Exact Sciences Corporation (EXAS) Earnings Call Transcript & Summary

Okay. Thanks, everyone, for joining us this afternoon. We have Exact Sciences with us. Very pleased to have the CFO, Jeff Elliott, representing the company. I think Megan from Investor Relations, she's in the background. Both Jeff and Megan, welcome to the conference, albeit in a virtual setting. I can't wait to get back to meeting you guys in person, but this is the best we can do for now. And I guess to kick things off, Jeff, 3Q was really busy. I mean, I had a tough time digesting the amount of news flow that you guys had between the deal, the equity raise, the deals and the print; it was a lot.

I -- let's maybe just start with the base business before we go -- get to the deals and the pipeline. How would -- the second wave has been topical. What, if any, impact have you seen from the second wave? Has it been in line with the expectations? Or perhaps any changes to that second wave assumptions for Q4?

Look, first, Vijay, it's great to be back here again, and I agree with you, I can't wait to get back in person and do these meetings live. So as far as the base business goes, we're very pleased at the pace of recovery in the base business. I think back to April, in that month, orders for Cologuard were down 60%. And as a society, things -- people were really starting to hunker down and try to wait out COVID. At that point, I thought our business would be suppressed for extended period of time. But because of the actions the team took and the convenient accurate nature of Cologuard, that business came back very quickly. And every month through the end of October, we saw a sequential growth in the business to the point where back in October we were up year-over-year despite a significant headwind from the pandemic. So we're very pleased both by the near-term pace of recovery, which is coming much faster than we had expected back in April. And we're also very pleased about the longer-term implications of COVID. We're confident that when -- down the road, say 3 to 5 years from now, when we look back at COVID, we think it would be viewed as a turning point for adoption of Cologuard, where we can help get many more people in this country screened and help clear some of the backlog that is being created by the depressed numbers of colonoscopy. Today, colonoscopy volumes are still down over 20% to 30% in this country. And anecdotally, I'm hearing more and more stories of health systems shutting down screening colonoscopies due to the rise in the pandemic. Yet, Cologuard taking share in this market, getting more people screened. And I think that longer term, Cologuard will be better off because of COVID, not worst off.

Got you. And I -- just maybe looking at these assumptions, I know Cologuard was down 2% in 3Q, Street's modeling a 1% decline in Q4, which sort of suggests a sequential step up. Is that consistent with sort of how you guys are looking at the business right now?

Again, we're very pleased by the recovery we're seeing in Q3. Well, revenue was down just a little bit, that's really the impact, the lagged impact from what happened in the second quarter. When you look at orders, which is more of the forward-looking metric, orders in Q3 were up year-on-year, although, albeit 03:36 modestly, they were up. And as I've said before, the underlying trends here are very positive. Through the month of October, we saw sequential growth kind of month-on-month and year-on-year growth. So things started off very well for Q4. When you look more broadly, there's a lot of exciting factors at play here. Whether it's an increase in electronic ordering today, up to almost 40%, or our ability to capture that 3-year rescreening patient, we continue to get better there. There's a lot of really exciting underlying drivers that set us up well for a very exciting year next year for the Cologuard business.

Got you. And then on the topic of electronic ordering, Jeff, what have been the trends year-to-date? Have you seen a sequential step-up as the pandemic progressed? Or -- I'm curious to see how that business has evolved? And I think you guys have mentioned in the past that e-prescribing results in a higher volume of tests out for you guys. So maybe perhaps talk about that electronic prescription.

Yes. Without a doubt, when a doctor can order electronically, she orders at a much higher rate, over 50% more than when they order via fax. So this could have a very profound impact on the business longer term. Today, nearly 40% of all Cologuard orders come in electronically, which is a significant number considering how many orders we get on a weekly basis. When you look back to, say, about 1 year ago, then about 30% of orders came in electronically. So this is a part of the business that has performed very well during the pandemic. In fact, we've had dozens and dozens and dozens of health systems convert to electronic ordering during this pandemic. And when you look ahead to the next 6 to 9 months, we expect that to be the most busy impact -- the busiest period of time in our history, converting doctors to electronic ordering. And that's because Epic, which is the leading health record system in the country and the one that we have chosen to use, Epic is coming out with a new version of software. This quarter they will enable electronic ordering of Cologuard for any doctor who uses that platform. Today, about 45% of primary care doctors use Epic, so we expect a meaningful increase in electronic ordering over the first half of next year as hospitals install that latest version.

