Exact Sciences Corporation (EXAS) Earnings Call Transcript & Summary

Okay. Thanks, everyone, for joining us this morning. A pleasure to have with us the management team from Exact Sciences. We have Kevin Conroy, CEO; Jeff Elliott, CFO. And I think in the background, we do have Megan Jones, Director of IR. We're going to structure this slightly differently. I think Kevin wants to run by a few slides, and then we'll get into the fireside discussion. With that, Kevin, Jeff, thank you so much for the time this morning. Over to you.

Vijay, thanks for having us. Today, we wanted to take an opportunity to step through and highlight some of the major programs in our pipeline and the great science that is -- work that's being done here at Exact Sciences. And so let me advance here. We'll be making some forward-looking statements. This is our safe harbor statement. The more complete one is on our website. So Exact Sciences' goal is to be the one-stop shop for all advanced cancer tests for health systems, health care providers and, importantly, for patients. We have 2 of the best brands in advanced cancer diagnostics with Cologuard and Oncotype DX and a deep relationship with health care systems with hundreds of health care systems, the largest health care systems, and well over 0.25 million of health care providers that order Cologuard, Oncotype DX and a deep infrastructure, enabling us to bring the most advanced test to the physicians who order them and the patients who need them. And also eventually layer in an incredible layer of intelligence that helps guide physicians which test to order, what's the next step that they should take in their treatment. We're excited about this pipeline that will add to the tests and capabilities we have today. The 3 largest programs are our multi-cancer programs, our colon cancer programs and also minimum residual disease testing. We have made a number of acquisitions that are targeted and support these efforts. And these efforts are making -- the teams are making great progress towards these goals. When you take a look at the amount of evidence that is going to be generated over the next 18 months, it's pretty eye-popping, and we're excited about the next couple of years ahead as our scientific teams are making incredible progress. There's a huge total available market. And that really is just another way of saying there's an enormous need. Our goal is to eradicate cancer and the suffering that it causes through earlier detection of cancer, earlier detection of recurrent cancer and also tests that help guide the right treatment to the right patient at the right time. Let's take a look at our colon cancer screening programs. We are tremendously excited about Cologuard 2.0, which is a significant improvement over the current version of Cologuard, which already is included in all of the major guidelines. Colon cancer remains the #2 cancer killer. In the U.S. alone, there are 110 million Americans aged 45 to 85 that are recommended to be tested. 45 million Americans today are out of date for screening. And if you look at the global opportunity, in Europe, only about 25% of people are regularly screened for colon cancer. In the rest of the world, the screening rates are even lower. We intend to be the leader not only in the U.S., also globally, and we have the ability to do that. Cologuard 2.0 will help us lead the way. So we will -- we have developed this advanced version of our test with new DNA methylation markers that we discovered in our partnership with the Mayo Clinic, painstakingly looking at millions of DNA methylation sites across colon cancer tissues, identifying markers that are even stronger than the current markers in Cologuard. Our goal of this program is twofold. We want to lower the false positive rate from 13% to significantly less than that, at least only 10%; and also to improve the pre-cancer detection rate, which today is at 42%. We have data that we will present at the ASCO GI meeting in January. That data is a combination of prospectively collected samples, people with precancerous polyps. All of those samples were collected prior to colonoscopy, and then retrospective samples collected after colonoscopy for patients with cancer. This is a well-powered study, and we're excited to present that data in January. This will set a new standard for colon cancer screening. And I think further separate stool-based test from the FIT test or colon cancer blood tests that are in development. And that performance is critically important. This is a part of a larger program in the BLUE-C study where we're collecting both stool samples for Cologuard 2.0 and blood samples. We expect to release top line data in 2022 and submit to the FDA in the first half of 2023. Our colon cancer blood test is also making good progress. It is important that there are multiple ways to screen for people. The evidence clearly shows that when you give people options, more people choose to get screened. And with only 57% of people screened in that 45- to 85-year-old population, we know that there is room for improvement. There are some people who, despite the performance characteristics of Cologuard and colonoscopy, may choose another test. We want to be a provider of that test to help more people get screened. So our colon cancer blood test, which is a multi-omic approach to testing with DNA methylation protein and potentially another marker class to help further improve performance. And our path forward will include releasing case-control data next year, completing our BLUE-C study enrollment next year and reading out and submitting to the FDA and seeking FDA approval and guideline inclusion thereafter. It's important to note the main guideline group is unlikely to update its guidelines until 2026, so we have some time to perfect this test and complete this study. Our colon cancer -- or Cologuard 2.0 program will complete before