Veracyte, Inc. (VCYT) Earnings Call Transcript & Summary

Good afternoon. My name is Dan Brennan. I cover diagnostic tools and life science -- excuse me, diagnostics tools and pharma services for UBS. Pleased to be joined on with me -- with me on this virtual stage with the senior management team from Veracyte, Bonnie Anderson, Chairman and CEO; and Keith Kennedy, CFO and COO. So this is day 2 of the conference, and hopefully, you've been enjoying it. Feel free, by the way, if you log into the webcast, you can pose questions that I'll see on my end, and I'm happy to try to put those forth when we go through the Q&A. I think that Bonnie is going to go through a slide presentation that's on the website, but she'll direct any of the audience online to find that. And with that, I'll hand it over to Bonnie and say welcome and thanks for joining.

Thank you, Dan. It's really a pleasure to be here. And yes, our first virtual health care conference so, so far, so good. I'd like to start by reminding everyone of the forward-looking statements that we're likely to make today as we walk through and answer questions on our story and also point you all to our website under veracyte.com, the Investor Relations portion of our website, where we have posted this morning a new deck that has the UBS Virtual Healthcare Conference title on it. So that you can either follow along today now or pull it as I walk through the story. It's really exciting time for Veracyte despite the challenges that we're all facing from the COVID-19 pandemic. We are seeing some tremendous opportunities unfold in the company. And our story has actually changed quite a bit over the last 12 to 18 months, and I'm going to try to illustrate that for you today. We are a global genomic diagnostic company, and we're focused on changing clinical care up and down the clinical care continuum for patients. We were founded only 12 short years ago. And when we started the company, we focused in on entering clinical indications at the point of diagnosis today. I'm going to point you to Slide 6 -- excuse me, Slide 5 in our deck as I spend a little bit of time illustrating, though, the incredible expansion of the business opportunity that we have unfolded with some of the strategic decisions that we have been able to make in the foundation for future growth that we have achieved. Despite the 3 product introductions that we were able to make and get through Medicare coverage in thyroid cancer, lung cancer and interstitial lung disease, which we refer to as Afirma, Percepta and Envisia through 2018, in 2019, we began to unveil a much broader and deeper story with actions that we have taken. Today, we are in 5 clinical indications, 4 of those oncology indications and our fifth in interstitial lung disease, all really important indications we're making -- where delivering answers to clinical care questions is incredibly important. In our lung cancer area, we are also looking at advancing an early detection tool through our nasal swab classifier, which I'm sure we'll talk about later, a very exciting innovation. And also wanted to point out that with our execution of the transaction to acquire the diagnostic business from NanoString Technologies in December of '19, we've expanded into breast cancer and lymphoma testing. The lymphoma test is in development. This broad base of menu and broad reach into the deep science that we use, which is the backbone of our testing, has also brought forth a number of very high-profile collaborations with biopharma companies that gives us the ability to not only have new avenues for revenue and value creation but also leveraging these opportunities to accelerate pipeline within each of these clinical areas of testing. On Slide 6, we show our progressive revenue growth in the company. But I will point out, we have pulled our 2020 guidance, and so we're showing here through 2019. But I'll point out that in 2019, you begin to see the layering effect of the multiple sources of revenue get -- now going well beyond just our Afirma test, which was our first clinical indication test, which launched back in 2011. And Q1 results, illustrated on the following page, also show very strong product and testing revenue growth. That's now the category that we use to capture the revenue and testing volume for Afirma, Percepta, Envisia and now our latest portfolio product, which is Prosigna. And in addition to that, we show the less predictable and lumpier revenue that comes from the biopharma segment. This gives you the basis for how both of these line up together. I speak to the value and best-in-class design system for the nCounter FLEX Analysis System on Slide 8. But besides the incredible attributes of this distributed platform that you can read about, I'll instead focus on Slide 9, where we're able to lay out for you a really exciting vision for the company as we move from being a CLIA lab only in the U.S. focused last year on delivering results on 3 clinical indications to today being a global enterprise and having our own distributed platform from which we plan to build our menu and expand menu even beyond our own tests in order to become a global-scaled enterprise over the next 3 to 5 and, certainly, 10 years. Just last week, we announced another exciting addition to this