Transcript: Helmy Eltoukhy at Vator Splash Health

Splash Talk Helmy Eltoukhy (Founder & CEO, Guardant Health)

At Vator Splash Health 2016,Helmy Eltoukhy (Founder & CEO, Guardant Health) tackled cancer with technology together with Bambi Francisco Roizen. Here's a transcript as well as the video version. 

[Introductory music]

Francisco: -- guess I’ll clarify that. Tell me, Guardant Health has raised nearly $200 million. Correct?                

Eltoukhy: Yeah. I think it’ll be over 200 million when we’re done. This last round was over a hundred million.

Francisco: Yeah, clearly impressive and especially what you’re doing. Just to familiarize the audience with Guardant Health, let me take a whack at this. Guardant Health has created Guardant 360, which is essentially a simple blood test for monitoring cancer in late stage cancer patients. Basically, it’s liquid biopsy.

Eltoukhy: Yeah, exactly.

Francisco: Liquid bipsy has been around since 1950. We’ve talked about this before. Why in the last five years has it become so much more popular? What’s happening that’s enabling you to raise $200 million to get this out into the market?

Eltoukhy: Yeah, well first off, thank you for having me here, Bambi. It’s great to be a part of this event. Let me backtrack a little bit. There have been different bio-markers that have been explored for looking at liquid biopsies and the concept behind a liquid biopsy is essentially trying to get a lot of the same genomic and sometimes, pro-geonic information you would from the tissue tumor itself from actually cutting a piece of the tumor, from a more non-invasive means, basically, sampling fluids, either urine or blood and so on. We look at blood.

The idea is that what’s happened in the last five years is that there’s been a confluence of different technologies, different extraction methodologies and different bio-informatic tools that have really enabled us to get much more sensitive and much more specific in terms of seeing these trace fragments of either DNA or proteins that are contained within a tube of blood.

The whole idea is that any tissue in cancer or in a body that’s growing rapidly is also turning over very rapidly and so, when those cells turn over, they die and they release some contents into the blood stream. Until recently, until we came to launch Guardant 360, off the self DNA sequencing technology were not sensitive enough to see these trace fragments of DNA in a tube of blood.

And so, we took a lot of the same concepts behind HDTV or rabbit-ear television, black and white television or DSL over dial-up modem, which uses the same transmission lines but through advanced digital communication, are able to get a hundred to a thousand times better error rate or better speed. And so, we used that same concept on existing DNA sequencing technology. It’s really a marriage between hi-tech and advanced molecular biology that’s enabled us.

Francisco: Well, you’re background is in genome sequencing, you’ve spent a lot of time there so talk about how the prices have really gone down to be able to actually sequence the human genome.

Eltoukhy: I started working at Stanford in the Genome Technology Center at the tail end of the Human Genome project and we won some of the first few grants for trying to build massively parallel DNA sequencer. I think it was first $100,000 genome grant and was a $10,000 genome grant now. I think in 2013, we’re in the area of $1,000 genome. The first human genome costs, depending on estimates, somewhere between a few hundred million dollars to $3 billion dollars.

Francisco: When was that?

Eltoukhy: That was 2001, essentially, just about 15 years ago. We’ve had this enormous decrease in cost per base.

Francisco: What will it be in five years?

Eltoukhy: I think $1,000 is the right packet size. I think our definition of the human genome will change so for that $1,000 or a few hundred dollars, we’ll get a lot more. The human genome we get today, I call it “ground beef” genome. It’s really like 91% coverage. It’s small little snippets. We don’t actually know the difference between the mother’s side and the father’s side. With higher resolution, there are areas in the center of chromosomes to the end of chromosomes that we’re not able to really map effectively.

And so, what will happen is there are technologies that are coming around the corner that allow us to get a more complete picture. The other concept that we’re exploring is the beauty of our blood tests and being able to go in non-invasively in these patients to monitor cancer is that cancer changes. It’s not static, essentially. Cancer is defined by alterations in the genome, essentially. It’s these mutations that happen in the three billion bases that we sequence inside of the cancer cell that really defines the cancer.

