GRAIL, Inc. (GRAL) Earnings Call Transcript & Summary

Hey, everyone. Good afternoon. I'm Tejas Savant, I cover Life Sciences here at Morgan Stanley. Before we kick it off, just some important disclosures, please see the Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to Morgan Stanley sales rep. So it's my pleasure today to host GRAIL. And from the company, we have Bob Ragusa, CEO. So Bob, welcome. Thanks for joining us.

Thank you.

Maybe just to start, given the recent separation from Illumina and the subsequent strategic review, can you provide a quick overview of GRAIL's mission for folks who are not as familiar with the story? And what are you focused on over the next 12 months?

Yes. So GRAIL's mission is to detect cancer early when it can be cured. We're really focused on multi-cancer early detection. We've had a product in the market for over 3 years now called Galleri. And so Galleri is to detect many cancers at once as well as providing a cancer signal of origin. So from a clinical standpoint, it's a very powerful tool to be able to look broadly across cancers. And if you think about it, our standard-of-care screening looks primarily at the cancers that are responsible for 20% of the deaths, product Galleri looks not only at those, but importantly, it looks at the other 80% of cancers -- other cancers that cause 80% of the deaths. And so we think the multi-cancer early detection approach is really the key way of inflecting the cancer mortality curve over time, and it has the best impact on public health.

Got it. Maybe on that point, right, so there's a bunch of other players going after screening. Most of them are taking a single cancer approach, I think, except for Exact. So as you think about your decision to pursue multi-cancer screening, walk us through how you arrived at that decision, why not sort of refocus your efforts on single cancer just to go live and then expand into multi-cancer over time?

Yes. So from the very beginning, our -- GRAIL has taken an approach to really look at multi-cancer. And so if you step back to the people who have rotated to single cancers, when we look at it, we just don't see a viable process to add single cancer tests together. Again, if you look at where they're going is that the 20% of -- the cancer that cost 20% of deaths, which is the standard-of-care screening, so colorectal has been the big one that they've rotated towards. Most of those single cancer tests focus on really high sensitivity, but also with that comes with penalty of high false positive rates and low positive predictive values. And so if you combine -- sorry, adding 1, 2, 3, 4, 5 of those tests, you get to a point where you have false positive rates that are probably in the 50% range. So we're looking at many, many cancers. And so we just don't see a feasible pathway to get to a practical test in the world. So on our side, the multi-cancer side, we recognized years ago that you could detect cancer from circulating tumor DNA. We did a large unbiased study to figure out the best way to detect cancer from the blood, hence we looked at chromosomal abnormalities. We looked at mutation [indiscernible] fragment length and importantly, we looked at methylation. And found that a targeted methylation panel gave us the highest sensitivity, and importantly, with that also gave us the ability to have a cancer signal of origin. So if you're detecting many different -- potentially many different cancers, you need that cancer signal of origins in order to be a practical diagnostic to be able to guide the follow-on diagnostic effort. So with our technology, we feel that we can deliver on the multi-cancer promise. We think that we can look at that, the cancer that causes the 80% in addition to 20%. And importantly, we think [indiscernible] cancer thing, the most important clinical measures are positive predictive value and yield. And so we've been able to show in our studies that we have a positive predictive value north of 40%. And then the yield is large because you're looking across all of -- many of the cancers. And so in our PATHFINDER study, where we showed both that we had a positive predictive value north of 40%. But we also show that, that combined with standard-of-care screening, more than doubled the number of cancers found compared to the standard-of-care screening alone. So just the sheer amount of cancer we're able to find, we think, is what's going to really make a public health impact, and that's the large driver of our focus on multi-cancer versus swinging at the single cancer.

Got it. I just want to double-click on that point you just made on PPVs because I do think there's a lot of like confusion around it as people sort of compare and contrast across assays. Why do you think comparing PPVs is the right approach for an MCED test?

Yes, Well, if you think about with the MCED test, so start with a single cancer. So if you look at like a Cologuard test. If you do that and you get a positive, you reflex, the reflex is relatively easier. reflex the colonoscopy and you can determine very quickly, but you know it's single organ, single reflex test. Second, in multi-cancer, we have 21 cancer signals of origin. And so if you don't have a high hit rate, which is your positive predictive value. If you have a positive, how often is it actually positive? If that's a relative low number, you're doing a lot of diagnostic procedures to get -- define cancers, whereas with a high positive predictive value, we're almost 50%. It's almost 50-50 in terms of -- we say there's a positive there, you're going to find cancer. And so we think, for a multi-cancer test where you have a lot of modalities where you might have to go look at, we think the high positive predictive value is critical.

