I'm Thomas Stevens. I work on the Life Science Tools and Diagnostics team. I'm joined by Aaron Freidin, CFO of GRAIL. Welcome. It's great to have you at our conference. It's an exciting time for multicancer early detection development. I guess just to kick off, there's been a series of developments since you spun from Illumina, large restructuring, lengthened cash runway significantly, new cheaper version of Galleri, and there's some MCED bills floating around the House and Senate. With studies beginning to read out beginning of 2026, what are you most focused on as a company today?
Yeah, thanks for having us. It's great to be here. Right now, we're focused on a couple, two, three primary things. One, as we said, we've launched the new version of the test, high throughput, lower cost. That launched recently. We're working on getting that introduced out to the market, working through the kinks of launches and so on, and really making sure that that goes well. No, the performance is great. Everything's going the way that we figured it would from a test performance perspective. Anytime you launch something new, we're going from doing 100,000 systems that's capable of doing 100,000 plus tests a year to a million tests a year. You got to work through that. Really excited about continuing to drive commercial adoption and access. You probably saw the Quest announcement. Now Quest, we're integrated in their EHR.
Physicians can order through the EHR. Patients no longer have to get a kit from us with tubes in it to then go to get a blood draw. They can just go to Quest and show up and get their blood drawn. It makes it much easier for everybody. With the Tricare announcement, we've got an opportunity here to really drive access into the military population. We're currently working on getting in contract with those primary carriers in the Tricare world. Beyond that, continuing to make sure we're executing against our operational goals, bring our burn in line with where we want to get to, and get to our PATHFINDER 2 readout at the end of this year and prepare for the NHS- Galleri readout next year.
Awesome. Yeah, I mean, I think with the Quest, you mentioned Quest and TRICARE. We might as well dive in there. They've integrated you into the system, network of over 500,000 physicians. Could you just break out how that works in a bit more detail? You send them the kits, they run it themselves. What kind of access do you see patients now having to the test that they previously did not have?
Yeah, so previously, every patient would get a little Galleri kit. We'd have to send them a box, essentially, with two tubes in it, requisitions for the physicians to fill out, and so on. They'd have to then take that to a blood draw facility. That would then have to make its way back to us. Somebody would have to open it piece by piece. We did 137,000 tests like that last year. That's 137,000 box openings, essentially, and people having to go get their blood drawn. What Quest integration does is it allows us, the physician, to order the test electronically and have all that come in electronically. It allows Quest and the patient to go to a Quest draw facility, and the tubes are already there. Quest can actually batch ship back to us as well.
They don't have to ship two tubes at a time or one patient at a time. They could batch five or ten or however many there are that show up that day. It makes it much more pain-free for the physician, for the patient, and also for our operational capabilities.
Awesome. Yeah, I guess to dig a bit deep on that, I mean, do doctors now, they just go through the Quest portal and they can order Galleri tests really easily? I mean, how much of a change is that in ordering versus before where you'd have to go through manually?
Yeah, so it's basically just a drop-down menu in their Quest interface when they're ordering tests.
Right. I guess on TRICARE then, could you just maybe size the TRICARE opportunity? And how did you manage to get coverage kind of prior to FDA approval, and how are you deploying Galleri in the TRICARE population right now?
Yeah, TRICARE looked at Galleri as an LDT. They have a series of steps that they can go through to determine if it's something that they want to offer on their menu, whether it's safe, whether their study's validating, and so on. The DHA, the Defense Health Authority, does that assessment. TRICARE then says, "We'll go ahead and cover this." You then work with the carriers who then pay for the test on behalf of Tricare. I think Tricare's population's around nine and a half million people. If you look at our test, it's built and designed for people between the ages of 50 and 80. I think there's about three and a half million people in that population. How many of those people will have access to the test? We'll see as we go through these carrier negotiations and so on.
Just on those carrier negotiations, is that like a state-by-state thing, or is that a purely federal thing?
are three primary carriers that cover the TRICARE population. There are other ones, but there are three primary. It is really three parties that we are working with.
As you kind of phase that in, is that a six month process, 12 month process? Do you kind of get those on board?
We'll see.
We'll see. We'll be waiting. Maybe just to switch back to kind of what the test actually is and your kind of go-to-market strategy there. Galleri covers 50 cancers today. It's a lot of decision trees to manage. Have you thought about go-to-market when you're scaling a test with so many downstream implications for clinical workflows? When we see final prospective asymptomatic data, would you consider trimming your panel to indications with the strongest performance and the greatest clinical need?