That's a fascinating stat, Jeff. So let's go through some of those numbers, right? Epic is in about roughly half -- 45% of primary physicians use Epic. And the map is once those customers convert to e-prescribing, the order volumes step up by 50%. What do you think that means for -- I guess, as Epic rolls out end of this quarter, should we be thinking about the Q1, Q2 sort of pretty big meaningful step up because those are some big numbers.

They are big numbers, and that's what excites us about the long term. However, I would say that the 50% increase doesn't happen overnight. It takes time. Like the 50% number where that comes from is when I look at all doctors today who order fax versus electronic. So some of these doctors have been ordering electronically for years, say, 5 or 6 years, that's where you see the 50% increase, electronic versus fax. It doesn't happen overnight. You do get some step-up early on, but it takes time. The other factor is that when a hospital converts to electronic ordering, there can be new doctors who have yet to order Cologuard that come on. So it may take time. Now the beauty of this is that we know all the systems that use Epic, our reps have been on this like a hawk for a long time. So they know who uses Epic. They know that who they need to target. When Epic goes live, they know -- look, to make sure those doctors know that Cologuard is now available electronically and really drive that message home. So we've been preparing for this for over a year, and the team is ready. This can have a very material impact on the business going forward. So this is one of the biggest drivers next year and beyond. And what's really exciting about this is not only does it help Cologuard, it helps other pipeline products as well. So an example of this is our COVID testing. We initiated COVID testing back in April. Well, already today, over 90% of the orders we get for COVID testing are electronic. So having that electronic foundation established, allows you to move much more quickly and more efficiently. So as we bring on other products, whether it's multi cancer, which I'm sure we'll touch on, or our liver cancer test, that foundation allows us to move faster and helps increase the adoption rate because it's a lot easier to order a test right through the native electronic record platform than having to log into a new portal or send a fax. So we're really excited about what this means for the long term for Exact Sciences.

That's great to hear, Jeff. The -- I guess you touched about 90% of orders for COVID coming electronically. I mean, I look at COVID revenues for you guys, pretty big step-up sequentially, 30 million to almost tripling up to 100 million. I mean so far, I mean, Q4 seems to be up sequentially for most companies. I mean, Thermo just preannounced, most others are saying demand is outstripping supply. Roche just said that. I'm curious on how we should be thinking about COVID revenues for Exact for Q4 or perhaps in the next 2 quarters?

Well, there's no doubt that demand is far outstripping the supply -- the market's ability to supply testing is far surpassed by the demand. And I see that as a good thing. Look, as a country, we decided for us to defeat this pandemic. Testing is one of the keys to making sure that people know when they have it and when they don't. So our ability to meet that is somewhat limited by well-known supply chain constraints. So for example, gloves and gowns, PPE supplies out there are somewhat restricted. So on our last earnings call, we said you should expect revenue flat to down from Q3 because of -- in part because of supply chain constraints. Another factor is that our lab capacity is limited in the short term. We launched our new lab in the summer of '19, with a goal in mind of being able to process Cologuard tests. Well, over time, as Cologuard has grown, eventually, we'll need to fill up that new lab with Cologuard testing. So as Cologuard comes back, that limits what we can do with COVID testing. In the medium term, over the course of next year, we may be able to bump up our lab capacity somewhat. However, in the short term, we're somewhat restricted, which is why I said, flat to down in Q4. So there's a big mismatch that's clear. There's far more demand out there, but there's only so much we can do in the short term to meet that demand.

Okay. So there's no change on the COVID aspect from your perspective versus your expectations on the 3Q call. Is that a fair summary?

Yes. Look, I'm trying to stay away from an intra-quarter commentary, but those supply chain constraints are real. It's not just the PPE, there's other things too that are somewhat limited. And then there's -- the demand for COVID testing is real. But there are parts of the country where it's bigger than others. A lot of our testing is in our home state of Wisconsin. There's other parts of the country where the supply-demand mismatch looks different than Wisconsin.

