Capturing Cancer Signals to Aid Early Detection - Dr. Josh Ofman, President of GRAIL

Shiv Gaglani: Hi, I'm Shiv Gaglani. We all understand the value of early detection in disease, and that is perhaps more true for cancer than many other conditions because it remains the second leading cause of death in the U.S. Today, we're going to learn about a new testing approach that leverages genomic technology and machine learning to detect cancer signals circulating in the blood, making it possible to identify a broad range of cancers beyond the handful that are currently screened for. 

With us to explain this is Dr. Josh Ofman, the president of GRAIL, a company that offers the first clinically validated test of its type that can detect a shared signal across more than fifty types of cancers and help physicians target locations for diagnostic evaluation.

Dr. Ofman also serves on the board of directors of CellBT, an immunotherapy company focused on the discovery and development of innovative cancer therapeutics. Previously, he spent more than fifteen years at the biotech company Amgen, where he was senior vice president of global value, access, and policy. 

So, Dr. Ofman, thanks for taking the time to be with us today.

Dr. Josh Ofman: It's great to be here today. Thanks for having me.

Shiv: So, I've heard quite a bit about GRAIL and I've been looking forward to this conversation for a while. Before we started the podcast, I mentioned that your company's work is featured heavily in Peter Attia’s book, Outlive, so I think there's a lot of buzz around what you guys are doing. But before we get into that, I'd like to start with learning more about you and what first got you interested in medicine and then digestive diseases.

Dr. Ofman: Sure. Thanks for asking. So, I was not a very science-focused undergraduate at Berkeley, and I was actually studying the history and philosophy of science. And through that journey, I just got more and more interested in science, and I thought to myself, what kind of career in science could be interesting for me? I came from a family of people in the mental health field, psychologists, psychiatrists, and I decided to try medical school. I had no

inkling early in my life that I wanted to be a doctor or anything like that, but I decided, based on advice from many mentors and my father, to try medical school and I just fell in love with it. 

Medical school was amazing. Young people are surprised to hear that from me, because they always hear how horrible it is, but it was just an unbelievable experience. I fell in love with human biology, human medicine, and completed all my training, decided to go into gastroenterology or digestive diseases, largely because it was a great way to study technology. I was very interested in how technology and healthcare worked and how you could assess its societal value, how you could figure out how to study it in a way that could tell you whether you were impacting human health or not. 

When I was at UCLA doing my residency and fellowship, I did a fellowship in what we call health services research, which is how do you study the healthcare system and how it works and operates and that finally led me into a career in academic medicine and then ultimately into the life sciences and biotechnology at Amgen.

Shiv: That's awesome. We love going deep into the thought processes of our guests because so many of our listeners aspire to be in positions like our guests, and so it's helpful to know how you've connected the dots along the way. Let's go into that shift from academic medicine to working for Amgen. How did you decide on that particular shift and what are some of the key takeaways you've had as you've shifted to industry? 

Dr. Ofman: I really enjoyed academia. You get to do research, you get to teach, and you get to see patients and I really did enjoy that, and I published a lot. But more and more as I was in academia, I got more interested in really the application of health services in clinical practice. How do you apply technology, innovation? 

We were doing some work on electronic medical records and clinical decision support. We had formed a for-profit company when I was at Cedars-Sinai and we ended up selling that company to the Cerner Corporation and that got me very interested in more the business side of medicine and how do you bring technology to improve public health outside of an academic setting, and I found myself increasingly getting more and more interested in that. My research interests were really more and more technology focused. How do you assess whether this technology is going to be a net benefit to society, a net harm, or a net cost? And how do you assess and measure its value? 

The more I published in that field, the more life sciences companies began to approach me and say, ‘Look, we're over here at Amgen or Novartis and we're trying to figure out which technologies, which drugs we should prioritize to bring to patients and the assessments that you're doing about measuring the social value, the cost effectiveness of these technologies could be very helpful for us,’ and so they began recruiting me and I decided after about seven years in academia that rather than publish another hundred papers, I would rather go try to bring technologies to patients, and that's when I joined Amgen.

