Exact Sciences Corporation (EXAS) Earnings Call Transcript & Summary

With that, the green light signals were live. Thank you all for joining the Exact Sciences presentation. We're lucky to have with us Kevin Conroy, CEO of Exact Sciences; and Jake Orville, the General Manager of the screening operation at Exact Sciences. I'm Dan Leonard. I cover diagnostics for Credit Suisse. Welcome, both of you.

Thanks, Dan, and it's really great to see you again.

Likewise, so we have a lot of ground to cover in 30 minutes here. But to start off, Kevin, you just reported Q3 results, I thought we could begin the discussion by circling back to the Q3 performance and highlighting anything you believe worth highlighting from those results to investors.

Well, and if I can just take a step back here, Dan, and talk about Exact Sciences, who we are and where we're going, Exact Sciences' mission is to help eradicate cancer through the earlier detection of cancer and better tools to help guide treatment. We have built this company. I've been a CEO for 13 years on a culture of performance, compliance, integrity and innovation. And when you look at the platform that we have built with a combination of Exact, Genomic Health, Cologuard and Oncotype DX, you have the 2 biggest brands in cancer diagnostics, both with great gross margin profiles, both with sustainable growth far into the future, and that powers what you saw in the third quarter. And there is something fundamentally important that we saw in the third quarter, which was top line 20% revenue growth, 25% Cologuard alone, with OpEx being flat. So we've been working on this quarter coming for years, multiple years to build the brand awareness of Cologuard and Oncotype DX to be able to continue to grow the way they're both growing, while we're now starting to -- able to hold our investments in sales and marketing constant. That didn't happen easily, and it didn't happen this year. It happened 3 or 4 years ago, the investments that we made in the awareness around Cologuard. And it continues today with the sales team and the marketing team and our customer experience capabilities, and Jake will be able to talk about that. So that's the key takeaway from the quarter is the idea of Exact Sciences as a platform company that many new tests that we develop internally and also probably tests that we either partner or acquire, get dropped into a 1,000-person sales force and amazing direct-to-consumer marketing and increasingly stronger digital tools. That's what that quarter just started to reveal, and we're excited about what that means for profitability and continued top line growth.

And Kevin, the key message there is that incremental revenue from this point forward should come at a very high incremental margin. Is that the path to hitting your 2023 EBITDA profitability targets?

We guided, Dan, to third quarter adjusted EBITDA breakeven. That is not a stretch. So it doesn't require heroics on the top line or heroics on operating expenses. We are still investing very aggressively in sales and marketing. And the customer experience required to delight the customers who are getting a Cologuard test, an Oncotype DX test, our hereditary testing, et cetera. This is just the engine is finally the financial engine is being seen now, and we expect that to continue into next year and beyond.

Is there maybe 1, 2, 3 levers you could highlight to folks that you really pulled to drive that operational efficiency? Or is it more the culmination of 100 different levers? How would you describe it?

Well, why don't I take it at a high level and ask Jake, who runs our screening business including Cologuard. It's really -- we came out of pandemic with exiting a partnership with Pfizer, and they laid off about 500 sales reps. We hired 400 of them. And then we brought our 2 teams together and resorted the territories, and there was some natural inefficiency during that process. Q2 was -- you had a combined team for 2 of the 3 months, and then Q3, it was a combined team for the whole period. And that had a positive impact. We also took control of our own marketing because we were doing all of our marketing buys through Pfizer. I think the efficiency we saw in the marketing line also was because now we have full control over this whole commercial engine on the Cologuard side. The Oncotype team has been doing an amazing job for years, and we know they'll continue to do that. But Jake may have some additional color to paint here.

Yes. I think when you think about 2 examples of primary growth drivers, you think about the age of 45 to 49. 19 million more people became eligible for colon screening and Cologuard. We can then be a lot more efficient at targeting our marketing campaigns towards that younger patient population. We also know that, that patient population is a lot more interested in a very convenient, highly accurate at-home test because of their younger age. And so we can, one, expand growth because it's a large population; and two, be a lot more specific and targeted. The second area is our rescreen opportunity. With the continued growth of Cologuard, the rescreens that we're seeing now from 3 years ago are at a higher level than before. And this patient population is very sticky. And we can reach out to them a lot more easier, more effective. It's a higher, more -- higher gross margin, if you will, for that patient population. And here, we can make things easy for them to get rescreened, whether it be digital options, whether it be telehealth options, whether it be actually getting their order in advance of when they're due and hold it in the queue and release it when their eligibility comes. Those are the things that don't just propel growth. We've now doubled our revenue in the rescreen population. It also makes us a lot more efficient, a lot more effective in how we get that growth, which is what you're seeing in these numbers.

