Personalis, Inc. (PSNL) Earnings Call Transcript & Summary

Thanks for joining us at the Morgan Stanley Global Healthcare Conference. I'm Yuko Oku, an associate on the life science tools and diagnostics team. It's a pleasure to have Personalis join us today, and representing the company is John West, CEO, Director Co-Founder and President; and CFO, Aaron Tachibana. Before we begin for important disclosures, please see Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. Okay. With -- all right, so to begin, for those that might be new to the Personalis story, John, could you give us a high-level overview of the company and the NeXT platform?

Yes. Be happy to. Thanks very much for having us here today. I'd say just to give you a perspective on it, my whole career has been involved in the DNA sequencing field. It's improved by a factor of 1 billion during that period of time. I used to run the sequencing business that applied Biosystems in the early 2000s when we sequenced the human genome for the first time and moved in 2004 to a little British company named Solexa that had a new technology for next-generation sequencing, is ultimately acquired by Illumina and has become the DNA sequencing technology that's used around the whole world today. We started Personalis in 2011 after I left Illumina, not to make a new sequencing instrument to be able to say, what's the unmet medical need here? How can we actually use this to make an important difference from a medical standpoint and what's the business opportunity behind that? And we've ended up focusing that on oncology. You'll see our NeXT platform is a leading platform in that regard. And Personalis also has a population sequencing business currently working with the largest population sequencing effort in the United States, which is won by the U.S. Department of Veterans Affairs. Personalis is one of the largest sequencing laboratories in the world. We estimate that last year, about 4% of all DNA that was sequenced anywhere in the world was sequenced by Personalis at our lab in Menlo Park. So we operate at really quite large scale, maybe partly because of the background in having develop next-gen sequencing. We've been very optimistic of what the future of that would be, and we've developed products that were often some years ahead of where the sequencing technology was with the understanding that the sequencing technology would catch up and we'd be in a good place with the right products and customer relationships and intellectual property to take advantage of that. So our NeXT platform for sequencing tumors, for example, whereas other groups looked at perhaps a few hundred genes, Personalis looks at all human genes. We look at all 20,000 human genes. So it's an enormously larger data set. But we felt that this was important, particularly for the next-generation cancer drugs in the category of immuno-oncology because classic cancer drugs attack the tumor. In immuno-oncology, the drug doesn't actually attack the tumor. The drug enables the immune system to attack the tumor. And the mechanism for that actually involves mutations that could be an anagen of the genome. So we developed our NeXT platform to be able to see that. And we also developed proprietary laboratory data and informatics that lets us interpret that in a way that nobody else could even if they could sequence the same things. So with that, I'd say that the company went public actually, with Morgan Stanley's help. And back in 2019, we've been growing quite a bit. I think the year before we went public, we were at $37 million in revenue. I think our guidance for this year is $85 million, so it's sort of more than doubled in that period of time, continue to experience a lot of growth in the company, and we've been able to grow every quarter compared to the prior quarter now for over 20 quarters. So all 3 of the pandemic, Personalis never had a revenue dip. We were able to manage our business so that we were able to keep the lab open. Where almost every genome center of the world was shut down, Personalis stayed open. We sequenced over 50,000 human genomes last year. We have now sequenced more human genomes than any company in the United States. So we've sequenced over 125,000 whole human genomes as not only a laboratory scale but a data scale that is almost unmatched in the country. So maybe with that I'll stop and be happy to take your questions.

Thank you for that great overview. Maybe starting with MVP, with expectations for task orders under the existing contract to complete by end of September and into October, have you heard from MVP if they would add an additional task order to the existing contracts? Given that it's mid-September already, is there a chance that you don't get an incremental task order in the next few weeks, creating an air pocket in the MVP contribution until they run an RFP process at some point down the road?