But not only is it different over time, it’s different spatially as well so it’s important to get a complete picture. I do believe that same concept. We’re going to learn more about the genome itself and we’re going to find that many diseases are caused by essentially de-synchrony in the different cells and the genomes. We have trillions of different cells in out body. They all have conceivably the same three billion bases, but we know that’s not the case as you grow older. There are errors that happen in different tissues.

I think our definition of the whole human genome sequencing is going to be much more complete than it is today.

Francisco: Right, I’m sure you’ll expand. Right now, you’re focused on cancer, late-stage cancer patients. Right now, you’re looking to displace all the tissue biopsies out there. What you’re offering right now is a blood draw, which is non-invasive. It’s also 25% of what an invasive surgery would cost. What about the accuracy of the findings?

Eltoukhy: We’ve done concordant studies. We’ve taken tissue biopsies from cancer patients across almost a dozen different cancer types, compared them to match blood draws. We have 99.3% diagnostic accuracy. It’s very accurate. When we find something, it’s a specific –

Francisco: It’s accurate with the current tissue biopsies?

Eltoukhy: Yeah, exactly. And so the idea is we’re not trying to replace the tissue biopsy at diagnosis. It’s going to be very hard to replace it. It’s going to be a 10 to 15 year process before the technology – and our ability to let go of physically seeing cells infiltrating tissue in a normal tissue and so on, doing this histological stains to identify what type of cancer it is, that’s going to take some time to displace and the technology has to evolve a little bit for us to be able to displace it, but we’re basically seeing that we’re displacing the repeat biopsies or when genomic information wasn’t able to be extracted from that initial biopsy.

And so, rather than subjecting patients to $10, $14,000 invasive procedures with 20% complication rates, we can come in with a simple blood draw, two teaspoons of blood, get that same information with a 99.3% accuracy and then, hopefully get that patient on to a tried therapy or another kind of therapy that hopefully is much more effective than they would have received.

Francisco: Right. Well, I want to talk about, in terms of the market and how may you’re already doing and how much tissue biopsies you’re displacing, but you talked about therapy. I’ve read an article about a  woman who in June 2013 had a liquid biopsy and the test found out a genetic mutation that suggested her disease would be treatable by a Novartis drug, Afinitor, yet that drug didn’t work for her. So, when you say 99% accuracy, does that mean that it falls into the 1% that didn’t work?

Eltoukhy: No. It was accurate in terms of identifying the mutation, but you have to remember, even when potentially actionable mutation is found, drugs have a certain efficacy rate, and so, even some of the blockbuster drugs that have just been improved have anywhere between 50-60% efficacy in terms of response rate, in terms of when that mutation is on, even tissue. Previous generations of drugs have between 20-40% response rates, and so, unfortunately, we improved the odds quite a bit, but I think the missing piece is we don’t have enough weapons in the armament in terms of drug effectiveness and so on.

That’s why this is a tool that is not only helping the clinical side of getting patients on to potentially more effective therapies, but we need to speed up the rate of how many therapies are getting approved and how effective they are. This tool is being used by pharmaceutical companies to do just that.

Francisco: So the genetic mutations that you identified, they’re only the ones that have a therapy or a treatment that can go – I mean, you couldn’t really say to somebody that you have XYZ and there’s no treatment for it. So you’re only looking for certain mutations that actually have a treatment that can –

Eltoukhy: We actually look a little more broadly than that, so there are two classes of mutations we see. We report everything that we find that’s in the tumor and obviously, the ones that the physician finds useful are the ones that are currently actionable that have a drug that they can get that patient on, that that patient is a candidate for. The mutations that aren’t actionable, today, are now a signature for that patient’s cancer, a personalized bio-marker by which you can follow that patient over time.

And so, one thing we’re finding with this test is we’re much more effective than other methods for monitoring progression, such as CT scans, serum protein bio-markers like PSA or CBA. We’re actually eight to 12 weeks past in terms of clinical response. We believe this is a much more general tool of combining the what of the CT scan, it just tells you if the cancer is going back or is going away, and the why of biopsy where it really tells you what’s going on in a molecular level. And so, with a single test, a single blood draw, we can combine both sides and give the physician a complete view.

Francisco: Right, and you can discover it sooner because you can diagnose it much quicker than traditional. That’s what you’re saying – than traditional procedures.