Shiv: Yeah, that makes a lot of sense. I can relate quite a bit to that because in academia, the incentives are around publications, which have a major impact, but I personally was getting a little discouraged with how few breakthrough publications were turning into treatments or things that students, in my case, or patients, in your case, would actually benefit from. The translational science seemed to be lacking. What are some of the things you're most proud of that you did at Amgen that you want to share with our audience before we go right into GRAIL?

Dr. Ofman: I think one of the things that I was very focused on at Amgen was how to get the company thinking more about the value of the products that it was developing -- not just the scientific value, but their social and economic value -- and really putting a public health orientation to what Amgen was doing, and then working with policymakers around the world to better understand why they should be making these investments in healthcare innovation, biotechnology products, and others. I think it took a long time, and there were a lot of people involved, but we were able to make that shift at Amgen and in the end, that's something I'm quite proud of.

Shiv: Yeah, no, clearly. I can see how some of that work has translated to what you've been doing at GRAIL. So, for our audience who doesn't know much about GRAIL necessarily, can you tell us about the founding story of the company and then how you describe its mission, as well as kind of the size and scope where you guys are at today? 

Dr. Ofman: Yeah, so, the founding story of GRAIL is quite interesting. So, flashback to 2013, 2014 at Illumina, the giant sequencing company. The chief medical officer there at the time was Dr. Richard Klausner, the ex-head of the National Cancer Institute, and they were conducting a study to determine whether we had to do amniocentesis in pregnant women to find the chromosomal abnormalities in the fetus. As you know, amniocentesis carries a risk of aborting the fetus, and they thought, you know, since the mother and the fetus are exchanging blood through the placenta, we can probably just look in the mother's blood at DNA in circulation and learn whether there are chromosomal abnormalities. So, they did a study in 125,000 healthy pregnant women and sure enough, they found out, yes, we can do this in the mother's blood - we don't need to do amniocentesis. 

But in that study, a woman came to Dr. Klausner and said we've got some DNA that looks really unusual. Dr. Klausner looked at that DNA for those ten samples and said, “You only see DNA like this in cancer.” And sure enough, they called all of those women back, and all ten of them had an occult cancer. 

Shiv: Wow. 

Dr. Ofman: So, the light bulbs went off within Illumina saying maybe we have just learned that we can look in the blood at circulating DNA and find signals for cancer. That was really the birth story of GRAIL. And so Illumina spun out GRAIL as a private company. GRAIL raised a huge amount of capital and undertook a huge number of studies to see if they could develop a technology using genomic signals in blood to develop, to find cancer. That was in 2015 and 2016. 

The mission of GRAIL was really to find cancer early when it can be cured, and since that time, GRAIL has now developed and validated this technological approach. GRAIL now has a fundamental platform of looking at the epigenetic signals or the methylation patterns on DNA that are very specific for cancer. As you know, methylation markers are a hallmark of cancer and these signals are shared across many, many cancers because these methylation groups are turning the tumor suppressor, the tumor promoter genes on and off and they do it for all sorts of cancers. So GRAIL really harnessed the power of human genomics and the power of machine learning, and they developed a machine learning classifier that could be trained, that could really decipher these patterns that are really only seen in cancer. If you look at a blood sample, they can look at this epigenetic pattern on the DNA on very specific regions of the DNA and say this is only seen in cancer and make a call. 

GRAIL's technology for early cancer detection now has two steps. The first step is they isolate the DNA from plasma and sequence the DNA using a bisulfite sequencing, which reveals this methylation pattern. The machine learning algorithm looks at it and says, “is this a cancer signal? Yes or no.” Most of the samples will be no, because most people don't have cancer. But about 1% of the signals in adults over the age of fifty have a cancer signal and then if the answer is “yes, a cancer signal,” then it goes to a second step because these same methylation markers are also among the most sensitive predictors of cell type or tissue type. 