strategy, which included the exclusive partnership with CareDx, who is a leader in transplant rejection testing. We have now partnered with CareDx so that they will exclusively have the ability to develop transplant rejection tests available on the nCounter. And this will be a win-win for both companies. For us, revenue from -- in terms of milestones, test kit sales and nCounter sales and service revenue as well, where we will control the expansion of that nCounter installed base to accommodate their menu needs. And what's really illustrated on this slide is the power of really combining a CLIA lab model, which I think is an essential way to innovate and bring advanced genomic testing to market. It's very important in the U.S. market to be able to control the adoption of the task during the process and building out the evidence to get these tests reimbursed and paid for. And as many of you probably know, that can take years to achieve. And if you enable a laboratory too early in the U.S., the laboratory bills for the test and it isn't fully covered, then they may stop adopting the test. So the CLIA lab model works very, very well in the U.S., where we can consolidate high volume of testing. We run all of our tests in our own CLIA lab today on a high-throughput whole-genome, whole-transcriptome RNA sequencing platform. And so every run that we run is at full capacity, and that allows us to do that with profitability in mind. Then when you look to the global markets where I've spent a lot of time over my years, I spent 18 years at Beckman Coulter earlier in my career, and what you learn about the global markets is that it's much more difficult to drive deep penetration and adoption of tests if samples have to be sent back to a U.S. laboratory instead of being able to be performed within the medical center that is closest to the patient and a medical center that's part of the infrastructure of health care within any particular country or geography. And so a couple of years ago, we tested a very long-term strategic vision for the company, along with our Board of Directors, and we determined that having built a very high-value menu of advanced genomic tests in very, very important and big clinical indications perfectly positions us to go on a search to acquire our own platform so that we would be in a position to expand globally using the platform as a vehicle for menu expansion and using the menu as the vehicle to drive installed base and get the stickiness of the business of pull-through on all that menu from medical centers around the world. And that's the journey we're on today, starting, of course, right now with Prosigna only as that menu item. But as you will see later, we plan to launch our Envisia classifier for IPF in 2021 as our first international menu expansion item. We'll have CareDx' portfolio coming along, along with our LymphMark test as well as our nasal swab test, which we plan to launch on the nCounter in international markets in 2022, a year following the introduction of that nasal swab test for early lung cancer detection in the U.S. market. So net-net, we are positioned to grow into a product line that will serve a $40-or-more billion market opportunity and are, I believe, the first company to expand a menu of advanced genomic tests such as these proprietary tests on a platform in which we ultimately have control. And we'll have the cost basis for the basis of our financial profile, which is very exciting because that is needed to drive scale and profitability. On Slide 11, I speak to some of the foundational strategy that we've used to achieve this level of success in the company. But Slide 12 really is the backbone of our science and technology, the multiple engines that we have built over the last 12 years in R&D that has been the basis of this multimenu and multiproduct launch success. We have a very deep discovery engine. In discovery, we build clinical cohorts that are very extensive, that collect every piece of data one can imagine from clinical data, to imaging data, to pathology data, et cetera, and follow patients over the course of time so that we can build a database that even has their outcomes in it. From that clinical cohort, we then take samples that we want to be able to use clinically and do whole RNA transcriptome sequencing. And from that data, we're able to do 2 things. First, we're able to build clinical classifiers using the truth labels from the clinical cohort and the features that are needed to get that clinical result -- highest performing clinical result, I will say, using machine learning from that data set. And from that, we've launched our Afirma, Percepta, Envisia clinical classifiers. But what's truly exciting is that this backbone of whole-transcriptome sequencing also gives us the ability to report the underlying content of the assay, perhaps at the same time our diagnostic tests are being run, and this allows us to inform on variants such as BRAF and NTRK and other -- EGFR, other important variants that we all know are critical as you're determining what pathway of treatment and what surgical decisions to make for patient care. And then lastly, and now the last leg of this really strategic framework is the addition of our own distributing platform where we can then build content from RNA