Got it. One of the other, I guess, trade-offs in a sense of multi-cancer screening versus single cancer approaches is in any single cancer, the single cancer test generally have better performance than an MCED, right? So there is that trade-off that you're going after multiple cancers at one go. Does that become a challenge to uptake of the test just given how physicians think and how guideline bodies might approach MCEDs?

Yes. So I think they are very different. And so if you look at, as I said, single cancer tests are focused on high sensitivity, but with that comes high false positive rates. With multi-cancer, again, by getting high positive predictive value and high yield, we're able to generate a lot of diagnostic value within that. And so we think that driving that multi-cancer pathway has a lot of value. And when you get the sensitivity, one of the amazing things that's remarkable with Galleri, at 99.5% specificity when looking at the cancers that cause 2/3 of the cancer deaths, we have a 70% Stage II sensitivity. And so it's not like you're trading really high sensitivity for really poor sensitivity, you're trading high sensitivity for good sensitivity, but across a broad set of cancers.

Got it. Fair enough. So I want to just dig into the reorder rate a little bit. Through the first half of this year, more than 215,000 commercial Galleri tests ordered across about 11,000 health care providers. Could you just provide a breakdown of the volume coming through to date in terms of the channel, the health care networks, the self-insured employers, life insurance, health pay, et cetera, as well as the reorder rate?

Yes, I'll give you some sense. So on the self-pay side, so self-pay for us is primarily clinical and clinicians, and then we have some patient initiated telemedicine. And that runs quarter-to-quarter about 2/3 to 3/4 of the volume. So the majority of the volume is in that area. And then what we call enterprise, so large self-insured employers, life insurance companies, some small payers and health systems makes up the other 25% to 33%. From an annual -- from a testing perspective, testing interval, we've done a lot of significant modeling and all our modeling suggests to date that annual testing is going to be the right optimal interval. When we get through our NHS-Galleri study across 140,000 people in the U.K., we'll get good data on that annual testing to get more data behind our modeling on that. And then from a reorder rate, we're just starting to collect some of the reorder rate. We've been out there for 3 years now. And so we're starting to get some of that data and not prepared to share details of it, but we are seeing reorder rates increase. And we do suspect -- or do expect to see differences by channel in the reorder rate as we go forward. So we probably have more than that to come later.

Got it. PATHFINDER 2, you recently completed enrollment in that sort of 35,000 patient study, which will be filed along with NHS-Galleri for your FDA PMA submission. Could you just remind us of the key objectives for PATHFINDER 2, the primary endpoints and what we learn at the end of that study?

Sure. So PATHFINDER 2 is focused on performance and safety. And so on the safety side, the test either generates a positive or negative result. And so if it generates a positive result, then we'll look at the number of -- the number and types of invasive procedures. And then if it generates a negative result, we'll look at the, what's the adherence to standard-of-care screening for people who received the negative. That's really on the safety side. On the performance side, we'll look at positive predictive value, negative predictive value and the specificity of the test. And then we expect, in 2025, we'll have the readout for that. We'll see the readout on the first 25,000 people on the safety and performance elements.

Got it. And that's in the second half of 2025, the interim readout?

Yes.

Okay, got it. On the NHS side, you've completed the final study visits for the 140,000 patients in the NHS-Galleri trial. Similar to the question on PATHFINDER, can you just provide an overview of the study design and the key data that we should be watching for in that study? Will you also present any interim data from that study ahead of the final readout in the first half of '26?

