Connected & Ready

How AI improves medicine from diagnosis to cure, with Chad Robins

Episode Summary

Immune medicine sits at the intersection of chemistry, informatics, computational biology, and machine learning. In this episode of Connected & Ready, host Gemma Milne talks with Chad Robins, co-founder and CEO of Adaptive Biotechnologies. He provides helpful context as well as specific examples of how his company is applying its technological innovations to real-world problems like COVID-19, cancer, and other immune-related diseases. Dynamics 365 is helping businesses of all sizes unify their data and create a digital-first culture. With next generation ERP and CRM business applications, employees at every level can reason over data, predict trends, and make proactive, more-informed decisions. Request a live demo of Dynamics 365 today: Thank you for listening to Connected & Ready! Do you have ideas of how we can improve the show? Want to recommend a guest for us to interview? We value your partnership and participation. Please drop us a note at We would love to hear from you.

Episode Notes

Gemma Milne talks with Chad Robins, co-founder and CEO of Adaptive Biotechnologies, about his company’s work in the emerging field of immune medicine, how they use technology to drive innovation, and the potential for their work to create further advancements in the area of personalized medicine.

Topics of discussion


About Chad Robins:

Since its founding in 2009, Chad has led Adaptive Biotechnologies in building a proprietary immune medicine platform that fuels businesses across life sciences research, clinical diagnostics, and drug discovery. Chad is routinely recognized for excellence and innovation. In 2019 and 2020, he was included in the Puget Sound Business Journal’s Power 100 and has been named a Goldman Sachs Most Intriguing Entrepreneur each year since 2015. He also named 2016 Ernst & Young Entrepreneur of the Year® – The Pacific Northwest Region Award.

Learn more:


Sponsor link

Dynamics 365 delivers next generation ERP and CRM business applications, helping employees at every level reason over data, predict trends, and make proactive, more-informed decisions. Request a live demo of Dynamics 365 today:


Contact us



Follow us on social media




Episode Transcription

Gemma [00:00:05] Hello and welcome. You're listening to Connected and Ready, an ongoing conversation about innovation, resilience, and our capacity to succeed. Brought to you by Microsoft. I'm Gemma Milne. I'm a technology journalist and author, and I'm going to be exploring trends around how companies are adapting to a disrupted world and preparing for tomorrow. We're going to speak to the innovators who are bringing products, operations, and people together in new ways. In today's episode, I'm chatting to Chad Robins, CEO and co-founder of Adaptive Biotechnologies, to talk all about immune medicine and the promise it shows for health care institutions and ultimately, patients. We also dive into how Adaptive Biotechnologies leverages artificial intelligence, machine learning, cloud, and data to bring about transformation from disease diagnosis all the way through to cure. 

Gemma [00:00:54] Chad, thank you so much for coming and joining us on the show today. Let's start with some brief introductions. Tell us a little bit about yourself, your role, and what you guys get up to Adaptive Biotechnologies. 

Chad [00:01:06] Sure. And Gemma, thanks for having me on the podcast today. My name's Chad Robbins. I'm the CEO and co-founder of Adaptive Biotechnologies. I founded the company with my brother in 2009. 

Gemma [00:01:18] Tell us a little bit about the journey that you've gone on over the past sort of 10 years. 

Chad [00:01:22] It's been quite a journey from a couple people starting a company to now having eight hundred and fifty people with products on market, products in the pipeline. And the journey has been really interesting because my original impetus for the business, my brother was incredibly smart and had a significant discovery, and I'm the business guy, so I thought there was a huge opportunity to start a company and the initial motivations, you know quite honestly and transparently, were probably financial for me. But very early on, when I recognized the power of our platform to impact patients’ lives and really use our technology to really make a difference in how patients could be diagnosed earlier and how new drugs could be discovered. You know, the motivations and what keeps me up at night, what keeps us driving is just this really profound impact that I know we can have to change medicine. 

Gemma [00:02:16] Tell us a little bit about what Adaptive is doing right now. I mean, what is currently with patients or with doctors? 