this because, interestingly, in the study to date, 20% of people who are willing to do a stool test are not willing to do a blood test. And so the stool program will complete and be submitted before the blood program. Turning to multi-cancer early detection. If there is one diagnostic test that has the possibility to impact human health greater than any other, it's multi-cancer early detection. Think one blood draw, many cancers, if not most cancers. And today, screening only encompasses while 30% of cancer, 70% of incident cancers have no screening test. So all of the screening efforts we're doing is for a fraction of our overall cancer diagnosis. And this is a test that potentially everybody between the ages of 55 and 85 would use and maybe younger than that. So at least 115 million Americans. And the difference between our blood program and our stool program in the past, our colon cancer programs, were aimed at a disease which already had multiple screening tests. Multi-cancer, there is no screening test. So we would expect to see significant uptake, and we're excited about the performance level that we are seeing. So one thing that you want to do with a multi-cancer test is control for -- and make sure that you have a very low false positive rate. The last thing that you want to do is send millions of Americans on an Odyssey telling them that they might have cancer with an array of follow-up tests. So our goal is to have a very low false positive rate and a relatively high level of sensitivity. There are 2 programs that have come together. Our own internal program and Thrive's program. As you know, we acquired Thrive a year ago and Thrive came out of Vogelstein's Lab at Johns Hopkins University, focusing on DNA mutations and proteins, curated mutations that span cancers. And our program, which was developed focusing on DNA methylation markers with the Mayo Clinic, bringing the best of the Mayo Clinic and the best of Johns Hopkins with our incredible R&D capabilities into one test. That's the goal. We expect to have data in the first half of next year. That will allow us to lock our algorithm, start a prospective study that would be between 50,000 and 100,000 patients in the U.S., potentially another study outside of the U.S. The Thrive technology is the only test that has been studied in a prospective study, the DETECT-A study at Geisinger. And it's pretty remarkable. What that study showed was the doubling of the screening detected cancers in a population of 10,000 patients. Powerful. I mean, there are patients that were diagnosed for example, one with ovarian cancer Stage 1. Ovarian cancer [indiscernible] [ detected ] at Stage 1. And just a woman who decided to participate in the study out of the goodness of her heart ultimately was found to have Stage 1 cancer. Surgery alone appears to be curative. It's powerful. And that's the goal of our program, is to combine these 2 technologies with advanced sequencing capabilities to have a powerful, highly sensitive and specific test. Our path forward, as I mentioned, is to lock the assay next year, launch a pivotal trial, submit to the FDA in 2023 -- after 2023, and seek USPSTF guideline inclusion. Because we're not the only company with a program, we think that is a very good thing for this space with lots of evidence being developed, which will increase the likelihood that payers will cover the test, including medicare, and guideline groups will include it. So we're excited about the MCEV program. We're also excited about minimum residual disease testing, frequently called MRD. You know there are tests that are being used today that really are changing the way that cancer is being treated. So people who are diagnosed with cancer and treated are left with 2 very important questions. Number one, did the surgery get all of the cancer? And two, is -- after that point in time, has the cancer recurred? Today, this is either left to a patient feeling symptoms or experiencing symptoms or potentially a patient periodically getting scanned. Unfortunately, that's just not good enough. And 85% of relapses are caught too late for curative surgery. The work that others have done in this space has been groundbreaking and has shown that there's an ability to see circulating tumor DNA in a patient's blood 9 months before recurrence is seen on a PET CT. We will study our tests, and we have 2 different approaches in, first, colorectal cancer with data next year, and then start a prospective study, a major prospective study in the breast cancer space, where we see about 50% of all breast cancer patients have their -- in the U.S. have their tissue examined by our Oncotype DX test. And methodically, deploy our capabilities in oncology to garner a significant share of testing in MRD. And we intend to take 2 approaches. One is a tumor-informed approach with -- that can detect up to 115 mutations and also a tumor naive approach using our methylation markers and also potentially the Thrive technology that emanates from Johns Hopkins. So we're -- we would be the only company with both a tumor-informed and tumor-naive approach, allowing us to test patients regardless whether a tissue sample was collected upfront. We also, and it's important to note, have made investments over the last decade in our commercial organization. It's the largest in advanced cancer diagnostics. Our IT infrastructure, which we've invested over $600 million to be able to connect directly with health systems. Our EMR platform. We have built the company on Epic to be able to engage with health care providers. And our deep regulatory and clinical capabilities along with our strong balance sheet. It's an exciting time at Exact Sciences. I'm tremendously proud of the scientific work that has been done by 450 scientists and the collaborations that we have with Johns Hopkins, Mayo Clinic and City of Hope. So excited to answer questions both Jeff and I.