transcriptome sequencing, port these tests over to the nCounter platform, which gives us up to 800 features to use in each of our tests, and that then allows us to take all of that work and build assays and menu that can be exploited in the global markets for individual patient care. It's truly an exciting time. Slide 13 just illustrates how each of these clinical indications have their own needs for actionable results for informing patient care. Many of our tests are designed to help patients that don't benefit from surgery to avoid it. We believe in this new environment of the pandemic and how more conservative decision-making is going to be desired as life goes forward that our tests are going to be very well positioned in this new era. But at the same time, as illustrated on Slide 14, we are providing comprehensive answers that can make really important decisions on care, not only diagnostic and prognostic cancers but inform on variance from this underlying data that can help inform treatment decisions as well at the same time of diagnosis and possibly without the need for ever collecting another sample. I will just quickly flip through slides 15 and 16, et cetera, because these are very self-explanatory. We lay out the number of patients for each of our tests: the clinical positioning of our tests, which is always designed to be part of the way the patient is worked up today; and then our classifier results, which report very high-performing clinical results that make next steps very clear to the physician and to the patient. For Afirma, over 70% of the patients that undergo our testing can be reclassified as benign. Prior to Afirma, these -- most of these patients would have gone on to a surgical resection of the thyroid. That's just one example of that strong value proposition. And then today, on Afirma Atlas, if the patient is not benign but likely malignant or suspicious for malignancy, we can detect these underlying important iterations that can guide surgery and treatment decisions right from the same fine needle aspiration biopsy. Percepta is -- and Afirma, just to reiterate, is today what we consider a standard of care product. We're in all guidelines. We've got virtually every payer with a coverage decision and, in fact, are in network with just about every payer across the United States. So now we just continue to educate patients, unveil our patient engagement, web-based campaign and make sure that every patient who can benefit from Afirma tests gets access. Percepta was launched a couple of years ago. This test too is broadly covered by Medicare, and we continue to grow the product. In the first quarter of the year, our lung profile -- our lung product, which consist of Percepta and Envisia, achieved about a $2 million number, which was very high growth over prior year. And Percepta's designed to help inform when a bronchoscopy is nondiagnostic. Without Percepta, many of these patients would have to go on for a costly and risky surgical biopsy or if moved to watchful waiting, could potentially miss the cancer. Our brushing is collected at the same time the bronchoscopy is done and allows us to restratify patients so that they can be safely moved to watchful follow-up or more quickly moved in for the biopsy test so that they could get on treatment. In 2021, we plan to launch one of the most exciting tests coming through our portfolio, and that is our nasal swab classifier that is going to be designed to help guide stratification of patients post nodule detection. I'm going to spend a little bit of time on this slide because this data are very exciting, and the opportunity is quite exciting as well. About 10 million patients in the U.S. are actually eligible for undergo low-dose CT screening. Since the detection of nodules today lead to a workup that could be invasive and costly and risky, many patients forego getting in for those low-dose CT screens. And so and at the same time, about 1.6 million patients a year are found to have nodules incidentally. This [ can defer the workup or ] surgical workup or from other procedures that they have done [ wherever the ] nodules show up. There's a number of guidelines that direct how large a nodule needs to be before this patient is put into the workup funnel. And what we know today is that about 750,000 patients actually are found to have nodules that meet that criteria. This group has a prevalence of cancer about 25%. So certainly, it's too high risk to not do anything and to watch the patients but not high enough risk where everybody is comfortable taking all these patients under an invasive biopsy to get a diagnosis. Today, there are calculators and clinical experience and all these sorts of nonobjective things that physicians will try to use to make the best guess. But we certainly know from data that we've collected and data that will actually be published on this that it is a guessing game and that most patients are not headed towards the right decision just based on the tools that are available today. So when we introduce our nasal swab test, we want this to be the first test that is done from a patient that is eligible with a nodule to be worked up. Our early data was unveiled at CHEST last year. And what we were able to show is that from our nasal swab data, we were able to