Heather Wetzel-The Personalis ACE CancerPlus Test A Comprehensive DNA & RNA Sequencing Solution for

Thanks so much, Judy. Hi everyone, good morning. Thanks so much for joining us. We're excited to talk about our newly launched ACE Cancer Plus Test. Without further ado, I'll go ahead and get started with the presentation. First, I want to go over the learning objectives. The purposes of the presentation are: first, to understand some of the current industry challenges with regard to next-generation sequencing and tumor profiling for clinical care; second, to appreciate how the use of Personalis' Accuracy and Content Enhanced Sequencing platform, or ACE, with DNA and RNA analysis improves precision and reporting in solid tumor testing; and third, to recognize through case studies some of the more common molecular changes that are identified that might assist in therapeutic management. What I hope to cover in today's session is a few of the topics listed here, starting with a review of the high-level overview of the Cancer Plus details, as well as a short background about Personalis as a company. Then we'll transition to talk about some of the current industry challenges that face clinical next-generation sequencing cancer testing, and alongside that discussion, we'll talk about some comprehensive solutions that are offered with the ACE Cancer Plus Test and what we've done to address some of those challenges. We'll follow that up with a couple of case studies, and then in the session, talk about some logistics of the ACE Cancer Plus Test, including some documentation that we have available for patients as well as physicians, and information on ordering the Cancer Plus Test itself. Then, of course, time allowing, we'll have some Q&A at the end. To first start, let's talk a little bit about Personalis and our company. Personalis was founded in 2011. We're a company based out of Mountain View, California, about an hour south of San Francisco. The services that we offer are all built on this ACE platform, so you'll hear us use the term ACE platform quite a bit because everything that we do at Personalis is really built upon this Accuracy and Content Enhanced platform. We have two different arms of services or tests that we offer through Personalis. On one side, we have a number of research services, such as large-scale sequencing analyses for complex disease, Mendelian genetics, and inherited disease research. One of our bigger focuses is cancer research, so we engage with a number of customers for clinical trials, translational research, as well as immuno-oncology, which is one of the hotter topics. What we'll talk about most today is really the clinical side of the products. Those clinical products are what you see listed here: our ACE Clinical Exome Test, which is our test for Mendelian diagnostics and the so-called diagnostic odyssey, and another clinical exome test that we've done through a partnership with Athena Diagnostics called the NeuroXome. For today's presentation, we're really going to talk about tumor profiling, and with tumor profiling, that's our ACE Clinical Exome Test, or our ACE Cancer Plus Test. Just to go over a few higher-level details of the Cancer Plus Test: the test itself, again powered by the ACE platform, consists of reporting of 181 genes of clinical relevance. That test is powered by a larger extended cancer panel of over 1,300 cancer genes that we're interrogating, and that raw data of the 1,300 cancer genes is actually available upon patient consent. However, the actual interpretation and reporting is based on a much smaller subset of those genes, that 181 list. The test includes DNA and RNA analysis of solid tumors, and by doing DNA and RNA analysis, we're able to identify a number of different types of variants, including base substitutions, insertions and deletions, as well as copy number alterations and gene fusions. Those 181 genes that we're reporting are genes that we report based on their relevance to having associated FDA-approved therapies or clinical trials that might be available. By looking at a smaller subset of genes with this technology, we're able to obtain a very high sequencing median depth of approximately 500x, and that really gives us the ability to detect some of these low-frequency variants, which we know are important, particularly in the cancer setting. Some of our early testing using the ACE Cancer Plus Test has been through our relationship with the Department of Veterans Affairs. We've had a very long-standing relationship with the VA. The VA was one of our very early customers in terms of partnering with them to do whole exome and whole genome sequencing as part of their Million Veterans Program. Along with that study, we've engaged with them with another type of test, one being our clinical exome test for diagnostic odyssey, and the other being working with the VA in Boston to pilot our ACE Cancer Plus Test. That's a test that we've been working with them on for a few months now, and we're in the process of a national rollout with the VA. What you see on the slide here is a quote from one of the directors of that program, Dr. L. Fiore, and Dr. Fiore has been very impressed with the findings from the Cancer Plus Test. He has said that the majority of the patients they've tested, which primarily are non-small cell lung cancer patients, have actually had variants that have been associated, oftentimes, with an FDA-approved therapy. They're really using this information to help further develop their precision medicine program and to make some of these drugs more available to veterans nationally. What I want to talk about next is really just reviewing and talking about some of the current challenges that exist in the clinical world in terms of next-generation sequencing cancer testing, and alongside that, we'll talk about some of the solutions that are offered with the ACE Cancer Plus Test. If we think about some of the challenges that are out there, a number of these are probably