Yes. So we continue to be confident that there will be a new task order. In general, this year, there normally would have been a complete RFP process to renew the contract. The VA has been sort of ground 0 for a lot of things to do with COVID. They were the site of many of the clinical trials or the vaccines. They have an elderly patient population of 9 million patients. Like literally, they've been vaccinating millions of patients. And when we talk with the people involved, they're very committed in the long term to the Million Veterans Program. They were thrilled that we were able to keep up data production all through COVID. And they said, "we don't want to change anything." We'd be happy to have you keep doing that. But frankly, they didn't have the time to put towards in the RFP process. And so we realized it would actually be possible for them to continue just by giving some additional task order, I'd say are all indications are that, that will happen. We don't know what the scale will be. Their budget has probably been turned inside out by COVID. On one level, they've had enormously larger expenses because of having so many COVID patients in all these clinical trials. On the other hand, Congress gave them an extra $19 billion to help pay for all that. So it's hard to know where that ends up. I think we'll just have to see what comes up. Usually, we hear, right, at the end of September or early October because they have to wait until the very end of the fiscal year to know what's the total amount of money they have available. So we'll hear in the next couple of weeks here what that looks like. I think the key thing has been that our cancer business has been growing so rapidly. We talked on our last conference call that I think we were in early August, we already had a dollar value of orders in Q3 that was substantially higher than our revenue in all of Q2. So with the oncology business has really been growing really fantastically. And so right now, that's probably the biggest thing we're focusing on.

All right. Well, maybe that's a great segue into talking about the biopharma business. Near term, would you comment on potential risks that rising COVID cases could slow down clinical trial recruitment in 4Q? And would you also comment on approximate split between prospective and retrospective work?

Happy to. Yes, so we have seen some slowdown in clinical trials, not just recently, but even as early as sort of Q2 of last year, there were groups where we knew they would be starting clinical trials, and the trials got shut down. We have seen some of the trials that were halted or that were -- didn't start last year finally actually begin to start. It is different. It depends on what part of the world the trial has been running. We tend to, I think, here, be very focused on the situation in the United States. But many of the clinical trials that are run are run in places in Europe or in Asia, and so where then the cycles of COVID have been different in those different geographies. I'd say the last 3 quarters in a row, we've reported revenue in the pharmaceutical side of our business has been up over 70% relative to the year earlier quarter. I'd say if it hadn't been for COVID, it would have been bigger than 70%. But we're still pretty happy with 70%. And we have been seeing very strong orders performance. And to the second part of your question, we have been saying in the early years, as people adopted our NeXT platform, they often initially use it to analyze samples that they already had. So these were retrospective analysis of samples from clinical trials that had already completed. What we're pretty happy with now is that, that has led to a degree of buy-in where pharmaceutical companies are now designing us into their clinical trials from the beginning. And so that leads to more in the way of prospective words, meaning that we are penciled in for -- not penciled in, we're mandated for all of the samples that will come from the trial right from the beginning, but it means that we'll receive them as the trial happens without waiting for the end. So it means that it will take some period of time for those samples to arrive. But as we build a bigger and bigger backlog of those because we've been getting orders that are so far above revenue, that in time will lead to a stream of revenue that comes from each one of those clinical trials. And so I think that will hold us in good stead as we go into 2022. We have trials that we already know we'll still be processing in 2023 and so forth. So I think it gives us a lot smoother revenue and a base. Usually, if you have a retrospective business, you process that business in the year that you got the order. And you start the next year and you have to go get a whole bunch of new orders. In our case, we'll be going into 2022, for example, with a substantial portion of our -- the revenue for 2022 already in backlog. So I think with that kind of -- we're building a base that we'll be growing on top of, so we see it as being a very positive outcome.

Great. And then I wanted to touch on the NeXT platform. What feedback are you hearing from adopters of NeXT platform? With additional orders following initial data delivery from pilot projects, what are the type of studies that are underway? And could you provide some examples of how customers are using it in practice?