create a 2-cutoff classifier that allows us to give very excellent results on the bookends without missing any patients or overcalling any patients that could lead to either poor management of cancer patients that are missed or over invasiveness procedures on patients that might be false positives. And so this early data showed that from the nasal swab, we could detect roughly half the true cancer patients with a very high specificity of greater than 94%. If these patients are found to have small nodules that are stage I cancer, the next procedure could potentially be surgical resection, which could be a cure. And this is a really excited advancement to be able to take these patients from a nasal swab to knowing that they likely have cancer. At the same time, though, on the other end of the spectrum, we showed the ability with about 46% -- I say here about 40%, but it was actually over 40% of the true benign cases were called out with a high degree of sensitivity. So by not missing any cancers, roughly half the benign patients could be called low risk for cancer from the nasal swab test. And in each of those bookend situations, the patients not classified as high risk or low risk end up in an intermediate risk bucket so they're not lost to follow-up. They're actually pushed forward and through the workup to get to a diagnosis. If bronchoscopy is used and the bronchoscopy is inconclusive, this is where Percepta would pick up this case. So really, a truly exciting opportunity. And then Envisia, which is next on the list, where we have developed the very first clinical diagnostic tool to help get an early and accurate diagnosis for IPF, these patients often are very frail and progress rapidly, and getting an accurate diagnosis and treatment decisions made the earlier, the better. We recently announced that we've licensed a 52-gene blood-based monitoring tool from Yale University to complement Envisia. We'll bring -- be bringing that test through full development and keep you aware of our plans to commercialize that test. But what's really exciting about it is Dr. Naftali Kaminski actually discovered and developed a classifier on the nCounter platform. So we're very excited to advance that. And then on Slide 19, the newest addition to our portfolio in December of last year, our Prosigna breast cancer test, which informs on risk of recurrence, similar to other tests that are on the market for informing the risk of recurrence. But in addition to that, outside of the U.S. -- wherein the U.S., the test is FDA cleared with the risk of recurrence score only. Outside the U.S., we are able to additionally use a 50-gene classifier to inform on 4 intrinsic breast cancer subtypes. Much data is now coming out about the potential benefit of these subtypes being able to be valuable in providing treatment decisions in the neoadjuvant setting. So a lot going on. We believe evidence is key to the march to standard of care. And each of these products is on that journey, as shown on Page 20, and a very strong pipeline that not only will expand our business in the U.S. market but be the vehicle for significant global expansion. As we expand menu, we will drive more installed base. As we drive more installed base, we will get more pull-through on the menu that's available on the platform. And to summarize on Page 22, we are addressing a very substantial global market opportunity here. You can read the numbers, well over $40 billion and well advanced in building out the portfolio to capture that market. We've added a slide that reiterates our business recovery framework, looking at what we call a U-shaped curve as we come out of the pandemic. I'm sure that we'll be answering some questions on that. So I'll save any commentary here. We did speak to that on our earnings call from Q1 that you can all listen to. But I just want to leave you with a final couple of slides. First is the really incredible, attractive financial profile we have built for the company. In the fourth quarter of 2019, we actually had achieved cash flow positivity, and that was even including the cost and expenses for transacting our diagnostic acquisition from NanoString. Strong revenue growth of 31%, very attractive gross margins, which do bounce around a bit depending on how much biopharma revenue we have in any given quarter. Selling and marketing leverage, this will become even more apparent as we move more and more menu to our distributed platform and engage institutions around the world to adopt these tests. And a solid cash position, we ended last quarter, Q1, with over $153 million in the bank, so have plenty of money to not only weather the pandemic impact but also to transact and execute on the business as we need to. And Slide 25 will be the last slide I cover, where we have updated our product and testing revenue to TBD for this year. But would love to point out here that all the rest of our catalysts remain intact. Our 3 product launches, our nasal swab test in the U.S., our Envisia nCounter launch to international markets and our Percepta Atlas to take advantage of all that variant data that we can report for lung cancer patients are all on track to launch in 2021. And with that, I'll just close with the great team that's helped us put this business together. So Dan, back to you.