familiar to you if you're a clinician potentially ordering some of this testing. One of the main challenges we talk about or hear about is having a poorly focused gene test. Some laboratories may have a large number of genes, hundreds of genes on their panel, some may have less than a hundred genes on their panel, but the idea is that oftentimes those panels either have too few or too many genes that aren't clinically focused, which is what you really need in this particular setting. Coverage is also an issue. Next-generation sequencing is notorious for having some areas of the genome that are really difficult to sequence, and that applies to exome sequencing, but of course it also applies to the cancer setting. Limited analysis of gene fusions is also an issue because not a lot of laboratories are doing RNA sequencing as part of their analysis, so the ability to detect gene fusions is somewhat limited and often not available or not able to pick up a large number of fusions or newer novel fusion events. Of course, FFPE challenges we'll go into in more detail. The last couple: assessing platform accuracy and how do we validate the testing that we're offering with the lack of gold standards out there in the field, and then how do we translate all of this information that we have at the laboratory level into a report that's helpful for a clinician to read and is clear and concise. What we've done is really taken some time to address these sorts of issues. These issues are not all-inclusive, but these are some of the ones that we've really taken the time to address with our analysis. We now have what we feel is a more clinically focused gene set, better coverage, we've incorporated DNA and RNA analysis, we've improved our capabilities for FFPE sample analysis, we've validated our test, and we've written and made a report that's really concise and intuitive, built for clinicians. We'll walk through each one of these and give a little bit more information. First, to talk about the gene set itself. I mentioned briefly that it's difficult for laboratories to have a well-focused or clinically focused gene set because there are so many different cancer genes out there, and trying to decide what should go on a clinical cancer test has oftentimes been difficult. The goal at Personalis was to keep the ACE Cancer Plus Test clinically focused. The test is based on a much larger panel, and that raw data is available for individuals looking for translational research, for example, wanting more information about a larger number of genes, but we've really honed in for the clinical reporting of those 181 genes to make sure those are genes that are clinically relevant. The vast majority of them are genes that are associated with active clinical trials or FDA-approved therapies. There is a subset of genes on that list, in our efforts to future-proof what we're doing, that are frequently mutated in cancer but may not have a current therapy associated with them, but perhaps they will in the near future. Some examples of that are what you see listed here: the CDH1 gene, oftentimes thought of as associated with gastric cancer, as well as JAK1 and RB1, just three examples of genes that are also included on that panel. Next, I want to talk a little bit about gene coverage. This is something we talk about quite often in the exome sequencing setting, but this also applies to panel testing and cancer testing as well. What I want to do is walk you through what you're looking at here on this particular diagram. In yellow, these rectangles represent DNA target regions or exons, for example. What you see in blue might be what a standard panel test could pick up. You'll often see that some exons might not be well covered, some may be incomplete, and some may not be covered at all. That's where our supplementation on the sequencing side really comes into play. We have developed in-house very specific chemistries and probe designs that have allowed us to capture a number of these regions that are poorly covered by standard technology. When you're looking at a more focused gene panel, you're able to obtain better sequencing coverage than you would with a whole exome, but oftentimes some of these areas that are difficult to sequence, such as areas where there's high GC content, are still going to be difficult to sequence with a panel. That ACE augmentation really supplements those areas where there's poor coverage. What you end up seeing is a combination of what might standardly be available plus what would be available with the augmentation provided by the ACE platform that builds our products. Those areas that are targeted are what you see listed here. I've mentioned high GC regions, that's one of the more common areas that is supplemented by ACE technology, but we're also looking outside of the exons as well. There are medically important variants that aren't in the exons, so we're also targeting known intronic variants or other biomedically important regions that are outside of the exons, filling in those gaps in those medically important regions. This ACE augmentation is really what builds the platform itself and improves the sequencing coverage for all of our products, including the ACE Cancer Plus Test. To show how that might look when you're thinking about individual genes, I've put here a couple of gene examples of genes that are part of the clinical reporting for the Cancer Plus Test. These are probably genes that might be familiar to you. AKT1, for example, is a gene in cancer that's associated with a large number of clinical trials, and CDKN2A is one of the more frequently mutated genes in cancer, so this is a very common gene for which we find variation on these clinical tests. What you're looking at here along the Y-axis is the depth of coverage, or how many reads that gene or coding base has, and what you're seeing along the X-axis is your coding regions of this particular AKT1 gene. This is really showing how comprehensive the coverage is for these