Yeah, it's a great question. I mean, there's a lot going on in the MCED space right now on test formats and how tests work. Galleri was always designed from day one to find cancer in asymptomatic people. The CCGA study and then the Pathfinder study all replicated that we were able to do that. What we really find is tumor fragments floating through the blood that have a specific methylation pattern that is highly specific to cancer. It's not looking at prostate cancer and then breast cancer and then colon cancer. It's looking for cancer. The way the test works is it's basically two bioinformatic decision points, right? Does the cancer signal exist? If it does, where's that cancer from? Highly specific methylation pattern for cancer.
If it's a yes, we then look at that same data and we determine where that cell originated from because methylation drives cellular from day one, like it turns into a liver cell or a lung cell or whatnot. That specific cancer signal is across all cancers. If you look at our stage two sensitivity in the 12 deadly cancers that make up two-thirds of cancer deaths, our sensitivity is 70%. That's localized cancers. They can be treated with curative intent. Just my view, our test is finding the cancer signal. You could return, why would you return results on just those 12 when you're actually finding a longer tail of cancers as well? The technology shows that.
In our view, if you've got a common cancer signal, the best thing to do for your patient and for your physicians is to tell them that you can find it. Yes, you might find one out of 100, but currently most of that screening sensitivity is zero today. With our ability to localize the cancer, we're about 90% accurate on that. Physicians, we've got 14,000, 15,000 now that have ordered the test, 137,000 tests last year. They know how to work these things up. Each of these cancers has a workup regimen for it. Image this, image that, and so on. Where we go look in a region for a cancer and they don't find it, physicians also there, they now know that since the test is so specific, they should pay attention to that patient.
They should probably see them again in three to six months. Some choose to do that and then they reimage. Other times, we offer a free retest. Since that test is so specific, if that second blood draw comes up negative, it's likely it was a false positive. If it comes up positive, it's likely that there's a cancer there and it's just not showing up on imaging yet.
Great. That was what I thought. Thanks for that. I guess maybe going to the big readouts coming up. Ahead of potential FDA approval, you have the NHS- Galleri study, which is kind of being run by NHS, and then PATHFINDER 2. There is lots of weedy detail we can get into, but I guess from a high level, what are the real primary endpoints investors should be looking at and maybe secondary endpoints which could drive commercial payer coverage after the fact?
Yeah. Both of those are our registrational studies that are going into the PMA submission, which is currently scheduled for the first half of 2026. PATHFINDER 2, we'll just go in order. That's essentially a safety study based on the US of 35,000 individuals. If you've seen the PATHFINDER study, it's going to be very similar to that. That was 6,600 individuals. This will be 35,000. Endpoints there are really safety. You will see you'll be able to look at how many cancers we found in total. You can compare that to the standard of care cancers found, very similar to what we did in Pathfinder 2. That data will be available to the back half of this year. We got the NHS- Galleri study, which is the three-year, 140,000-person study in the case-controlled study in the U.S., a randomized controlled study, sorry, in England.
That's fully enrolled. We're basically just waiting for the one year follow-up to expire, which will be in July of this year. Endpoints there are going to be a stage shift between the control group and the interventional arm. Do we find less late-stage cancers in the folks who got Galleri in years two and three than in the control arm? If we do, I mean, that's going to be the first look at clinical utility that we'll have. That will also go into the PMA registration in the first half of 2026. The full data readout for NHS will happen sometime in the middle of 2026. We'll essentially be submitting to the FDA the first-year results because the FDA is looking at performance. They're not looking at clinical utility.
We will have the clinical utility available if they're interested in it, and we'll share it with them. For them to begin their assessment, they're going to be looking at performance.
Yeah, yeah. I think one area investors have maybe dug a bit more into is there's FDA approval and the safety and effectiveness, and then there's what gets commercial payers on side and the health economic data that follows up there. It seems to me the really big readout is potentially the downstream health economic assessment the NHS does. What kind of timelines are around that, and kind of what do you think they'll be assessing if you were to speculate at this point?
Yeah, I mean, I think folks will start looking at the economic impact based off of just the stage shift. I mean, there's lots of data around what it costs to treat cancers in stage one and two versus stage three and four. There will be an economic readout from the NHS study as well. I don't recall exactly when. It's not going to be in 2026. Also, we'll have currently we're doing the REACH study with the CMS population, which is 50,000 people a year for three years. That'll also have an economic view as well in the CMS population. That likely won't read out until 2028, I believe, somewhere in there.