Jake, you've anticipated my next question on the 45-plus opportunity and then also rescreening. I appreciate that you've doubled the opportunity from rescreening. Can you -- maybe talk about what inning we're in, in penetrating rescreening compared to your own expectations, where you think you can go from a penetration standpoint and how you get there?

Sure. We'll do about 220 million of rescreen revenue this year. That's up about 100%. But to be clear, 3 years from now, that should be about $500 million. That's a recurring revenue stream from what we build from that we'll be able to continue to tap into. And so that's the growth that's there. That's the recurring revenue, if you will, from these rescreen patients. We also know, Dan, that the experience is so sticky that we see a 20% increase in compliance or adherence in patients that have done Cologuard before. So they're very motivated, not just to get screened again, but also to get screened with Cologuard.

And then what about the 45 to 49 age group?

Very early there. 19 million Americans that are now eligible. We are winning in that segment, as you would expect, because of the patient demographics more interested in getting something like Cologuard. And so there, we're just scratching the surface. That will be a continued growth driver for the next few years until we saturate that younger patient population. But also when we get them at 45, now we've got, I think, it's 6x more value than we have had if we've gotten them at 65. And so there is where the rescreen shows back up again. We get them at a younger age. We offer them these very easy tools to continue to get screened and rescreened and then that builds to our rescreen patient base, that recurring revenue that I mentioned.

That was a 45-year-old has 6x more value than a 65-year-old.

Right.

Okay. Moving on to the pipeline a little bit. BLUE-C, your big clinical trial underway to augment your efforts and screening. Kevin, can you remind us what are the performance objectives of BLUE-C and what is the anticipated timing we should learn more?

So the BLUE-C study is a unique study in that it is a study of both the next generation of Cologuard, a stool DNA protein test and then also a blood-based colon cancer screening test. We have enrolled over 20,000 patients to date. We expect to complete enrolling patients next month. It takes typically 2 to 3 months for the blast people into the study to complete the study and have the Cologuard test, the blood test, the FIT test and a colonoscopy. So they actually get 4 different colon cancer screening tests. They'll be the most screened people on the planet. And then we will lock down and run our Cologuard next-generation version test, we call it Cologuard 2.0 internally, and also our colon cancer blood test. And then report out the top line results. Cancer sensitivity is a key endpoint. Pre-cancer or advanced adenoma sensitivity is a key endpoint and specificity is a key endpoint. And so those 2 products, we think, have different roles in colon cancer screening. We believe that we'll be able to lead in both of those categories. And over time, continue to have a huge impact in colon cancer. Just as an aside, when we ran the DeeP-C study, which was the pivotal study upon which Cologuard was approved, we started that study in 2012. It only took us 10,000 patients to get to the number of cancers needed to be able to power the study. Today, we needed more than 20,000 patients, which tells you something about the underlying prevalence of colon cancer. And what it tells us and many others is that with a decade gone by of screening and precancerous polyps found and removed, the prevalence of cancer is actually lower. And the only way that the prevalence of cancer goes lower in that population is if you're finding and removing precancerous polyps. If you only have tests that detect for -- detect cancer, you don't change the prevalence of cancer. You find it and then you treat it. But the way to move prevalence like the world did with cervical cancer, the only cancer that has been nearly eradicated because the Pap smear's target is not cervical cancer. It's the precursor lesions, which over time turn in to cervical cancer. And that's the power of colonoscopy. It's the power of Cologuard. It's even the power of the FIT test, which detects 1 in 4 pre-cancers. And so that has been really neat to see because we know what comes next if prevalence maybe hasn't been cut in half. If prevalence has been cut in half, what comes next is mortality. And that you're playing a role in that is incredibly gratifying.

Well, that's really neat. So precancer will be very important, precancer detection as an endpoint for BLUE-C. Can you remind us what your objective is, what you're aiming for from improvement points perspective?

Yes. So for Cologuard 2.0, the main goal actually is not to get worse in precancer or cancer detection, but to improve the -- to lower the false positive rate. So the false positive rate right now of Cologuard is 10%. So it's 90% of people with no polyps, no colonic findings whatsoever, there's a 10% false positive rate. We'd like to reduce that by 10%, 20%, even 30%. We would also like to see an increase in the precancer detection rate. And any improvement is really helpful in terms of generating greater life years gained. In terms of cancer detection, we already detect 94% of Stage 1 and 2 cancers with Cologuard. So it's a lot harder to move the needle there. That's probably within the error bars. You just don't know if we, of course, want to stay above 90% sensitive for cancer. Can you move that to 95% or 96%? That's probably more luck in statistics than anything.

Sure, especially with lower prevalence.