particular genes. These are just two examples; we certainly have examples of all of these genes that are on this panel, but this is really speaking to that comprehensive coverage that is addressed with the ACE platform. We've also published on our technology. We released a publication in August of last year in Genome Medicine. The publication was written by Personalis and co-authored with the NIST Genome in a Bottle Consortium group. What we were looking at was comparing our ACE platform to a number of off-the-shelf or conventional exome sequencing kits, as well as looking at whole genome sequencing. We were looking at that comprehensiveness of coverage, or gene finishing as we call it, and comparing our ACE platform to a number of other different tests or platforms that are out there on the market. What we found is that ACE has very superior coverage and an ability to finish a number of medically relevant genes. The figure that you're looking at here is one of the other areas that we focused on for the publication, and that was looking at the issue of GC content. I mentioned before that high GC content is often a reason why next-generation sequencing actually fails and is unable to sequence certain areas of a gene. What this is really showing is that if you're looking at gene finishing or coverage on the Y-axis and the degree or percent of GC content on the X-axis, ACE in green has a really high level of gene finishing all the way across various levels of GC content. Even at a lower level all the way up to 70-80% GC content, the ACE platform is really able to continue to sequence and cover these difficult regions, whereas some of these other platforms, as you can see, as the GC content starts to increase, the sequencing coverage really starts to fall off significantly, which is the case with some of these exome sequencing platforms and of course with something like whole genome sequencing. What I want to move to next is to talk about gene fusions and how we've addressed the issue or difficulty of accurately assessing gene fusions in clinical cancer testing. I briefly mentioned earlier when we were talking about the high-level details of the Cancer Plus Test that we do include DNA and RNA analysis of those 181 genes. We're not for this test selecting a certain number of genes where there might be common gene fusions or known fusions; we're actually doing full RNA analysis of all of those 181 genes. As a result, we're able to detect gene fusions in all of those 181 genes on the panel, which would allow us to pick up new fusion events. It also allows us to be able to pick up gene fusions in those 181 genes but in those genes and other partner genes that might not be covered on that 181 gene panel. This is a great example of what I'm saying here. ALK, of course, is a common cancer gene and it's associated with non-small cell lung cancer, but this is a very common fusion event involving the EML4 gene, which isn't a known cancer gene. It's not on any sort of common cancer panels, but it is a fusion that would be detected with our ACE Cancer Plus Test because we're including ALK and RNA sequencing to detect that fusion, and we would also pick up the fusion partner of this EML4. What you see here are a number of other examples of some of the gene fusion partners that have known targeted therapies available. Some of these, of course, ALK we've just talked about with non-small cell lung cancer, RET fusions in thyroid cancer, ROS1 fusion events and other mutations with lung cancer, and then the NTRK1 fusions as well. These are just some examples; there are a number of others, but this just gives you a flavor of why really complete RNA analysis is important, particularly in detecting gene fusions, which are oftentimes associated with hematologic malignancies, but we're also really quickly learning that there are a large number of solid tumors that also have gene fusions as well. I also want to talk some about challenges with FFPE. If you're a pathologist ordering this sort of clinical testing, or even an oncologist who might be submitting this testing, oftentimes what happens is you may submit a sample from FFPE tissue because most of the patients who are undergoing this type of clinical testing aren't necessarily patients with a newly diagnosed tumor where you have that fresh frozen tissue that you can submit. Oftentimes, these are individuals who are much further along in their cancer diagnosis, and this might be a third or even fourth tier test after some other treatments might not be working. Often, clinicians are going back into the archives and getting samples from FFPE that have been preserved, and oftentimes what happens is that those samples fail either at the QC step even before sequencing, or we're just not able to get really successful DNA sequencing just because some of the issues with preparing and fixing with FFPE lead to degraded DNA. Rather than just accepting that as a problem in the field, at Personalis we really took a step back and said, let's think of some ways to improve our capabilities with getting good quality data from FFPE, since this is such an important tissue type in cancer. What this is showing here is some of the published data that's been looking at success rates with FFPE. Those studies are showing an average success of around 30%, and this is one of the citations here from Heydt et al. a couple of years ago. What we've seen through our own experiences with FFPE, due to the specific technologies that we've developed in-house to improve our capabilities, is that providing a good tissue sample, we've been able to successfully extract and sequence approximately 90% of the samples that we've received that have been FFPE tissue. The take-home message in this case is that while FFPE tissue samples might be really difficult for some laboratories, we really haven't had any difficulties as compared to those laboratories. While fresh frozen is always the preferred method or preferred sample type, we don't shy away from