The leaders in cancer care have always understood that the genomic architecture, the somatic mutations, if you can identify them, then you're much more likely to optimize care for the patient. You now know the genetic makeup of a specific type of cancer. How do you get the connection to the efficacy of the drugs to that? Like, that's gotta be a third party research. That's a really good question, and this is just gonna accelerate with AI. There are a lot of people, investigators, as you know, are doing cancer research. Pharma is looking at all of the studies being done worldwide, and they're looking at potential targets for intervention. But I don't really believe in supervised learning. I like unsupervised learning, going to first principles and letting the data identify opportunities that are missed by human intervention because there's bias there. What parts of the process has it sped up? A more practical consideration is the economics of it and the ability for the hematopathologist to review cases and sign out cases. With our tests, it takes a good hematopathologist maybe fifteen to forty minutes on a very complicated case to decide whether there's residual disease or not. With AI, it takes three minutes. Fast forward to twenty thirty. What does your world look like? The whole hope here is you don't. Welcome to the twenty thirty podcast. I'm Keith Jensen, president at CADRE AI. We're here with Jeff Miller, founder, CEO, chief scientific officer at InVivoScribe. Well, thank you for the invite, and thank you for the opportunity to speak with you today and your viewers present the prep quad system and some of our capabilities of our products and services. They're international. Yeah and you guys have been you bootstrapped this company for thirty years. Yeah we've we've had great success because our standardized products have been adopted by the key opinion leaders internationally and they're really shown to help optimize the care for cancer patients, specifically in the area of hematologic malignancies, so leukemias and lymphomas. Okay, so I am not a chief scientific officer. I want to make sure I understand what we're talking about here. If I break it down, simply, are we saying that when you have a type of when you get diagnosed with cancer and you have some form of cancer, the standard methodology might be for a doctor to say, there are a handful of drugs, chemo and others out there that can, you know, treat this. Hey, you have this form. I happen to like this one. I've seen this be more effective, and they prescribe it. But what we're talking about is an additional layer of testing on the cancer, the tumor, whatever it is, to look at more of a genetic level of that to to figure out the makeup and determine if there are drugs that would be more targeted. Exactly right. Yes. And have you found I mean it's been thirty years so adoption obviously you guys have have gotten to a level of adoption that the company has been successful, profitable, etcetera, you haven't had to raise, that's amazing. What have the adoption barriers been that you've seen and have they changed over time? I don't think there are adoption barriers. I think the leaders, the key opinion leaders, the leaders in cancer care have always understood that the genomic architecture, the somatic mutations, I'm sorry, I'm getting a little technical but it's appropriate, the mutations are the variants that drive the disease, if you can identify them, then you're much more likely to optimize care for the patient. And I'll give you some examples. So chemo is kind of like napalm, it's indiscriminate. Whereas targeted therapies, you take a pill and you can knock the cancer into remission. And chemo is going to knock out more than just the cancer, right? Like you're harming the Quickly, dividing cell is kind of knocked out by chemo. Just to give you some background, so on one of the many leukemias lymphomas that we really concentrate on, acute myeloid leukemia, when a patient presents at the clinic to an oncologist, sometimes it's their primary care before they get to the oncologist, that's often the case, They have the equivalent if you took the cancer cells that are circulating in their bloodstream and their bone marrow and you packed, removed the liquid and just had packed cells, it would fill a liter bottle circulating through your body. We now have tests, first of all it's very easy to identify those cancers initially, but now we have tests that we've developed that can identify less than one cc of packed cells throughout the body. And so what that does is when a physician oncologist is treating a patient they can monitor the efficacy of the treatment. It's like looking at milestones along the way. So they take samples at intervals of say three to four months for what we call measurable or minimal residual disease testing and they can detect like one cell in a million and so if you see a cancer recurrence you can catch it early, it's the canary in the coal mine, right? So you can intervene and keep the cancer from coming full blown and killing the patient, and you can modify treatment individually for the patient, and it's really great because now, you know, pharma companies we've gotten partnering with three pharma companies we partner with like seventy five pharma partners but we've gotten three drugs approved for the most dangerous form of leukemia which is AML with the Novartis, Astellas and Daiichi Sanko. So are you guys the product or products that you guys are creating? Is it the testing capabilities? Is it the drugs? Is it both? So we're not in the drug side. We're on the, say the CDRH side of the FDA in the in this. We also work with regulatory agencies around the world. So the three drugs I talked about we got approved. Those are targeted therapy for FLT three positive AML. So FLT three is one of those biomarkers that we talked about targeted. But your product is the testing capabilities. Yeah. Is that right? And so so in order to get that adopted, so you you come up with it back in what? When did you start the company? Well, going back thirty one years. Okay. So you you're back thirty one years, you got a you come up with a clinical test. Is the is the motion to go directly to doctors and hospitals to try to get this adopted or is it like is it you have to go through FDA, what's the method? So initially we developed, let's go back to the