Yes. Well, with the cancers that you have in the population, I don't -- prevalence probably won't impact. What impacts the sensitivity more is the number of cancers you enrolled in the study, and we already have more cancers enrolled in the study than we did for DeeP-C. So that narrows the error bars and increases the likelihood that the test performance we see is the test performance that we expect. One thing we know is that Cologuard 2.0 performs better than Cologuard 1.0 because we've studied it enough to know that. So we're confident going into this study that, at a minimum, that specificity is going to be improved and most likely precancer detection as well.

And then from a timing perspective, you said that the study should finish enrollment by year-end, 3 to 4 months to complete all of the work up after patients finishing enrollment. And then remind me, the stool test and the blood tests are on different timelines, correct? There could be a difference in timing and the readout for the 2 of them?

Yes. We continued the enrollment of the study to better power the blood colon cancer test. And so interestingly, this is actually fascinating. There's been this big debate is a blood test going to replace, is going to be more patient-friendly and widely used than a stool test? Well, in our DeeP-C study where you're paying a patient to participate in the study, 20% of them, 20% of those people willing to do a stool test just won't go get a blood draw. They're either afraid they don't like getting blood drawn or it's too time consuming to go get a blood draw. And that's something that we knew going into this. So we have actually had to over-enroll in the study to make up for this 20% gap. That's okay, but that's the reason that we extended this out. So we will ultimately -- we would like to see 100 cancers or thereabout 90 to 100 cancers in the blood test, in part because the performance of the blood test is, as we know, is going to be worse than Cologuard. And it's -- you're comparing both of these tests against the FIT test. So you want to be at least as good as the FIT test, which means for blood test, you need more cancers. I know it's a long explanation. But these are the important considerations when you decide when to end the study.

So they are on similar timelines then, because the studies being driven -- concluding the study is being driven by blood enrollment?

That's exactly right, Dan. And we will then do the testing in probably Q2. So after we lock down, do all the quality checks on all of the data that we have in the study, probably test first stool and then blood. We're going to separate those 2.

That's more a decision on your part as something driven by the requirements of the trial?

That's correct.

Okay. And are both assays locked down at this point? Are both classifiers locked down? Because I know at least the blood, you had IDA at that times around what's the right technology to deploy that blood test on. I'm not sure where we stand on that, if that's...

So Cologuard 2.0, the stool test is totally locked down. The blood test, we recently have seen data that, at some point, we will likely release for a different approach to blood-based testing where we've seen even better performance than the performance we've already had, which was pretty good. And so we are looking at the possibility of modifying the test, which had been locked down, and now we're looking at potentially adding another marker class that would further improve our performance. So those are the typical things you consider. With Cologuard, we locked down -- we finished the trial, we finished development of the automation platform and lock down the assay within each about a week of each other, and it all came together at once, and we test it. You test all these samples at the same time, not everybody takes that same approach, but we'll test them on the automation platform that we intend to use because that platform becomes part of the PMA. And anyway, we brought all of that together at the same time way back in 2013, which led to ultimately to -- or 2014, which led to approval on that whole platform. A lot goes into these, into the product development. A lot goes into then commercializing. And we can't wait for the day when the debate is who has the best colon cancer screening sales team? Who has the best colon cancer screening marketing team? Who has the best customer experience for tens of millions of people. I think that our -- the last 8 years of investing in all of those things puts us in a leadership position, and we won't have to keep investing the same way that we had in the past on a relative percentage of revenue. This is just -- this team is ready to go with Cologuard 2.0, with colon cancer blood and beyond, multi-cancer testing, et cetera.

That's actually a great segue into what I want to talk about next, which is multi cancer. And you mentioned multi cancer, but I was thinking the segue would be that you're adding another marker to your CRC blood test. And one of the takeaways I had at the ESMO poster because you added another marker class to the multi-cancer test. So please talk to us about your approach for multi-cancer detection, how it's different? What gets you excited about your data? And what are the timelines we got to look forward to on that development path?

Yes. If you go back years ago, we developed a partnership with Mayo Clinic, and we early on, we were focused on colon cancer, but behind the scenes, the real passion that the Mayo Clinic team had was identifying very specific DNA methylation sites for each of 16 different cancers. And it took a decade, but we went and curated the specific locations in DNA for each major top 16 cancer killers and identifying where that DNA was methylated, in tumor versus no tumor. Then fast forward, we acquired Thrive, which was based upon Vogelstein's work, which curated DNA mutations, cancer versus no cancer. And when we put these -- and the theory was when you put these 2 technologies together, DNA methylation and DNA mutation are different pathways, and you're going to increase your yield in finding cancers. And we didn't know what we would find until we got the data that was shown in the ESMO paper. And what that showed was DNA mutations added about a 40% bump increase over methylation for stage 1 and 2 cancer detection. It was huge. I know it may not sound like a lot. But that was a big, big, big deal. And we didn't know what would happen. It was a big bet that, that is going to pay off, and it's not only going to pay off in multi-cancer, it's going to pay off in any cancer that you want to test in blood. That's fundamental intellectual property that we have. It took a long time for Dr. Vogelstein and Dr. Ahlquist from their respective institutions to do that work. You can't replicate it overnight. And that puts us in a position going back to our mission of cancer eradication through earlier detection. We now have the tools that nobody else has to do that. And the problem -- tell me to stop if this isn't interesting.

