Okay. Good morning. We're going to kick it off. I'm Tycho Peterson from the Life Science team. We're going to kick it off this morning with GRAIL. Let me turn it over to Aaron and Harpal to do a quick presentation, and then if we have time, we'll do some Q&A.
Yeah, we'll do a quick intro. That's our presentation right there. Yeah. Thanks, Tycho Peterson. Jefferies, thanks for having us here. Really excited to be here. GRAIL, find cancer early, when it can be detected, when it can be cured. We just launched, we just announced all of our NHS-Galleri data at ASCO. Had a really good reception there. You'll hear a lot from Harpal about that. We had a great quarter. We had 50% volume growth. We had 37% revenue growth. We've now ran over 500,000 commercial tests. We're really penetrating deeply into our physician offices, brick-and-mortar offices, and we're enabling further awareness through things like Epic, which we'll be launching, Quest integrations, and so on. We're really excited to be here, and we're really excited about where we're at right now. Harpal, do you want to-
No. I'm sure we've got questions that we'll get into detail.
Yeah. We'll spend a fair amount of time on the ASCO stuff. Maybe just before we do, you mentioned the 1Q numbers. Anything kind of coming out of 1Q? Volume's up 50%, obviously. How should we be thinking about momentum, sustainability of what you're seeing now?
Yeah. It's a great question. We've guided for the year 22%-30% revenue growth, 32% revenue growth. We didn't update that. We've got a lot of irons in the fire. The digital health channel's a newer channel that we're in. We've got some successful partners there with Function Health, Everlywell, and so on. We've also recently launched with folks like Hims & Hers. We'll see how that grows over time. We're excited about where we're at. We're enthused by the volume growth. Last year, we really leaned into discounted pricing for where volumes are higher. We offer a lower price, so the ASP has come down a little bit. We're prepared for that. I think our volume trajectory's looking good, and we're still comfortable with our guide.
Maybe just shifting over to Galleri, obviously we were talking a minute ago about the 14% reduction in Stage 4. Just talk a little bit about the reception you had from the physicians at ASCO. Obviously there's been noise on missing the primary endpoint, maybe just touch on that as well.
Happy to. Yeah, for those of you who didn't see it, yeah, we just announced actually results of two key studies at ASCO over the weekend. The NHS-Galleri study, which is our randomized controlled trial, by the way, the only randomized controlled trial of a multi-cancer early detection test that there has been or is. We also had full results from our PATHFINDER 2 study. Yeah, just touching on NHS-Galleri, this is a huge study, 140,000 people. We had a number of really important conclusions from that. I'll just run through three or four of those, and happy to dive into more detail. One of the things that people say about this kind of technology is it's really good at finding late-stage cancer, but not so good at finding early-stage cancer. I think we hit that one squarely on the head.
We found more Stage 1 and 2 cancers than the entire NHS screening program finds cancers of any stage. We really did find lots of early-stage cancers. 40% of the cancers we found were at Stages 1 and 2. Just under 70% Stages 1 to 3. That's the sort of first message I would just leave you with. The second message, as you rightly said, Tycho, is overall, we saw a 14% reduction in Stage 4 cancers. By the second and third rounds of screening, that was more than a 20% reduction in Stage 4 cancers. I just think this is incredibly important because that was across all cancer types. If you just think about the implications of that at a population scale, reducing Stage 4 cancer by more than a fifth across all cancers has profound implications for the future of cancer care.
I've been in the cancer world for more than 25 years. I have never seen a single intervention that can have this kind of impact. Added to that, we found four times as many cancers through screening. We had very good test performance compared, very consistent with what we've seen in our other studies. I think we've really got some very, very strong messages from the trial. Yes, we were disappointed not to hit the primary endpoint, but that primary endpoint was designed several years ago, and the treatment landscape has changed fundamentally in that period.
Most importantly, the biggest changes we've seen in treatment have been in Stage 3 cancers, where we now have many effective treatments and many curative treatment options for people with Stage 3 cancers. The opportunity now to really have, as I say, a profound impact on cancer survival and cancer mortality is very much within sight. Those are just some of the headlines, Tycho. Happy to go into more detail.
If we could flip a little bit to the FDA. You've been kind of confident there, coming out of the trial results. I guess, what does an FDA label look like, with a pan-cancer approval potentially unlikely here?
