not to mention even my experience, while it does help me and actually outperform some of the risk calculators, I still need something more. We need something more to enhance performance. I'm going to bring this to light with a real case, a case that I just saw recently, a patient that I'm currently working with to sort out what he has. So a little bit of background, he's 77, he's Caucasian, he's a former smoker, significant former smoker, does a lot of traveling and hunting to Arizona, That will become relevant in a moment. And how did I even get to how did he get to me and how did I get to him? He went to his primary care with a shoulder pain. He got a shoulder X-ray. By serendipity, it caught a glimpse of the lung. They saw a nodule. They asked for more imaging, which led to a CT scan. Lo and behold, this very detailed image here highlights a nodule in the right upper lobe that he would never have had a symptom from. So an asymptomatic incidental finding someone who has 2 risk factors, risk factor for infection, the travel history, a fungal infection in Arizona, very prevalent there versus a cancer, a lung cancer in a person who is of risk factors. I did a coccytopomycosis panel to see if I could sort that out better. It came back positive, so could he have 2 diagnoses? It's possible. And the workup is ongoing. It's still being worked out. But you can see that there are patients who have competing clinical issues that need to be further helped in decision making beyond what we have currently. And this really exemplifies that. So imagine here or don't imagine this exists, this is real, an accurate non invasive genomic test that could be a real game changer in adding to this decision making. A nasal swab after the scan is done to help further implement stratification to high risk, particularly or low risk and take more and more people out of that intermediate limbo land. So that when you have a high risk test that correlates with the other things I've mentioned, then that's the person you move on to diagnostic, more invasive techniques to make a diagnosis and treat and cure. On the other hand, we might stratify them to low risk, therefore, watchful observation and no invasive intervention is necessary. What a wonderful way to build trust and support of a patient in a very patient friendly way and a meaningful way that we all can trust together. It's a wonderful opportunity. I thank you so much for your time and I look forward to engaging you later with further questions. Thank you. So I want to thank you for the chance to speak this afternoon. I'm Mike Bernstein. I'm a pulmonologist in Stamford, Connecticut. And I always think when we're speaking online in one of these video forums, you don't really get a chance to meet people quite as well. So I'll give you a little bit of my background. I went to both college and med school at Duke University and I'm clearly a little bummed we canceled all of our non conference basketball games. And after training at Duke, I went up to Mount Sinai in New York City, where I trained in internal medicine and pediatrics as well as pulmonary critical care and ultimately interventional pulmonary. And like most pulmonologists since March, I've been most of the time spending taking care of COVID. I'm very excited on Thursday to get my vaccine and get back to really doing what a lot of us love to do, which is take care of and evaluate patients with lung nodules and lung cancer. So since coming to Stanford, Connecticut, I've been the Associate Director for Pulmonary and Critical Care as well as the Head of the Interventional Pulmonary Program here in the Lung Cancer and Lung nodule program. We're sort of a mix between private practice and pulmonary medicine, sorry, private practice and academic medicine. We have residents and students from Columbia, as well as seeing a lot of patients clinically. We do about 500 lung cancer screening CTs annually. And personally, I see about 250 patients with lung nodules a year. We had a pretty big bronchoscopy program and performed about 200 interventional or advanced diagnostic bronchoscopies. And I work with a great team of pulmonologists and thoracic surgeons and oncologists where we have a weekly multidisciplinary tumor board and are now affiliated with the Dana Farber Cancer Institute and have access to their tumor boards as well. I know this slide has been shown a few times this morning and sort of showing the paradigm for lung nodule and lung cancer evaluation from lung nodules to larger masses and ultimately trying to find out for patients for treatment. And what I really want to focus on is the unmet need here for lung cancer patients, which is really thinking about how we frame the question about treatment already at the time of diagnosis. So for lung doctors over the last 15 or 20 years, our principal goal had always been to just diagnose lung cancer. It didn't really matter where we got the tissue from, but ultimately over the last 10 years or so, the focus has really been on properly staging patients at the time of diagnosis. This occurred in the setting of having better access to PET scans and certainly in the bronchial ultrasound, making staging more important. But really, that's not the issue anymore. We have to think about the treatment and in particular thinking about what driver mutations or molecular markers are available for these patients to help them set up for the best treatment when they go on to the next step just after diagnosis. So in the United States right now, there are just shy of about 3 quarters of a 1000000 people who either have been diagnosed or living with lung cancer. And by the most common way of thinking about this, about half of those patients or 350,000 of them are eligible for comprehensive genomic profiling. Ultimately, this number is going to grow through time here. Over the summer at ASCO, the major oncology conference, it now recommended for patients who have all the way down to 1b disease that osimertinib or Tagrisso could be used for these patients. So finding driver mutations is really going to help what's going on. We know that patients with targeted therapy can live longer and this is part of the major guideline from the NCCN and other major international societies. So we know we have to find these markers, but how good are we as pulmonologists at doing this? And the short answer is not very effective. One large study showed that about 20% of patients actually had molecular mutations set for evaluation at the time the patients would see a medical oncologist. We also know that having the mutation set early on shortens the time to treatment and ultimately can help the patients get on the right course. I know that patients want to go directly from the bronch suite to an oncologist's office. And while it's really not critically important to wait those few hours or even a couple of days, it's really important to patients. If your family member had a cancer, you'd really want to know all the information upfront. A second even larger trial showed that only 60% of patients had EGFR and ALK, which are the 2 most common mutations done properly and less than 10% had the current recommended NCCN gene guidelines. And we know for these patients, having the diagnosis can help set them on the right road to having the right treatment and better survival. Patients who have molecular markers or targeted mutations have a survival that's almost 3 times better than patients who don't. This occurs from molecular markers and it's also increasingly occurring for PDL markers, which is more widely known as immunotherapy, won the Nobel Prize just a few years ago. And they all go hand in hand in getting the maximum amount of information for patients all the way early on. So why are there a series of molecular why are there a series of barriers to getting the right molecular markers for patients? I think we can identify 4 of them early on to figure out what's going on. The first one that really hits me the most is inadequate tissue. And every week I go to Tumor Board and I'm often faced with the question of we didn't have enough tissue. And it's kind of important to understand why this can happen. As a pulmonologist, the best way I like to describe it is we throw a big net into the water and hope to pull a fish out. We throw it up on the dock and a pathologist looks in the room and says, sure, you got a fish there and you have a cancer diagnosis. And then what we do is really throw the net back in 3 or 4 times. So until the entire case is over that we know that we have rocks and seaweeds and clams mixed in with the fish and we don't really have effective material to make the right diagnosis. 2nd issue is that the pathologists themselves don't really have standardization about how much material you need to be able to send on from molecular markers. The only way I think about it is how much fish do you need to have to be able to cook dinner properly. The second issue for us is that the number of logistical challenges go into this. As a describing FISH, it's really a similarly complex situation when doing a diagnosis that we're putting the material in different test tubes and different slides. And this creates a number of challenges as we try to paraffin embed or place informal in the material to get out for the right testing in the future. The 3rd issue is that there's a lot of rapidly changing biomarkers. Every year, new drugs are coming out. These new drugs require new molecular markers to be sent. It's really a dizzying letter storm to remember NTRK and ALK and EGFR and ROS and KRAS to put it all in place for pulmonologists to get it in the right situation. Having somebody keep track of this would certainly be advantageous to us. Certainly, there's a number of just logistical delays along with the system. What we do now because tissue is so precious, we're often sending reflex testing to try to maximize the amount of tissue we have. Often we need to get more tissue and the only way to do that is to have the patient have a repeat biopsy. And these are not optimal for the patient or for the team trying to take care of the patient the best. I think the best way to describe this is to give an illustration of a case of a patient I saw just within the last week or 2. And this shows the dilemmas we had along the way. The patient was a 78 year old retired 8th grade teacher here in our town and she had quit smoking right around the time she had retired from teaching in middle school. If quitting smoking is difficult, being a middle school teacher is probably even more difficult. But this patient came to the hospital with shortness of breath that was from an unrelated cause. But at a CAT scan, we found she had a 1.6 centimeter nodule tucked behind her heart on the left side. Ultimately, she was also found to have a series of mediastinal and hilar lymph nodes. As part of our practice, we had the patient have an outpatient PET scan, which confirmed the findings here. The bright areas in the center here are a series of lymph nodes that are PET added or highly suspicious for cancer and a patient had an endobrochial ultrasound, bronchoscopy as well as navigational biopsying. What happened in the room is the patient's the first needle goes into that lymph node in the middle of the chest. The cytopathologist in the room says, yep, you have enough information for cancer. We don't want to do a series of more biopsies to try to get more tissue. We tried to even more maximize the right amount of tissue. I navigated down to the mass at the bottom of the left lung. Again, that also showed evidence of malignancy. But lo and behold, 2 or 3 days later, we have a diagnosis of adenocarcinoma for the patient. But what we have is the dreaded pathology report of inadequate tissue for molecular markers. As pulmonologists, we're really often in the business of delivering bad news. Telling patients they have cancer is never easy, but telling patients they have cancer and I don't have enough material to meaningfully make diagnosis with molecular markers is really something that's tricky and difficult for the patient. So how would the Perceptive Genomic Atlas really fit into the spectrum and how could it solve some of the problems we have? So if we can see along the top way, the patient would undergo a biopsy and you make a cancer diagnosis. But we're still left with the issues of inadequate tissue. Well, here the Atlas sort of solves the problem. Instead of the net thrown in the water and waiting to see if we have fish, we put the material directly into the sampling and take and be able to utilize the precious little tissue we have to help the patient. 