Got you. No, that's clear. The -- I guess, the other part of the business, Precision Oncology, looks like there was some timing element when you look at the 2Q to 3Q progression, maybe talk about what happened. And what's the current state of affairs for that part of the business? That seems to be lagging screening. Any reason why that market should lag screening?

The Precision Oncology business is a very sticky business. And longer term, what's happening here, I view it mainly as a timing shift in the business. The reason why -- the primary product there is Oncotype DX Breast. So this is for women recently diagnosed with early-stage breast cancer. And the question we're trying to answer is, will that women benefit from chemo? Most of them will not, in fact, about 80% of women will not benefit from chemo. However, before this test came to be, most women, about 80%, chose to use chemo. So we're trying to help align the use of chemo with women that will need it to help cure or most benefit from it. So in the early days of the pandemic, mammogram volumes were down significantly, over 90% in the month of April. And what that meant is a woman that was going to be diagnosed with breast cancer didn't find out that. Her diagnosis was delayed. There's typically a 1-2 month delay between a breast cancer diagnosis or mammogram and an Oncotype DX order. Because during that time period you have to confirm the diagnosis and do surgery. A portion of the tissue that's removed during surgery is what's sent to us for testing. So as breast cancer diagnosis are delayed and ultimately that surgery is delayed, that led to a temporary decline in Oncotype DX Breast orders. The reason why our third quarter was hurt more than our second quarter is because of that 1-2 month delay. So a decline in mammograms in the second quarter hurt orders in the third quarter. However, mammograms have come back. Today, I believe they're down between, I'll say, flat to down 10%, which means we're far better of today than we were back in April. And as we head into next year, I think, the business will normalize over the course of next year. And again, a woman who was going to get breast cancer and didn't find it in April, she'll eventually find it, whether it's through a self exam or mammogram, eventually she'll find it. So it's really more of a timing impact. In the fourth quarter, we expect a sequential increase as mammogram volumes have started to normalize.

That's good to hear. And just so we have some of the numbers. What was mammo volumes in 2Q on an average? What was it in 3Q? It looks like it's trending at 90% of prepandemic in Q4. Do you have some sense on what those numbers were, Jeff?

In April, down probably 95%; for the quarter in total, probably down the 35% range. They did come back quite a bit over the course of May and into June.

Got you. And when you think about the base business normalizing to prepandemic levels, right, now that you have the vaccine that's about 90%, 95% efficacy. A lot of your peers are talking about just don't look at the headline numbers, there are logistical challenges in getting people vaccinated. So I'm curious on how you guys are thinking about the base business normalizing to prepandemic levels? And Cologuard is back, I guess, prepandemic, right? Those were the numbers in 3Q. Were there any timing dynamics, any backlog sort of impact in 3Q when you -- do you feel like Cologuard is back to prepandemic levels?

Cologuard orders are clearly back. In fact, they're above where they were at the start of the pandemic. That said, we expected more growth, right? To start the year, we had guided to revenue of over $1.1 billion. Today, consensus for Cologuard is around $800 million. So it's clear that the business in total has been impacted by the pandemic, temporarily, otherwise, I said before, longer term, I think COVID will be viewed as a tailwind to Cologuard, not a headwind. The exciting part for the business is that there is a significant backlog of people needing to be screened. Well over 1 million patients have deferred colonoscopy this year alone because of the pandemic. And colonoscopy volumes are still down at least 20% to 30% in this country, which means that backlog continues to grow. Cologuard on the other hand, it can be done safely from the convenience of your own home. There's no face-to-face office visit required to order a Cologuard or conduct a test, which positions us to really help alleviate that backlog. So relative to colonoscopy, we are taking significant market share. And over time, I think we'll keep a lot of that share as we come out of the pandemic. On the Oncotype side, the backlog here is, in dollar terms, is well over $20 million. When you look at the revenue that we haven't yet captured because of the pandemic. However, I think most of that will come back to us eventually. When you look at the market share for the breast business, of the tests that have happened in this country, at least 90% of the breast cancer prognostic testing is an Oncotype DX test. So it's a very, very sticky business. I expect to get most of that back eventually. The one exception here is that there are some women who, because of the pandemic, their cancer has progressed to a point where chemo was required, let's say, they're into -- well into Stage 3. At that point, chemo is likely required and an Oncotype DX Breast would not be used.