Sure. So the NHS-Galleri study is a randomized controlled interventional study across 140,000 people in England. In the study design, it's longitudinal. So there's a baseline draw off followed by one year later another blood draw and followed by another year later a third blood draw. So 3 blood draws spanning 2 years with a year follow-up after the last blood draw. As I said, it's a randomized controlled trial. So in the control arm, we're banking those samples for further evaluation later on. We'll look at each of the time points later. In the intervention arm, in the Galleri arm, we're, again, looking at if you get a negative in that area, then we'll just have you come back for your next year's blood draw. If you get a positive, you get referred to the National Health Systems rapid diagnostic pathway for valuation of workup. From a performance metric standpoint, so the key measure there is a clinical utility measure, which is a reduction in Stage III and IV cancers in the intervention arm, the Galleri arm versus the control arm. There's also an element where we're looking at reduction in Stage IV cancers, again, preparing the 2 arms. And then from a performance standpoint, we're looking for performance and safety. On the performance side, we're looking at positive predictive value, CSO accuracy and specificity.

Got it.

And then from a readout standpoint, so for that main objective of the trial, you really need to go through the entire trial, not just the prevalent round or first round. So we won't see any of -- any readout until the full readout in 2026.

Got it. One of the questions we've gotten recently is just given some of the critiques in the British Medical Journal article and so on, there's just a lot of noise around NHS and their relationship with GRAIL. Has Galleri now become a bit of a political hot potato in the U.K. where, irrespective of how the results from the trial readout, the broader rollout is just unlikely and complicated?

I think the NHS has had a long history of being very data-driven. They have incredible rigor that's -- I described the study. It's a large -- very large study, high rigor on the endpoints. And so I think that's really what guides them. There's a lot of -- always a lot of noise around it, but the -- I think the rigor of the study and the science is what guides their decision-making. And then what's also useful is because they've had that demonstrated rigor, the rest of the world looks to their studies as well. So when we have the end result of that, that data will be useful as we go and talk with other countries around the globe just because of the reputation of the rigor.

Got it. You're also planning to release some data at ESMO next week, I think. So could you just give us a preview of what we should expect there?

Sure. There'll be a few things there. So one is we'll show some additional data in our clinical pathway workups. We'll have some health economic modeling that we'll show. And then in the precision oncology space, one of our biopharma partners will present some data in the precision oncology world.

Got it. I want to get to stage shift, right? So one of the ways that screening approaches have shown clinical utility state shift, which is something you're hoping to show in the NHS trial as well. First, where is the field today in terms of accepting stage shift in lieu of long-term survival benefit, which obviously takes longer to demonstrate and then entails costly trials?

Yes. So I think there's more and more buying, so if you look at mortality versus stage shift, I think there's more and more buying, and that stage shift is a good proxy for it, and certainly what the NHS is relying on in this. And I think people just recognize the realities of the time duration of mortality studies versus the rate of technological change. So if you wait for a mortality study to finish, the technology has changed over 2 or 3 times since then. And so people are looking for rigorous end points, but ones that are actually also more timely. And I think stage shift is a reasonable way of doing it.

Got it. So I guess just a corollary to that. But just a couple of years of follow-up, do you think the NHS data would be sufficiently mature to see a statistically significant difference between the control arm and the intervention arm? And if you don't achieve that stats say, difference with 2, 3 years of follow-up, are there other end points that you could point to that also demonstrate Galleri benefit?

Yes. So one of the things we think we do have to wait for the full study results. We don't expect to see it in any interim looks at the data, which we're not going to do because the study is blinded, so we wouldn't be able to do it anyways. But we do expect to be able to see that at the end of the study. In either event, whether we see it or not, there is just an enormous wealth of data. We're going to have 140,000 people, had over 3 rounds of Galleri. And so again, when we think about the most important clinical measures of positive predictive value and yield between the NHS-Galleri study, between PATHFINDER 2, which will be 35,000 people, we'll have data -- great data on the performance across 175,000 people. So we think that will be a really rich data set.

Got it. Going back to that earlier point you were making, Bob, about early-stage sensitivity, right? That's been a point of pushback in the KOL community around sort of the MCED approach, the limited ability to detect early-stage cancers, I think even your sort of prostate data that you showed at AACR, I think the number was in the single-digit range, right? And then you compare that to some of these FDA-approved blood-based sort of CRC screening tests, those are sort of in that 55%, 65% range for Stage I disease. So is Galleri's early-stage cancer sensitivity sufficient enough to justify the benefits versus the cost or risks?