basis of our company. Our company is focused on improving lives with precision diagnostics because that's really the foundation of precision medicine. You don't get the proper diagnosis, you can't treat the patient optimally, right? So over these thirty years we've developed tests and analytical tools that allow oncologists and pharma partners to optimize the care for their cancer patients, right? Along the way we've developed generic tools, assays, that lets hospitals and cancer centers screen samples for malignancies for cancer, but also we've developed the tools that help us monitor those cancers we talked about, right? And we're an ecosystem, so in order to provide molecular diagnostics that's actually precision, You've got to be able to provide your tests internationally. So what we've done is established clinical labs and companies in five countries so that we can provide not only as kits and products but also test services the tests that that we're talking about. The oncologists and the pharma companies don't have the infrastructure to provide the tests internally. So you need to stand up brick and mortar locations for people Oh, So we have, like, two hundred and sixty employees worldwide. We have labs and companies in France, Germany, Japan, and China. How many actual locations, physical locations? So that's that's five That's five. Locations. And are they typically standalone, or are they inside of a facility? No. No. They're wholly owned by us. It's our employees. Of course, in China, we're hiring Chinese employees in China and Japanese employees in Japan. And and China, we have we actually have employees in Portugal and all over the world. So it's it's it's really been a very fun journey. And the teams we've developed are remarkable. Have you seen AI Yes. Make a the answer is yes. Have you seen AI make a big difference? And I think I think that's Agency to the method. So Yeah. That actually is a really good area that we could have an hour long conversation about because this is something we've been pursuing for years. As a matter of fact, we have now machine learning tool for flow cytometry that, you know, one of the things you want to get into AI is kind of fun. I don't really believe in supervised learning. I like unsupervised learning. Going to first principles and letting the data identify opportunities that are missed by human intervention because there's bias there. If you have a training set that's defined by say gating on flow, you're already introducing human bias on the way they're trained, populations of cells that they're looking at. I'm getting a little technical, but what we've done is gone to looking for aberrancy as opposed to cell populations that are identified and using a normal dataset, a series of normal samples. The more normal samples, the better. The better the end, the better reference database. But then when you have a new sample, machine learning algorithm can identify aberrancy because it's not it's separate from that training set. But the training set was on normal populations. It's as close as you can get to kind of, first principles. And so you guys started implementing AI on the would imagine it's like research and development side, is where you'd be focused on it, right? Yeah, actually we were lucky because we were already offering the best, like again MRD, some measurable residual disease, the most sensitive flow assays for say AML around. Twelve color flow looks at twenty one different biomarkers. So we were getting data already from our own samples that were sent to us, And so when we have our data set and we look at these samples, we have, of course, you know, very sophisticated expert doctors who are our hematopathologists and have a big background in training on flow. They can identify populations, but then we can use our machine learning algorithm and we can see how close we come to matching their assessment. And one of the things that's really exciting on the machine learning side is we're going to be able to identify cells in transition. So when you treat a patient who's got cancer, they have a clonal architecture of that cancer. When you intervene, you're perturbing that and you're hopefully wiping it out. But in fact, normally you're not wiping it out. There's some survive one percent or so, and those grow up and proliferate. With our machine learning tools, what we're seeing is we're seeing aberrant populations that are getting escaping the treatment and kind of moving towards the next dangerous clonal type that's going to expand. And that's going to be very exciting because we're seeing new populations that have never been identified before. And the idea here is that AI is AI actually helping to do any of the the data analysis of this? Or is it like you I guess I'm I'm wondering how are you using it in the sense of you've got this massive amount of data on healthy and unhealthy population cells, etcetera. Like what parts of the process has it sped up? That's a good question. So a more practical consideration is the economics of it and the ability for the hematopathologist to review cases and sign out cases. So with our tests, it takes a good hematopathologist maybe fifteen to forty minutes on a very complicated case to decide whether there's residual disease or not. With AI it takes three minutes for them to get at least a quick pass and identify a population that then the hematopathologist, who's the expert, can go back and review and say, yes, I agree that that population is real or not. So it's really a crutch, it's not eliminating jobs, it's not making the call independently, but it's allowing the expert kind of a crutch to look at the data first. I would imagine do more reviews also. Exactly. A lot of flow companies that offer flow cytometry, they have a room full of hepatopathologists, and these are very expensive professionals, four hundred, five hundred thousand dollars a year. And if you can increase the number of cases they're able to review in a day, you save quite a lot of money. Right? Have you seen well, how big is the company actually, Yours, in terms of employees? About two sixty worldwide. Two sixty. Okay. And we're growing. We're hiring, by the way. Have you implemented AI at all in the the non r and d? So in like the, you know, efficiency of the organization? Yeah. You know, this is a great question because this is exactly what we're