Fantastic. Thanks for that presentation, Kevin. Maybe if I could just start with -- fiscal '22, that's a big, big year for you guys, right? We have the Pfizer sales rep being productive, COVID is COVID, and we'll see what happens on the pandemic or perhaps the -- what's now become an endemic disease, and a lot of clinical updates, right? And I think you went through some of the timelines. But maybe just for '22, talk about -- recap to us, there was a lot to go through in the presentation. If you don't mind, just what can we expect in first half versus second half from a clinical perspective? And then we'll go from there.

Well, if it's okay, I'd like to touch upon something that you mentioned there, was the strengthening of our commercial team this year with the addition of 450 highly talented people from Pfizer that have been part of our mission for the last 3 years, coupled with the team of about 1,000-strong in sales and marketing that we have built over the last 7 years. This is a team that is totally -- they wake up every morning, ready to go out there and get more people screened, more people tested, with tests that actually change outcomes. And we have seen the impact of the addition of these professionals to our team, coupled with the performance of the teams that we've built. And you saw that almost immediately as we got all of our people back into the field for the first time in 18 months. We've been fighting this cancer fight with at least 1 hand tied behind our back for the last 18 months. And it feels great now as we're getting more access into offices with more sales calls leading to a greater uptake of both of Cologuard and Oncotype DX tests. And we're proud of all of the laboratorians that have made this possible during awfully challenging times. In terms of the first half versus second half, maybe the highlight in the first half is right upfront in January at ASCO GI, the Cologuard 2.0 data. I'm really focused in on that because the performance that we expect to see is powerful. It sets a new standard in colon cancer screening. And one thing we know is that we are perfecting the ability to get people tested at home, and the Cologuard brand is so powerful. So the ability to educate physicians about even better performance, a lower false positive rate, higher pre-cancer detection, the 2 things that docs really care about. If there are 2 critiques of Cologuard that -- what they want to see is a lower positive rate and a higher pre-cancer detection rate, and I think the R&D team has nailed it. There's -- in the second half of the year, we would expect to release colon cancer blood data, and that's powerful. In the first half of the year, multi-cancer detection data. And then in the second half of the year, we would expect minimum residual disease testing data. So that's highlights. We don't have to go through all of the data that will be released, but those are 4 key highlights.

Fantastic. And I do want to dig on the commercial aspect, but I want to start with the ASCO GI, if you don't mind. My understanding is the data coming out in Jan, ASCO GI has said that's a case control data, right? That's not a prospect or a real world. But I think Kevin, you're hinting that this could perhaps highlight what the new 2.0 test could look like. Maybe talk about why your case control data or trial is different versus other case controls, and why this should be a precursor of what perhaps we could see in your pivotal?