Keep going, please.

When you look at patterns of DNA methylation, which is an approach that people in the field take, and you're looking at maybe 5,000 to 20,000 different sites, which add up to some type of signal, and then you use machine learning on that, those are all attributes which are variable patient-to-patient cancer to cancer. And you can use machine learning to help say this signal's cancer, this signal's no cancer. The problem is when you're looking at 20,000 variables or even 5,000 variables, machine learning doesn't work so well because you -- the general rule of thumb is you need 10 cancers for every variable, for machine learning to really hone in on the right signal. I've often described this, if you took 1,000 marbles and rolled them across the stage and you saw 10 of them clumped together, say, Dan, pick up those 10, mark them with a little white paint mark and put them back into the bag, shake them up and roll them on the floor. Do you see 10 together again? Yes, of course, you do. They're right there. But are they the same 10? No. So the problem, it's just statistically unlikely that you're going to see no signal. You need a huge number of data points for machine learning to really add a lot of value. And that's we're taking the curated approach where you know the biological thing you're finding is cancer is different than looking at patterns. And again, this was a big bet when we decided to go by Thrive and the IP that they had and -- I mean, Bert Vogelstein and his team are working every day this week on this project. And it's very different. What we're finding is clearly cancer. These are cancer driver mutations. So this applies to multi-cancer early detection because what you want is a test that with a very low false positive rate, you go into a population and you find 30%, 40%, 50% of all cancers. That changes everything. There's no cancer therapy in the world that is better than that test, that is more impactful to human health in that test. Because what -- if you shift cancer detection from, let's say, on average stage 3 to on average early stage 2, it's a game changer in terms of outcomes. That's why colon cancer has been so impactful screening. That's why cervical cancer screenings that shift to early detection is massive. So imagine if you get a test that every 2 or 3 years, you get a blood draw, and no, it doesn't find all the cancers. But it finds a big percentage of them, 40% of them over time in a population. You're doing a big shift to the left in terms of stage. That's what we're seeking to do with our multi-cancer early detection test. And we're really excited where that leads.

If I could interject, is 40% the right number? Is 30% the right number? I get this question from folks, what is the sensitivity for a multi-cancer test need to be to drive broad adoption? What is your thinking on that?

Well, if you take a step back and look at societally, part of it just depends upon the cost of the screening because it's things aren't done in a vacuum. But what we spend hundreds of billions of dollars on cancer therapies. What is the right cost to do a stage ship? Most early cancers are treated with either surgery alone or surgery and conventional chemotherapy. It's the late-stage disease that you're spending a ton of money on. So it really depends. Is 30% good enough? 30% will change everything. 40% changes it even more. And 50%, cancer will never be the same again. So it's a function of cost. It's a function of how frequently you're doing the test. Over time, that yield increases and mortality decreases.

There's always been a mismatch between the amount of money spent on therapeutics and the amount of money spent on the diagnostics.

Yes. Well, I think that is changing. I give GRAIL so much credit for the work that they did to bring and they continue to do to bring attention to the space and the work that they and we are doing in D.C. to bring attention to the legislation that will help unlock this space. A lot of people years ago said, come on, a stool-based colon cancer screening test, you're just not going to get that many people to do it. But it was over and over and over. And here we are this month, we'll hit our 10 millionth person screened. Screening is where it's at. It truly changes outcomes. And I think that the next decade of cancer impact is going to be seen in stool and blood-based multi-cancer screening.

We only have a minute left here, so I'll wrap it up on something I was trying to discuss earlier on multi-cancer, which is the number of markers you're looking at. You mentioned DNA methylation, mutation, Thrive also threw some proteins in there for good measure. And I think in your most recent poster there was maybe chromosomal abnormalities that you included in the signature. And to me, that all just sounds very expensive. So tell me how you're going to get all this within a cost envelope that the market will bear.

It's a combination of, interestingly, a PCR assay and next-generation sequencing assay. It's not overly expensive when you do it at scale and considering the impact. So I think over time, you can get that down to a reasonable -- I don't want to provide guidance, but it's hundreds, not thousands in terms of cost of goods. And you can have a healthy margin even with that kind of full range of different marker classes. So that's DNA methylation mutation, protein are the 3 big drivers.

Got it. Well, with that, we're out of time. Kevin, Jake, thank you so much for being here with us today. The passion clearly comes through.

Thanks, Dan. Great to see you.