doing right now. We're working internally to use all of the instruments and all of the ARIES data points to get standardized data from the instruments, from the people who are working in the lab. So we have, for instance, we have R and D, we have manufacturing, we have, quality control, we have all those things. And a lot of it is managing time on instruments and core facilities. So if we get APIs from the instrumentation, we can use software to maximize the efficiency of use of the equipment, we can look for early failures on the equipment, lots of material or reagents. There's a lot of this that goes on. And this goes to prep point too, by the way, the platform, which is we're consolidating three separate processes where right now everybody who does nucleic acid if you do any testing molecular for with nucleic acids DNA RNA CF DNA for liquid biopsy you have to first extract and then if you want a sizable haystack to look for the needle for MRD, you need to concentrate, and then you need to quantify the material before you can even start testing. Well, right now, if you use the existing workflow, there are fifty points of contact of a technician in the lab to get to a final sample that they can start testing on. Fifty opportunities for contamination, fifty opportunities for mismatch or, you know, miss assigning a sample to another patient or another subject. And ours brings that down to nine touches, and it saves about sixty three percent on the overall cost. Let alone the opportunity cost of, you know, someone getting a misdiagnosis. Right. So good or bad. It's really exciting. So we think we're really gonna streamline testing, and anybody who touches nucleic acid who wants you know, not just clinical testing, but R and D and and, you know, academic labs. So you free up your graduate students, your post docs, your fellows in the hospital so that they can concentrate on doing the tests and the clinical trials that they're interested and been trained to do instead of the manual parts of labor that are going to the extraction or getting the sample in the first place. What is the growth lever of the business? Meaning, like, is it more hospitals you need to get into in order to to continue to expand? Like, what's the what's the limiting factor right now? Like, if something if a barrier could be removed and blank happened, you guys would triple in size. Like, what's that thing that's preventing Probably getting the message out. That's a very interesting question because we have seven hundred clinical labs that use our products in a hundred and sixty countries. And we have more than seventy five pharma partners using working with us on clinical trials. But because we're never venture capital backed, we all did this internally. You know, VCs have a habit of networking, and they have multiple companies, and they find relationships between companies. You know, there's a there's a boys network. Boys and women, fortunately, are getting more into this as well. But, you know, they get a lot of play that way. We've been kind of, I won't say in stealth mode, but the only people who know us are the seven hundred academic and cancer centers around the world. We're known by the key opinion leaders. They publish in JAMA Oncology and Blood and the leading journals. But, just making ourselves more known, getting out. I've never had to go out and raise money. You know, and that's one of the questions I wanted to to ask which was you didn't take money. Right? You could have, I'm sure. You mentioned VC at this point, profitable thirty years like you'd be in a you'd be an acquisition for a very solid private equity firm that would have a lot of connections and a and a board that would likely be able to help you expand. Is there a thought of of taking on any outside capital? We have a very good partner, a minority investor in Hitachi came in about three years ago, and they saw the opportunity. We got into liquid biopsy testing with our CF DNA extraction. This is the prep point. And they saw the opportunity there too. And so they've been in a minority investment in us. And there are a lot of companies that are interested in acquisition. You also want it's your baby, and you wanna make sure you're giving your baby over to, you know, a company that's gonna grow it and make it uber successful. Right? And, you know, we're we're happy where we are right now. I'm sure there's no like, oh man, I wish I would have done it differently because here you are and things are working and that's great. But looking back, knowing that getting this technology more proliferated across the world will help people, right, will allow people to not have to go get some massive, like you said, nuclear problem of chemo treatment that's going to wreck their body, Allow them potentially to get a pill or something that's gonna be very targeted to the type they have. Knowing that this could be helpful, do you have any, like, get I don't wanna say regrets, but like, is there an alternate world where like if you went back you'd go, no. Yeah. Let's let's sell some of this thing. Let's take some money and let's get this out faster. If you take money from VCs, then you put yourself in the situation where now you're beholding to whoever the board members are they put on the board. And it's nice to be able to grow organically because you can concentrate and completely focus on the science. And we do get out and, you know, we attend conferences worldwide now. We we have ASCO that's coming up in Chicago, end of next month. EHA, I'm sorry for the but, you know, European group in Stockholm in same month. AMP, which is where we started Association for Molecular Pathology, that's in in the US, and they also now do it in Europe. And then ASH, which is the big conference that generally is in San Diego, actually. You know, fifty thousand of your best friends, right, show up for this conference. So we do have worldwide presence, and we are expanding. We we're seeing double digit growth and so we're we're not unhappy with where we are and we're very happy to have remained private because we don't have quarterly numbers even though we're doing really really well. We can concentrate on the science and not get distracted by, you know, worrying about the stock and that sort of thing. Yeah. I was internal at two private equity firms and I've have seen it play out where founders are thrilled. