Adaptive Biotechnologies Corporation (ADPT) Earnings Call Transcript & Summary

All right. Good morning, everybody. It's our pleasure to welcome Chad Robins from Adaptive Biotechnologies to the Cowen Conference. Thanks for being here, Chad. Adaptive's immune profiling platform is well-validated and its product pipeline continues to advance. Chad and his team have successfully demonstrated their regulatory and reimbursement prowess via clonoSEQ and also via many clonoSEQ and immunoSEQ-related partnerships. Over the next 30 minutes or so, we're going to focus on the outlook for 2020, progress across the base business as well as pipeline developments and the associated time lines. So again, thanks, Chad, for joining us.

Thanks for having me here, Doug.

I wanted to spend a few minutes just going back to basics with Adaptive. So maybe to try to set this up. Immune profiling is very exciting, and I think that's as obvious to this group as anyone. It's also really complicated. And by extension, I think many view Adaptive as occasionally, both of these things, exciting and complicated. To try to simplify, would it be fair to describe Adaptive, and this is just coming from an analyst who likes playing with Excel models a lot. Would it be fair to break Adaptive down to really 4 things: one, a company that's built to unique database that associates specific T and B cells with different diseases; two, a platform technology that you've developed called immunoSEQ, that really builds off of that knowledge; three, advancement made by essentially augmenting your knowledge base via the partnership with Microsoft and others; and fourth, a technology that's yielded specific diagnostic signatures that help detects minimal residual disease in blood cancers, which you branded as clonoSEQ. I mean is that a fair way of describing Adaptive?

Yes. I think there's a lot of elements in there that are certainly fair and indicative of what we do. But I think if we bring it up one level, it would be helpful to start with, which is to say that most companies are focused on the biology of disease, and they're looking at one disease at a time. What Adaptive has done is, we've built an immune medicine platform that's capable at looking at how our bodies respond to disease or what we call the host response. And so that platform is comprised of kind of really 4 different components: the ability to both sequence, which is the core of the platform, map your receptors and then characterize receptors. There's some other nuances within there, but that's essentially what we do. And then you can map the different product lines onto that -- those basic concepts. So just in terms of using that platform then to generate this rich data set is something that we've been focused on for a very long time. And we're starting now to exploit those -- that data into developing clinical applications, which will have revenue streams associated with them.

And the way you drive those revenue streams is diagnostics, pharmaceutical partnerships and actually developing your own pharmaceutical products?

Yes. Well, on the third point, we look at ourselves as more of an enabler of drug discovery at this point as opposed to bringing in-house pharmaceutical product. But certainly, from a diagnostic standpoint, developing our own products. We also have products in the life science research space that basically uses this powerful technology as a tool for research, and to the biopharma community and to principal investigators and the academic community as well.

So there's ways to monetize the database via diagnostics and pharmaceutical targets, principally, for partners as you just described. And one of the things that we've looked at in the space, and this is true of Adaptive, it's true of companies we used to cover like Foundation Medicine, other companies we cover today like Guardant, just name another one. It's really the database value. When it comes to profiling a sample, profiling a patient, how much -- I guess, what I'm trying to get at is how much value we should put on the Adaptive database, and essentially, the question is, how much are you able to learn from every sample that you profile and are you entitled to that data?

Yes. So this is a great question. And I'm reading through your past reports, I think you were the first one to try to put a -- to quantify data on Foundation's database. We -- so, first of all, we have access to right now 37 billion immune receptor sequences in our database. This is probably most applicable as we look at our partnership with Microsoft. And what we're doing with Microsoft, is essentially, we're mapping disease by disease, the relationship between your T cell receptors or your immune repertoire and the clinically relevant antigens for each one of those diseases. Basically, we're mapping -- so each disease has its own essentially genome and we're mapping the receptors to that genome. We're able to -- essentially, how we do that, we can actually start generating that map in vitro. We can expand massive pools of T cell receptor by antigen to start developing that. Then Microsoft applies a machine learning model, a series of algorithms that essentially creates more connections between our immune receptors and these antigens. But on top of that, when we find these connections, we then can go back into our database and find those receptors that are in the database associated with people, and then pull additional receptors from that set to improve the diagnostic. It's called bootstrapping or machine learning, it's called pseudo-labeling. So there's actually a real value to the data. So now in terms of the ability to quantify what exactly that value is, one good example is, as you guys know, that we released kind of last week in the earnings, is how we were able to, in a 9-month period, develop a test for Lyme disease. It was 2x more sensitive with very little false positives and false negatives coming out of the test. So the long answer to your question is, the data has a tremendous amount of value. I think we and the community is working on ways that you can truly quantify that. And some of that is, okay, time to next product is one. And then the second thing is, we -- for just receptors that are in our database, we will be able to, from a research radical, go back and offer the ability to annotate those receptors and say, "These are what these immunoreceptors bind to in the future." And there's going to be a value stream that's associated with that coming out as well that you really -- you're not doing any additional development for us. So interesting to think about how that will be valued.

That makes sense on the Microsoft and immunoSEQ front. Can you learn from even clonoSEQ samples?

ClonoSEQ is a different -- so everything I said about the kind of immune system being the host response. It's actually a little bit different for clonoSEQ. So clonoSEQ is our test for a minimum residual disease in certain, what we call lymphoproliferative blood cancers. What this means is that it's actually cancer of the immune system cell. So it's a cancer of that T or B cell. And that one -- with that one B -- most of these are B cell drive. That one B cell winds up kind of proliferating. And with it, it carries the immune receptor on its tag. And we're able to tag it at diagnosis, and then after the patient gets treated, myeloablated, they're trying to knock down that clone, we're able to then find that pre-identified sequence in relation to a patient's total other cells, very accurately, very quantitatively. So you're not -- it's not the same type of learning. This is -- at its essence, we are counting cancer cells, which is also very different than like a -- other companies in liquid biopsies, like take a Guardant or Foundation or whoever that is trying to look at stuff that's shed from the tumor. We're not looking at anything that's being shed, we're looking actually at the cancer itself and team will account it very accurately. So it's -- I would say, a very -- it's a different application of learning.

Okay. Super helpful. Let's talk a little bit about 2020. So last week, you said 2020 revenue guidance at $114 million to $119 million. You laid out your vision for the year in terms of key milestones. So just correct me if I'm messing any of these up. But the plan is to launch clonoSEQ for blood-based CLL, post-FDA approval likely in June. The plan is to file clonoSEQ for blood-based ALL in multiple myeloma, FDA clearance this year.

ALL, yes. Multiple myeloma, we're waiting on data there, so we'll see.

We'll see what happens. Okay. Plan is to launch a new immunoSEQ research use-only kit.

Yes, correct.

Submit immunoSEQ Dx for Lyme disease for FDA clearance.

End of the year. Yes, sure.

End of the year. And then the goal is to aid Genentech in its IND submission for the first share of T cell therapy discovered under your partnership. So those are the key milestones for this year.

Yes.

And -- so that's a pretty good list. I think the thing we hear from investors, and it's not a critique. It's just a lot of these don't lead to 2020 revenue. It's more about 2021 and 2022, I believe. So if -- I guess, really my question is, if you succeed with most or all of these targets, does this position you for a further inflection in the revenue trajectory as we look to 2021 and 2022?

Absolutely. I mean it's a huge setup here. I -- the one counter I would have is, right, I mean, you're growing -- if you look at 2019 to 2020, growing revenue at 37%. But it's not...

No. But it's not so...