Chad [00:02:23] Yeah, we currently have a test on market for blood cancer patients, and this is for doctors to be able to better manage their patient care. And it's called Clonoseq, and it's a test for detecting minimal residual disease, or MRD, we're actually changing the term in the field to measurable residual disease. So what it actually does is it uses the immune receptor as a tag at diagnosis, meaning these are types of cancers of the immune system cell, of the B or T cell themselves. And so that one cell that B-cell winds up metastasizing or making clonal copies of itself. And with it, it carries the immune receptor that we sequence on its surface so we can essentially tell you at diagnosis, we can determine that high flying or lymphoproliferative cancer clone. And then the goal of therapy is to knock down or myoplate, or essentially kill that cancer clone. So giving that information to a doctor has many different implications. One is, did the treatment that the therapy or procedure did it work or not kill the cancer cells? If we can't see any cancer in the body, then the therapy worked, right? And then over time, different decisions made on the patient, given the level of MRD, should you escalate or de-escalate therapy? 

Gemma [00:03:44] Amazing. So let's dive into that a little bit. You guys work on immune medicine. What is that? And specifically, what are you doing at Adaptive that's kind of different than this broader term that talks about whole fields. 

Chad [00:03:58] Yeah, sure. It really starts with the premise of your adaptive immune system and your adaptive immune system fundamentally does two things: it detects disease and it treats disease. So you've got these immune receptors that kind of float around in your body and they look for pieces of disease and that's the detection methodology. And then once they find the disease, they go spring into action and mount an immune response to go kill that disease. So the premise of adaptive is if we can learn to read how the immune system is naturally seeing disease, then we can create a whole new class of diagnostics based on just reading how your immune system is naturally seeing disease. Correspondingly, if we could harness the power of these immune receptors is targeting molecules for fighting disease, then we could really develop a whole new class of therapeutics based on these immune receptors as targeting molecules. 

Gemma [00:04:56] So tell us a little bit about what does that look like in practice? You know, what does it mean to try and I guess, create an artificial version of the immune system and be able to spot things? And equally, you know, is it drugs, is it software, what is it? 

Chad [00:05:11] Yes. So one of the things that's most interesting is that it's very much a convergence of many, many different disciplines and fields. So you need chemistry, informatics, computational biology, and then machine learning. And so what we're doing and this is why we partnered with Microsoft on a franchise called T Detect, and this is where we're building a map of how our bodies essentially see disease and more specifically, there's these specialized cells in your body called T cells that have on their surface call it the scanner, and that's the T cell receptor and the T cell receptor's job is to find pieces of disease that get inside of cells and are presented on the outside of cells called antigens. So what we're building is a T cell receptor antigen map, ok? In order to do this, we need chemistry to be able to sequence T cells. We also needed informatics to be able to connect T cells to antigens, essentially filling out this massive jigsaw puzzle so we could start essentially seeding the map with our technology, our chemistry, our informatics. Some of this really sophisticated kind of address mapping technology, but we recognized early on that we needed a partner that had massive computational power and really machine learning capabilities to impute essentially the connections, the additional connections that were made between these T cell receptors and antigens. So that's why we partner with Microsoft. So they brought this very sophisticated machine learning capabilities on. And just actually to give you a sense of scale - the amount of data created by this map is larger than all of the data on the internet, right? Just think about that for a minute. It's just an extraordinary amount of data, but this is a very distinct machine learning problem. So it's not some black box. We know what we're looking for. It's just a very, very big, big, big problem. You've got this map that connects T-cells to antigens. And the key here, which I might have left out, is antigens or the signals of each disease, they're specific for that disease and only that disease. So therefore, if you have a T-cell receptor, you know it binds to an antigen, then you know, essentially that your body is seeing that disease or it's in memory and has seen that disease. So that's a concept that we're doing for diagnostics in our T detect franchise. 

Gemma [00:07:36] Can you tell us a little bit about what you're doing with Microsoft with respect to machine learning and data? 

Chad [00:07:41] One of the really interesting aspects that we're working on now that really scales machine learning is we're co-developing a machine learning workbench. So all of the models, algorithms, models that we're developing, how do you make them so they're easier to use and more kind of off the shelf tools? And so we've really put a lot of effort into developing this workbench so that essentially you don't have to be a machine learning expert to be able to use this set of tools and models that we built based on all the data we're aggregating, all the knowledge that we're collectively putting together to develop this really unique set of tools that applies to how we analyze the immune system in our data. 