That's interesting. And I think in general, people are expecting first half of next year to be below trend and normalization perhaps mid of next year or back half. Would those assumptions make sense for you when you're thinking about things like physician office visits, mammograms. These are all basic screening services. When do you think that volume gets back to prepandemic levels?

I think that's probably a safe assumption, Vijay. First half's still somewhat impacted by the pandemic. But ultimately, it's going to depend on how effective is the vaccine in reality and how quickly will people decide to use it and then get back to their lives. When you look at mammogram volumes, we're getting close to where they were prepandemic. On the primary care side, physician office traffic is probably still down 10% to 20%. So there's still ways to go there. But people are getting closer. I think each day that goes by, people are getting more confident to return to their daily routines.

Got you. No, that's helpful. And I'm going to switch around things a little bit here, Jeff, because the one you mentioned back on electronic ordering. Right now, 40% of your volumes are electronic prescription. What should -- is that the cadence 40% goes to 50%, next year it goes to 60%. How should we think about that fax to e-prescribing conversion?

We'll try to move it higher as quickly as possible, Vijay. If you look at some of the big labs in this country, they end up at around 2/3 of orders electronic, 1/3 fax. Hopefully, we can get to at least there, considering that we've got this Epic platform, which is so important. We also have a physician office portal that any doctor can go to, regardless of what platform they use, they can go to our portal and place the order in that way. And we continue to enhance that portal to make it a more valuable tool. For example, we're rolling out COVID testing functionality through that portal. We're also -- we've added enhanced rescreening capability through that portal. So any doctor who logs in can see on that portal all the patients they have eligible for colon cancer screening again or 3-year retesting. And through the click of 1 button, once they've validated those patients are in fact due, they can click 1 button and reorder Cologuard for all those patients at once. So we've made that portal more valuable. And over time as we launch other new tests, whether it's liver cancer or multi-cancer testing, that portal, I think, will become a bigger share of our orders.

Got you. Is there a margin implication of e-prescribing versus ordering on fax, I'm curious?

Longer term, electronic ordering is a much more efficient way to receive orders. Think of a typical fax. If you've ever gone to the doctor and looked at their handwriting, no offense to any doctors here, but doctor's handwriting is often difficult to read. Or imagine a fax form where they forget a field. No, it doesn't happen all the time, but when it does, it requires somebody in our end to pick up the phone and either call the doctor or the patient to retrieve that information, and that takes time. When you get an order electronically, you typically have all that information upfront. So we avoid that upfront step, that phone call. And then when we bill, we bill after we complete a test. When we bill,having all that information updated is important because insurance information can change year-to-year. And if you get the insurance information right from the electronic system, it's more likely to be accurate. So it saves us time and energy and money on the upfront and in the back end. So it is a more efficient way to do it. And then as we grow -- this Epic platform is so powerful, as we grow, we can grow much more efficiently. Before we install the Epic internally, we had basically been building an EHR platform organically, which I think is over 40 years into doing this, they are a leader in doing this. It's a lot more efficient for us to leverage all their expertise than for us to try to build it on our own.

You sound pretty excited about Epic, Jeff. What's the closest analogy for me to think about what Epic means to Exact? Is this like your Pfizer where you had the agreement with Pfizer and what that did to the business in terms of accelerating revenues, could Epic be a similar sort of impact on your business?

I don't want to try to guess on what it could do next year, Vijay, but it is a very exciting tailwind for us next year and beyond. Again, not only for Cologuard, but for all of our tests. We're also working to bring the Oncotype DX Breast test, the prostate test onto this platform. And as we grow, again, it'll help us grow a lot more efficiently. So I'm excited. It is one of the top few drivers as is Pfizer. So it's in the same ballpark, but I don't want to try to compare from a revenue standpoint, which one's bigger.