Well, first of all, the majority of those samples weren't case control samples; they were prospectively collected samples. So all of the advanced adenomas were collect -- the stool sample was collected before colonoscopy. So that is prospectively collected samples exactly like in the prospective BLUE-C study. The normals too were collected prospectively, so the vast majority of the samples -- the only samples that would be considered retrospective or case control are the cancers that were -- a stool sample was collected post or during the colonoscopy. So that data -- the powerful part of that data is we've set a specificity. Our goal is to be at or above 90%. And we feel pretty confident that we'll achieve that goal given these prospectively collected samples that we are testing them on and how we are setting our algorithm. And then the pre-cancer data, we use powerful machine learning capabilities, coupled with improved markers and improved chemistry to be able to improve that pre-cancer detection rate. All from the same sample collection device that we currently use with Cologuard 2.0.

That's helpful perspective, Kevin. Maybe on a couple of points out there which ties in with a lot of noise around competition. I thought you made a couple of points which are fascinating for me. In your prospective trial, 20% lower sample collection in blood. I think the market [ we have ] -- look, if you have a blood-based option, people want to give a tube of blood. Is there something about how your trial was designed why you're seeing 20% lower? Or is this a real consumer phenomenon? You might have confidently come in and the market has this completely wrong, this perception of what is easier.

The market isn't always maybe tuned into some of the finer points here. And all of us, I mean, this surprised us to some extent. But one thing you have to realize, when you deliver a screening test to somebody's home, and it's simple, and Cologuard has -- 7 million people have gotten a Cologuard test. So there's a buzz out there, hey, this thing is so easy to do. And you do it -- and you don't have to get stuck with the needle. I mean, I don't know about you, but my favorite thing to do in a day is not to get stuck with a needle. And the other thing is practical is that you need to get in your car and go get a blood test. And we all know that you don't go to your primary care physicians' offices typically to get a blood test. You actually have to get into your car and go someplace else or walk down the hallway or go into a different building. And all of that is -- represents a barrier to screening, whether it's a fear of needles, or whether it's taking an hour out of your busy day to go to some site to collect blood, and there's a 20% gap there. And so we think that's indicative of what the real world would be, not taking into account performance because we're not telling these people that a blood test is not likely to perform anywhere close to a stool test. So I think that's a realistic dynamic here. And you see it too when you talk to the large labs in the U.S. Only about 2/3 of people given a test order to go and get blood drawn actually follow through with it.

That's fascinating stats. On the second point on outperformance, Kevin, I think you've dropped tentative hints. I think, even on your recent 3Q call, you did give an explanation on why perhaps a blood-based approach might have a lower sensitivity than a stool-based approach. If you don't mind, I guess it's a 2-part question. One, recap why you think early cancer sensitivity, a stool-based approach would be better. And I think related to that was, I got this question from investors on sample size has increased. Does it mean there's something wrong with the test or something changed in the market, right? So maybe address both of those topics when it comes to sensitivity.

Yes. When I first joined Exact Sciences, the first people I met with were Dr. Dave Ahlquist from the Mayo Clinic, and Bert Vogelstein from Johns Hopkins. And they both told me -- and I came in saying, "We've got to get away from stool and move to blood. It's only logical." And they both sat me down and said, "Well, let's look at the biology and also the data." And the biology is that with A small -- a Stage 1 tumor may be 7 millimeters. It may be 1.5 centimeters. They're small, and they hyper-exfoliate cells into the colon. And so there is a sufficient amount of cells and DNA from those tumor cells available on a stool sample. However, they are not hyper-exfoliating into the blood supply. So there's just a tiny -- for those Stage 1 cancers, there's a tiny tumor fraction that is available for testing, and it may or may not be in 4 ml of blood. And you see that. So there's no doubt there's a detecting Stage 3 and 4 cancers from blood, you could do it today. And if that was the goal of screening, we would stop all these other types of tests. It's not the goal of screening to detect metastatic disease. Stage 2 is a little bit easier to detect because there is more circulating tumor DNA, but Stage 1 is the challenge. And it's especially a challenge if you collect samples prospectively versus case control. I don't need to go into all the reasons for that bias, but we've seen it time and again over the last 12 years.