Cool. That's very helpful. I guess on that point of stage shift, I mean, I think that's been a big, again, another big area of debate between a diagnostic test, which maybe gets you at stage two versus stage three, and what maybe an exact science is called an early cancer detection test where it's super early stage one. I guess, could you speak to the clinical benefits of stage shifting from a stage three to a stage two, or even a stage four to a stage three, and kind of what health economic data you can point to?
Yeah, I mean, it's pretty accepted the earlier is the better, right? And stage two is localized, similar to stage one. You always want to find it earlier, right? The key thing here when you're thinking about single cancer companies or single cancer screens is they're focusing on one cancer so they can afford to give up specificity for sensitivity. Current cancer screening is only finding 14% of cancers. So 86% of the cancers that are showing up are not being found by our current standard of care. Any level of sensitivity better than what we have for one test that's also going to pick up interval cancers in that other population is going to make a difference. I think as we go through the clinical utility data with the NHS, I think there will be several different views that people will take.
As long as we're saving lives, there's generally a value associated with that in a stage shift. We'll be able to quantify that.
Yeah, I guess, and again, maybe this is too much speculation, but the ultimate bars for test performance for private payers and the USPSTF, if and when they eventually adjudicate on it, what do you think those bars are and what would you be satisfied with from maybe a PPV perspective?
Yeah, I mean, the FDA has been, we've been in breakthrough with them since 2018, I think. They've reviewed our NHS- Galleri study, our PATHFINDER study, PATHFINDER 2 study, then there's a CMS REACH study. They've done all those under IDE. They're familiar with our performance, and we're in regular conversations with them. We believe we have got an approval package we'll be submitting as long as the data is consistent with where we've been, which we've kind of showed between CCGA and Pathfinder data was pretty consistent. As far as USPSTF and reimbursement, I mean, we know USPSTF looks at mortality. We're not going to have mortality data because it would take 10 years and probably $500,000,000 to get there. By then, the technology is no longer likely relevant.
We will be engaging the USPSTF, working with them on using stage shift as a corollary to mortality.
Cool. Very helpful. Maybe if we could just move on to the kind of competitive landscape. You mentioned the difference between a multi-cancer test and a single cancer test and what the benefits are of casting that wide net. I guess with the likes of some companies iteratively rolling out additional single cancers on their blood tests pre-existing, which is maybe more tuned and specialized to those cancers, how do you compare your competitive position to the kind of wide net you cast with your multi-cancer test?
Yeah, I mean, I think MCED is going to be a huge space, right? It's 100 million people in the US. You quickly get into 350 million people when you add other countries around the world. We don't expect to be the only people in it. At this point, we thought there would be other people in it helping us build awareness, build this market, engage KOLs and so on. It's going to be a big space. Looking forward to seeing the data that anyone else is going to release that's going to join the space. Currently, we haven't really seen any from a locked assay validation, analytical, clinical, and whatnot. You said we did 137,000 tests last year. Our physicians, we've learned a lot from them on what they want, what they value, and so on.
Something like the tissue of origin and that ability to localize is very high up on their list of features that they need because they can really direct a workup and not send somebody on a journey. I think that the market will definitely get more crowded than it is today. I'm looking forward to what data they're going to have. I mean, and I kind of mentioned it earlier, and you did too, you're either finding, you're either taking single cancers and adding them up, right? That likely has a pretty high false positive rate. Or you're finding a common cancer signal today and you don't know you're finding those other cancers. We started at the needle in the haystack, finding cancer in asymptomatic people. I think it'll be interesting to see what the data shows.
Yeah. Great. You mentioned a tumor of origin there and how it's built into the assay kind of from the beginning. Just from a healthcare cost scenario, how does that change the downstream costs and follow-up availability that GRAIL provides? Yeah, let's just start there and maybe we can dig in.
Yeah, I mean, the ability to localize really gives the physician an opportunity to go into a directed workup, right? You get to results faster and those treatment options versus sending somebody to a whole body scan or image, which I don't know how many of those are in the world, but there are a lot more ultrasound capabilities and so on. It's what we find and what our physicians tell us. It's a great way to work things out because it fits into an existing downstream infrastructure that already exists. From a cost perspective, I don't know what that analysis looks like, but usually the faster you get to a diagnosis, the cheaper it's going to end up being.
On specificity, so 99.5% specificity, I think one comparator is 98.5%. Just how much of a difference does that one point make when you're screening across a massive population?