Fair, Jeff. There's some work left for the sell-side to do, I guess. I guess turning on to the pipeline. Next year is going to be pretty busy on -- let's start with the colorectal cancer screening market. I think there's data coming out for Freenome and Guardant Health, I think, is around at some point. Where are you on your own blood-based colorectal cancer screening test? Is that -- the enrollment complete? Or when are we expecting the enrollment to complete?

So our BLUE-C study is both for Cologuard 2.0 and our colon blood product. That study is enrolling. The enrollment for that had been put on hold over the summer because of the pandemic, but we're now enrolling now. We expect that to complete either late next year or early the following year. I gave a range because of the pandemic. The way that the enrollment works is, a patient has to first provide a blood sample and a stool sample, and then follow that up with the screening colonoscopy. That colonoscopy, those volumes are somewhat restricted because of the pandemic. So enrollment timing is a little hard to understand right now given the pandemic. But late next year -- the following year enrollment should complete. We'll analyze the data and then publish that data and use that for FDA approval, again, for -- both for Cologuard 2 and our colon blood product. We also, at some point, will publish our case-control study. We've had multiple case-control studies for our colon blood product. We'll publish that at some point before we go to the FDA. Right now, we're working to further enhance the performance of that test. When you look at the Base Genomics technology we recently acquired, we think that can further improve the performance of our blood-based assay. So we're really excited about what that technology can mean, not only for a colon blood product, but for all of our pipeline products.

Understood. And just so I understand, Jeff, the -- your study right now is looking at both the 2.0 as well as the blood based test. And that data will be submitted at the FDA at some point in fiscal '22, looks like it. And I think you mentioned Base Genomics will improve your blood based test. Is that a 3.0 version test? Or would your FDA submission fiscal '22 include some of the algorithm changes because of Base Genomics?

Base would be included. We're working to prove that out, but we plan to include that both in Cologuard 2 and our colon blood product. The way Blue-C works is you enroll all the patients and you bank the samples as you enroll. You don't run the samples through your assay until enrollment is complete. So you have time to further improve the performance of a test up until kind of near the end. You have to lock down the performance of the test before you start running any of the prospective studies, before the prospective samples.

Got you. No, that's helpful to know. And I guess, CMS came up with their proposal for the NCD, 74% sensitivity and 90% specificity, how confident are you of meeting those thresholds when you think about your blood-based offering?

Look, we think it makes a lot of sense what Medicare did. And again, this is an example of the agency being very forward thinking. So we're pleased that they went ahead with a -- coming up with an NCD for blood-based testing broadly. We think that was the right thing to do. From a performance standpoint, for an annual test, we think the metrics need to look something like the FIT test, which the 74% metric you cited, that is from the FIT test. In fact, that number comes from our DeeP-C study, which included a FIT component. The specificity cutoff for an annual test likely needs to be in the 95% range. And if you do the simple math here, a blood-based test almost certainly has to be an annual test. If you have a 90% specificity, which means a 10% false positive rate on an annual test, you'll send somebody unnecessarily to a colonoscopy at least every 10 years, which then the health economics of the test don't work very well because what the guidelines also says do a colonoscopy every 10 years as the frontline test. So a false positive rate for an annual test probably needs to be in the 5% range. So that's what we're working towards. Our goal is to have something that works similar to the FIT test. So I think 90% probably proves too low for an annual test. For a 3-year test, which is how Cologuard was designed, 90% makes sense. In fact, that 90% number was based on the guideline specificity, the U.S. PSS specificity for Cologuard, which is 90%.

And just on the annual versus once every 3 years, I'd just remind us on -- is that CMS, is that looking at an annual -- I thought, for some reason, CMS proposal was for once every 3 years? Or am I mistaken?

They did, although they did refer to the guidelines to say, look, it's 3 years unless the guidelines say, something else. That's the draft. I'm not sure what they say in the final, but the draft said 74% sensitivity, 90% specificity, 3 years. But then there's also this guideline inclusion component of the draft coverage decision.