And...

On your second question you asked, is there something wrong with the test. I would just remind you that in our BLUE-C study, we're enrolling patients and collecting stool and blood and banking those samples. We do not run the test now. Those patients then go on to get a colonoscopy. Once we have a sufficient number of cancers to power the studies, at that point, and only at that point will we run the test -- run the samples through our test.

Yes. In other words, it's colonoscopy that binds the cancers in our study, and then we label that patient as cancer versus no cancer or cancer versus precancerous polyp. So no, those cancers are found by colonoscopy.

Thanks for clarifying that because when I went through clinical trials at [ gFOBT ], it wasn't clear the trial flow. Why -- maybe if you could -- why are there fewer colon cancers? I mean, in one way, like you have to take a pat on your back, Kevin. Maybe Cologuard made a difference here. But I'm curious what's changed. Is this some pandemic impact?

No, I don't think so. I mean we know that there are going to be about 150,000 total colon cancer cases in this year total, and there will be about 50,000 deaths. So the numbers haven't changed. It's probably a function of during the pandemic, who you could get to enroll in clinical trials. They were the extra-compliant people, people who have probably been regularly screened, had colonoscopies, et cetera. And so with an extra focus on making sure that we get to populations who haven't been screened in the past. We have confidence that we'll be able to get to the appropriate number of cancers to be able to close the study. And the team has done a great job, but the pandemic has caused some changes in behavior. And whereas in DeeP-C, we were -- we saw a much higher incidence rate. We think that's the main reason, is the types of people who participated in the study.

That makes total sense. Maybe just to close the loop here on BLUE-C and Cologuard 2.0, we know specificity, it will be at 90% or something above 90%. Do we -- what is considered as significant, Kevin? Let's say, in January, we saw the numbers, right? Let's say it was 91% specificity, 92% or 95%. Is there a barrier or a hurdle which our clinicians would say, "Ooh, this is different versus 1.0." And versus competition, that early asymptomatic or, I guess, the precancerous polyps, right, how big of a deal is that from a sensitivity perspective? Because certainly, it's not part of the guidelines. So sometimes it becomes hard for us to put that into context.

Right. Well, Cologuard has an intended use of detecting both cancer and precancerous polyps today. At 42% detection of precancerous polyps though, it's -- we -- at the time that we launched Cologuard, we hope that, that would be over 50%. And we think that a significant improvement there is possible. And with the significant improvement, then it's really hard to put space between colonoscopy and Cologuard, as -- especially with Cologuard being done every 3 years. So -- and the dwell time of a precancerous polyp being 10 to 15 years. If you're able to get over 50%, and I'm not saying that we're going to be able to, but if you do, now Cologuard is an every 3-year test, with this automated compliance engine behind it, the only national colorectal cancer automated screening program that exists. It's awfully powerful. I think it would be hard for the FIT test to make -- anybody to argue that the FIT test is appropriate or a blood test -- let's say a blood test, the evidence is that there's only 50% to 70% Stage 1 detection. I mean those are the cancers that you simply can't afford to miss and let progress. So we think that Cologuard 2.0 is going to be the most important data that comes out next year. And that the launch from the time we have data to the time that becomes available, will be relatively short with probably a shorter FDA approval process than existed previously.