But you're also kind of backing out the fact that, that development revenue is essentially, you're not getting much growth in that because you're straight lining the amortization of the Genentech upfront. So your sequencing is actually growing at a significant higher clip than development because you're starting at kind of 50-50 going to a 60-40 split in sequencing. The other thing I would point out, as you and other analysts and investors had modeled in that we would get the Genentech milestone in 2020, where those numbers exclude the milestone. So essentially, we came in -- we're coming in, essentially, like $10 million higher in terms of our forecast, right? So if you think about it that way. But that being said, we have said since the IPO in June last year, especially with respect to the clonoSEQ business, that your changing medical practice is hard. We're doing all the right things to set it up in terms of account activation, working at different departments, working the investigators within departments we've got this land and expand strategy of where -- where our -- where the clinical community is getting utility out of using the assay for MRD. In addition, doing that peer-to-peer education, continuing to develop additional data sets that show that -- that they really demonstrate where clonoSEQ can be used throughout the care continuum. And then you'll see a concerted effort, both in branded and unbranded marketing, in particular, to the -- directly to the patient that should have really a stake in making care decisions with our doctors. So all that stuff is going on. We're executing against it. And yes. And I'll say it's an ambitious set of goals and we clearly intend to hit them. I mean that's -- we've looked to it in the eye and put -- we're ambitious and said, and if we do that, we are set up really well.

But all ambitious, but clearly, doable?

Doable. Yes.

Yes. Okay. Let's talk a little bit about clonoSEQ in a bit more detail. So I think this is viewed as about a $4 billion to $5 billion market opportunity in terms of what you're going after. Roughly 4, 4.5 million patients that you're targeting. Currently, clonoSEQ's covered by CMS for blood-based CLL and bone marrow-based ALL and MM, multiple myeloma, indications. ClonoSEQ also received FDA approval for bone marrow-based ALL and multiple myeloma indications as well as New York CLEP approval for all blood cancers recently. You expect FDA approval for blood-based CLL, I think in June, as we talked about and Adaptive plans to file a blood-based ALL in multiple myeloma. Well, we just talked about that ALL by end of the year, multiple myeloma, we'll see on the data. So clonoSEQ's been integrated into 150 accounts out of the, I think, you're targeting 250 Tier 1 and Tier 2 accounts. How much -- if you just stopped at the 150, how far -- how well are you positioned to address the addressed market opportunity?

Yes. So just to be clear, the...

Yes. Please correct me if I don't get this correct.

Yes. No, No, No. I just -- let's get on the same page. The $4.5 billion market opportunity is a global addressable market opportunity. Where we're starting first is attacking the U.S.-based market, which is about $1.2 billion to $1.3 billion market opportunity. If you look at how -- just looking at overall how the trajectory and where that market opportunity is available to us, about 25% of it is available through ALL and multiple myeloma. An additional 25% kind of rolls on with multiple -- sorry, with CLL. So that gets you up to 50%. And then the remaining 50% is in this basket of NHL diseases, which comprises follicular, diffuse large B cell, mantle cell, et cetera. So that's really how you look at it. So if you look at kind of the 150 or to the 250 tiered accounts, you do some multiplications. But it gives us essentially access to about -- over time, about $500 million in addressable market opportunity with ALL, multiple myeloma and CLL.

And in terms of just driving market penetration, what are the -- I mean, beyond, obviously, just looking at revenue. What are the other metrics that you look at that demonstrate you're in a position to actually get after this market. Is it market penetration as I started to ask my question?

Yes. Yes. Yes.

I mean account penetration? Or is it something else?

Yes. Yes. So literally, if we look at it, it's kind of as a waterfall, where, first, we look at how many accounts of our targeted Tier 1 and Tier 2 accounts can we penetrate? And then within that account, there are departments. How many of the departments within that account have we penetrated? Then each department has a certain amount of doctors that are associated with treating that indication. Let's say, there's 8 multiple myeloma docs that are treating that patient. Okay. Then how many -- have they found a use case for every patient? So one of the big pushes is to get the identification sample or the fingerprint of the cancer for every patient. So that's a big use case. And then the first MRD, where they're using it, is it post induction, kind of pre-transplant? Is it looking post-transplant? Is it a confirmation of CRO by a traditional method? Is it a maintenance therapy decision, where you can either increase your maintenance therapy or take a patient off maintenance therapy based on the MRD status? And it's using it once and then getting them to say, "Hey, this data is so critical to my thought process of how I'm going to treat the patient that we're going to use it at every point or multiple points throughout this patient's care continuum." So if you multiply kind of all those things together, it's the number of times for patient, it's the number of patients, it's the number of doctors, it's the number of accounts, then you start to kind of build your model as to how this rolls out. And just to be clear, I mean, we firmly believe, and we now have 30-plus publications on this matter that clonoSEQ will be the gold standard. I mean this is how these types of blood cancers will be treated because it is more accurate and more sensitive and you're simply counting cancer cells. It's like using digital photography as opposed to analog. And you're starting to see that switch over happen. We're starting to get to that inflection point, and when we do, and again, that's why we've been very careful about what needs to happen to change medical practice, where you actually have this inflection point and why we're doing all this kind of peer-to-peer education. Because even if you've got the best tech out there, if you -- I was reading something like of -- there were a lot of doctors making hundreds, that said, "I don't need a thermometer." Essentially, this is a thermometer, if you will, of your blood of how many cancer cells you have. In the 1880s, this -- Paul Gaynon from Children's Hospital, L.A. actually was talking to me, he's like, "In the 1880s, like these doctors were saying, hey, we don't need a thermometer to treat. We can just feel, feel the patient's head and know exactly what's going on." And that's kind of exactly people like, why wouldn't you want to know exactly how many cancers cells are out there?

Right. Get to the level of percentage.

And that's also the push to patients is, "Hey, know your status, right? And what's my number, right? What's my clonoSEQ number?" So these are kind of stuff that you'll see in some variation of this rolling out towards the third quarter -- third and fourth quarter of this year.

In terms of the move from blood marrow -- I'm sorry, bone marrow to blood, how do you -- do you think that there's going to have to be some compromise when it comes to sensitivity? Or do you think that it is going to be pretty consistent with what we've seen thus far?

So I think it depends on disease state. I think in CLL, no. I think the sensitivity in concordance is there. ALL, same thing. I mean, frankly, we would have gotten approved in ALL, it's straight to the blood, we would never even done bone marrow. It's just -- there was no comparator. So we had do that. There are others -- like most of the myelomas are a little bit tricky in the sense that it's a disease of the bone marrow. And we're doing a lot of work on it right now. But what I will tell you is our technology is 1,000x more sensitive than flow, so -- in that bone marrow. So we are actually going to be more sensitive in the blood than flow is in the bone marrow.

So may not be as good as you are in bone marrow. But it could be, in one way or the other, it's going to be better than existing flow.

Correct. So it might be -- and again, the pathway might be that you -- basically, if you can pick it up in the blood, you can follow in the blood. So might be, you take the Ig test in the blood, you can't find it, then you go to the bone marrow. So it might be kind of like a Cologuard, right? Like you can't find it in the Cologuard screen, then you go to colonoscopy. That might be the pathway that -- one of the things that we're looking at from myeloma. For everything else, we're truly looking at blood-based testing. And just to be clear about blood-based testing, I think it removes a barrier. It doesn't -- if you're not on MRD yet, it's not going to all of a sudden, say, "Okay, I'm testing every patient for MRD." But it certainly removes a barrier to entry. It allows you to start trying.

Very quick on NHL, non-Hodgkin's lymphoma, is -- what's the rate-limiting step? Is it really at this point just making sure you're confident with the right TCR signature?

No. No. I wouldn't say that. I think -- rate-limiting step is we've got data trial readouts that are coming. And we are quite confident that we'll be able to -- from what we've seen so far, it's going to work really well in a whole variety of NHLs. There's not -- there's some tech dev that needs to be, whether we pick it up in the blood or a cell-free DNA, we've got a little bit of tech dev on cell-free DNA to do. But there's no real rate limit on how this is going to work. It's going to be a standard test across NHL. We have no doubts about that.

Okay. Just last question on the clonoSEQ section. Biopharma, could you just provide a little color on how having biopharma companies as partners impacts clonoSEQ revenue in terms of ASPs and milestones over time?

Yes. I would -- let's -- I'll address ASPs in a second.

Yes. Do you want to frame it first?

Yes. So companies that are trying to get a drug approved in an indication that our technology applies to, we're the only FDA-approved test and we're now a trusted partner. So almost all of them have now come to clonoSEQ and to Adaptive and say, "Hey, we want to incorporate clonoSEQ." Now the FDA is looking at whether clonoSEQ MRD status in general can be incorporated as a secondary end point and hopefully moving, at least, in myeloma to a primary end point where you can get a drug approved based on a patient's MRD status, which would expedite the approval of the drugs. Now this does provide revenue. How we actually characterize our revenue, we would consider this as part of our kind of research revenue. Not -- we -- unlike, for example, we don't put that in the clonoSEQ revenue. But it does provide, in many cases, these deals have a set of upfronts, a set of milestones associated with approvals. And then -- and that's how we would get revenue. In terms of ASPs, it doesn't really affect the ASPs directly.

Because of what you see. Right.

Because it's not a clinical -- these aren't clinical tests. These are research tests that we run out of trial. I'll give you an indirect method. I mean if you have an investigator that's running a pharma trial that's using clonoSEQ, they're much more likely, when that trial ends, to say, "Hey, I had a really good experience. Why wouldn't I continue using it in my clinical patients?" And then use it and then -- so indirectly, you could affect ASPs. But it's not a direct -- hey, Pharmaco is using it, therefore, ASPs go up. It doesn't work that way.

And you guys have mentioned you're eligible for -- I think it's around $130 million in milestone payments from biopharma. Is that all clonoSEQ? Or is that more broad?

Yes. That's mostly, mostly clonoSEQ. There are maybe some translational milestones in there. But that's non-Genentech milestones, clonoSEQ. And frankly, that we haven't modeled any of that into...

That's not in the guidance, yes.

Into 2020 because a lot of it's going to be brought into play when -- with FDA approvals of allowing for MRD status to be used as end points in trials, which hasn't happened yet, but we do think it will. So we think that those -- that cash is available to us over time over the next 5 years, but it's not necessarily going to hit this year.

Okay. Genentech, just to close on that. So you recently announced the partnership with Genentech on CLL, where MRD is being used as a primary end point in their study. Is this deal structured similar to your other biopharma deal? And then what are the key milestones here?

Actually, this one is not structured because there's -- we have some -- I can't go too much into it, but we went up kind of taking up some of the -- and we've done this a lot in the past. But this goes to your first question on value of data. We have some really intriguing opportunities to access and share the data coming off of that, that we thought was more important than near-term revenue. This one is -- that one is a little bit different. And we've got such a deep, as you'll see -- continue to see over time, we have such a deep relationship with Genentech and Roche folks as a development and what we're looking at doing in a variety of different areas, cell therapy and other areas, and so.

ImmunoSEQ, we talked about it a little bit earlier.

Yes. Dx.

Well, let's just focus at the core platform. So the platform is used by, what, over 2,200 researchers, over 165 biopharmaceutical partners at last count. It's been used in, what, over 600 clinical studies. How -- these are big numbers. It's impressive. How do you frame the market opportunity? How do you think about the market opportunity for immunoSEQ? Just the platform itself?

Yes. Yes. So we look at the total addressable market opportunity of just kind of the core, using immunosequencing for profiling in -- of many diseases is about $1 billion. Most of that's comprised by the biopharma community, about $850 million of the $1 billion is really biopharma getting on trials, having them do research. $150 million, we believe, is the academic market. And again, just to frame it, we, in the past before we did the Microsoft deal, a lot of that -- we did a lot of trade-offs by saying, "Hey, let's get data access. Does the 37 billion sequences that we have access to, let's get data access and trade-off some revenues to be able to get that data access." Now going forward, because a lot of -- so that was our incubation engine for products. But now our true diagnostic product portfolio pipeline is coming off this -- the map that we're building with Microsoft. So we're not as -- so we're -- long story short, we can be more focused on saying, "Hey, let's get this out of the community as a true tool. Let's put some marketing behind it. Let's put some sales and development behind it to try to go sell this product." Because it's the richness and kind of the speed to which we can do our products from that versus what we can do now with all the machine learning and the connections between your immune system and the relevant antigens is going to be, we think, a more valuable incubation engine for product.