It's three times the false positive rate, right? It impacts the PPV. When you're screening that many people, you're going to have three times more false positives with a 1% change essentially in specificity. I think in our modeling, Galleri is essentially 45% PPV. It's almost a one-to-one, one false positive for every positive. That ratio changes based off of your specificity, really. I think by the end of the day, we'll learn what physicians really appreciate the most.
Cool. Just maybe again, attacking screening from a different angle, some have advocated for pushing MCED into a high-risk population first before pushing out to an average risk. How much do you think MCED becomes a one-size-fits-all market versus having these more specialized tests potentially for a high-risk population? Is that a population that you'd consider going into yourself?
Yeah. As we designed our assay, again, it wasn't an asymptomatic population. Sure, there were folks in there who had other risk factors, but they looked more like the population. It wasn't people who didn't look like the intended use, which is why the performance is kind of maintained between CCGA, PATHFINDER, and so on. Again, I'm just the CFO. I'm not the scientist. If you can find cancer in healthy people, you can probably find people who are more at risk for cancer better. From a technology perspective, I imagine it's the same underlying technology that's doing that. Just a person who's at more risk likely has more tumor DNA floating through their blood. There are definitely folks who have asked if we are going to launch a test in that way.
I'm sure there are physicians who have used the test in that way on off-label use. But it's a very interesting path to go down from a reimbursement perspective. Proving somebody doesn't have cancer because it kind of becomes a rule-out test is a challenge. How do you do that? Is it time? Is it imaging at a certain point in time? And then you've got to change physicians' perspectives. I think it's definitely an opportunity. I think there's 14 million-15 million people in the U.S. who'd kind of fit into that at-risk for cancer demographic. We'll see.
Cool. We'll get back onto the financials. Looking at your 20%-30% Galleri revenue growth guidance in 2025 in the U.S., today the majority of those are self-pay, right? Kind of how do you ensure that you generate recurring testing volume in what is currently a rapidly growing consumer market? What is the typical profile of people getting screened with MCED today?
Yeah, so about two-thirds of our volume today is self-pay, which is either directly through physicians. We've got an online patient-initiated telemedicine capability where folks can log on and talk to a telemed doctor to order the test. The rest is through self-insured employers, life insurance, or enterprise-type business. I mean, as I said, there's 100 million people in the U.S. who can take this test. We did 137,000. There's, I think, six and a half million people who make more than $250,000 a year in the US. There's a lot of opportunity there prior to reimbursement. Learning from your physicians on getting them comfortable, what they find useful, how they implement, what a physician who has the test and their protocol that they offer to everybody, what makes them do that versus a onesie, twosie order.
Building that ecosystem is really where we're focused and how we make sure we continue to drive growth in addition to things like Quest, Tricare, and so on. We're really focused. In the last reorganization we did in the last year, we focused the commercial team on creating basically more white space across the U.S. where we don't have reps calling on doctors to really focus on the areas where we're seeing more success. That's more from an efficiency perspective why we did that. We're not chasing every last Galleri sale. I mean, there's a lot of opportunity to grow and even broader reimbursement.
I guess, yeah, just on that, I mean, I think if you look at other single cancer tests in the space, they've generally been quite sensitive to kind of Salesforce reps and/or spend. I guess how are you approaching that from a multi-cancer perspective? I imagine doctors are still very sales-sensitive in terms of the more money and people you throw at them, the more likely they are to prescribe the test. How are you approaching that in a more disciplined manner? You mentioned targeting, but is there anything else that you can do differently with a multi-cancer test that you perhaps could not do with a single cancer?
Right now there's no one else calling on these folks. We don't have a ton of experience on what it's going to look like when there's two people calling on them. I think the depth and the richness of our clinical validation programs, of our studies and so on, in addition to just the commercial experience we have, have put us in a really strong position. We've really learned that once these physicians find that first cancer, they become believers, they become champions and so on because they see it work. It's one of those, it's a love. It's numbers though, right? You're going to find one out of 100 patients, right? When does that one come?
Building that atmosphere around them, that ecosystem around them so they can talk to each other, so they know what each other are doing. Continuing to just put out high-quality data. I think, as I said, we'll learn more about the competitive landscape as this evolves.
Have you ever pushed out any data on, it's very early still in its commercial launch, but in terms of reordering rates and kind of what they look like and what doctors who reorder, where they're situated?
Yeah, I think we called this out at JPMorgan. Our reorder rate is just over 20% with, I think, 70% of those happening in the first 16 months. For an unreimbursed test, we're pretty proud of that. That's 20% of those people deciding to pay for it year after year out of their own pocket. ASP last year was about $800. It's not a small amount. People are really passionate about cancer. It touches everybody. We think working on broader access and reimbursement, we've got our COGS profile down lower so we can support a lower ASP as that happens. It's just really going to make this thing tick.