Yeah. Look, we haven't got into detailed label negotiations with the FDA yet. We've had a Breakthrough Designation for a number of years now. We completed our submission to the FDA in January of this year. We are in very active dialogue with the FDA. They are intensively reviewing our submission. We have a sort of not quite daily, but almost daily back and forth with the FDA. We know that they're very active in the review. We said at the time we completed the submission that we would expect it could take up to a year. That remains our baseline assumption. The critical determinant of that will be whether they decide to hold an advisory committee. We don't yet know the answer to that. We're waiting to hear. We hope to hear in the next few weeks whether they'll have an advisory committee.
If they do, then, as I say, that remains our baseline assumption, and that would mean the sort of 12 months expected-ish timeline. If they don't hold an advisory committee, it could be shorter than that, but we don't know that, so it's not our baseline assumption at the moment. Yeah, we feel we're in good shape, and we're in good dialogue with the FDA. To your specific question about labeling, that will come after the decision on an advisory committee. We're confident with what we can do.
One of the conversations we've had consistently with the FDA over years now is that we will be able to report any cancer we find, and that's really the most important thing about the performance and opportunity of this test, which is that we really do find many different types of cancer, and we want to be able to report any one of those. That's the conversation we're having with the FDA.
How about additional data? You've committed to following the NHS-Galleri population for another 12 months. I guess, what would this look like? What can we expect out of that? What's the right solution longer term, since a mortality study would probably take too long?
Yeah. We're not going to do a mortality study. We made that very clear. A mortality study would take probably three times as long, probably cost more than three times as much. Actually, more importantly, we just don't believe it's necessary. The reduction in stage 4 cancer is a very meaningful proxy for what will happen in terms of survival and mortality. As I say, doing a mortality study would have taken at least 10 years, probably closer to 15, and with the pace of change of technology, by the time you've completed that study, the technology's probably obsolete. I just think we need to be thinking about a different paradigm of these kinds of trials now. We very deliberately decided on this late-stage reduction as a meaningful endpoint.
To your question of what additional data, yes, we're pretty confident we'll have an additional 12 months follow-up data, which we should get Q1, Q2 next year. Hopefully, that will show an even bigger effect, but we'll see when the data comes through. We have reason to believe that that will be additionally positive. After that, there are a number of different things that we will be doing as a follow-up, a passive follow-up for mortality. It's not powered for mortality. I don't want to set any expectations about the sort of significance of a mortality result.
Just thinking about PATHFINDER 2, 71% of new cancers in stage 1 through 3 that were detected. Just talk about the rationale why this is a strong outcome for you guys.
Again, the corollary of a stage 4 reduction is a stage 1 to 3 increase, if you think about it in practice, what you want to be doing is finding those cancers before they progress to stage 4. A diagnosis at stages 1, 2, or 3 would come in advance of a progression to stage 4. The opportunity to find more than 70% of cancers at those early stages is really what we want to be achieving, as I touched on earlier, at those stages, we have many curative treatment options for most types of cancer. The opportunity to really make a difference for patients is very substantial. Just to put that into context for people, right now, we're finding around 20% of cancers at stage 4, they represent about half of all cancer deaths.
That's where we really need to make the biggest difference, and there is now a huge survival cliff between stage 3 and stage 4. In the U.S., overall stage 3 survival is about 64%. Overall stage 4 survival is about 26%. That's really where the massive difference is in terms of cancer survival. It used to be between stage 2 and stage 3. Now it's really shifted to between stage 3 and stage 4, with all of the treatment advances there have been over the last decade or so.
Maybe just thinking about CSO, 91.3% accuracy, 48-day median diagnosis resolution. How do you think about the benefit of this over competitor assays or even your own cases where CSO is undetected?
Yeah, look, we think that a CSO for a multi-cancer test is an absolutely critical element of the offering. The notion that you have a test that it doesn't really matter how many cancers you're looking at, whether it's 10, 50, or in our case over 150. If you simply say, "Yeah, we think you've got cancer," but you can't give any guidance to the clinician as to what that might mean and where they should look, the potential challenge is a real diagnostic odyssey for patients and their clinicians. What you want to be able to do is not just say, We think you might have cancer, but you want to be able to guide a really efficient diagnosis.