Gemma [00:08:27] I wonder if you could talk a little bit about where these massive datasets come from and how do you utilize it? 

Chad [00:08:33] Yes. So our data comes from a variety of different sources. One is there's large academic groups and principal investigators that are trying to solve specific problems. Covid is a great example. We got datasets from eight different countries, especially as variants were emerging in different parts of the world at different times. But we also got data from vaccine manufacturers. We got data from different institutions and societies like the Leukemia and Lymphoma Society, all looking at different aspects of the problem. So we were able to aggregate all these different data sets and into this system that we built and with Microsoft to be able to capture the data and then analyze the data and then make use of the data for, you know, different purposes, whether it be for, the show vaccine efficacy to the pharma companies or to diagnose past infection or to help the immunocompromised population. 

Ad [00:09:26] Dynamics 365 is helping businesses of all sizes unify their data and create a digital-first culture. With next generation ERP and CRM business applications, employees at every level can reason over data, predict trends, and make proactive, more-informed decisions. Request a live demo of Dynamics 365 today by following the link in the episode description.

Gemma [00:09:56] Amazing. So tell us a little bit about, you know, you launched earlier this year T detect COVID. So obviously this is the, I guess, the specific version of this map or parts or, you know, country existing in this map, that's specific to Covid. 

Chad [00:10:11] Yeah, that's actually a great analogy, Gemma. It's like probably like one little municipality on this massive, massive map, which is this one disease called COVID, which is - so what we did, and I called up Peter Lee and some of my counterparts at Microsoft, right when we knew that COVID was starting to be a major problem and said, Hey, we have an obligation because we were working and we still are on many different diseases in autoimmune disorders, for example, and IBD and IBS. We're working on rheumatoid arthritis, we're working on Lyme disease, but we said, Hey, let's direct our resources and make sure that we're kind of being part of this global solution to look at the cellular immune response. And by the way, that's what T-cells do as opposed to antibodies, or neutralizing antibodies, which everyone was looking at. There was this whole other half the immune equation called the cellular immune response, which is how the T cells were responding to the SARS-CoV-2 viral proteome. And so what we did was we created this map data public for research related purposes. So we essentially mapped the T cells to the SARS-CoV-2 antigens and created this map. And now that we had this map, it could be used for a couple different purposes. The one is T detect COVID in which is we went to the FDA and got an emergency use authorization and it's the only T cell test approved. And we're expanding that to really looking at studies of correlate of protection. Meaning if you have the right T cell profile, are you protected against this virus? And we're doing a quite a bit of work with that. So one is the T detect diagnostic. But there's also we're working with the vaccine manufacturers. So what's public in us is J&J and Moderna getting on all of their trials to determine kind of vaccine efficacy and did their vaccine elicit a broad and diverse T-cell response? The reason that we have coverage against the Delta variant is because the T cells actually happen to elicit a broad and diverse T-cell response. And we were able to show that.

Gemma [00:12:15] I wonder if we could zoom out a little bit because, you know, obviously it's amazing to have this really specific example. But I think for people who are not familiar with the sort of pharmaceutical space, although I suppose it's difficult to escape it considering we're in a pandemic to some degree. But having the understanding of the dynamics of the field in terms of why things take so long to get to market or, you know, this idea you're talking about of mapping the immune response, I mean, surely that's not a new idea, right? But a lot of this stuff has not been doable until now. And I think trying to contextualize the role that technology, specifically AI and machine learning, has had in changing the way that we think about diagnostics and creating medicines. And, you know, playing around with that luck, shall we say, in order to be able to actually move things forward. I wonder if you could contextualize that a little bit for people who are not as familiar with the space. 

Chad [00:13:08] Yeah, I'm happy to, and I'm going to start with an analogy to really the technology industry starting with hardware. You know, if you look at the initial big supercomputers, these big mainframe computers that are in massive rooms over time, they got smaller, faster, cheaper, and much more powerful. And then Microsoft came along and made it so that these computers were accessible and you obviously went from mainframes to desktops to laptops to mobile, and they got smaller, faster, cheaper. But it wasn't until you had kind of an interface, it was Microsoft and Apple, et cetera, that you really had an interface where you could make these machines accessible. And then obviously you had the internet and further phases that really changed the way in which we did everything in our lives. We communicated, we shopped, we worked, et cetera. Really, if you look at the next generation sequencing field, right? It started with the hardware. And sure, you're aware of a company called Illumina. It was kind of the 800 pound gorilla in the space that makes the machines that do next generation sequencing. 