I see. And that's where things should move based on what the guidelines recommend. The one -- I guess, given that you have 3 blood-based tests hitting the market at some point over the next 2 years, hopefully. One, what -- is there a number to put on the clinical performance, Jeff, or you can look at these tests objectively and say one test is better than the other. I guess, there is one on the analytical side, right, one test perhaps is 90% sensitive, and another one is 88%. I'm not sure if that matters clinically. So when you look at the market from a practical clinical standpoint, patient perspective, what kind of numbers do you need to see before the market -- physicians start differentiating one test is better versus the other? Is that a 5% delta on sensitivity? Is that a 5% delta on specificity? I'm curious.

Well, really to compare these tests, you have to have prospective -- large prospective studies. Up until then, comparisons, you really can't do because the design of each of the case-control studies has been different. And there's questions like, do these studies include precancer. In the real-world setting, you have to be able to find precancer. That's -- you just have to do it right. The impact on patient lives and patient outcomes from precancer is arguably more important than it is for finding cancer. Probably 10 times more people will have precancer in a prospective setting than cancer. So you have to find precancer, and nobody's published any data on precancer. So I would say before you can really compare any of these companies head-to-head, you need to have that prospective data. I think both Guardant and Freenome are both very talented companies. I understand why they're pursuing the market. We have been working on a blood-based test for many years. Ultimately, I think the FDA will end up having a combined panel, just as they did back in 2014, they did a combined panel with Cologuard and another blood-based test. They reviewed both of them at the same time and then ultimately issued the decisions. I think that's what they'll do probably in the '22 time frame for any test that makes it through a prospective study and for companies that submit around that same time. As part of your question on the difference between performance, look, it's -- I think there's more than just a couple of points here and there. That's probably not enough. There's other things like the precancer data, how do you do on high-grade dysplasia, there's a lot more to it. And then ultimately, how do you market the test, what kind of relationships do you have with physicians? We're confident in our positioning, given that the primary care call point is something we do today with Cologuard. We do today with COVID testing. We've got the foundation, whether it's the lab, IT, sales force. And I think that gives us a good head start on the launch of a blood based test.

Absolutely. No, that makes sense. I do want to get into the infrastructure part a little bit, but just sticking to the trials itself. From your perspective, have you seen any differences between how these trials are designed when you look at these 3 companies? Anything that's take out from an inclusion/exclusion criteria?

They're all pretty similar, right? And every company is talking about at least 10,000 people, which is the right way to do it. They're all looking at an average risk prospective setting. I think they're all looking at including a younger person. When you look back to DeeP-C, which is our study for Cologuard 1, we enrolled certain age 50. Now we're enrolling certain age 40 given that colon cancer incidence rates have started to trend younger. So I think all the companies are looking at a younger population. So they're all fairly similar. There are some small differences, but the studies are pretty similar.

Got you. And when you think about the infrastructure, and this is where I think -- I'm not sure how well it's been understood by the market. It's very hard going after primary care docs. How easy or difficult is it to order the patient building the awareness? And I know you guys have made a ton of investments in those areas. Maybe highlight those capabilities. They can be meaningful, but I think it's harder for the market to understand or assign what -- the significance of those barriers.

You're right that the barriers are significant. There was a study public almost 20 years ago now that said, on average, it takes 17 years for new procedures, new tests to get fully adopted in primary care, 17 years. Cologuard has been in the market 6 years, and we often talk of our longer-term ambitions of 40% market share. Well, there's a long ways for us to go to get from here to there. We're at about 5% share today. So 17 years on average. And it takes a significant investment. We've invested all told north of $3 billion in our infrastructure; things like our sales force, on a total basis, over $2.6 billion; our IT infrastructure, over $400 million; our labs, over $300 million. So it's a significant investment going forward. And I can go on, there's a lot more than that. And that's looking backwards in time, forwards in time with inflation and competition, the investments will likely be higher. So it's a significant investment to attack the primary care market. We can do it a lot more efficiently going forward now because we've got Cologuard 1 and eventually Cologuard 2, layering in a blood test and a liver cancer test, other tests, whether it's colon cancer or multicancer, that same foundation can apply. That same Epic foundation, that same sales force would market new test. So that gives us a big head start. We see how hard it is selling Cologuard, and that gives us a lot of insights. We know the early adopter physicians. We know the early adopter patients. So we know who to go after. And also when you look at the commercial insurance company landscape, there are several very large, very difficult to penetrate insurance companies. We've got deep relationships with them that have improved with time. Early on when you launch a new test, it's hard to get the attention of a company. You want to go after somebody like UnitedHealthcare, it's a huge company. Where do you start? Well, with Cologuard, we've got a great relationship with the big payers. With Oncotype, we also have a great relationship. That gives us a deep foundation to build on, to bring new tests to market.