Yes. So in general, the customers that we have are using this for analyzing patient samples from clinical trials. This is not mostly discovery research. This is mostly real life clinical trials on human patients. Generally, we are not used for inclusion and exclusion criteria. We're usually used to understand why certain patients responded and other patients did not respond. But that's the role in pharma. In general, when we start working with a pharma company, they often begin with a pilot order. Years ago, those would have been small. We've often had pilot orders that were $50,000 or that kind of thing, and it could take a long time for the whole process to go forward. I'd say, starting -- maybe during last year, we started to see pharma that understood the power of the platform and where the initial orders could be hundreds of thousands of dollars or even $1 million at the beginning. As we've been seeing good results at -- with the clinical trials at many of those pharma, we see them broadening out. So if our platform was being used for testing on drug and they have good results, that word gets around inside that pharma company and other drugs, other groups inside the same pharmaceutical company that may be testing other drugs say, well, we should use that in our drug development program. And so we often start at 1 level, but then we kind of build out and broaden out our footprint in those pharma. We are now working with the majority of oncology pharmaceutical companies. Some of them are further along. Some of them are really quite large customers for us. Some are at the earlier stages but working on ramping up the penetration. I'd say the next phase of growth in those is also moving from a single time point, like sequencing tumor biopsy, to something where we're getting orders now for sequencing at many time points on each cancer patient. So for example, on our second quarter call, we highlighted a multimillion dollar order we received from a biotech company for a group of clinical trials they were planning where they were going to be using multiple time points of both tissue biopsy and multiple time points of liquid biopsy and on both liquid and the tissue looking at all 20,000 chains over time. So it's a fair number of time points. It's really leveraging the good interplay that you have between there's things you can only see in tissue, there are things you can only see on the liquid biopsy. Because we have both, we have the ability to do a really comprehensive job of it. I think we're really the only company in that space. So I think that will be a pretty nice trend going into 2022 as we begin to see more and more pharma companies adopt this combination of tissue and liquid biopsy.

Maybe just following up on that. Could you elaborate a little bit more on what you can see perhaps on the tissue side versus the liquid side? I mean I understand liquid, you can -- there's an optionality to do the serial samples, but why not just use all liquid, for example?

Sure. Well, so part of what we -- our customers are interested in looking at is the RNA. And so the DNA, obviously, you can see the cell-free data, and that's what a lot of the cell-free platforms including ours look at. But to have a comprehensive view of what's going on inside of cancer, you need to see which genes are being expressed and how much and whether that's going up or down. And so the RNA typically would get from a tissue sample. In addition, you'd like to understand the immune cell repertoire. And so only a few cancer platforms actually look at the immune cell repertoire. And most of them that do only look at the immune cell repertoire. We realize that you actually need to look at both what's happening on the cancer side, which is why we sequence all 20,000 genes, but you also need to see the complete immune repertoire. And the immune cells that are inside of a tumor are different from the ones that are in circulation. So if you look at with a blood sample, there's lost of immunes cells. And all the white cells are mostly immune cells, but they're not the ones that are inside the tumor. And so in order to be able to see the immune response to the cancer, which in immuno-oncology is central to understanding whether the drug is working or not, you need to have a physical sample of the tissue so you can get out the immune cells that actually had made their way into the tumor. So the ability to see the RNA, the ability to see the immune cells is crucial from that. But then obviously, as you put it much, if you want multiple time points, it's possible with tissue. We have people who are doing that, but it's much more practical with a liquid biopsy sample. And so the -- in pharma, our approach all along has been, and we now see increasing buy-in of this approach, that you really want to use the combination that there are certain things you can see in the tissue you can't see in the liquid, but it's -- but you can probably get more ability to track the mutations over time with liquid biopsy. So by being a company that's expert at both, we have a capability to do that. Some companies have only ever done tissue and never developed the liquid biopsy capability. Other companies started really focused on only doing liquid biopsy and they end up missing the things that you can only see if you have the tissue. So Personalis gives the capability to do both.