Eltoukhy: Yes. Our time to treatment can be much shorter than the traditional biopsy-based approach and so, we now have multiple publications where we see that we cut in half the time it takes for a patient to actually get on to a target therapy over biopsy-based approaches.

Francisco: Right, I want to open this up to the audience because there are two – wow, we already have an eager participant. Okay, go for it. There’s a microphone over there, so -- [pause] he has really intelligent questions.

Audience member: What do you see your technology being useful for screening for cancers that are typically not found until Stage Four such as pancreatic cancer?

Eltoukhy: Yeah, that’s a great question. When we started the company, it was our goal and I have to say that we have versions of the technology now that are, almost in order of magnitude, more sensitive than what we’re using for late stage cancer today where we’re seeing some fantastic results in just those cancer types. When you think about it, ovarian, lung, pancreatic, those are cancer types that are typically diagnosed late stage. We don’t have bio markers and protein bio markers that are very effective there, so if we can essentially dial up the sensitivity of this technology and give those patients and the physicians a shot of detecting them in Stage One or Two, where hopefully, they’re more treatable, then it’s going to be fantastic and it’s where we’re going in the very near term, actually.

Francisco: Before we get to this, which cancer does your technology work most effectively?

Eltoukhy: We look now at {inaudible} – wow –

Francisco: That was to wake us up.

Eltoukhy: Yeah, I know. We looked at maybe 50 different cancer types – {inaudible} wow –

Francisco: I think that was Crystal.

Eltoukhy: [laughs] We looked at 50 cancer types. I think the universality of this technology is that it’s based on a property that really defines cancer, which is it grows fast and so, any tissue that’s growing fast is also turning over. The one area where we see less sensitivity are brain cancers, blastomas. Our sensitivity, instead of being in the 80, 90% range, is down at the 30, 40% level because of the blood-brain barrier that prevents DNA that’s being shed from circulating in the flow of blood.

Francisco: But you’re placing a lot more –

Eltoukhy: Yeah, if you look at our basically rate or sequencing, I believe blastoma, it’s much higher than the incidence of glioblastomas which is only 1.4% and that’s because the alternative of doing a biopsy is pretty gruesome there in the brain biopsies. I think there are applications, you could be looking at cerebral spinal fluid, CSF, and sample that and get much higher sensitivity. But all in all, about 95% of cancers, we’re able to see with this technology with very high sensitivity.

Audience member: Yes. Since tumors evolve and progress and change their patterns, therefore, you’re biopsies are going to return to a heterogeneous population. How does the clinician use that information in an actionable form, because you’re going to get data that’s going to have a different display of markers, etcetera? How do you make that actionable?

Eltoukhy: No, that’s a great question. When you take a typical tissue biopsy, you’re looking at one microscopic section of the tumor. You’re not getting a complete picture of the whole cancerous systemic view of what’s going on in these patients. I’ll give you just maybe a few examples of real patients. We had a patient actually in Israel that we tested about a year ago. We found basically an EGFR deletion in that patient. The patient was put in a [indiscernible 00:15:23], had an amazing response, and then about a year later, as many of these patients unfortunately do, he relapsed, but relapsed very suddenly and his oxygenation levels went down to 85%, lungs were failing.

They told him that, “This is happening so quickly, why don’t you put your affairs in order?” So he flew to Romania to buy a gravesite right next to his sister and they had done another Guardant 360 test. When he came back, the results were ready. We found a med amplification now that was not present initially. Crizotinib is a drug that works for that so he’s put in Crizotinib, we got the PET CTs back a couple of weeks ago and the patient is basically at performance stat of zero which means they already given a clean bill of health, they can go back to work and so on.

It’s pretty fantastic to see this new era of multiple shots on goal. That’s the temporal evolution aspect where you can come back and resample and figure out what the next best shot that patient has. There are other examples where we had another patient who had an (ALK) Fusion and so the patient was put on Crizotinib again and then, there’s a mutation that basically causes a resistance to Crizotinib that we detected.

When we tested again, the patient was put on Seratinib. And then, we tested the patient again after they progressed and then, they had another mutation that was resistant to Seratinib, but actually re-sensitizes the Crizotinib. So the patient was put back to Crizotinib and they started working.

But the beauty of the test is we see a superposition of all the mutations there and it’s quantitative. When you look at the results of that patient, you see the original driver at the highest percentage and then you see these follow-on resistance mutations at lower and lower percentages. It’s really a fundamentally new tool of seeing a complete picture of these cancers and really designing what the next best step is. A part of what we do is we’re building a database, one of the largest databases of metastatic disease. Because it’s quantitative, we’re able to build what we call ways for cancer. Can you predict the road map that these cancers are likely to take and put these patients in treatment pathways that hopefully avoid some of the traffic jams down the road?

Robinson: Hello, Jaime Robinson, professor at Berkeley. It wasn’t clear from your comments how far you’re beyond a non-investigation, clinical use in the United States. Assuming you are. Could you briefly discuss the regulatory pathway you went through at the FDA to get in comparison with the pathological markers and assuming you got through that, where you are in terms of reimbursement, coding at Medicare and at/or the private plants?

Eltoukhy: Sure, it’s a great question. The test has been commercial now and been in clinical use for almost two years. It’s an LDT so it’s gone the Clea-Kappa route. We are in discussion with the FDA right now, because we work with a lot of pharmaceutical companies in terms of them building their companion diagnostics on top of Guardant 360 so that patients can have better access to advanced drugs without having to go through a biopsy.

Obviously, the test has been one of the fastest adopted tests in oncology history. We went from zero to 20,000 tests a year in under 18 months. We have seen almost 20% of oncologists having used the tests now and growing. We sequence the most lung cancer patients of any center in the United States or any company in the United States through blood biopsies, and that’s even in comparison to people who do it through tissues so we’re just seeing amazing sweet spots in terms of lung cancer and being able to get patients to the standard of care that they deserve.

In terms of, obviously, at a a diagnostic lab, the most important thing is reimbursement and getting paid for the test. With private care, we work with patient's insurance to essentially to recoup as much as we can on their behalf and we’ve had pretty good success there. With Medicare, we’re still in discussions with our MAC, basically, Palmetto and Uridion, those conversations are going really well. I think something that’s important that I found is this concept of DRIP, data rich information poor. We somehow conflate information and data with utility and our test is very targeted.

We’re not trying to boil the ocean and find every possible marker that’s there, which may be interesting from a research setting. We’re really going after bringing that patient to the guidelines that the NCCN have dictated. And so, we’ve done studies where we found that something like only 50% of lung cancer patients have EGFR and (ALK) which are the two basic bio-markers that are there. There’s seven bio-markers that are recommended by the NCCN. EGFR and (ALK) are two of them.

If you go to the next five, it’s only about 10 – 15% of patients that have had those markers genotyped. The beauty of our test is that it’s the only liquid biopsy on the market today that’s able to get all seven, everything that’s dictated or recommended by the NCCN with two teaspoons of blood. And so, our value proposition is essentially not trying to boil the oceans, just getting those under-genotyped patients up to the standard of care that they deserve, whether they’re at Memorial Stone Kettering or in a community oncology center in the middle of the country.

Francisco: Now, your technology really applies to the one million late-stage cancer patients. When do we get to a point where you can address the 15 million cancer survivors or you can get a blood draw if you’re healthy and get early detection of a potential cancer?

Eltoukhy: Yes, so we have active programs in both those applications and the results are very promising there, but we’ve had some use-cases where the test has been used for breast cancer survivors and so on, and we’ve found some amazing results where we’re able to detect occurrence six months before any other means. And so, our test essentially allow that physician to be very vigilant because we started seeing markers coming back up, and as soon as – one patient in Philadelphia who – we essentially found a mutation that had come up and then three months later, they did a scan and they found a small metastases on her spine.

They were able to remove it, but you never know if that eventually metastasizes to the brain or really becomes untreatable at that point. And so, all those patients have been very grateful for this technology and for this opportunity of active surveillance that we can afford their physicians and themselves.

Francisco: I’m sorry. So, you said that you ran a test, a blood draw, and you were able to detect breast cancer six months earlier than an alternative test?

Eltoukhy: Then a CT scan or a PET CT, yes.