So the second step would be cancer signal detected, predicted origin -- ovary, pancreas, liver, stomach, blood -- and we can now predict where we think that cancer signal arose in the body. So that's the two-step test that is what we call Galleri, which is our multi-cancer early detection test that can now detect over fifty different types of cancer from a simple blood draw. 

Shiv: Wow, that is incredible, and that backstory reminds me of one of my favorite quotes from Pasteur, which is, “Chance favors the prepared mind.” So, the fact that everyone up and down that chain who noticed the DNA signals in the amniocentesis study, and then obviously the physician you mentioned from National Cancer Institute was able to separate signal from noise and find something really valuable out of that. So really, really cool. 

I'm curious, some of the things you've mentioned align with what Peter Attia calls “Medicine 3.0.” The big realization coming out of Outlive is that people who live to be centenarians delay the onset of certain diseases, whether that's Alzheimer's or cardiovascular or cancer. It's detected much earlier. All of his patients get colonoscopies five years younger than their preventive screening guidelines, so about forty versus forty-five. 

Can you tell us a bit about how many people you've screened with Galleri? What's the guidelines? How often should they be getting this, do you think? And then do you recommend Galleri in addition to say, Prenuvo full-body MRI screening? We had the founder of Prenuvo on the podcast a couple weeks ago.

Dr. Ofman: Well, maybe I can take a step back from it and just remind our listeners why this idea of multi-cancer early detection is such an important public health idea. We've been fighting a war against cancer for fifty years, and it's not one that we're winning. We're losing about 2,000 of our loved ones every day to cancer. Now, why is that? It's because we're finding most cancers too late when they've already spread, and treatments are largely ineffective. And why are we finding most cancers too late? Well, we do have some screening tests. We have screening tests for three cancers in women -- breast, colon, cervical -- and two cancers in men -- colon and prostate -- and then one screening test for heavy smokers, low dose CT. 

In total, those screening tests are only finding about 15-16% of the incident cancers in the adult population over the age of fifty and as you know, age is the strongest risk factor for cancer. People over the age of fifty have about a thirteen times higher incidence of cancer than younger than fifty. So, let's just use fifty as a marker for elevated risk. That's not going to bend the cancer mortality curve by finding so few cancers, right? We have an opportunity now to dramatically improve the number of cancers found in the population and that's multi-cancer early detection testing. We need to add that. It's a compliment to single cancer screening. 

We are never going to single cancer screen our way out of this problem, right? It will never be cost effective. There will be way too many false positives because for each single cancer screening test, you have to add the false positives. For me, I'm over fifty years old. I'm a man. I have to add the 10% false positive rate from Cologuard to the 10% false positive rate or 20% false positive rate for my PSA test and you add them all up. 

For a woman over the age of fifty, for example, she might have a 50% likelihood of having a false positive test. That's unacceptable. Nobody would ever approve that. But that's what we're doing with single cancer screening. We've estimated that if you add ten single cancer screening tests, that any individual's false positive rate would be well over 70%. That's unfeasible for the healthcare system. The beauty of a single blood test like ours, Galleri with a multi-cancer early detection, is we can look for over fifty different types of cancer because there's a shared signal with a false positive rate of one half of 1%. So that's a game changer, right? We can do that and now that can scale with the population. 

Now our ability to detect cancer is based on how much DNA is being shed from the tumor and that's variable, so the sensitivity will be different for different kinds of cancers. But we can have a single specificity because it's around that shared cancer signal. There's a huge unmet need and now we think our multi-cancer early detection tests can meet that unmet need. It's been published, that if we add Galleri to standard of care screening, we can find and intercept about 70% of the cancers and that could result if we find them one stage earlier than they otherwise would have been found based on symptoms or serendipitous findings. We could reduce the death rate over the next five years by almost 40%. 

Shiv: Wow. 

Dr. Ofman: Yeah, and that's a big impact. That's about the same impact as everything we're throwing against cancer today. 

Shiv Wow, that is incredible. Very exciting. I know you can see why GRAIL is being covered by so many different people and so many groups. What is like the biggest obstacle right now to take that really compelling data and scale it out to everyone above fifty in the US? Is it cost? Is it insurance reimbursement? Like what's the limitation? 

Dr. Ofman: Well, it's a combination of things. First, it's a new technology and it's a whole paradigm changer about how to think about screening for cancer -- similar to whole body MRI -- where doctors are just getting their heads around this idea. Cancer screening, you know,

it's got a very checkered history, right? A lot of screening has been attempted with technologies that are very poor. Think about the past. We look at images and shadows. We visually inspect tissue, but now we're talking about DNA that comes right from the tumor, so it's very biological. And we're learning a lot of new things about cancer and how fast it progresses. It's really a paradigm change, so that's part of the obstacle. We have to really educate the clinical community, the guideline community, and doctors about that. 

Second, it's a new technology, so right now it's not covered or reimbursed. It's in the market as a lab developed test. We're in the process of looking for FDA approval -- and that will come as our data mature -- but we don't have that right now. A lot of payers are very hesitant to reimburse for a test that's not FDA approved yet, especially one in cancer screening. We do have many customers who are reimbursing for the test or providing it to their employees or their health system partners and their patients, but largely it's self-pay. The test right now costs about $949 if you walk in off the street to get it, but we have line of sight to getting those costs way down as we get to the next version of our test and so we're very excited about that opportunity. 

We think once we have FDA approval, we'll get a lot of commercial payers providing coverage. We're also working with Medicare and Congress to give them the authority to pay for a test like this. Stakeholders in Washington have introduced legislation to give Medicare that authority to pay for it. So, once we get FDA approval, hopefully Medicare will have the authority to provide coverage and that'll be the real inflection point.

Shiv: Yeah, absolutely. That'll be an exciting future to see and hopefully not in the too distant future. One of the themes we cover a lot on Raise the Line is direct to consumer healthcare and providing the power to patients to get informed and to choose. We've had folks like Eric Topol on the podcast who you probably know and has written the book, The Patient Will See You Now. Also Vivian Lee of The Long Fix, who's written about value-based medicine. I could see this really fitting into a lot of value-based frameworks, detecting cancer early, so then you don't have to spend tons of money on resections and biopsies and chemotherapy, not to mention the emotional toll of all that on the patient and their family. 

As far as the educational efforts go, we've done some great work with 23andMe, as an example, to educate healthcare professionals around what happens to an NP or a PA or a physician if someone comes in with a 23andMe test. What is some of the feedback you're hearing from an average clinician who maybe reads JAMA once in a while or New England Journal or The Lancet? What is their feedback on this? Are they worried? Generally, I think people get worried about too much direct-to-consumer because patients will come in with a test and be like, ‘Oh, I have brain cancer for sure.” How are you guys addressing that and what's your educational strategy? 

Dr. Ofman: Well, importantly, in the initial phase of our rollout and our launch -- which is happening now, we're in the market -- we've done over 100,000 tests commercially now. The initial phase is that this is a physician-prescribed test. So, unlike 23andMe or these companies that provide information about your risk of developing a disease, which are direct-to-consumer, this is pretty serious business, right? This is going to tell you right now if you have a positive test, you are likely sitting in front of me with cancer and that needs to be done with the physician, right? There is a telemedicine option, but it's a physician-directed test right now. 

Until we get FDA approval, I don't think we're going to be in a position to market direct-to-consumer or anything like that, right? This whole initial phase is very much focused around the physician and the patient because those who have a positive result -- given our positive predictive value -- have a 40 to 50% likelihood that they're sitting there with cancer right now. 

As you know, with single cancer screening tests, that's not the case. This is the positive predictive value. In other words, if you have a positive mammogram or Cologuard test or low-dose CT of the lungs, the likelihood that you have cancer with a positive test is in single digits. Okay? That's how many false positives there are. It's very different with Galleri. Because we're looking for so many different cancers with a single signal, we can have a much higher predictive value. So if you have a positive Galleri test, there’s a 40 to 50% likelihood you're sitting there with cancer right now. You need to hear that from your doctor. So, we are educating doctors a lot about that. 

I think what people fail to recognize is that there's a huge difference in survival if you find cancers when they're localized versus when they're already metastasized. If cancers have already metastasized, two out of ten people will live ten years or five will live five years. If they're localized, nine out of ten people will live five years or longer. So a huge difference. And so the opportunity to make a big public health impact by finding these aggressive cancers early is really important and huge. 

Now, single cancer screening tests like mammograms and PSAs and low-dose CTs find a lot of things, but they may not always be the aggressive cancers that are going to kill people, right? And so there's a worry about over-diagnosis of these indolent cancers and that's what happens when you look at shadows or you look at proteins that are nonspecific, like PSA, or you look at low-dose CT. You find a lot of things, but, you know, it's hard to know, are they going to kill you? Are you going to die with them? Are you going to die from them? Like encapsulated prostate cancers are a good example. Turns out that those types of cancers, the slow-growing ones that are unlikely to kill, are also unlikely to be shedding a lot of DNA into the blood, so we don't tend to find those. 

It's very unlikely that Galleri is going to contribute to this problem of over-diagnosis and so that education to primary care providers is really important because it's fundamentally different than what they're used to with current cancer screening. Current single cancer screening, they're used to very high false positive rates, and no ability to discriminate indolent cancers from the deadly cancers and yet our test now has the opposite effect. We are unlikely even identifying indolent cancers. We have a very low false positive rate and a very high predictive value, almost the inverse of what you see with single cancer screening.

Shiv: Wow. That's extremely compelling and, you know, I'm already going to follow up and see if I can get one or both of these for my parents after this. My dad turns seventy next week, so it's certainly top of mind for that. 

Dr. Ofman: It's a really important point you're bringing out. Who should get this test?  In the early stages, we are offering the test to those who are at elevated risk for cancer, not just anybody. Obviously over fifty is the biggest predictor and the biggest risk factor, so we're using age. So, any adult over the age of fifty is at elevated risk of cancer and could benefit from Galleri. As for people younger than that, if you're a thirty-five-year-old smoker, if you're a thirty-five-year-old cancer survivor, you have about the same absolute risk as a fifty-year-old. And there are other environmental exposures and other risk factors like obesity and diabetes germline mutation carriers such as Lynch syndrome and BRCA carriers who are at very elevated risk of cancer. So, this test could really benefit adults who are at elevated risk of cancer. 

Shiv: That's good to know. Another big focus area of ours at Osmosis and Elsevier is ‘zebras’ or rare disorders. You know, the fact that you guys screen for fifty-plus cancers...I assume some of those cancers are what qualifies as rare diseases. Can you talk a bit about that, but also where Galleri goes for here? Do you think it'll go from fifty to a hundred? 

Also, you talked about the cost curve. I saw some of the big news from Illumina earlier this year that it was getting whole genome sequencing down to about $200 a pop, which is incredible. So I assume that cost curve, hopefully, will apply to Galleri. But yeah, any commentary on zebras and then, you know, how many cancers do you think you'll eventually be able to screen for?

Dr. Ofman: Well, we started with fifty or more cancers, but we already know it's grown. We're already finding cancers that we've never trained our machine learning algorithm on because it shares a very common methylation pattern with many other cancers and so we're training our machine learning algorithm to get better and better all the time, which is very good. That's one of the real benefits of being out in the market, collecting a lot of data, is being able to do that. 

We're also grouping cancers, right? So breast cancer is not one type of cancer. It's four or five different types of cancer. If we really break it all down, it's well over fifty. I would expect over the next several years, we'll get to seventy-five or a hundred different cancers. Now, most of these cancers, as you know, are uncommon, and some of them are rare. The deadliest cancers, like pancreatic and ovarian and gastric, are uncommon cancers. The most common cancers, breast and prostate and colon -- colon gets very well screened for and we can also find precancerous lesions using colonoscopy -- but many of the most common breast and prostate cancers are not the deadliest cancers, right? They're very common, but they're not particularly deadly. There are deadly breast cancers and deadly prostate cancers, but those are not the most common types. 