When you think about what you just outlined in terms of the market opportunity and also where the demand has been from customers and partners, is it mostly cancer, mostly oncology?

Oh, yes.

And is that evolving a little bit beyond cancer and oncology?

Yes. It has been mostly oncology. And -- but I'm not sure that's necessarily the market as much as it is. It's just, practically, we made some hires who had great relationships with oncology. So naturally, they sold the product within oncology from a biz dev and got in all the trials. Now however, we brought in and have a focused effort to look at areas such as allergy and infection, autoimmune disorders, infectious diseases, neurology, like all different areas that are immune-mediated. And so we're going to go kind of go prosecute this technology into areas outside of oncology as a tool.

Makes sense. All right. Let's talk for a couple of minutes about immunoSEQ Dx. So you've generated some encouraging data on immunoSEQ Dx for Lyme. You intend to conduct clinical validation studies this summer and the data is going to be used for, hopefully, an FDA filing by the end of the year. You've also generated some early data on Celiac disease and early detection of ovarian cancer. And you have an ongoing R&D program with Microsoft for additional disease indications. In Lyme, you kind of referenced this earlier, but just to be clear. How do you think the sensitivity of the test compares to ELISA-based test for Lyme today?

I think it's 2x more sensitive already. And I think it's -- the data is getting better because of that concept I explained. This is a self-improving diagnostic. The more data we have, the stronger the test gets because you're adding T cell receptors that enhance the signature to Lyme disease. And so I can't -- it's hard to just maybe taking our hats off for a minute. This is a theory, right? This is -- and we had some data in a disease state that no one really cared about other than as a proof-of-concept in CMP, right, that we published in nature. And then went to Microsoft and Google and whatever. We want to partner with Microsoft and said, "Hey, we think we can essentially link our body's immune response to the antigens they bind to for the disease and then use that to diagnose disease " So this was a theory, even right up to our IPO. And we think, in a 9-month period of time, we're talking about having what we believe is a definitive diagnostic in a disease with an unmet medical need that we also believe we're doing a significant amount of market assessment with respect to kind of pricing and application of the test. But if you're talking about kind of the development time frame and again, as we mentioned, it's aggressive, but we're talking about last June to -- we're trying to have a product on market by the end of 2021. So if you -- going back to your question on the value of the data, if the infrastructure and leverage that we're going to be able to get out of the system, if we can generate signals in these and then kind of leverage that infrastructure to stamp out these diagnostics, the idea just to frame it, is to go one by one, to start with, to say, "Okay, we're going to have unmet medical need where there isn't a diagnostic or there's a highly invasive diagnostic, where we know that earlier intervention is going to lead to a better outcome." And so that's step 1. We can do -- you mentioned Celiac, Lyme, et cetera. But step 2 is really this differential diagnoses approach, where a patient comes in with a set of symptoms. I mentioned CLL, actually, let's keep it in GI. And as patient comes in with stomach issues, and we can definitively tell them, "Okay, you have Crohn's, ulcerative colitis or Celiac disease." Now that's real value because the diagnostic odyssey that these patients are going on is really superior. And then the final component is to say, our real goal is to have what we'll call population immunomics, where you can walk in and we have hundreds of diseases mapped. And we can take one blood test and get the results and what you're diagnosed for across disease states. And so we're -- right now, we're able to work at 5 of these at a time and get the engine moving to turn these into products.

That's great. All right. Lots more to talk about, unfortunately, we're out of time. So we'll -- I think we're doing a breakout. If not, we'll catch up later.

Yes. Cool. Thank you.

But one way or the other, thanks for taking the time to -- we are doing a breakout around the corner in the Nantucket suite.