sequencing any samples that are provided through FFPE tissue. What I'd like to talk about next is a little bit of background about validation. Anyone who has looked at other clinical cancer testing laboratories knows that validation is a really difficult concept because there aren't any gold standards out there for us to really assess our platform accuracy or compare them to. Oftentimes, what you're left doing is establishing your own gold standards in-house and using that for a comparison. What I want to talk about here is just some of the numbers behind our validation of the ACE Cancer Plus Test, keeping in mind that while the Cancer Plus Test is really looking at those approximately 181 genes, the actual testing itself is based on that much larger panel of over 1,300 cancer genes, and that's what the validation is based upon. To actually validate that product or that test, we ended up procuring close to 40 different cell lines to be able to look at different cell lines from different tumors and also looking at various allelic fractions anywhere between 5% all the way up to 95% to have a very comprehensive or robust validation. We were really happy with the results that we found from the validation study. Some of those numbers are listed here. What we found for base substitutions was over 99% sensitivity at approximately as low as a 5% allelic fraction. For insertions and deletions, also over 99% sensitivity at an allelic fraction as low as 10%, and also really great numbers for copy number alterations and gene fusions, and high specificity as well, over 99% for base substitutions and indels. Last but certainly not least, in terms of talking about some of the challenges, I want to also talk some about reporting. Once you've gotten to the point of really improving the sequencing, improving your ability to detect fusions and get good DNA from FFPE tissue, and having that validation, all of that's really important, but at the end of the day, these are analyses that are being reported to a physician who is then going into the room to translate these results to their patient to determine if there might be any clinical trials or therapies that would be available in terms of their treatment program. It's really important that a report that any laboratory provides really makes sense and is as concise as possible. Oftentimes, what we see is that these reports from this sort of clinical cancer testing are often 20 to 30 pages long because there's so much information to list and detail, so it's really hard to hone in on the important parts of a report. At Personalis, we've really taken the time to do that both with our clinical exome test that we do for Mendelian disease, and we've sort of translated that here to our Cancer Plus Test with having a much more concise report. One of the parts that I want to hone in on here, this is just a snapshot of the first page of the report. Obviously, I don't expect you to be able to read every single word on it, but I just want to point out on the first page, of course, you've got your summary of your patient information, and what I've got circled here is a summary of the findings themselves. Right there, kind of in your face on the first page, is a nice summary in this case of the relevant variants that were detected in this particular patient who had a diagnosis of melanoma. In this list, three different variants were detected in three different genes on that first page. What you see on the following page is really a short summary about those particular variants that were detected, just to briefly go over what was found before going into a ton of other detail about the actual variant itself and its association with disease, as well as all the clinical trials. To point out what you're looking at here, this is that short summary that I mentioned of these three different variants that were detected. What we've provided on the report is a table that lists out the availability of therapies that are associated with this particular variation. To walk you through a couple of these, this is the very common BRAF V600K mutation that was detected in this case, again a diagnosis of melanoma, and this is telling you here that there's an FDA-approved therapy that is associated with that particular tumor type. If this were no, then the next part is, is there an FDA-approved therapy that's maybe not available for melanoma but available for a different indication and that may be able to be prescribed off-label, for example. Then, are there any active clinical trials that are available for this particular gene or particular variant? In this case, again with BRAF being a very common variant that's detected in cancer and particularly melanoma, the answer to all of these questions is yes, so you would imagine that on pages following the report, there's probably a pretty long list of some of these FDA-approved therapies and clinical trials. A little bit of a different scenario with something like a PTEN mutation in melanoma where there's not a therapy approved for that particular cancer type, but there are FDA-approved therapies for other cancer types and active clinical trials. Following this summary, there's a lot more information on the following pages, but again, this is really meant to be a quick glance on that first couple of pages of what might be available for these different mutations, particularly when we're reporting more than one variant, it's important to have that summary information up front. Now that we've talked about some of these industry challenges and what we've done at Personalis to address some of these challenges, what I want to talk about next is just a couple of case studies that we've been able to identify from some of our earlier testing and how important this clinical cancer testing can really be. In this particular case, this is a very short, very straightforward case study. The clinical presentation was a 65-year-old male who had been diagnosed with a lung adenocarcinoma. The sample that