Okay. Great. So moving on to the MRD recurrence monitoring solution NeXT Personal, expected to be available later this year, would you describe how this product is differentiated from existing MRD recurrence monitoring tests out there? And then what are the key attributes that you focused on when designing NeXT Personal? And why choose to take the tumor-informed approach versus tumor uninformed approach? And what are the potential advantages and disadvantage just of each approach in your view?

Yes. So it's a great question. So the approach we're taking is a tumor-informed approach. We think that it's the only way to achieve the sensitivity goals that we have. I think we started by looking at the scale of the market. The American Cancer Society has done a nice job of quantifying the scale of that. And they report that in the United States, there's something over 17 million people who are considered cancer survivors. And if you look across that data, you see that almost half of them are people who have survived either breast cancer or prostate cancer. So if you look at those tumors, those are cancer types that tend to be detected early, thankfully. And since the mid-1980s in the United States, most breast cancers are detected with mammography, and typically at stage 1. So the amount of DNA that's being shed by the tumor is quite low at stage 1. The goal is surgical resection. Almost very few people die of a primary breast cancer. Usually, it's the recurrence of the cancer that is the serious medical problem. And so the question is, can you detect that, that cancer is coming back or not? And so we looked at if you had a tumor that was that small to start with, it's just a Stage 1 tumor, typically it might be about a centimeter across. So it's really small. It's not shedding much DNA. We looked at the data, and we said, that DNA in the blood plasma would only maybe about 1 part per million after the surgical resection. None of the tests that we saw in the market could come anywhere close to that. We said, what would it take to look at 1 part per million, because that's really the market is people are starting with other things. They're starting with colorectal cancer and other areas where the shedding is higher and so forth. But those aren't the biggest parts. The biggest part of the market are these cancer types, particularly breast and prostate, which have low mutational burden, low shedding rates. They're detected early. So as we said, we have to design something that can get to the sort of part per million range. And when you look at the DNA and the blood plasma, there are only a few thousand copies of the genome. That's all the molecules there are. So for each place in the genome that you look, there's only a few thousand molecules. So you can't possibly get to a part per million if you've only got a few thousand molecules. The way to get around that, you can't sequence more deeply because you just sequence the same molecule again and again. You're not seeing anything new. The way to see more is to look at more places in the genome. And because these cancer types have such a low mutational burden, we said, if you want to see 1,000 places in the genome, which is kind of what we expect is needed, if you see 1,000 places in the genome and you've got maybe 3,000 copies of the genome at each one, now you've got 3 million molecules. If you can detect it, now you have the possibility of actually seeing 1 part per million. And we think that's the goal. So you need to look at 1,000 or several thousand locations in the genome. On these cancer types, the only way to find that is with the whole genome. And so Personalis has enormous expertise in whole genome sequencing. We have sequenced over 125,000 whole human genomes with our population sequencing effort. We've also been sequencing cancer genomes now for an extended period of time by being able to bring that capability into the cancer arena. It gives us the potential even with low mutational burden cancers to be able to see 1,000 or more mutations and then to be able to aggregate the signal across those. So we've developed NeXT Personal to be something that will be a test specifically for each patient. And the goal is to be able to be sensitive enough, not to detect the cancer eventually. Ideally, we would like to be able to see the cancer, if possible, within a few weeks right after the surgery. And that will be the challenge. We expect to show that. But the product has been designed with that in mind. We think that this will be an enormous issue for the insurance companies because within the last year, starting last December, the FDA has begun to approve a number of cancer drugs, which had previously only been approved for late-stage cancer. They have now been being approved for patients who have had surgical resection of early-stage cancers. So this is terrific for the cancer patients because the drugs really do make an enormous difference. The difficulty for the payers is it almost doubles the number of people who may be qualifying for those drugs. And at $150,000 per patient per year, this is a pretty serious issue if you're an insurance company. And particularly, many of those are patients that don't have cancer. The surgeon got it. There's nothing left. And the idea that you would be giving the patient an extremely expensive drug that may make them sick, for a patient who no longer has cancer seems kind of crazy. And so what people want to have is a test that can be more definitive in saying who really is going to have cancer coming back or not. And you can't say there could be a cancer that comes back 5 or 10 years later, and it's down at such a low level that nobody can detect it. But at least for people where the recurrence might happen in the first couple of years, which is most recurrences, you'd like to be in a position to be able to say can we be sensitive enough that we could actually detect it within a few weeks after the surgery. And so that's what we've been aiming at developing. We think that will be transformational for medicine. And we think the payers will think it's a pretty good thing too because, otherwise, the FDA has approved giving these drugs if somebody -- if you have somebody who has just had cancer surgery, they're very likely to want to go ahead and say, well, the FDA says I should do this, I should do that. That's pretty expensive for the payers. And it's actually not good for patients because a lot of the patients will be getting a drug they don't need.