Bonnie, thanks for the helpful overview with all these businesses evolved and added many, many levers since the IPO that I was involved with many years ago. So glad to see you guys are obviously tackling a lot of good opportunities here. Yes. So we've got -- yes, we still have a good amount of time here to go through some Q&A. So maybe I'll go -- I'll ask maybe 1 or 2 short-term tactical questions on COVID, and then we can focus on some of the important long-term drivers. But I think at the time of the 1Q call, you kind of characterized this U shape. I think you discussed maybe business down 50% or so thereabouts in April. Any -- just maybe the -- when you say U shaped and kind of the business normalizes in Q1, Q2 of 2021, did you give any flavor for is that down 50%? Is that what's kind of expected through like Q2, and then kind of it begins to get a little better from there? Can you -- were there any numbers attached to that? Or kind of how do we think about just the overall kind of level kind of impact as we work our way through this year?

Sure. Yes, I mean, I know everyone sort of has their own thoughts and ways of thinking about this, and we sort of laid out these 3 scenarios. One scenario, it could be a rapid recovery. Some believe that, that will happen. The third scenario is this never recovers, it goes way, way out before we get back to normal. And I think what landed us in the middle as a U-shaped curve, and we did not give any numbers. And we did specifically say it's hard to predict how wide the bottom of that U is. At some point, we will begin to see us come up the other side. But some of the factors that influenced us to kind of position the recovering in this way has as much to do with what is going on economically out there as it does physicians, offices and hospitals opening their doors to start to see patients. Because when you think about it, all the testing that goes on in the space by us and by other companies as well, has a funnel that patients go through to get to our testing. Very rarely does the patient go to the doctor in order to have an advanced genomic test done to make a diagnosis or inform on a treatment decision. It's much more nuanced than that. And each test has its own paradigm. So for example, in breast cancer, our testing as well as other breast cancer tests in the market, are performed following a patient getting diagnosed with breast cancer, which usually follows a mammogram. And so as you sit back and think about the economic situation that we're in as a country and the 30-plus million people out of work and people losing their health care and JCPenney is going bankrupt, I mean, who would ever think. And so our worry is that, that level of impact is going to mean that when women wake up following this and businesses and doctors and hospitals are starting to see patients, we doubt that the first thing a woman's going to do is go get our mammogram. The sad side option of that might be that 2 to 3 years from now, we might find more women dying of breast cancer because they put off the mammograms that could have gotten them diagnosed earlier. The same is true in lung cancer. I mean patients don't know they have nodules. The nodules are detected because the patients are undergoing CT workup, for example, presurgery. Well, with surgeries greatly slowing down and even with some surgeries starting to open up, what is happening is patients are being seen much fewer and far in between. So the surgeries are taking place at a slower pace. And that might mean it could take twice as long to get the lung nodule patients identified that are the feeders into our test and doctors' workup. So I think we just felt it was prudent to be very thoughtful and not get overly excited that this is going to have a quick bounce back and then have to come back to you all later on and tell you we were wrong. It's going to take longer. And by thinking this through, we were also able to position our spend. We did some furloughs and cut spending so that we're going to be able to preserve about $20 million to $25 million of spend through this while we work through the next quarters. And then if things happen a little bit faster, we're very well positioned to pivot on a dime and get people back on board and get back to business.

