00:33:00 · Helmy Eltoukhy, Mark Jacobstein
Autoimmune, they don't change the genome, but they do change the epigenome. And so that's what's been exciting about marrying up, now having this early detection test where we're testing hundreds of thousands and soon millions of people a year that fortunately very few percentage have cancer, but all those individuals have many other diseases. And because we've built this architecture where we can see the claims data, we can see the outcomes data of almost every patient we test in a de identified sort of manner, we can now throw that into all the AI machine learning sort of scripts we've created, and we have labeled data now. So we can say, okay, these ten thousand people are in HUMIRA. They have rheumatoid arthritis. Are there any signals there we can sort of like, and so we're actually doing all of that in parallel now in a massive way, and we're seeing exquisite signals for some diseases that essentially would just be a software update on top of the existing tests that we have. [00:34:06] S1: That's amazing. And what's so interesting is that a lot of those disease areas have never had any kind of diagnostic. It's like they put you on Humira until it doesn't work. And [00:34:17] S0: there's [00:34:17] S1: not a biomarker, there's no understanding really, I think, of the mechanisms of action and why somebody is a likely responder or not, and presumably therapeutics will fall out of that also. I don't know if that's on the roadmap or if you can talk about that.
00:34:17 · Mark Jacobstein, Helmy Eltoukhy
[00:34:17] S1: not a biomarker, there's no understanding really, I think, of the mechanisms of action and why somebody is a likely responder or not, and presumably therapeutics will fall out of that also. I don't know if that's on the roadmap or if you can talk about that. [00:34:31] S0: Yeah, so I think the idea is like, we've talked about going, like the next thing is essentially thinking about non oncology applications, basically using our blood test as a sort of everything test to detect hundreds, if not thousands of diseases. We have about 500 labeled diseases in our database now that we're seeing, so it's really exciting in terms of like, I'm as excited now as I was 14 ago just because we can do some pretty massive things potentially. Then I think what's exciting is working with pharma in non oncology applications because there really is a lack of precision medicine in most areas and if you can essentially build these surrogate endpoints where you can measure disease more quantitatively at a much faster scale, you can decrease the cycles of learning and improve the progress considerably.
00:35:23 · Mark Jacobstein, Steven
[00:35:23] S1: There's an interesting macro lesson here for everybody in the room sort of thinking about AI, is all these things scale with data and the challenge is how do you get your data and Guardant has managed to sort of unlock a way to actually not only get the data, get paid to develop the data, which certainly makes it an awful lot easier. All right, well, thank you. I'm going to open it up now to other questions, but I would be remiss. So for those of you who don't, Steven is a South Park Commons member. He's also a former Guardant employee. I think you wrote the first bioinformatic pipeline that Guardant ever had, but also always had the first question at every all hands meeting at Guardant. So I'm very happy to have Stephen ask the first question. No pressure. [00:36:03] S2: So one of the things that I took away from Guardant Health was that we had the core value of putting patients first, and that's always been important to the company, it's been important to me. One of the things that happens as companies mature is you have the opportunity to democratize access to technology. How are you guys thinking about that?
00:36:19 · Helmy Eltoukhy
[00:36:19] S0: Yeah. It's a good question. We've always made sure that no matter which technology we launch, that everyone has access to it, so we think about Medicaid patients, we think about those that don't have the ability to pay. It's also why I think there have been different business models for commercializing sort of multi cancer early detection tests, and we never really wanted to go the sort of concierge kind of route, because bending mortality curves means testing everyone and testing the most disadvantaged, And that's really where, like, I think we're very proud with the sort of strategy we have with Shield where we have Medicare coverage, we just got UnitedHealthcare coverage, so now we have about 70,000,070 million of 120,000,000 eligible now with coverage, and hopefully over the next few years we'll get to one hundred percent sort of coverage. So we just want to make these tests as accessible as possible, and then when we think about it globally, I think this is where we have the opportunity, and I think we have the first deflationary sort of healthcare innovation in the field in the sense that our tests, like three sixty today, is 100 times more powerful than it was ten years ago. It has 100 times a bigger payload, it's more sensitive, yet it's about the same cost. Same thing with Shield, we launched it with just colorectal cancer. We're now detecting 10 different cancers. Soon if we add other diseases, we'll be detecting other things, and all for around the same cost. So I see a future where could you get to a test, yeah, let's say it's around $1,000 but instead of detecting one thing, it's detecting $1,000 and you can get to a point where it's $1 per diagnosis and we can essentially take a page out of tech leapfrogging where essentially cell phones leapfrog to Africa in the sense and sort of replaced wired infrastructure, and could you do the same thing here where we can get to a sort of per capita healthcare GDP sort of price point that actually makes sense for the system because it's doing so much?
00:38:39 · Mark Jacobstein, S3
[00:38:39] S1: Thanks. All right. Who's next? I see a hand. [00:38:45] S3: We have a similar background, E major, undergrad, masters and PhD in information theory. So during my PhD, got heavily influenced by John Wheeler's It From Bit thesis. And now comes the question of what part of biology is conquerable through the instrumentation of or the project of computation itself. What part of biology can be conquered and what answers can be answered using the abstractions afforded by the philosophy of computation, right? So there, I'm curious from your vantage point, do you see that this compute maximalist or compute reductionist attitude can take us further? Because I'm trying to understand, like if isomorphic lapse, which now claims that if biology and hence disease is mod level and computable and you can help climb on GPUs, then where is the residue from a company such as yours coming from? What is it that you can solve using your data assets and your compounding flywheels that a computation reductionist research lab like Isomorphic cannot?
00:39:54 · Helmy Eltoukhy
[00:39:54] S0: Yes, I don't know much. I mean, I don't want to speak sort of out of turn in terms of what they're doing, what they're not doing. Obviously, a lot of smart people there, a lot of funding and so on, but just from seeing other sort of what exists right now in terms of the world of biology, there aren't amazing data. I mean, are some data sets. Some of it was, I'll give you an example. We detect things down at one in a thousand, some of the newer tests, one in ten thousand, at lower levels. Imagine you said that I tested a million people and I detected at levels of one in ten or one in one hundred. So the claims all sound the same. I have a million blood samples and so on, but you may not be seeing everything that's there because you're not actually seeing what's unique, like the unique biology that's being captured by some of those rare variants. It's the same thing. Like a lot of the data that's out there that's RNA is they're doing like 10,000 transcripts per cell. But it turns out like, you know, most of that is capturing housekeeping things and so on. It's not actually capturing the low level sort of transcripts and so on that actually differentiate that. So, the devil's in the details in terms of like, and I think we've seen this, there's the Gartner hype curve, right, of peak of inflated expectations and science because we create a new tool, it allows us to see something that we haven't seen before, and we make the assumption that we're seeing 100% of what's there with that tool, when in reality we're still under sampling 1000X from what's actually there. Yeah, we improved the 10X, but unfortunately there's 10000X sort of necessary. And so the question is there'll be advancements there, but it's unlikely to be like solving all of human disease either because we're still heavily under sampling. I think our sort of unique advantage is that we're sequencing more and more data. Like, we have a data collection engine. Next year we'll be testing probably a million sort of individuals across different disease types. We're getting all that data. We're improving the payload of the tests, so sequencing goes from 3,000,000,000 to $1,000 to $100 to $10 to $1 Our tests will be collecting 10x, 100x, 1000x more data per test. And so we have sort of all of that working in our advantage.