Yeah. Just thinking about that kind of COGS trajectory, I think you're fitting four more samples on a flow cell now. I guess, can you help us understand what the pathway on the kind of fixed cost leverage looks like? Is that, do you have to eat a ton of upfront COGS in the interim period and it gets way better as you scale? And kind of what ASP ballpark does that new COGS outlay satisfy?
Yeah, so the platform that we launched, this version, is going to support our 50%-60% margins at scale. Our adjusted gross margins last year, which takes out the intangible asset from the Lumen acquisition and then spend of GRAIL, were about 46%. We expect those to increase through 2025. We've got capacity with the infrastructure that we've built today to run a million tests a year. The economics of the company, the cash burn profile of the company can change quickly as volume materializes because the CapEx infrastructure is there.
Yeah, I guess we have kind of half- arse on kind of cash burn and the balance sheet now, I guess. Could you detail how you got to your 2028 runway under your new Salesforce and OpEx outlay and how are you thinking about your burn and financing needs as we go forward?
Yeah, so we've guided no more than $320 million for 2025. We see cash runway into 2028. That's achieved through just subsequent years, cash burn coming down as we continue to grow at the rate that we are with the higher gross margin from the new version. Also, some of our development programs, we're still submitting our PMA. We're still going through the launch of the new version. As those programs roll off, those dollars could be freed up. We see subsequent step-downs in spend going through into 2028.
Just financing needs-wise, is that really an event that happens after we have the full data? You've got a bit of cushion now, right? What kind of timeline should investors be really thinking about there?
Yeah, so we were again funded into 2028. We're in a position now where we don't have to raise immediately. As a CFO, I'm always looking and interested in folks who would like to provide capital. Right now, it's execute against our plan and get out there and really be a public company. We kind of got spun out in June of last year. We're almost a year removed from that. I think the further we get away from that event, the more that GRAIL engages and executes. I just think the better that overall next fundraising story will be.
I have to ask about Stargate. I think GRAIL has been caught up in some AI thematic trades, given the mention there. The way the assay was initially designed, it sounded like you leveraged quite a lot of LLMs to titrate on fragmentomics and to kind of set on methylation as the approach. To what extent, as you get more and more and more patient data in the field, to what extent can you use that to iterate on maybe a V2 or V3 in future?
Yeah, I mean, that's work that we're actively doing. We've launched this new version of Galleri. That doesn't mean the innovation stops. Leveraging what's happening in the AI world, we've got, as I said, 137,000 commercial tests. I think we've ran over 600,000 tests total since we were founded. We have a large methylation data set, which is definitely an indicator for cancer. It's an indicator for a lot of things. Our R&D team are working through what's possible there. Is it possible to renew x amount of fragments now of cell-free DNA to make a call? Can you take less fragments and have it predict what it's going to be based off all the data we have? We'll see. That's one way that you could lower the limit of detection and increase sensitivity bioinformatically, really without even changing the assay. It's really exciting.
It is something that I hope to be talking more about later.
Just to squeeze one more brief one in because we haven't really touched on DC at all. Just to what extent did your guys in 2028 really rely on any kind of Medicare reimbursement and MCED bill? And kind of what should investors be focused on? Is the covering of follow-up really important to scan after the test? Or is covering of the test sufficient, you think, to widen the market massively?
Yeah, so on the follow-up, we've already ran a bunch of tests when test isn't being reimbursed. We haven't ran into issues of payers not paying for the follow-up or for the procedures and so on. As far as the MCED legislation goes, I think it's very well positioned. It's still extremely bipartisan. All the head folks of all the key committees are still in their positions. As far as our plans go, I think that the way the bill was written, coverage should start in 2028. I think the one that made it to the reconciliation bill that was terminally killed in December started in 2029. Really, we're looking at it as PMA submission the first half of 2026, assume 12 months to get an FDA approval. That's the first half of 2027.
If that date slips, as long as it doesn't slip past 2028, 2029, we're okay. It's not an expectation. As far as reimbursement goes, it's planned to start phasing in in 2029. Even before then, we'll have the NHS- Galleri utility data, which is a single payer system, highly regarded, highly respected single payer system. I think we'll be able to take that data outside of the U.S. and inside the U.S. to start talking to payers about it.
Awesome. We'll leave it there. Thanks very much, Aaron.
Yeah, thank you.