That's really helpful. Yes, there's a -- that long funnel creates further separation from exactly your own test opportunity -- your own test kind of trend because you've got to work through layers back. So that makes sense. Maybe since we have not too much time, maybe just jumping over to the nasal swab classifier for a moment, if you don't mind. I know you've been now talking about this for a bit, but it's still relatively new to me even though I've heard the pitch several times. Maybe what's been -- when you think about the launch of this going forward, maybe just walk us through the key milestones to watch between now and then. And can you again give us some feedback about what the experts are saying, kind of what the interest is from the field? So when we try to think through what the slope of this launch could be maybe through the milestones and the feedback, that will help us get a sense of that.

Sure. Yes. So it's, I think, a very -- it's going to address a very high unmet need. And if we come even close to reproducing the preliminary data we unveiled at CHEST, I can tell you when the pandemic hit, we got calls from some of our Percepta customers asking us if they could have early access to our nasal swab test. And of course, we couldn't do that because it's not yet final and validated. But that just tells you that there is a real interest in having a tool that is a lot less invasive, that we have the potential eventually of moving to be collected really in any site in health care. I've often said like my dream is that someday, our nasal swabs are going to be able to be collected at CVS or Walgreens. And that's very, very feasible. These nasal swabs could potentially even be collected from home. So I think that being less invasive but not giving on the data has really been key here. And we've accessed very deep cohorts. When we showed the preliminary data, we started with one very well curated lung cancer cohort that we had built and then through our J&J collaboration have been able to access thousands of additional samples. In those cohorts, we have clinical truth labels on all of them already. So we weren't slowed down on the nasal swab with the need of our clinical trial because our clinical cohort is already so advanced that we have it. We split these cohorts into training and test sets. And right now, we're in the mode of training. We need to train and work through all the little nuances of the assay and making sure we get the ability to get results from almost every single sample we get. Some samples can be low-quality samples. We don't want to have to not test on patients that send us samples that aren't super high-quality RNA, for example. We've achieved that. We've got to work through a lot of comorbidity that are present in lung cancer diagnosis. This includes many patients that have lung nodules and lung cancer that also have COPD or interstitial lung disease or asthma or other biological things that are impacting the genomic makeup of their airway. So our team has been working through probably literally hundreds of machine-learning models looking for the exact mix of genomic content or other content, clinical content. We have access to imaging content actually in all these databases, but picking the right feature set that will give us the ability to reproduce those preliminary results and have very high confidence on the patients we call high risk and low risk so that we can get the high-risk patients to a treatment really quickly and move low risk out of workup. We are in that mode of sausage-making right now, and I'm sure we'll see preliminary results from this training cohort as that machine learning moves along. We may have some data that will be shared in conjunction with perhaps CHEST later this year. But the final validation data, the analytical verification data, all are critical data sets that we'll have to have completed before launch. And once we launch, we'll go to a select set of sites to be able to collect the cost effectiveness and real clinical utility of the test. We hope to reproduce what is shown on Slide 17 in how we want positioned and how we want it to create a new standard way of doing workup on lung nodules. Perhaps we can get in guidelines someday to drive that use. But we'll go to a number of sites, reproduce this, collect that data and get it published, and then we'll be off to try to get our early reimbursement. The cool thing here is that our customers are going to be customers that we're already calling on with Percepta, and many of these customers are also customers for Envisia. So we're going to get a lot of leverage in the pulmonology space as we continue to build out our pulmonology portfolio and answer new questions that create new opportunities to improve patient care and advance these tests into adoption and reimbursement.

And maybe just one. Since we're almost at the tail end of this, maybe I'll just stick on the nasal swab classifier, just one more question there. In terms of the data that was presented here versus what you're trying to achieve now in the pivotal data set, and you're kind of working through all the machine learning classifiers. You've done this many times before, obviously, in different indications. So I guess what's the [ bar like ]? How much is the hurdle to go from? I figured how big this data set was, but was this such a small data set that there's still a lot of wood to chop to reproduce this in a much larger data set? Or just how do we think about that jump?

Yes. I mean I think that we used a preliminary patient cohort that was a little bit rich and prevalent. Funny enough, most people that use a set that is not exactly similar to the market prevalence usually go the other direction because ruling out cancer and providing a low-risk result in a population with high cancer is actually harder. The higher the prevalence, you're -- the lower your negative predictive value, and we show very, very high performance here. But we wanted to tap a broader set of clinical patient samples that are more similar to the whole cohort that we'll actually be testing here with the nasal swab. I will mention we actually have quite a bit of confidence in the work that we do. We've never had a program fail once we put it into R&D. And I think it's because we use an unsupervised way of collecting all the whole transcriptome data. And we have a machine learning team that has been building clinical classifiers using these kind of models now for 12 years. So a ton of experience doing this. We've had really great success. Our J&J collaboration is helping, obviously, as well to advance the nasal swab test, and we think we're going to have good success here.

Well, we're at 3:40. Obviously, the prepared remarks and the slide deck was extremely helpful. I just hoped we could get to other questions, but obviously, it was really informative, Bonnie. Keith, you're on the phone as well, so thank you for joining us today. And obviously, hopefully looking forward to the next one to being in person. So -- and thank you all for participating...

Well, I hope so, too, Dan. Thanks, everyone, for joining us, and thanks, Dan.

Okay. Thanks, Dan.