Makes sense. And I also wanted to touch on the NEOPS. With the recent publication describing NEOPS for predicting response to cancer immunotherapy in late-stage melanoma, how would you describe the value of the composite biomarkers for IO response monitoring? How do you see incorporation of NEOPS into clinical use over time?

Right. So we see this as something -- this leverages our understanding of what are called neoantigens. These are something you can only see if you look at all 20,000 genes in the genome. And the mutations that occur in each gene create the mutated DNA. It can lead to a mutated protein. The mutated protein looks like it doesn't belong in your body, and the immune system can pick up on that and attack the tumor. But only about 1% of mutations actually cause an immune response. And in general, nobody has known which ones are the 1%. And so we undertook a research project starting, I want to say, just began looking -- working on this back in 2015, we said the only way to generate the data to figure that out is to actually genetically engineer a set of cell lines so you can actually put specific mutations in specific, what are called, HLA types into the cells and then you can measure with a mass spectrometer which ones are being -- would be likely to issue an immune response. Nobody had that data. It's quite a challenging program that genetic engineering was complicated. In the early days, when we first genetically engineered the cells, they all died. So you couldn't get any data, so we had to get past that. We eventually solve that problem. We were able to keep the cell lines alive. We were able to get the mass spec data, train the neural network algorithms. So it's been quite a multidisciplinary, actually nothing to do with DNA sequencing. This is all required for interpreting that. But now that we have that, and now we have a peer review publication on that and so forth, this would now let you begin to say -- I mean checkpoint inhibitors are fantastic drugs for many people with solid tumors but not for all patients who have it. Unfortunately, many patients do not respond to those drugs, or they begin responding and then there are escape mechanisms that the tumor uses to get away from the immune system. And so unless you have a biomarker that can help you understand which category your patient is in, the assumption will be, well, if you have melanoma, it's approved by the FDA. That's probably the drug you're going to get. But not all melanoma patients will respond. And so making sure that patients do get the right drug is going to require the right biomarker. We will work with pharmaceutical companies who are interested in that. We anticipate that in time, our NeXT Dx test will be able to offer this. We expect to have the only diagnostic test that will be able to have this kind of a biomarker because we're the only ones with that, all the background and data for that. And we think that this may be very helpful in deciding which patients should have checkpoint inhibitor drug versus other opportunities, other therapies that may be available to them.

So with -- while the focus near term is going to be on biopharma applications for NeXT Dx and NeXT Personal, what is your current plans for moving them into the clinic? In light of the reimbursement momentum that we're seeing in these settings, what do you want to achieve in the RUO setting and before bringing them into the clinic?