Gemma [00:14:12] And this is for DNA, right? For people who are...

Chad [00:14:14] This is for high throughput sequencing of DNA, which really one of the most incredible inventions of our century is that we can actually use machines to sequence down to the base level of DNA and seeing that very exquisitely and precisely and sensitively. But really, so think about that is 1.0. But then 2.0 is, OK, now that you have these machines, what are the clinical implications? How do these machines impact patients’ lives? And so adaptive and others, we now can use these machines to look at a patient's immune system, and then we can apply that technology and platform to be able to diagnose disease and to discover new drugs. And so think about that is 2.0, if you will. And I'll use another term in tech, think about it like content, right? And in the immune system, because it applies to almost every disease, a category, whether it be autoimmune disorders, infectious disease, cancer. This is one of the largest areas and areas that we can have the largest impact on patients because if we crack the code of the immune system, then we can diagnose disease. And what I was talking about earlier is that one sample why it can see COVID, we can also see every other disease you have just from that same blood sample, and that's really the power of T detect. But continue with the analogy. If you think about 3.0 and the next generation sequencing category, 3.0 is really about machine learning and how you use data and information to really amplify the answers that you're going to generate in this data with the chemistry. But now you're overlaying on this machine learning to be able to impute the correct answer and to be able to sift through this massive amount of data. So unlike in the software industry, where you can iterate on software and just put a product out to market. In the medical field, whether it be in diagnostics or in therapeutics, you have, I'll call them externalities. There's a regulatory environment, whether it be the FDA or laboratory developed tests. And that's why there's different phases: phase one, phase two, phase three, et cetera, of different drug trials, to be able to effectively get a drug on market. Covid, through the emergency use authorization through the FDA, wound up expediting some of those processes, recognizing that we're in the middle of a pandemic and we had to take the additional level of risk to move faster to get both therapies and diagnostics on the market. 

Gemma [00:16:47] Amazing. Thank you for contextualizing that because I think sometimes when you talk about, I guess, the future of health care and so on and so forth, I guess the power and the real change that technology had is sometimes difficult to understand. What sort of challenges still remain in terms of adoption and use of this approach and these kind of technologies, these kinds of approaches to diagnosis and obviously also to treatment, in terms of curing disease more broadly?

Chad [00:17:13] Let's take diagnosis - we'll separate those out from diagnosis and then treatment. So one is you have to develop the body of evidence in each disease. Let me actually take one step back and tell you what we're trying to do. And then I'll give you some of the challenges. The idea, as I mentioned, taking that one blood sample and be able to extract the immune information out of that and then map it to the map we've built to diagnose disease. The idea is that it's one test that you can get many test results, many diagnostic from the same test. But in order to do that, you have to start with one and go one at a time. So we started with COVID, we're working on Lyme disease, IBD, rheumatoid arthritis, multiple sclerosis, many different diseases. So that's phase one. Step two is this concept of differential diagnosis, meaning you walk into a doctor's office and you have a set of shared symptoms. So let's say you walk in with a stomach ache, GI gastrointestinal related issues. You don't know whether that is that Crohn's disease, is that an IBD, IBS, Crohn's, celiac? There's something else. There's many diseases that it could be. And T-cells are specific. And so that's the power of T-cells. So once we've mapped that disease and what we want to do is be able to differentially diagnose and provide a clinician with a tool that can differentially diagnose and say, you actually have Crohn's disease. And by the way, it's not just Crohn's, it's ileal Crohn's, a specific type of Crohn's disease, and we can rule out that you don't have IBS. So that's the differential diagnosis category. And then the third category is this concept of an immune scan. So that being part of your primary care channel, just like you go in and get a complete blood cell count at your physical checkup, this would be you'd get an immune checkup. And so that's to get to kind of phase three. Now the question is why is this hard? Why doesn't this happen faster? One, there's science that needs to be done. So we proved that we could do it with infectious diseases and COVID and Lyme disease, we've got a really twice a sensitivity of anything else out there. And then the second category for us is in autoimmune disorders. We've got it in Crohn's and multiple sclerosis is emerging. But you have to develop the evidence that this is the analytical validation and the clinical validation that says this is absolutely working and we actually set our specificity level at like almost 100 percent. So 99.9 percent.