Yes. So there's a couple of different things. One, if you use prostate as an example, one of the things that paper is showing is that what you don't want to detect is indolent cancers. And so one of the things that is really good with Galleri and circulating tumor-based test is that the amount is related to the aggressiveness and all the things we're finding we have a paper published on showing that being as relevant to stage. And so what you don't want to do is detect cancers that you're going to die with rather than from. And so that was part of that paper. If you look at the goals of single cancer tests, as you said, they're geared towards high sensitivity. But with that, you get high false positive rates and low positive predictive values. And we contrast that with MCEDs, MCED is really important since you're looking for a broad array of cancers in, again, against all the cancers. You want to make sure that, again, your specificity is very high. We have a false positive rate of less than 1%, which is really critical when you think about the public health implications of it. And again, we just talked about it before, but the positive predictive value, when you do get a positive in it, when you have a high percentage of those actually being people presenting with cancer. We look at -- so we look at it as the positive predictive value and the yield. And again, when we add our single test, our 1 test to standard-of-care screening in PATHFINDER, we more than doubled the number of cancers found. So I think that's -- from a public health standpoint, that's really the big element.

Got it. Going back to the PMA filing, per your discussions with FDA, what is the benchmark that they're going to use or the key performance criteria that they will be looking for in PATHFINDER 2 and the NHS-Galleri readout to grant approval?

Yes. So we anticipate being the first MCED to go through the whole PMA process and do the final filing. We're in process of a modular submission. And because we're the first, the FDA doesn't have any precedent and they have no external guidance that they're showing right now on MCED requirements. So we've really -- with our breakthrough designation, we've been in discussions with the FDA over the past several years. And we're planning on submitting the final module with the clinical validation data coming from PATHFINDER 2 and the NHS-Galleri. So it will be performance, safety and clinical utility data across 175,000 people.

Got it. Turning to the CMS statute, the Medicare MCED Screening Coverage Act, it's working its way through Congress. If passed, runs the ability for CMS to cover Galleri once you have FDA approval. Can you just provide an update on where that bill is in the review cycle? And what the next steps are? How does the whole election cycle dynamic sort of change that sort of time line?

Yes. So we're really thankful that a broad coalition of stakeholders brought forth the bill. It has Bipartisan, Bicameral support. It's one of the few things that really has Bipartisan support these days. And literally everybody, almost every adult has someone close to them that's been impacted by cancer. That's no different when you get to Congress. And so I think a lot of them also have a personal sense that this is an important thing to do. And so with the recent lack of progress or lack of productivity of Congress in terms of getting legislation through, it's difficult to predict the exact timing. But the recent markup of the bill in the House Ways and Means Committee got through the markup and then had a unanimous vote. So it's really -- relatively unusual to have a unanimous vote on pretty much anything. So that's a really encouraging sign. And then we're just -- we're hopeful that the bill will pass before we get FDA approval because that's really the gate that we would need it for.

Got it. In case it doesn't, right, given all the sort of logjam that we see in D.C. these days, what happens in that scenario? Obviously, you've got the USPSTF pathway following FDA approval. How quickly following FDA approval could the USPSTF review Galleri? And just as a point of clarification, assuming receipt of Grade A or B from USPSTF, will CMS then automatically issue an NCD?

Yes. So it's a great question. So the USPSTF pathway is relatively unclear from a timing perspective. So it's difficult to comment on the exact timing for that. But upon FDA approval, we would engage with them to bring Galleri forward and try to get either Grade A or Grade B determination. With a Grade A, Grade B, then CMS would have the authority to pay for MCED, for Galleri in that case. It would not though convey a national coverage decision, that will be a separate decision that CMS would have to go through.

Got it. The Galleri 2.0 version, could you just provide a progress update on where that launch stands? What drives your confidence that a narrower panel will be noninferior to the current version?

Yes. So a few things. So as we look at the next version of Galleri, we've designed from the beginning, this is a population scale screen. And so we recognize that to really hit population scale, we need more scalability as well as lower COGS. And that's -- so that's what the next version is really geared on. We really used a flywheel effect of the data that we've generated. So we've generated over 400,000 samples, about a little over half is on the commercial side and about half on the research side. And we use that information to generate this more optimized panel for the next version. So we believe we've run a large number of samples at this point through the new version. So we still have to show noninferiority, but we'd be able -- we believe we'll be able to show noninferiority. And then from a COGS perspective, we expect to see a step down in the variable COGS. If you look at the variable and fixed COGS, the variable COGS will use less sequencing space. So we expect variable COGS improvement there. To get the fixed COGS leverage because that's a system really built for scale, we're going to have to wait until we get to scalability. So with reimbursement and broader scale, we should see the COGS move favorably in that at that time.