Yes. So almost all of our business with pharmaceutical companies is under what's called GCLP, which would be similar to CLIA lab developed test. Our Lab has been a CLIA and licensed and cap accredited lab since at least 2014. So we have a lot of the pieces of that. Our expectation is that we will begin -- we have a version of our NeXT platform we call NeXT Dx, which actually comes with a complete clinical report. That's been set up specifically for working with pharmaceutical companies to begin with, but we're interested in beginning to bring that to clinical practice. And I think we would begin with high-end cancer centers there because they typically have the hardest cases. The more comprehensive information we can provide is more useful in those cases. And they're also often have a dual mandate, both to provide clinical care for today's patients but also to do the research that will save future patients. And so we have the ideal platform for that kind of thing. So I think going forward, you should expect to see us working with a number of high-end cancer centers to bring this kind of technology to bear both for their research but also for clinical care. We are encouraged by the improvement in the reimbursement market. It clearly takes work. But compared to where we were 5 years ago, that's an enormously better situation. I think previously, it was unclear what the path was. Now I think the CMS, in particular, with the Multi X program has enabled pathways that make it practical for companies to bring breakthrough technologies to reimbursement relatively quickly.

Okay. And in particular, given NeXT Personal is -- going back to NeXT Personal. Given that you are taking a more broader approach using whole exome sequencing, how do you ensure keeping prices competitive in the clinic against existing MRD tests?

Yes. So the -- we actually would use whole genome upfront for our MRD test. We expect to be able to be quite competitive. And part of the reason is that since we started Personalis, we've always been looking to the future where DNA sequencing was going. And as a result, if you came to visit our laboratory today, almost everything we sequence is on Illumina's highest end sequencing platform, the NovaSeq 6000, almost all of it is done on Illumina's largest flow cell, which they call their S4 flow cell. This also has the lowest cost of sequencing of any of the Illumina platforms. We also anticipate that the cost of sequencing will continue to come down. Illumina themselves, as part of their FTC filings, has said that they expect their cost of sequencing to come down by 43%. I think it's by the year 2025. Most of my career has been in the DNA sequencing field. I would say that's probably a very good conservative estimate. I'm anticipating the cost of sequencing will come down even more than that. People have talked for some time about the $100 genome. I think it's not far away. And we've been developing the products that can actually take advantage of that. So we've been building products that are larger and more comprehensive and require a lot more sequencing than most of the competitors in this space, which, based on yesterday's cost of sequencing, would make us noncompetitive. Based on tomorrow's cost of sequencing, we'll be fine, and we expect to have higher performance products, taking advantage of the reduced cost of sequencing in the future.

In the last few minutes here, could you summarize what you're most excited about wrapping up '21 and heading into '22?

Yes. I think certainly the growth in the cancer business has been fantastic. We commented on our second quarter call about a very strong order volume. But we've been seeing these strong orders ever since we introduced the NeXT platform. It is gratifying to see it now coming together with about a year ago, we launched the liquid biopsy side of that, which was, again, kind of a forward-leaning product. We were -- everybody else is looking at 100 genes or that kind of thing. We were looking at 20,000. So that wasn't something people were used to. But people have done pilot studies. They're beginning to realize what the power of that is. We're getting -- we talked about it in the early Q3, receiving a multimillion dollar order for clinical trials that we're going to combine both liquid biopsy and tissue together. I think this is the wave of the future with our pharma customers. And now it puts us in a really strong position to grow the clinical diagnostic side of the business. And as I said, you could expect we'll start with some of the high-end cancer centers there that we'll work with them. We'll have more to talk about that, I expect, as we go into 2022. But I think it's going to be a terrific time. Our plan all along has been that the strongest business will be one that combines both diagnostics and the pharmaceutical side. Because of our platform being more advanced, we actually started on the pharma side of that. But I think we now have a really solid foundation for beginning to build out the diagnostic side. And I think that's going to lead to a lot of growth, and that's part of what we've been aiming at all along. So thank you for the time today.

With that, thank you so much, John. Thank you, Aaron. And thank you, everyone, for joining us on the webcast.

Super.

Thank you, Yuko.

Thank you.