That's helpful perspective, Kevin. Maybe 1 more on the clinical topic before -- I did have some near-term questions for Jeff. But Jeff, I'm not letting you out, just so you know. But on MRD, Kevin, one of the questions we get is obviously with Illumina and GRAIL, some of the headline numbers are harder for us to -- how these trials are run is different. I mean, there's a lot of moving factors, right? But the headline numbers are 27% sensitivity. And the question I get asked is Exact is trying to combine 2 different methodologies, right, the Thrive method and the Exact method. What should performance be? Should Exact be -- whatever the industry standard is, right? Should it be at those levels? Maybe talk about your confidence on the NCD side of things?

Yes. I think the important question is how is a test that is highly specific and not tremendously sensitive, what impact would it have? Well, if you could test 100 million Americans and find 27% of cancers earlier, would you? And would you do that every couple of years, that over time, what you're doing is you're calling out and you're shifting the curve to earlier detection. Yes, you would do that, especially if the positive predictive value, the likelihood that a positive result is truly cancer, is north of, say, 40%. Well, that changes everything. And if you can bump that to north of 30% sensitive or even approach, someday, 50% of cancers detected by blood, this is a game changer. And I saw an article just a couple of days ago that critiqued the GRAIL approach. And I strongly disagree with that. The ability to detect early cancers, even if it's only 27%, which, by the way, that was the first version of this test, so we expect performance improvement to occur with all of the investments we're making. It's -- you'd want that test. You'd want that test for your parents. You'd want test for your friends and family members because it will shift cancer detection on a population scale and also for many individuals. And this is especially true going back to the salient factor that 70% of all prevalent cancers had no screening method. So I think, ultimately, this technology doesn't have to be 80%. It doesn't have to detect 80% of all cancers. And the key reason being is that 1% to 3% of people walking around over the age of 50 have cancer and they don't know it. Let's go and find at least a fraction of those people with the inexpensive, relatively inexpensive approach compared to treating them with late-stage metastatic cancer.

Just 1 more question on this topic, Kevin. The sensitivity of whatever the number is, right, 30 or 35 or 25, does it matter that we need those kinds of sensitivities in Stage 1, Stage 2? Or is that an overall sensitivity, including Stage 3, Stage 4 cancers. Maybe we'll start with that, and I had 1 more follow-up.

Yes. So in the DETECT-A study, again, which was with the first version of CancerSEEK, they detected 27% of all cancers. However, 2/3 of the cancers that have found were early stage. So that's powerful. And we think with improvements, we'll see better than 27% overall detection. But the important thing was there are about 100 cancers found in that study, half of them were found symptomatically, unfortunately. About 1/4 of them, 24, were found through current screening method. That's all -- that's colon, breast, cervical screening. We doubled that with this one test that detect 27% of cancers. That shows the power of this is the result of how many different types of cancers that are out there that we don't screen for. So I think that the evidence rule the day. Hopefully, Medicare will cover the test, and the guideline groups will say this makes a huge amount of sense.

That's a helpful perspective. I know I thought about 24 doubling up with this 1 test. But it's a good perspective. The asset lock on this NCD test in first half, did the timeline get pulled forward, Kevin? I thought timeline was a little way in the past, but it seems you have more certainty around the asset lock now.

There will be 2 case-control studies done next year with the prospective study starting after that. We are still collecting samples in the second case control study. So that's the 1 variable in locking down the final assay. We're on a good trajectory right now in terms of collecting that sample. So we're pretty confident with that -- with projecting that, that will be completed at or around the middle of the year.

Got you. Understood. And Jeff, maybe a couple for you just to close the discussion here. Your Q4 guidance of 265 to 270 of Cologuard revenues just with this new wave, Omicron, is there -- I mean, my understanding is a lot of Q4 was dependent upon how orders flew in -- came in, in 3Q. What are your current thoughts on the Cologuard numbers here for Q4?