2nd issue is it's going to relieve a lot of the logistical challenges. If we send the molecular markers off right from the beginning, don't have to wait for reflex testing and other testing along the way. If tissue is gold to us as bronchoscopists, we can really maximize the value and the amount of gold we have. The third issue is there's issues with rapidly changing markers. As I said, it's an alphabet soup of new genes that lead to a whole series of medications and chemotherapeutic options for the patients. If we're working with a company that's really focused on lung cancer, that's most of their game, that's really going to help us focus in on getting the patients the right testing from the beginning. And by streamlining all of this, we hope to get and remove the delays for the patients. We're not waiting 2 or 3 weeks for all the results, but having everything within 1 week from there. So you can see the bottom here looks at the way of working with the genomic Atlas that the cancer diagnosis and the Atlas would actually move in sync together. And what this will allow us for faster personalized care for the patient. Ultimately, what we can get is a comprehensive genomic profiling with a very high degree of sensitivity, will take the tissue right at the time of diagnosis and that will help minimize the time for the patients and maximize the accuracy. And ultimately, getting the same tissue out is going to get the patients the answers they need. So I think as a pulmonologist, I'm really excited about this. It will get rid of some of the dread of not getting as much tissue and helping our patients get the right answers right from the beginning. So I look forward to answering any of the questions or things that come up during the Q and A and the roundtable that will come up next. Thanks so much. Well, thank you so much, Doctor. Bernstein. And as a big thank you to all the rest your physicians today for sharing your insights and the challenges that you face every day in trying to deliver the best care possible to patients. Your remarks really underscore the fact that patients are really at the center of everything we do at Veracyte. One cancer, in fact, is personal to me. My husband was diagnosed with stage 1 lung cancer and was able to be successfully treated with surgery fortunately before it spread beyond that early stage. Thanks to screening and for surgery for sure. My father died of lung cancer after being diagnosed with metastatic stage 4 disease, the brain. He actually died not from the disease, but from complications of the procedure performed to try to get tissue to diagnose it. So the challenges that these folks brought up today are really real. And there is not any one of us that probably has not been touched by cancer and probably many by lung cancer specifically in our lives today. So a big thank you to all of you, and I'm excited to kick off the roundtable. I also want to remind the listeners, our investors and analysts on the call today, please submit your questions to us through the Q and A box and I will move some of those in to the round table. And then we have about 45 minutes to get through both of these sessions. There should be plenty of time to get your questions answered. And I'm going to start with Doctor. Lam, who just did such a great job of talking about the upfront piece. But Pearl, I wanted you to just highlight one more time why you believe CT imaging is such a great tool for screening and filling the top of that funnel with every potential patient? Well, thank you for that opportunity. And I think the take home message for CT scanning is that, again, it needs to be done in a very responsible way. I mean, those patients who need to either the NOST or the in situ guidelines for screening other patients who should be screening and we're learning that there'll probably be a broader reach for patients who will qualify. But the thing that makes CT scan so important is that it's highly sensitive. It can identify even the smallest of nodules within a gut test to contextualize where to go with that nodule that I mentioned earlier in the presentation. Carla, do you have some insight on how many patients you believe were actually being screened today from those 10,000,000 plus that meet the eligibility criteria? Well, sadly, this is where we have a lot of work to do. Throughout the country, throughout the world, it's really significantly lower. I mean, there's 10,000,000 patients who are eligible today in the U. S. Alone. And I would say less than a tenth of those are currently being screened actively. So I think it's a mission for us as providers to educate ourselves, each other, our primary care physicians, and most importantly, educate our patients to be advocates for themselves as well. So there's many things that we're actively working on to enhance and really make it an imperative that we are screening all the right patients, you know, and getting them to early treatment and diagnosis when it's appropriate. Because it changes everything. It saves lives. It's a game changer and probably no other screening that we do for prostate, colon, other diseases, breast even breast cancer for example, lung cancer screening has the potential to save more lives than all of those combined. Powerful. Yes. Thank you. And with that, Julia, you had mentioned in your prepared remarks that you considered the types of samples that would be best to answer this question following the identification of nodules by CT. You chose nasal swab over blood. Could you elaborate for our listeners? Sure. It really is a remarkable opportunity because the same biology that is in smokers that is predisposing and then ultimately resulting in a malignancy in the that starts in the airway is propagated all the way up through the nose. And in fact, it goes the other way around. The smoking comes in this way and then the damage progresses downwards. But we're really, really fortunate to be able to get that biology in a non invasive sample. Not all indications have that opportunity. And because we're getting so close to that smoking damage, we're and it affects the gene expression, we're able to pick it up with great sensitivity. That's great. Thank you, Julia. I'm going to ask all of our physicians to comment on the likelihood of the non a good portion of those true benign patients to a good portion of those true benign patients to low risk, what would it take for this type of test to become part of your screening program? And Doctor. Saffy, let's just start with you on that since and then we'll go to Carla and Doctor. Bernstein. Sure. It wouldn't take a lot to be honest. I'm waiting for it. The reason is, is I actually live in Hisco Belt where Ohio is. And so we have a lot of fungus that's called histoplasmosis along the Mississippi Ohio River Valley, which means we get a lot of nodules that are benign, probably more than the rest of the country. And unfortunately, they all light up on PET scan. And all these patients come in with a lung cancer scare and then it's up to us to work them up. So I'm waiting for it anxiously so that I can start it in my clinic. Well, thank you for volunteering to be on the early wave 1. Doctor. Lam, how likely do you believe this is to become a key part of the screening program? I think there's a real good opportunity for this. I think the minimally invasive part is very appealing. The fact that it's easily accessible, very appealing. And I guess a silver lining to COVID is that we're now very used to having our nasal passages swabbed. And so I think that what might have been perceived as a challenge before COVID, this is a very familiar entryway for testing. So I think it actually opens the door in a way that we unexpectedly, frankly. But I think it actually