Yes, that makes sense. You touched upon pancancer screening. The acquisition itself in and of itself, Thrive was really interesting. Maybe let's start with that philosophically, when you look at that market, right, is that a right approach to that market? Because when I look at those different companies, some people are looking at targeted methylation, some are looking at mutation analysis. And those are all -- and these companies, they have data to show like we've tested all the other things, and this is what works. So I'm curious, philosophically, is there -- how do you look at that market? Can different approaches survive in that market? Is there eventually a one approach, which ends up being the right approach?

We think there'll be multiple companies who are into this market. GRAIL is obviously out there. GRAIL is a very well-run company. It's done some really nice work here. I believe the approach they're using today is primarily methylation markers. We've been looking at methylation markers for over 10 years as a company. We think that is probably the best class of markers out there. However, I think different approaches could survive. We think, eventually, there will be multiple different companies out there. With Cologuard we found combining multiple classes of markers, multiple orthogonal classes is what guide us to a high level of performance. So Cologuard 1 includes methylation markers, mutation markers and protein markers. When you look at what we've done historically in multicancer testing, we've looked primarily at methylation markers. We shared some data several months back, using just a handful of methylation markers, we could achieve a high level of sensitivity and specificity. Thrive historically has looked mainly at mutation markers and protein markers. On a combined basis, we think that we can get to market with a more accurate test and more quickly. We're working to prove that out right now. So we plan to publish data next year showing a kind of bind basis when you combine the legacy Exact work in methylation, the legacy Thrive work in mutation and protein, plus adding the other technologies we've acquired over the years through Base Genomics, and our Biomatrica blood tubes, we think that can get us to a very good point when it comes to overall test performance.

Understood. And would that -- the data next year, Jeff, is that a retrospective analysis? Or will you be running an algorithm in a new, I guess, patient cohort sample?

It's likely to use case-control samples that we've collected. We've collected over 200,000 blood-based samples over the years. So we plan to use those bank samples to validate the combined assay.

Okay. And is there a timing, Jeff, on when perhaps we could see this data next year?

Well, the first step is next year is to close the combination with Thrive. It's -- look, that's well underway right now. So we expect to close that next year. And then the teams are working together to plan out what we'll do on a combined basis. So it won't be early in the year. It will be at some point, likely at a scientific conference that we publish the data.

Got you. The one, I guess, I wanted to ask, Jeff, is on asset price valuations, right, you had 2 similar companies, pancancer screening markets. One was valued at $8 billion, the other was valued at $2 billion. And some people have asked the question that disparity in asset prices in the same space, going after same market, is that reflective of test performance? And how would you answer that? And the reason I ask that is people look at different metrics, right, sensitivity, specialty, predictive values, and I think predictive value for Thrive was 19%. GRAIL was 40% oddish. There were differences in those trial designs. I'm curious how you would answer that question to people asking is there a difference in test performance, which would suggest why the valuation levels are different?

Well, I can only comment on the valuation we paid for Thrive. We think it's a fair valuation. Internally, we had worked on in multicancer test for years. And we knew that going forward, there's additional investments that we would have had to make, things like additional talent and bioinformatics and sequencing capabilities. And so we had a pretty good idea of how much more we would have to invest to get to the endgame. We know that based on the relationship with Thrive, which we had invested in Thrive over a series of different findings and rounds. And we knew that team well, and we knew that they had deep experience in the areas that we needed to bolster our own talent. So based on that, we could make that trade-off analysis between build versus buy. And given the price that we agreed on with Thrive, we thought that was a fair valuation. At the same time when you look ahead, there's significant investment that still needs to be made here, whether it's the prospective study, which will be over $100 million study or bringing the test to market, a lot of the incremental investment we've already made. So we feel good about that. But there's a long ways to go here. We think it's worth it, though, given how big and attractive this market is. We see this as at least a $25 billion market in the U.S. alone. I know others have put a bigger tag on that. We think it's at least $25 billion. So we're excited to combine with Thrive. We think it's a great combination when you look at what they have and what we have. And I know both teams are excited to get moving forward.