You really want to be able to say to the clinician, "You should look at the ovary," or, "You should look at the pancreas," or, "You should look at the liver." We want to be able to do that with very high accuracy. Our CSO accuracy across all our studies is consistently around or above 90%. That gives a really high level of confidence to the precision of that guidance, if you like, to clinicians as to where they should be looking. This is very much consistent with what the FDA has been saying. There was an advisory panel back in 2023 where they also said that having a CSO prediction capability, they called it tissue of origin at that time, but it's the same point, is a critical component of a multi-cancer test.
We have integrated this into our test from the very beginning, and we've designed it such that it has really high accuracy for the reasons that I've just said. We think it's fundamentally important, and I think the data from the studies bears this out. You touched on PATHFINDER 2. We know that with a true positive, the time to diagnostic resolution, the median time was 37 days, which is a really rapid time from saying someone might have cancer to when we can actually confirm a cancer diagnosis and where it is. Without that CSO capability, it would be substantially longer. On average in the U.S., based on the best data we have, it takes more than 100 days to get to a diagnostic resolution for patients who are presenting with symptoms, for example.
To be able to bring that down to just over a month means you can start treatment much faster, much less anxiety for the patient, much lower levels of diagnostic resource utilization. Actually, if we just think about the potential impact on patients, there was a study published some years ago that showed that for every four weeks delay in starting treatment after a cancer diagnosis, mortality increases by 7% for every four weeks delay. Being able to get to a confirmed diagnosis and starting treatment quickly is really fantastically important. For all those reasons, that's why we believe the CSO capability is fundamental.
You noted on 1Q that you've done an Epic integration for Galleri. I guess, how do you think about the lift you'll get from that in the context of the 22%-32% revenue growth guidance?
Do you want to take that?
As I said, we're in the middle of the implementation. We'd expect it to be implemented by the end of the year. For the year, it doesn't really have very much impact. After that, we expect it to make it easier for physicians to order, increase the breadth of physicians who are ordering. Still, remember, it's a self-pay test. It's not reimbursed. Will it have a huge reflection point until it's reimbursed? We'll have to see. For the year, it's going to go in at the end of the year.
Are there other similar integrations on deck for the next year or two with Flatiron or others?
Currently, we're looking at rounding out the physician customer ecosystem. At this point, we're really just focused on getting Epic in there as it's the biggest piece. As you know, there's other players in that space as well.
What are the other remaining barriers to nationwide physician adoption, I guess, beyond Epic?
We probably could have picked this. As I alluded to, broad access, the test being reimbursed is one of the larger barriers in the U.S. The MCED legislation has been passed, which gives CMS the authority to pay for the test or FDA-approved tests starting in 2029, and that will phase in year over year. Along with that, we'll be looking at USPSTF and those pathways and so on. Really getting to broad reimbursement and broad access will take reimbursement. We'll have the sales force, we'll have the ecosystem built. I think we've shown that there's a pretty large and growing pre-reimbursement market here, given what we've done over the last four years. I see that continuing to expand, given what you're seeing with the consumerization of healthcare and digital health and so on. I think in my view, those are some of the big barriers. Harpal?
Yeah, I think you've touched on it, Aaron. Just sort of reiterating a few points. Now we're going to be able to arm our sales force with actually the key messages that flow from that and really being able to go out and talk about that data. They've been a bit hamstrung over the last few months, not really being able to talk about the data in detail until we presented it at ASCO. That's kind of a step that's now in hand. We just touched on FDA approval. We know that will make a significant difference for a number of physicians and a number of commercial payers. Medicare reimbursement, Aaron just touched on, the Medicare legislation requires an FDA approval. That's a critical step that we're well in hand with.
Being able to open up Medicare reimbursement as well as commercial payer reimbursement are both going to be critical things. The self-pay market, as Aaron just touched on, is absolutely a substantial market and a growing one in our view. Getting into guidelines. I think, again, after an FDA approval, we'll be working hard to make sure we get into ACS guidelines, hopefully NCCN guidelines, and so on. All of these are steps along the way to opening up what is a very substantial TAM
Can you maybe just touch on the sales channel? You had an expansion earlier this year. You obviously had the Quest and Hims partnerships. What does steady state look like for you guys in terms of the mix between your own in-house sales force, where you capture all the economics, and then the partnerships?