will help break it knocks down a barrier that was seen as potential. But I think it's wonderful potential and I think appealing to patients to think that they could have something done in an office based setting, minimally invasive, time efficient, hopefully affordable, so that all people could have the same level of care that it wouldn't be segregated in any way. There would be no disparity in the option. Great. And Doctor. Bernstein? I think patients would love to see something like this. It's a lot of people to do bronchoscopy. So when you plan it out, it's a whole day, it's a big process. This can easily be adapted in an outpatient office pretty quickly. I think the other issue is that when we're trying to give patients information, if we're going to do another CAT scan or do a biopsy, what we're really doing is looking at epidemiologic evidence for the patient, as we talked about, their smoking history, their age, the size of the nodule, if you give some degree of personalized information to the patient, they really love that. And I think it will help the patient feel more comfortable and ultimately help the medical team feel more comfortable as well. Thank you for that. Let's talk a little bit about the barriers to adoption. How do we build the confidence as of you all as physicians in using these tests on rural patients and what will be required for us to be able to get these tests reimbursed so that we can have broad patient access. Doctor. Bernstein, why don't you begin? Well, I think there are a couple of different issues. One is it takes a lot for pulmonologists to understand all the statistics of all these. We all sort of mentioned that none of us are early epidemiologists or biostatisticians and understanding pretest probability, posttest, positive predictive value, negative predictive value. Over the last couple of years, I think the Percepta gene classifier was the first step for helping us understand these. And as we roll out new tests, it becomes a little bit easier to understand these for the physicians to adapt it into their practice. In terms of payers, it's always an issue. I think it works best when it's adapted through a hospital system and ultimately seeing where the cost changes can be. And that occurs both with the decreased utilization of CT scans and the decreased amount of non diagnostic bronchoscopies is obviously very appealing to any payer along the way. Thank you for that. And Doctor. Lam, you spoke earlier on the importance of evidence, clinical evidence. What role will that play? And how do you envision the follow on studies that will need to be collected Julia's presentation about how this was presented in 2019 at CHEST and that really building up the test so that it has staying power and increasing validity in a wider range of populations that are relevant to discern the high risk, the low risk and try to move these people out of that intermediate limbo that we all struggle with. I think that's a very hard place for a patient to live in that space of limbo. And so by being able to use a tool like this, again I think it takes several things. It takes the medical community to partner with the Veracyte teams and the researchers, putting it together in a registry like way. So we put this in the real world where we want it to live successfully. Prove again in a consistent way that it's working for the purpose intended. And that's really meaningful. That's powerful for providers to really want to utilize the test in their algorithm with the patient. And the other side of it is when we have these conversations with patients with the lung nod, with the CT scan is only the tip of the iceberg, it doesn't necessarily leave a patient definitively with certainty. But if we can couple our clinical knowledge and our acumen and the CT scan and another test that's minimally invasive that may prevent an unnecessary procedure and on the flip side push a patient to a more direct diagnostic procedure when needed, that's incredibly meaningful. And so I think it takes all those things partnered together to work meaningfully. Okay. Thank you for that. Julia, any comments from the company perspective on what we do to help with the physician confidence in adoption curve? Yes. Well, first of all, we start to engage clinicians very early on in the process because it's really important to get the right. We can build all the great models and algorithms in the world. And if we're not asking the right question clinically, it doesn't really work in the long term. It's got to be a test that is going to the answer to that test has to change what the clinician does with the patient. And so we start with that, the right question. And then, we pull out our engines and our tools that we've been building up for over a decade. And it's really important to use the right for the right questions. So you saw sometimes machine learning is the answer and sometimes we just have to measure certain things. And then finally, that's just beginning, that's we'll get to launch the test, but then we have to do all the clinical studies as Doctor. Lam mentioned, what does it look like in the real world and in the hands of everyday clinicians. And then we have to publish, publish, publish. And I actually have a question for you guys, which is we spend a lot of time working to try to publish our work in high quality peer reviewed journals. And we know that's really, really important for payers to be able to cover our test that evidence in the literature. But you're all really busy clinicians. How important is it to you all to see these papers in the published domain? Doctor. Sethi, do you want to start? Sure. Peer reviewed articles are actually the foundation of our scholarly publication system. So what it allows is an author to be scrutinized by the peers or other experts in the field. And essentially what that means is that you're going to get really high quality research. It holds everyone to a certain standard. Peer reviewed articles is what everybody looks for because it guarantees that high quality and it allows anyone able to speak up and review or bring up any concerns that they may have. Absolutely. I was going to ask Doctor. Lam about the same. Go ahead. The same question to Doctor. Lam and Doctor. Bernstein about how important are the actual peer reviewed publications to you all. You're all really busy. Do you have time to read those? Yes. I think we prioritize the areas that we work in every day because we want to be current. We want to be state of the art. We want to be able to ask answer to questions that patients have. And I'll tell you patients are quite informed. They do a lot of research and now more than ever we're using media, the Internet. Everyone's become savvy in Internet and social media use very accelerated way in the last year with COVID. And I would just say that patients will challenge and they look for their own solutions. And so as a clinician, we want to be able to meet them where they are and give them some type of endorsement based on peer review that we can trust. Validity is important. Honesty and integrity in the literature is very important because it does change clinical practice a lot of the time. I think ultimately, using a test is what helps gain a lot of clinicians' confidence in the test when you've had a patient, you've had a good experience. But you won't get to that place of bringing it into your practice without first seeing it in major peer reviewed journals. That's sort of where we find our information, where the first steps are. And it's the foundation of academic medicine and really scientific medicine that all of us hold ourselves to no matter where we're practicing. I appreciate your comments on that. And I want to shift now a little bit to the role of the pulmonologist. There's several questions coming in on our Q and A line. And one that I wanted to tee up in our roundtable. So we'll just bring it all together. And that is we tend to think of the different steps of the clinical process as segregated, one physician on the front end, a different physician like an oncologist in the back end. But what I've learned in working with all of you is that the pivotal role of the pulmonologist is greatly changing in all of this. And so as Veracyte looks to bring a portfolio that includes aiding in screening and diagnostic workup and treatment decisions, I'd love to hear you tell our listeners how the role of the pulmonologist is changing and why you regardless of where your focus area is, will have an opportunity to influence and bring the other stakeholders together to decide how to adopt and integrate these tests into your clinical practice? And Doctor. Bernstein, maybe you can start with that. I think there's a couple of different issues here from a nodule perspective. Most pulmonologists see patients across the spectrum from COPD to nodules to oral fusions and all kinds of other issues. And so we have an obligation to have an understanding. And nodules are a big part of practice. Obviously, for the 3 physicians on the line here today, this is what we'd spend a lot of our time with. But many of my partners or friends from fellowship who are not clearly what I often call noduologists, they're all seeing a lot of nodules as well. And that's really the focus. While radiologists are involved in reading these cases, thoracic surgeons ultimately in biopsying, the pulmonologist is really the person in the middle dictating what's going on. I think in the for lung cancer diagnostics and some of the unmet needs here, we at our program serve a very clear delineation that somebody is the quarterback at all time for the patients. And for us, for the majority of patients early on in their process, the quarterback is the pulmonologist or the interventional pulmonologist who's making the diagnosis. And we have a very clear mark at our tumor board when we say the point person from the patient should go from the pulmonologist to the medical oncologist or thoracic surgeon in that case. And so as a pulmonologist, we're not just really one of the side fringes for lung cancer, we're really the person driving a lot of the diagnostic choices in the beginning and driving those diagnostic choices will help set up the therapeutic choices just a week or 2 after we do a procedure. And Doctor. Sethi, how about Cleveland Clinic? Sure. We've actually looked at our data of how our lung cancer, lung nodule, lung masses, lung cancers actually come through the door. And looking at our own data, we actually were surprised that 75% of all patients are actually hitting us first. So they come to the interventional pulmonologist, that includes the ones that are sent to thoracic surgery that are, you know, then sent back to us. So we actually are the gateway for all of these patients getting into our system. So it starts with us, that's why we have a dedicated group, that only does this. We're very consistent about how we do it amongst our group. We meet often. This is where our long tumor board actually is very important. Everyone understands the steps and where all these tests fit in that we do and we're also aware of all the studies that the oncologists might be doing, that patients could be enrolled in to make sure that we're getting appropriate tissue and whatnot and what how that may affect surgery going forward. So these patients are constantly discussed and as Doctor. Bernstein said, we are the quarterback, the majority of them are coming to us first. We are then making the next decision. We're presenting the majority of the cases at Tumor Board with follow-up decisions taking place and then the team continuing to follow through the entire process. That's great. And Doctor. Lam, is it similar at Leahy? You focus very much on leaving the screening center, but how influential are you in these downstream decisions? Well, I think we actually get together as a group, multidisciplinary from the very beginning. So any in our lung screening program, any patient who is a high risk lung reservoir, which is basically a nodule of certain size that makes it more actionable for at least the pulmonary consultation and further evaluation. So we at that gateway or that funnel that we mentioned earlier, we meet there in the very beginning as soon as the nodules identified and we have a collaborative effort to really get that patient again to the right physician at the right time for the right intervention whatever that may be. And so I think that's highly valuable. It really sets a model for how we should manage incidental nodules also. So I think they really go hand in hand. If you're going to have a lung screening program, you need to have a other nodule, pulmonary nodule clinic where there's, you know, a number of pulmonary nodules, not quite as many as through lung screening, but enough so that that's where the missed cancers really could be at highest risk because they're the nodule who had a CT scan for chest pain in the ED, nodule is found, who owns the nodule? And I guess that's what I should probably bring home point. We must have ownership to the nodule. Someone from the clinical side must own it and walk with it until there's a resolution because that for us is very uncertain and uncomfortable. And the worst case scenario is to have someone come to the ER for another reason, have their scan, but the result was never given to an owner and the patient presents a few months later with an untreatable unresectable lung cancer. That to me that's really nobody wants that. And so that's where we need to find that that's a real vulnerability that can be remedied with the right program. That is really, really helpful. Thank you so much. Julie, I'm going to bounce the ball to you. We've got a couple technical questions coming in, and I want to make sure we get addressed here live. The first, could you just go back and explain the prevalence of cancer in the original CHF cohort on that data, which was pristinely selected to show best case scenario, as you've often said, if we can't answer the questionnaire, we're not going to be able to answer it in the real world. And then, a little bit about the prevalence that was in the various test cases and models that were shown today. Sure. So the original cohort that we reported in CHEST had quite a high prevalence of malignancy. These were smokers with heavier histories, larger nodules. They had nasal swabs were collected, but these patients were also slated to have a bronchoscopy procedure. So they were generically at higher risk. So that's one of the first reasons we wanted to bring in the other cohorts is to basically get cohorts that had lower prevalence of cancer as well as different types of nodules, different sizes in a variety of different smoking. So another point I'd like to make is prevalence is very important for obviously computing NPV and PPP. We as in our machine learning and R and D tend to focus on specificity and sensitivity when we're doing these measurements because if you can nail that, you can be sure that you have a test that could work across a range of preference. So it's important to have that spectrum. And so by adding the cohorts, we've been able to enrich the training examples to patients with a variety of risk profiles. Thank you for that. And then also, could you just clarify, there's still some confusion on how the NASER risk test works to answer both the roll in and the rollout test. And one of you are asked, do you run the test twice? So please clarify how NASER RISC will actually work. No. It's only run once. So the benefit of having all that great transcriptional data is that we can build the classifiers are built using the same input genomic data and clinical features. And we just the algorithms pick out what features they want to use. And we haven't settled on an algorithm yet. It hasn't been locked. We haven't chosen it yet. But the way it works is the nasal swab test will just be run. The RNA will be extracted, run-in the lab, and then whatever model is pedal on, we'll grab the features that we that are part of the locked model. And those classifiers might one work first and another classifier might work second or different cutoffs could be applied. But that's all under the hood. That's the IT part of everything, and the touch just gets run there. Thank you so much for that. I want to go next to a question relating to the lung cancer space is a really, really busy space. I mean, we as service providers know that there's many, many companies trying to address these problems. And long term, hopefully, all of that will be great for patients. That's who we ultimately serve. But we have chosen some unique approaches. We've chosen to go with a nasal swab on the front end to complement screening. And on the back end, we've chosen to use the biopsy that was actually taken for the diagnosis as opposed to say introducing another blood test to be drawn on the back end for treatment decision. So I'd love to hear from each of you on this question, How will our nasal swab and Percepta Atlas fit into what you do today versus say, have we chosen different types of samples to offer these tests on? And Carla, you're a been a user with a Percepta sequencing classifier now for some time. So let's just start with you and Envision, the moving upstream with the nasal swab and downstream with answering those Atlas questions on that biopsy that was sample, even if it is a blood sample, which we hear so much about today? Well, I think it's a great vision. I think you all had such a forward thinking vision from the very beginning, as you've alluded to. And I think for us, I think that it really is a nice spectrum where you take the patient from the original nodule and what if you can get them to a decision tree that's more definitive where they don't even need another biopsy? That's meaningful. If you can allay a person's risk and prevent an unnecessary procedure. And on the flip side, if they need to move on to a procedure, you're not delaying that. So I definitely think that for all the reasons we've mentioned, there's significant game changing capability here with that. I think with the Percepta, I do find it useful. Even though we have all this navigation technology to get to the small nodules, it's not 100%. And even with highly high volume centers, with lots of expertise, with the best technology, we still are challenged by that to some degree where we're going with bronchoscopy. And so to have something that we can add to what we're doing during the one procedure, we can actually tell a patient that even if the biopsy does not render something definitive, we'll get enough of a signal that's meaningful and even more so that's meaningful to treatment. That to me is the Holy Grail in some ways. Okay. And, Doctor. Sethi? I agree with what Doctor. Lam said a lot. I mean, it really attests to the dedication and commitment that you guys have really had in this field with lung cancer. I mean, starting from the beginning when the patient shows up in the office all the way to the end of getting them to treatment. It's huge. For me, the nasal swab would be fairly easy to do in the office. We logistically have had trouble getting the blood tested in our office for various reasons working at the clinic and having centralized labs and who's drawing the labs and sending it off and the paperwork and whatnot. So, incorporating the blood work into our clinics has not been as easy as it sounds. And the nasal swab would be a lot easier because I could do it myself, you know, just sitting there in the office going to. Now I've decided to do a bronchoscopy, you know, quite honestly you make my bronchoscopy better, like I said in my talk. And it's very frustrating, like Doctor. Lam stated, we go through these procedures and for whatever reason we come out on the other end not being able to get an answer no matter how fancy the tools are that are handed to us including robots now and you know there's failures for many reasons that we can sit and speculate about and we talk about amongst ourselves and, you know, a lot. And we're in that situation quite honest. What are we doing? It's really helpful having something that helps us decide what to do next. And then, you know, it's the final step of someone had a bronchoscopy somewhere else and then it's being, you know, sent back to me to get more additional tissue. I can't tell you the number of referrals I get for we don't have enough tissue to do second procedures on patients. And that is something we absolutely don't want to be doing. I don't want to be having to put a patient through another procedure again because we weren't able to get enough tissue and that happens as well. So I think the commitment from beginning to end of putting the patient through this pathway is just phenomenal with the tools that we're now able to use. Thank you so much. And a reminder before I come over to you, Doctor. Bernstein, on that same question, I just want to remind and clarify for our listeners that the Percepta Atlas sample is actually going to be a tiny piece of the biopsy that is taken, whether it be by bronchoscopy, transthoracic needle biopsy or even surgery. It's a very small piece of that biopsy because our assay has been optimized to work on very tiny samples, which is what allows us to do this. And so taking a very tiny piece and preserving it for the Atlas. So if that biopsy gets diagnosed as cancer, we already have the sample. So Doctor. Bernstein, with that backdrop, I'd like you to go back and answer that same question on how this fits in and as opposed to a different sample being taken even if it were a cold blood sample? Right. I think there are 2 different ones. One is everybody thinks about liquid biopsies and blood based tests. But certainly in terms of the nasal classifier, I think everybody's perception of nasal testing has probably changed in the last 8 months. Asking somebody to stick a small toothbrush up their nose, probably everybody on the call or probably half of you on the call have this done. And this is actually less invasive than a COVID test. So I think last January, trying to explain this to patients would have been a little bit tricky. We're now in a new world and that's easier. In terms of the Atlas and how it works, the phrase liquid biopsy is always like a really it seems like a really great idea. But really at the end of the day, you still need a biopsy to tell patients they have cancer. You can't do it based on a blood test and you can't even do it on the Atlas. The Atlas works hand in hand with a tissue diagnosis. And so by tying it in with the tissue diagnosis in the Bronx suite at the time, we're not changing anything that we obligatory have to do. We have to do a biopsy to show that they have cancer. We have to have a cell slides and things that prove that, but then we can immediately go on put material for the genetic testing right then and there. So they're paired together. It doesn't really waste much time, and it fits in the flow pretty easily. That's fantastic. And the comprehensive nature of the data, we believe will be more comprehensive than really anything else that's out there. Julia, let's shift gears a little bit and talk about the studies that are needed for reimbursement, both on the nasal swab, but also if you could talk a little bit about the types of studies that Percepta Alice will need to undertake, because it's a little bit different kind of tool than those that are going to change clinical care decisions on a diagnostic type call. So if you can speak to both of those. Yes. I think the post launch studies will be very, very different in nature. Both are utility studies. So once we do the clinical validation on the nasal swab, we have to unblind our test set. That test set already has the truth of label associated with it. It's locked away in a clinical database that we can't see. And so we have to show that the answer that our nasal swab classifier gives a global risk or high risk, how concordant that is with the truth. That's clinical validation. Once we launch that product, the studies will be clinical utility studies that will be designed to show how that answer impacts the patient journey and does it improve outcomes. On the Percepta Atlas side, it's an analytical question. And so the validation is can we pick up mutations, variants, whatever that have already been clinically validated out there to be useful. Then the name of the game then becomes putting that test, then the name of the game then becomes putting that test that's analytically valid into the hands of clinicians who can put it into their practice and we can allow multi center studies to bring data for and provide a question as to how useful this is for picking up these variants in their particular practice. And all of this, we work with the key opinion leaders to publish. And so all of that data has to be published in a peer review journal. Thank you. That's very, very helpful. We've heard a lot over the last decade, the innovation has been incredible on our ability to collect large scale genomic profiling. But most of the biorepositories that have been built to date have used patients that were typically late stage cancer, which is where the treatments were being targeted to benefit those patients. And Doctor. Bernstein, I know you spoke to a particular case in your talk, but I want to come back and have you elaborate. If we are able to truly do as was proposed by Julie and my talking points on building the first whole genomic profile from whole transcriptome data, which I will remind everyone doesn't just include those variants that are known that we're looking for, but the variants that may be unknown that we will find over time because we run that test on every patient with a diagnosis of lung cancer. If we're able to build this biorepository and have included early stage 1 and stage 2 and all types of lung cancers, Is there an appetite, do you believe, in the world of lung cancer treatment and oncology to benefit from this data being run on those earlier stage patients? Yes. I think we're ultimately seeing this importance coming down in stage, from stage 4 all the way down all the way to stage 1. Look, the patients who have the highest risk of developing lung cancer are patients who already had a lung cancer. And with patients being ideally found at screening programs early on and if they're diagnosed young enough and with the population getting old enough, we see a lot of people with recurrent second stage 1 lung cancers or another lung cancer. If we know about their first lung cancer, that helps us even more on how to properly follow these patients up, how many scans do they get and whether some sort of adjuvant or post surgery chemotherapy might reduce their risk of developing a different lung cancer along the way. So there's a lot of avenues there that having all the information, I think, would help the oncology community, the pulmonary community, really the, we say the thoracic oncology community together. We are really excited about being able to build this earlier stage lung cancer biorepository. And we're really looking forward to engaging some early access sites that have equal enthusiasm about building that. Doctor. Bernstein, I'm sure you'll be part of that and we'll look forward to expanding that group out. I want to kind of go to a question that perhaps you all never think a lot about. I'm sure investors that are on this call think a lot about and we certainly do as well. And that is, what is the benefit from a physician perspective on a single company taking on being a provider of answers across many different questions in the clinical care continuum. Often in our space of genomic testing, you have companies that focus on one part of the clinical care continuum. And we believe we're taking on a big commitment here by being a company that will try to serve multiple needs and multiple stakeholders as we spoke about. So, Doctor. Duffy, let's start with you. Maybe you can describe and explain how the benefits of the drawbacks of having a single company to partner with on this journey? It goes to show really the dedication and commitment that the company has to lung cancer. I commend you guys for taking on one project that you think really needed to be worked on, which was lung cancer and then trying to address all aspects across the board with regard to that one disease entity. I think it makes it a lot easier for us to work with you guys because there are different parts in the patient's journey through this process that what you have we can utilize and it's very easy to know who to go to, we know how it works, We've got access, we've got standardized things between all your different tools that you're providing us and using methods are very similar to be able to get it. So I think the education piece for everyone around us, meaning the nursing education that's needed and everything for the collection and the mapping out and all of that, makes it a lot easier for the entire team. Thank you. Doctor. Lam? I think that I would echo what Doctor. Sethi and Doctor. Bernstein have said, I think, in prior statements. But I also think that when we think about nodules, not all nodules are equal and not every patient is the same. There's so much individuality that we have to take into account and individuality of cancer. We now know that there are many different phenotypes of cancer. Some behave more aggressively, some less aggressively, and we only have to imagine their signals must be their fingerprints are different. And how we manage them and how we address them over time is different. And I think to have a comprehensive understanding of that, you know, as I heard Julia discuss all the different signals and things are being able to be identified with this latest development of testing is very exciting to me as a clinician because I think it just reinforces that we have more work to do. We can always refine what we're doing. And I always mention this and you all know this from all of our meetings together is that the round glass opacity, that less dense nodule that has meaning and can often be cancer behaves differently, can we evolve to assessing that with a signal that's meaningful and can be utilized in a non invasive way also? So I think that I am encouraged by the comprehensive vision and the drive to ask the questions and really engage clinicians early on, I really would give you all the credit for that because you continue to go back and ask us what's needed. Are we getting this right? And I feel that you are getting this right because we're collaborating early in the game of these efforts. So thank you for that. Thank you. And Doctor. Bernstein? I think Doctor. Lam mentioned that not all nodules are created equally. I think it's also important that not all cancer centers or diagnostic places are created equally. I'm humbled that represented by Doctor. Sethi that Cleveland Clinic is probably the premier place for diagnostic bronchoscopy and the Lahey Clinic is really at the forefront of everything for lung cancer screening the rest of us do. And while I'm at a pretty big community based hospital, That's not even where most lung cancers in America are diagnosed. They're mostly diagnosed at smaller hospitals all across the country. And I'm honored and blessed to have a great team here who are trained at really big academic medical centers in the Northeast. But that's not the case everywhere else. And having one standardized approach can really bring the lab technologies and the sort of academic rigor of a place like the Leahy Clinic, place like Cleveland Clinic, clearly, even Stanford Hospital out to all kinds of communities. And that's what Veracyte can set up. It creates a standard quality control that will help down to the level of the patient getting the best evaluation and diagnosis. Thank you for that. And we look forward to partnering with you to make that happen. Julia, we're coming down to the last couple of minutes here. Are there any final comments from you, from a company development perspective or any other perspective, I guess? We're really excited about building this biorepository. One of the things I like to tell the team, and we all feel really excited because we're a data driven company is that we'd like to believe that the future targets for treatment are swimming around in our database right now. These are all the way from early stage all the way to the late stage cancers. And we have all the transcripts in the genome that are expressed in those tissues is the data that we have plus the clinical data and then we collect the outcomes. It's such a rich biorepository. I'm really excited because personalized medicine means coming up with more targeted therapies. We've gotten great success with the therapies that we've seen over the last couple of years. Let's just keep doing that and let's keep enabling that so that new therapies can be developed for more and more patients. So it's a database that we're so excited about and also launching the products as well. And we really, really enjoy and are enriched by our interactions with all of you and your colleagues because it's not we don't just throw a test over the transom and say, here, hope you have fun using it. We really feel it's important to join hands with the clinicians from the very beginning to help inform these questions and kind of course proactive work going in a little bit of a different direction than we need to. So the journey you're with us on this journey, and we really, really, really appreciate it. Thank you so much for that. We're coming down to the wire. It's been a fantastic 2 hours. We want to especially thank Doctor. Lam, Doctor. Sethi, Doctor. Bernstein for an incredible commitment of time, but also from your amazing expertise and thought leadership. We look forward to working with all of you and additional colleagues as we bring these products to market and make it a reality for patients everywhere. Thank you.

• Press Release
• Slides