I think I even saw $100 billion number on that. So yes, it's a big number. It's going to be a big market for sure. I guess, on the pancancer market itself, is there a single metric which would matter because some people are focused on predictive values, sensitivity, specificity. And the question for me is, without a 2-arm trial, a control arm, how do you know that any cancers that you find in the device arm or the test arm if that is incremental or not, I think it's going to be hard. Would you agree that you need a 2-arm trial to validate these results?

I think that would be really helpful, Vijay. A number that I think everybody agrees on is specificity. I think specificity probably needs to be at least 99%. We're comfortable with that number. And then the goal is to maximize sensitivity given a fixed specificity of at least 99%. We're confident that we can do this with the team we have and the team that Thrive is bringing. PPV that is now you're starting to do math between those numbers, the positive predictive value, we think, needs to be high as high as possible. When you look at the incidence in this overall market, typical incidence in this population is probably 3% or 4%. So if you have a 1% false positive rate, and let's say you can find 50% of cancers, which we think we can do better, but let's say we can find 50%. Well, now you're talking about a positive predictive value of north of 50%. So that's a very attractive outcome. Considering that's far above where you'd see it for mammography or even colon cancer screening. So we think that given those numbers, a pancancer, really a multicancer test, could be a test that not only helps patient outcomes, it could potentially help the system to save money given that many of these cancers are found at late-stage today, typically, oftentimes Stage IV. If you look at colon cancer alone, 60% of colon cancers are found late stage, where it can cost $300,000 per patient to treat. So if you can shift that earlier, where oftentimes you could quick do a surgery and avoid chemo or expensive late-stage therapies, you can really save cost of the system and improve patient outcome. So that's why we're so excited about this.

Makes sense. And when you think about post transactions closing, Thrive closing, how should we think about expense ramp? I think -- and you put out the $100 million on the trial expense that is going to ramp at some point. But outside of the trial expense, any thoughts on what Thrive does to the P&L for you guys?

Well, the biggest incremental expense would be the trial -- on the trial, I think it's at least $100 million, spread over multiple years. The ultimate cost of the trial depends on the design. What Thrive had looked at is a study of about 80,000 people. Ultimately, it's probably at least that big. As far as how much it costs, we have to finalize the design first. So we'll share more thoughts on that as the -- over the course of next year as well as the overall impact of the P&L. That said, we've done the math internally. We think that given the size of this market and the strong foundation we have, we think it's a very high return on the investment.

Got you. And then one, I have a couple of minutes left, but Street's modeling 55% Cologuard revenue growth for next year. That's -- basically it's getting back to what you guys should have been for this year, right, prepandemic guide. Do those numbers make sense to you that next year should be a normalized year for Cologuard? I mean, it feels like first half physician office visits will still be anemic, I'm curious to -- any thoughts? I know you guys don't like to comment on fiscal '21, but thoughts on how -- do the Street assumptions are normalization for next year? Does that make sense to you?

Well, look, we'll provide guidance for next year probably in our fourth quarter call. Next year, I think will be a very exciting year. When you look at the drivers we've talked about today, whether it's electronic ordering with Epic or 3-year rescreening, we're just getting our reps back into the field. Both we and Pfizer are starting to see more and more doctors face-to-face, which should be a big driver. That said, the pandemic is here. It's still real. And I think at least through the first half of next year, we expect an impact on the business. You talked about 50% growth. Really, the way that works out is closer to 20% growth relative to Q4. Given the size of this market, yes, those numbers are possible, but we'll guide to 2021 on our fourth quarter call.

Fantastic. I think with that, we're at the end of our allotted time here. Jeff, thank you so much for the time this afternoon. This was great. Have a wonderful rest of the afternoon.

Thanks. You too, Vijay. Thank you.