Yeah. It's a great question. We've got an expanded sales force. We're training them now, getting them out there to focus on the brick-and-mortar physicians offices. We also have self-insured employers and enterprise-type partners where you don't really have a rep out talking to a Google or an NVIDIA and so on. They've got benefits managers. They make the software available. There are these different avenues of really expanding your sales force, your sales capabilities without expanding your sales force. Various partnerships, other opportunities to put our test in somebody else's bag. Those are things that we'll look at as we're growing. I think as Harpal's alluded to, there's some big milestones here. We've just got the ASCO data out there, FDA approval in the near term.
I think those are really going to be gate opening to new groups of physicians, moving from beyond the early adopters into a different part of the segment of that population. We're open to those ideas. We do want to continue to grow as capital efficiently as we can. If there's the right economics or the right opportunity for us to drive volume, of course, we'll be looking at those things.
How about just the competitive environment? Obviously, some new entrants in the MCED space. How do you defend your first-mover advantage here, or is there enough of a rising tide, it's less of a focus right now?
I'll start, Harpal can really hammer it home. We're the only test out there with the data set that we have. Being able to find cancer in the intended use population, so a group of asymptomatic people for cancer. No one else has that data. Before we even go into NHS data, we're the only company that can actually show a stage 4 reduction in MCED. You pair that with the CSO and the test performance capabilities, even when you look at case control data, which case control data is important, and that's where we started years ago. In 2018 is when we released CCGA data, I think. What we've continued to show is we can re-perform or do better in the intended use population. No one else can do that.
I think we're very comfortable with the level of data that we have out there and how important that is to physicians, because you can actually trust us to find the cancer in the population it's meant to be used in, not a case-controlled group. Harpal.
Look, first of all, this is an enormous market opportunity, not just in the U.S., but globally. The notion that there would only ever be one player is fanciful. We've always expected that competition would be coming in, and frankly, up until now, we've been plowing a lone furrow here, and having some other noise in the market is actually helpful to us. Look, we think we're years ahead of where the competition is. To reiterate a couple of the points that Aaron made, we are the only company that has data in the intended use population, by which I mean people who are not diagnosed with cancer or under active suspicion of cancer. We're the only company that has done studies in the intended use population, people who are not symptomatic.
That's the first thing to say, and we've done three studies of varying sizes in North America and in the U.K. We're the only company that's thought about doing a randomized trial. Secondly, we're the only company, I believe, that is exploring an FDA approval for an MCED. That, as I've already touched on, is critical when it comes to broad-scale reimbursement. The Medicare legislation requires an FDA approval for an MCED. There are a number of things that give us that confidence that we're ahead. I'll just add one other point, which is, as Aaron touched on earlier, we've now run over a half a million commercial tests. We've run something approaching 400,000+ clinical tests. We're getting to about a million tests now that we've run.
With every additional test we run, we gather more data, and we're developing a data set that we believe is unparalleled and which will help us to continue to evolve and improve our technology over time. That data asset is an incredibly important part of our overall continued development of this kind of technology and approach.
How about R&D? You're seeing some of the newer entrants come in with methylation profiling, protein biomarkers, whole transcriptome, whole genome, whole exome. How do you think about improving the assay over time?
Yeah. It's a really good question. First of all, the assay will improve over time, and I've just touched on how we'll use data for that. We're very much focused on constant innovation, constant improvement of what we have already. We've looked at many of these alternative analytes or approaches. So far, none of them has been additive in terms of performance for us. We've looked at proteins, we've looked at RNA, we've looked at mutation analyses. We've looked at all these things, and we will continue to do so as technologies evolve. We do that constantly. So far, none of them have been additive to our targeted methylation panel, which is producing really very strong results. As I would reiterate, technologies will continue to evolve, and we'll continue to look at them very actively.
If anything looks like it could add substantial performance advantage without adding a massive amount of complexity and cost then we'll absolutely look to pursue those.
Maybe just, you could touch on the importance of sensitivity and keeping a low false positive rate, and the trade-offs between sensitivity and specificity and why being tethered to a high specificity, 99.6, is a winning strategy versus some of the peers that may be optimizing a little more around specificity.