Well, Vijay, as we said in our earnings call, we saw very strong momentum during October. If you look at the last week of October relative to the first week, orders grew 10%, and we were reaching record levels of orders. I said at a conference recently that, that strong momentum had continued into November. So we feel good about the momentum. Obviously, we're watching the latest variant closely. At this point in the quarter, we're getting close to the end of when orders this quarter affect this quarter. Typically, there's a 30-day lag between an order and revenue, the time it takes to ship the kit out, get the sample back and process it. So at this point, the -- if there is another big spike, obviously, we'll watch it closely and do everything we can to offset that. If there is a big spike, it would likely affect more next year and beyond.

That's helpful, Jeff. And just on -- because I know there are a lot of numbers flying around, but if I go back to like June, with the physician office visits were at 50% pre-pandemic level. I think it went down to down 70% or down 80% versus pre-pandemic sometime in 3Q. Where is that right now in terms of physician office visits versus pre-pandemic base.

If you look at the number of face-to-face details our sales reps are doing, it's about half of pre-pandemic levels. So it's improved a little bit. It's obviously better than it was in September. Recall in September, at that point, there was some disruption as the team from Pfizer transition. But we're a little bit better now, around 50%. I expect over the course of next year, that number to improve pretty nicely. But at the same time, wellness visits are also improving. This is the number of patients going to their physician for an annual health check, that number is improving as well. So all told, I think our -- probably our biggest growth driver for Cologuard and something that we're very excited about is that sales force that's energized, expanded. They're eager to get out there. That sales team is probably our biggest driver next year. In addition to other things like 3-year rescreening, Cologuard 45 and electronic ordering, there is a lot to be excited about next year for Cologuard.

[indiscernible]

Yes. As proud as we are of our R&D team, we're equally proud of our sales and marketing teams. And not a lot gets said about our marketing teams, but we've continued to increase the amount that we invest on TV, on digital and social. And it's having an impact. Really, Cologuard has become iconic out there. The box, people know about it, they chuckle about it. It's no longer this kind of weird thing to do. People do it in the privacy of their home and they freely talk about it. And it's funny people said, no, that if you go back 10 years ago, you're never going to get people to do a stool test at home. This really clever, creative and impactful marketing we think, next year is going to get even better. All I can say is wait and see, but we have some exciting things that we're working on to now really connect -- go from awareness to providing more of an impetus for people to call their doc and say, "I want to get tested. And I want to get tested with Cologuard."

Interesting. No, it certainly is. I think, one of the -- I think a key strength for us has been the commercial organization, which I think the addition of Pfizer sales force was -- it was massive. I mean, Jeff, maybe on the topic, how incremental is Pfizer, right? Because I think whatever numbers we had for Pfizer contribution in 2019, should 2022 be similar to 2019 levels? Or should that contribution be higher? I know the math is a little tricky because now it's one family. It's not Pfizer and Exact. But maybe any ballpark on how to think about what it means, this doubling of sales force?

Well, Vijay, we have really high expectations. This is a really high-quality team across the board. Now we have 1,000 people out in the field, educating physicians on Cologuard. That team is unmatched in diagnostics, and we have very high expectations. I expect the productivity to grow next year as sales force access improves, as wellness visits improves, as we add in things like electronic ordering, we have high expectations for this team next year.

Fantastic. And I think, Jeff, you've spoken about the addition of Pfizer sales force reps as being cash flow-neutral. Like what is the math behind saying it's cash flow-neutral or neutral impact?

Yes, the math is pretty straight forward, Vijay. So we expect even more productivity of this team now, focused entirely -- back out in the field focused on Cologuard. So we expect their productivity at selling Cologuard to be even greater next year, which more than offsets the incremental investment we're making in this team now that they're full-time employees. Recall that also our fees to Pfizer, now that this partnership has won down, our fees to Pfizer for promotional efforts will also decrease next year, which further improves the kind of the financial return next year in total.

Fantastic. I think with that, we're at the end of the time. Kevin, Jeff and Megan, thank you for the time this morning. It's exciting. Certainly, I think '22 is going to be a big year for you guys, and I wish you guys all the best.

Thanks, Vijay.

Thank you.