Got it. And Bob, can you just dimension how much of COGS improvement you'd hope to see with 2.0 versus the $500 for the current version? And at what scale?

Yes. So we really haven't gotten into the details. We think it will be significant. And just to give you some sense, to get the scalability, we've nearly fully automated the entire assay. And so with that large amount of automation, we do expect, again, with scale to get to a good COGS point.

What about the bridging studies needed to support PMA approval for the 2.0 version? How big does that trial need to be? And would you likely need an outcomes type of end point?

Yes. So we don't think so. So we're running Galleri right now and the clinical validation for that with CCGA and PATHFINDER. We're also -- now we're running PATHFINDER 2 and the NHS-Galleri study, but in parallel, we're developing this new version of the test. So we're in discussions with the FDA based on our breakthrough designation to understand the exact bridging studies. So we're really finalizing what are the key elements of that bridging study.

Turning to just the follow-up workup needed in case an MCED has a positive signal. One of the key concerns that KOLs often bring up is the ability for a primary care physician to follow up on a positive result in an appropriate manner, right? In a health care system where we're pretty reliant on specialists to confirm cancer diagnosis, how difficult has it been to follow up on the positive cases so far for Galleri? And conversely, how long are physicians following up a positive signal in order to confidently deem them as true positives versus false positives?

Yes. So when we look at the PATHFINDER, so Galleri test returns either a negative or a positive. If it returns a positive, it also generates a cancer signal of origin. And so that's very useful in the diagnostic workup. Also upon a positive, our medical science liaison will reach out to the ordering provider. And so we have a lot of experience Galleri ordering providers at this point. So there, it's usually pretty simple. If it's a relatively new Galleri ordering provider, we'll often connect them with more experienced providers that have just been through it. We also provide guidelines for the workup. So we have 21 cancer signals of origin. So we have kind of standard workups for each of those cancer signal of origin. So today, it really hasn't been an issue in terms of physicians being able to follow up. And then on the timing, it really varies by physicians. Some physicians look -- kind of do the first test and then we'll call it false positive. Others will dig deeper. We do offer -- in a case where the initial workup doesn't show a cancer, we offer a follow-on free Galleri test to be able to do a reconfirmation of that to help in the diagnostic workup.

I see. Got it. So the NCI has multiple ongoing efforts to answer questions around MCEDs, including collection of blood samples to establish a reference standard for all the different tests that will come on the market eventually. They're also running a randomized study called Vanguard, just to look at performance of these assays in head-to-head fashion. Can you just elaborate on the extent of your involvement in those efforts?

Sure. So first of all, we think the NCI has done a lot of work to progress towards their MCED research goals. We really value the work that -- what they've been doing. And we have a lot of shared goals around pushing the MCED field forward from a research perspective. When you look at the actual studies, we've already done our CCGA study, PATHFINDER. We're well through PATHFINDER 2 and NHS-Galleri. And so with that, we've amassed a large database already. And that database is roughly where the NCI will look to be in about 5 years. So we didn't feel it made sense to participate in the Vanguard study. But going forward, we will continue to deeply collaborate with them on our shared goals around MCED.

Got it. I want to spend a little bit of time on the restructuring. Could you just elaborate on the streamline of commercial sales channel? I know you had pretty robust physician support for Galleri. Have you cut back on that service as well?

No. So we wanted to make sure as we go through that, there's no patient impact. And so we size the business to make sure that the -- there's kind of a ratio of sales to medical science liaison. So that piece, we didn't touch because again, the patient side of this has -- and patient experience has to be good with this and the provider experience.

Got it. And walk us through the thought process and scaling back on that employer and life insurance channel.

So we look broadly, really a commercial -- our overall commercial footprint. And over the past several years being in the market, we've learned a lot of things, what works, what doesn't work. And so we've just really cut back in the areas that weren't as productive. And so a lot of those areas are productive. So we're staying engaged in those, but the areas we're less productive, so for example, we would more blanket the U.S., and now we're much more concentrated. So we're not trying to chase after every sale, but trying to go where it's more efficient, especially in the non-reimbursed space to get sales.