Yeah. It's a really good question, and one that we've considered very carefully. What we have today is a number of single cancer screening programs, and they do optimize for sensitivity over specificity. The reason is you know exactly what you're looking for, so you can home in on, for example, a breast cancer if you have a mammography program. In a multi-cancer context, we strongly believe that specificity is going to be key because, as I touched on earlier, what you don't want is lots and lots of unnecessary investigations of individuals, and that's what a false positive gives you, right? Because you want to be sure that if you're saying there might be a cancer there, you want to be more confident than not that there really is a cancer there. So we've anchored on specificity.
The target we set ourselves was to have specificity higher than 99%, so in other words, a false positive rate of less than 1%. The important philosophical and clinical reason for that is that in the population, the cancer rate is about 1%. What you don't want is a false positive rate that is higher than the cancer incidence in the population. That would be clinically really difficult for patients, for clinicians, for health systems, for payers. Having a false positive rate less than 1% we think is critical. Some of our competitors are quoting false positive rates 1.5%, 2%, 3%. That's 6 times the false positive rate that we have. We think that's not going to be tolerated in the market.
You touched on the assay iteration. Just touch on a little bit on how you're integrating AI, as well, both into the workflow and maybe assay improvement over time.
Yeah, look, we're already adopting AI in a number of ways across the organization, within the organization, in our software development, for example. We already use AI in our technology. We have very sophisticated machine learning algorithms that underpin our test, and we'll continue to look at opportunities to build on that further over time as those technologies evolve.
Aaron, anything you'd add on COGS, revenue cycle management, anything, where AI can really help you get some more leverage in the model?
As Harpal said, we're looking at implementing GenAI in several different places, and specifically around revenue management. Nothing novel that others aren't doing. There's things you can do around customer service and billing and so on. We're building for population scale, so those are all things that we're thinking about as we're thinking about the 100 million people in the U.S. who could access this test. How are you going to scale to take care of that? It's going to really look different than I'd say other screening companies in the cancer space have looked. It's going to need to look different.
You're finding it effective on sales force targeting in terms of getting to customers?
Yeah. We've been in the market now since 2021. Just for everyone's general awareness, we're not blanketing the U.S. trying to find every single Galleri sale that we can. We started that way, and it was a very expensive proposition, being the only folks on the market having to educate so many physicians and find those customers. We're really concentrated in areas where we're finding success. There's still a lot of white space to grow into, and we've definitely, I think as our growth has shown, we've found success in those places. Our targeting has improved and gotten better. I think as anyone who's implementing AI, as you build out your data set, it's going to get better and better and better. You've just got to make sure that you're building it out appropriately.
Maybe in the last minute or so here, we can just hit on Congress obviously passed legislation to enable reimbursement of MCED tests starting in 2029. How are you thinking about the phasing of age eligibility?
Yeah. Quickly, it's, again, starting in 2029 with FDA approval. The CMS will have to make a national coverage decision as well. We have a REACH study, which we're doing with them that'll hopefully enable that. That phases in from 50-65-year-olds in the first year, and then it adds a year, so 66, 67, as every year goes by. USPSTF is another path that I'm sure you're all aware of. We'll be pursuing that. If we get that in an A and B rating, it kind of supersedes all that age phase-in, and the whole Medicare population would then become accessible. Even though I think the 50-65-year-old population is like eight million people, so it's a pretty big TAM just to start with.
we've got a lab that we've built that can run over one million samples today, so we're prepared for that volume.
Just last one, I guess, just on the profitability side. If we assume Medicare rate is likely $500, $600 given crosswalk, how do you think about that in terms of the longer-term margin target of 50%-60%? At what scale does Galleri become profitable for you guys?
Yeah, with the current version we have that's on the market, we see 50%-60% margins at that pricing. We've got a lot of fixed cost leverage to grow into. As I said, our lab can run over one million tests. We're running them 50,000-ish a quarter right now. We can definitely continue to improve margins at the current ASPs. As they come down, we'll still feel comfortable hitting the 50%-60%. That's on the current version. As Harpal said, we've got R&D programs where we're continually looking to innovate and improve the test. We're also looking to decrease the price of the test because in our view, for this to be a globally adopted test accessible to people to actually change cancer outcomes, it's going to have to be a more affordable test. The price is going to have to come down, not go up.
I think we'll leave it at that. Thank you.
Thank you.