was submitted to us for analysis was, of course, the popular FFPE with 50% tumor content, and we were able to successfully sequence that particular sample and resulted in an exon 19 in-frame insertion or indel that was found in the EGFR gene. This particular gene is another pretty common cancer gene, and this gene has a number of targeted available therapies from FDA-approved therapies as well as a number of active clinical trials that are associated with that particular variant itself. The second case study is interesting really from a sample preparation and quality standpoint. This was another typical presentation: a 56-year-old male who'd been diagnosed with an esophageal adenocarcinoma, and we ended up having to get a couple of different specimens from this particular patient. The first sample that was submitted to us was from the primary tumor, an FFPE sample, and we had extracted and sequenced a number of different times because we were having trouble at the QC step. We do a QC component even prior to sequencing just to do a check of the quality of the sample itself, and it continued to fail at the QC step for DNA. We ultimately checked back with the laboratory who had prepared the sample, and the specimen itself had reportedly been processed by microwaving by the submitting pathology lab. For a non-pathologist, this was something I had to look into myself because at first I was thinking, microwaving sounds crazy, why would you do that? It's not unheard of in other preparations for processing samples, it's just not something that is great to do in terms of sequencing because it's going to really affect the DNA quality itself, but it's a great method of processing samples for other things like immunohistochemical staining. Anyway, we were unable to sequence with that particular sample, and we had another sample that was submitted later on that was a sample from the metastasis, submitted as fresh frozen tissue. With that sample, we were able to successfully extract and sequence the DNA and RNA. What you see here are some of the results that we had identified in this particular patient: again, that common CDKN2A gene that has associated clinical trials, and an IDH1 splice variant, as well as a TP53 mutation. Still some important findings for this particular patient because this identified some clinical trials that might be available for him. Just a follow-up FYI, we did take a look further at that sample that had failed the QC step, and even though we didn't submit results and put this on a clinical report, we did complete the sequencing for this particular case, and those three variants that we found in the metastatic sample, we did identify them in that sample from the FFPE tissue. So even though the sample had failed the QC metric step, we still were able to have the same results that we had found with the secondary tissue. The last part of the presentation is really just going through some logistics of the Cancer Plus Test. We've prepared a number of different documents in an effort to provide as much information to a clinician that might be needed to really talk about the Cancer Plus Test and talk about tumor profiling with a patient. We have an ACE Cancer Plus Test sheet that provides some of those higher-level details that I've mentioned earlier on. It does include the list of those 181 genes that are part of the clinical test. We also have clinically focused specimen preparation guidelines both for FFPE and fresh frozen, so particularly if you're preparing samples, it gives you guidelines for how to prepare those samples but also how to send those samples to our laboratory to maximize our chances of success about getting reports from these samples. Requisition forms are available and will eventually be available online, but they are part of our sample submission kit that we provide to the clinicians. I have a picture of that specimen transport kit here on the next slide. These are specimen transport kits that are sent to the clinician who is ordering the ACE Cancer Plus Test, and in that kit, all the information that you would need, including the requisition form and ordering form, to be able to submit a sample to the Personalis clinical lab. A little bit more detail in terms of logistics: this is a clinical test, it has to be ordered by a physician. At this point, we're set up for billing as a self-pay through a patient or institutional billing. Right now, the Cancer Plus Test is only available in North America. We do have plans to extend our offering globally, but with an early launch of the test, right now it's in North America only. In terms of what's needed, everything's in that kit: requisition form, transport kit would be mailed to Personalis. The turnaround time for the test is approximately three weeks. Last but certainly not least, we've also developed a patient FAQ document. The FAQ document is really just thinking about some of the more common questions that might come up in a patient setting. This is meant to be something that can be provided to patients to take home and read more about the ACE Cancer Plus Test, or it can be used for a clinician as a guideline of what you might talk about with your patient in terms of ordering this type of testing, how long it takes to get results, what we're actually looking at, and what that information might be able to provide in terms of additional access to clinical trials and therapies. With that, I'll conclude the presentation portion. Thanks a lot for your time today. I've listed here our media contact, Robert Power. That is not me, but Robert is listed here should you have any additional questions that we either don't get to today or questions that you might think of later. Certainly jot down Robert's email and send him a message. If you'd like more information about the test that we offer at Personalis, feel free to visit our website, and we do have a virtual booth here at LabRoots today, so feel free to visit that and ask questions there as well. Thanks so much.

For raw data, I mentioned that we report on those 181 genes and could provide raw data for over 1,300 of those genes on the larger extended panel. This is something that we have a consent form in place that would be provided to you that the clinician and the patient would have to sign to be able to access that raw data. What we provide on the DNA side is a FASTQ file, BAM file, and then the variant annotation report. For RNA analysis, we're providing the FASTQ file as well. That can be submitted either via hard drive or FTP, whatever works best in this particular case.

It's a question that I thought might come up. At this point, we're not accepting liquid biopsies. It is something that we're certainly looking into at Personalis to be able to possibly offer in the future, but at this point, the sample types that we're taking are FFPE and fresh frozen tissue.

As I mentioned before, one of the documents that we provide is an information sheet about the Cancer Plus Test, and on that sheet, the list of genes is actually included there. If you're looking to see if a gene is included on that panel, that would be a good place to look. If you're more asking what are the coverage statistics, is this a gene that's well covered on your panel, you could certainly submit that question to us internally, and we can provide some information about the coverage or maybe even one of those gene coverage plots that I showed earlier in the presentation to really illustrate the sequencing coverage of that gene.

This one's an easy question. We take FFPE and fresh frozen tissue. If at all possible, we love a sample from fresh frozen tissue, but also are very happy to take FFPE.

We certainly do have a sample report. We don't have one posted on our website quite yet, but if you're interested in browsing through the full report and not just the couple of pages that I showed for the presentation, again, certainly let us know, and we're happy to share one with you.

I don't think I have time to go over the whole process here, but just to give a little bit of detail, the process is really this ACE technology, and this is a patented technology. You could certainly read through all the nitty-gritty details if you wanted to read through the patent itself or maybe some light bedtime reading. Essentially, what's happening is we're optimizing some of the probe design and chemistries that are done prior to the sequencing process to really maximize our ability to sequence some of those difficult areas like GC content. That's one of the main areas of focus for obtaining that comprehensive coverage.

Yes, yes there is. For this particular test, we require 20% tumor content.

We don't call CNVs in every single of the 181 genes. We call them in 41 of those 181 genes. If you were looking at specific genes and wanted to see what genes we call CNV in, that would also be provided on that information sheet for Cancer Plus.

I think it cut out for just a second, but I think the question was where do you get your gene list from. The gene list, again, is a really focused gene list. What we're looking for is testing and reporting on genes that are clinically relevant. The vast majority of those genes are genes that have active clinical trials and therapies associated with them. There is a smaller subset of genes that are frequently mutated in cancer but might not yet have an active therapy associated with it. The genes themselves are pulled from a number of sources.

Nope, I was finished, Judy.

Another question I thought might come up. This is certainly something that we are investigating. We don't have a release date or an actual time frame for when we might offer a test for non-solid tumors, but it certainly is something that we're considering.

I don't want to give specific pricing information right now, but I will say that the test itself is very competitively priced. If you have ordered a similar tumor profiling test from another laboratory, the pricing would certainly be in line with what those are being offered at. If you'd like to get more specific pricing information and talk about pricing for your institution, again, feel free to contact our internal media contact, Robert Power, at the address listed below, and we can be happy to give you more information.

Thanks, Judy. No, I don't have any further comments. Thank you all for your time, and enjoy the rest of your day. I hope the conference is great.