We find a lot of these deadly and aggressive cancers that are uncommon. Some of them are considered rare and so we find all of those, which is a wonderful thing because there's no other way that it would ever be cost-effective to screen for the uncommon or rare cancers one at a time, so that's a real benefit of Galleri.

Shiv: Wow, that's fantastic. We'll definitely hashtag this podcast episode with all the work we're doing on the Year of the Zebra, which is dedicated to rare diseases of which there are 7,000 that we know about right now. 

I'm aware of the time, so I have two last questions for you. The first is, we have a lot of early-stage healthcare professionals in our audience. Given your interesting career ranging from clinical to academic to now industry, what's your advice to them about approaching their careers?

Dr. Ofman: Well, I think for someone coming up in medicine now, there are so many wonderful options. I think that healthcare is an enormous part of our economy and the intersection between clinical medicine, public health, and innovation, I find incredibly interesting. So, I would encourage young physicians who are interested in that intersection to really consider a career in industry. You can do it in academia -- and I think that's a great place to start -- but for me, I found the application of the research into actual healthcare delivery and public health the most interesting. 

For young physicians, I don't want you to think that the traditional model of either going right into clinical practice or going right into academia is a necessary requirement. There are so many options now about getting into biotechnology, other technology, bioengineering, devices, sensors...and they all need doctors who understand how medicine works and how the practice of medicine works to really make sense of all this technological advance that's happening. 

I would encourage physicians to really think about careers in industry. It's very fast-paced, it's very innovative, it takes a lot of creativity, but you need to learn different things, things you were not taught in medical school about leadership, about innovation, about the economics of healthcare, and how it all fits together. I found that very energizing and would encourage others to think about it. 

Shiv: That's awesome. I couldn't agree more, having taken a break to launch a company myself and then now back in med school. I think it's really gratifying, being able to connect those dots in creative ways. You can have an outsized impact like you guys clearly are having at GRAIL.

My last question for you is, is there anything else you want our audience to know about you, about GRAIL, about cancer in general? Anything else you want to share?

Dr. Ofman: I think just keep in mind that cancer is projected to become the number one killer of men and women worldwide. We have an enormous opportunity to make an impact there. But we have to be able to understand and embrace these innovative kind of paradigm-changing approaches. If we wait fifteen years for mortality studies in every new technology in cancer...you know, there's no technology that's going to be around that long to study. By the time we do those studies, while very important to do -- I'm not saying we shouldn't do them -- but we just can't wait for all of that, right? We have to move with a great deal of urgency. 

So, when we can validate our technologies, when we can show the clinical benefit, we need to have some urgency to be able to provide these technologies in an accessible way to a broad swath of the population and make sure that all the inequities in cancer care that we see between ethnic groups, racial groups, and underserved communities can be addressed. 

So, we at GRAIL have a lot of urgency. We're doing some of the largest studies ever done in cancer screening. We have a clinically validated assay right now out there, and we're trying to make sure that we can get that covered and approved so that it can be broadly accessible to all the relevant populations who are at elevated risk for cancer.

Shiv: Well, that's a very inspiring message to end on. I'm inspired by the work you guys are doing. I would like to really thank you, Dr. Ofman, for taking the time to be with us today, but more importantly for the work that you and your team are doing at GRAIL to hopefully bend that curve and help us win that war on cancer. 

Dr. Ofman: Well, the new front in the world of cancer is multi-cancer early detection testing. Hopefully that message came through, and it's been my pleasure to be with you today. So, thank you for asking me.

Shiv: Absolutely. And with that, I'm Shiv Gaglani. Thank you to our audience for checking out today's show and remember to do your part to raise the line and strengthen our healthcare system. We're all in this together. Take care.