Francisco: Okay, and that was pretty accurate?

Eltoukhy: No, it was. We’ve had now a number of these cases in the research setting where these tests are being used in clinical trials. Obviously, we’re not going to find every case before other methods, but if we can at least provide an additional safety net for those patients and give them quantitative peace of mind, as we call it, then it’s much better than what exists today.

Francisco: So, you raised nearly $200 million, certainly have a war chest, about a $100 million is relatively new, I think it was earlier this year, but there are other companies out there, BioSET, went public in 2014, Radiance Technologies, this is a pretty big market and in some ways they can validate what you’re doing and/or could be a huge market and we need all of the groups, all of the companies to satisfy the demand out there, but how do you stay ahead?

Eltoukhy: That’s a great question. I think if someone had told me three years ago that two years after we launched our tests, we would still be the only company with a comprehensive liquid biopsy in the market, I would be very surprised. There’s a lot of noise out there. You look at Genome Web, you look at the other publications, it seems like there’s a new liquid biopsy company coming out every year and for good measure. All the JP Morgan, Piper Jeffrey, all the banks are estimating this to be a $20-$40 billion market opportunity. Everyone wants to get a piece of it.

We estimate we have a 100% of the comprehensive liquid biopsy today and at least 80-90% of the overall liquid biopsy market, so there’s a lot of what appears to be noise in the media, but not a lot of actual tests that are in the market yet.

Francisco: You’re creating a lot of that noise, by the way. You publish all the time. Tim, you’re doing a great job. Tim, the publicist. So, last question, there’s a lot of entrepreneurs out here. I think the healthcare industry is one of those industries where there’s a lot of challenges as a startup. What was the biggest challenge for you? What can the audience take away as they’re building their companies?

Eltoukhy: I think it comes back to, especially when – I’m an engineer by original training and I think what’s great about Silicon Valley here is we’re finally seeing a lot of tech minded folks thinking about healthcare and jumping into the space, but it’s not just about technology. If you look at most of the successful healthcare companies, technology has probably been the thing they’re the least competent at. It’s really five or six different things that have to go right. As someone brought up, it’s the reimbursement, it’s the regulatory, it’s the commercial, it’s dealing with medical affairs and clinical utility, all these studies that you have to run and technology is one of five or six different segments that are important.

It’s really having a holistic picture there and I think the second thing is humility. I don’t think that this company would’ve been successful if I didn’t fail multiple times during my PhD, doing different ideas and different things. There’s a certain amount of hubris sometimes that people from the tech space have in terms of going into healthcare. It’s good, we should have more ideas, borrow as much as we can from the tech side to solve some of these important healthcare challenges, but these challenges exist because they’re hard challenges, and so we should remember that cancer is not something that is going to go away any time soon.

We’re going to move the needle, we’re going to make things better, but we should understand that the more we find out about cancer, the less we actually know. This isn’t even Physics. 93% of the universe is dark matter and dark energy which we don’t have any clue what it is. And so, it’s going to be the same thing with cancer. Maybe the final kind of take home message is what I call PPP, make sure you have the patient, physician and pair value proposition in mind in alignment. They’re both from a utility aspect and from an economic aspect.

There are a lot of tests that are good, but if they’re going to take money off of the hands of either the payer or the physician and so on, then it’s going to be hard. You’re basically moving uphill there, so it’s important to have that holistic view about where your technology fits.

Francisco: Great final lesson. Helmy, thank you so much.

Eltoukhy: Yeah, thank you.

Audience: [claps]

[END]

Thanks to our main sponsors KPMG,Javelin Venture Partners, SAP Startup Focus, Bread and Butter Wine,Artis Coffee,Scrubbed, Stratpoint and Healthiest.

Editor's Note: Our annual Vator Splash Spring 2016 conference is around the corner on May 12, 2016 at the historic Scottish Rite Center in Oakland. Speakers include Nigel Eccles (CEO & Co-founder, FanDuel, Andy Dunn (CEO,Bonobos), Libby Schaaf (Mayor, City of Oakland), Mitch Kapor (Founder, Kapor Center for Social Impact), Oisin Hanrahan (Co-Founder & CEO, Handy), and more. Join us! REGISTER HERE.