Gemma [00:19:47] Yeah, please do elaborate. Tell us a little bit about treatment as well. Because obviously this is an area you're working in as well. 

Chad [00:19:53] Yeah, happy to do so. But just to pick up on what you were talking about, if you look at the regulatory bodies, the FDA, although I would say they're really thinking hard and trying to put teams in place that understand data, real world data, real world evidence and kind of machine learning. Remember, these are relatively new discipline and relatively new applications of the technology to medicine. So really being able to essentially diagnose something that is imputed from an algorithm, you really have to prove the algorithm is accurate. How you prove that and what the standards are is something that the FDA and the industry is kind of working together on really putting that infrastructure in place. Now what Adaptive is doing with Genentech is we're looking at not a chimeric antigen receptor, Part B cell part T cell, so we're actually looking specifically at T cell receptor cell therapy. And the difference is I said this earlier, but I'm going to highlight this is that T cells can actually see inside of cells so they can see intracellular pathogens. So we partnered with Genentech in a two part strategy. The first part is to find and characterize, and when I say characterize, you know what I mean, look at properties of T cells that make them a good therapeutic, that make them a good targeting molecule. But the second strategy is where we're really going and what's going to be revolutionary is each patient's cancer is entirely unique. It's an N of one. And so the idea is for each patient to create a personalized cancer therapy based on how that specific patient's immune system is reacting to that specific patient's tumor. Hopefully, with the goal of being 30 days what we call vein to vein, from taking an immune system out to be able to put it back in souped up to specifically have a specific designer or bespoke therapy for that specific patient. 

Gemma [00:22:02] Incredible. I want to end by kind of coming back to, I guess, talking a bit more specifically about AI and kind of the power that it brings in order to kind of do all these things that you're talking about both now and moving into the future. I would love to just kind of hear a bit of context as to what the technology really unlocks as we move forward. 

Chad [00:22:23] Yeah, I think one of the areas, it's really interesting that it'll unlock, it's this whole concept of personalized medicine and whether a therapy will work on a specific patient, either drugs that are being developed in the pipeline or existing drugs out there to be able to have a set of I'll call them biomarkers and to be able to combine biomarkers. I think we're right now at a point where machine learning, at least in medicine, is incredibly powerful. If you have a very complex problem that you need machine learning for, that can solve the problem. But where the future holds is kind of more of what I'll call this black boxy machine learning where you're combining the different omics along with different datasets. And you can overlay on this really powerful machine learning for this concept of personalized medicine, where you can tailor a treatment based on a patient's biomarkers. And those biomarkers could come from many different and perhaps disparate, but also interconnected systems of the body. Where you have these biomarkers in machine learning can help kind of really answer those questions about which biomarkers are playing different roles and really driving disease or essentially helping with therapy selection because that information all being put together, sifted through and machine learning, helping to get to an answer to provide the right treatment for the right patient at the right time is where I think machine learning is going to really, that's what the future of medicine holds. 

Gemma [00:23:54] Amazing. That's a lovely point to end on, Chad. Thank you so much for coming and sharing your expertise, both at a high level in terms of the future of medicine, but also some of the amazing work that you're doing at Adaptive Biotechnologies and of course, your partnership with Microsoft and what it's been unlocking for you guys over there. Thanks so much for coming and joining us on the show. 

Chad [00:24:10] Thanks for having me, Gemma, it's been a real pleasure. 

Gemma [00:24:15] That's it for this week. Thank you so much for tuning in. You can find out more about Chad's work and indeed, some of the broader themes we discussed today in the show notes. If you enjoyed the episode, please do take a few moments to rate and review the podcast. It really helps other people discover the show. And don't forget to hit subscribe, and tune in next time to continue our conversation about innovation, resilience, and our capacity to succeed.

Ad [00:24:45] Dynamics 365 delivers next generation ERP and CRM business applications, helping employees at every level reason over data, predict trends, and make proactive, more-informed decisions. Request a live demo of Dynamics 365 today by following the link in the episode description.