Got it. And then on DAC and MRD, you've decided to substantially scale back on your investments there. But could we see you perhaps partner with another vendor to commercialize those assays down the road, particularly in MRD? That is a pure -- and now opportunity with a large TAM and there's broadening support from the payers as well. So why do you prioritize it?

Yes. So we think we have a lot of capability in the MRD space, particularly so in prognosis and in MRD itself. And our DAC product, we think if you look at the SYMPLIFY study, we showed pretty amazing data to be able to really take people who had actually a suspicion of cancer so they were symptomatic and really divide them into people who are very likely to have cancer, people who're very unlikely to have cancer. And so the work there was really, really strong. When we look at our business, though, and you look at Galleri compared to all those other elements, Galleri really stands out as the breakthrough opportunity. And so we wanted to make sure that we had the runway -- the financial runway to get us through to some of the key inflection points for Galleri, and based on that, we had to deprioritize both MRD and DAC.

Got it. Biopharma partnerships that does remain a priority for investment for you. Can you elaborate on what sort of revenue generating work you envision doing here?

Yes. So if you look last year, we generated roughly $17 million of revenue in the biopharma space. We're doing work. We've been public with our work with AstraZeneca. So a lot of work with them. We've had other biopharma partners also in the pipeline. And so when we're looking at the prioritization work, those partnerships generate revenue and gross margin for us. So we have a little bit of profitability there. So it made sense to continue on with that work.

Got it. You've talked about having a really rich internal data lake based on your CCGA efforts. Can you elaborate on whether this is multimodal data that is paired with EMRs and outcomes as well?

So it depends on the study, but some of the studies have very strong match databases. So we think that's one of the key advantages at GRAIL as we've generated over the years, just a tremendous database of both Galleri results, but also with the matched medical histories. And so we'll continue with that.

Got it. When we met Josh earlier in the summer, he said, the purpose of data lake has helped with internal test development and improvement. But there are examples of companies like Tempus, for example, that have looked to license that data, have done so successfully to drive pharma drug discovery. This could potentially also solve your near-term top-line problem as you bridge to that PMA outcome and then commercialization. Walk us through the strategic thinking behind that.

Yes. So we try to find, or try to see how much value we can really generate with the data. To date, as Josh mentioned earlier, we found using our -- the data internally has been the highest value generator. And in fact, the next version of the assay is really -- really comes directly from that data work. But over time, we will look and see if there are any opportunities to utilize the data and be able to monetize data in a responsible way.

Got it. So you're not sort of ruling out the possibility of working with pharma companies in that capacity. Is that fair?

That's fair.

Got it. As you think through that cash runway on the restructuring, you talked of extending it out to 2028 at this point. Can you just walk us through the underlying assumptions there?

Yes. So we did a long -- we always do a long-range plan. We then updated our long-range plan looking at a number of things. The things we talked about deprioritizing. If you look at commercial, for example, in the early days -- when we're not reimbursed, that's really an investment. And so we recognize, and I think we've talked publicly about the fact that we view that's an investment we can modulate. And so we did the work to understand how much we could invest, how much sales we could still generate. And the reality is we can still grow well with the modulated down commercial sales force, but we just won't be able to grow as fast as we kind of originally inspired to. So we looked at that and then we looked at things like MRD, DAC and with the deprioritization of those and taking the funding out for those, we were able to take a runway from the 2026 time frame into 2028. And we thought that was critical to get through against some of these key inflection points.

Fair enough. Last question, Bob. What do you think is the most underappreciated aspect of the GRAIL story today?

It's probably just multi-cancer early detection. I mean just the general concept. I think it's new, it's novel. The medical community, the people who have really dug in and understood it, I think we have a lot of -- we call them cancer enthusiasts. I mean people who have really been able to dig in and find cancer in their patients, and the patients are -- obviously have generated many, many amazing stories in that. But I think, if you think more broadly, it's a new concept and new concepts take a period of time to really digest, understand, to go through all the ramifications that we've been talking about today. And I think the more knowledge is our friend in this. And so the more people understand what MCED can do, I think the better off we all are.

Got it. That's all we have the time for. Thank you so much for joining me.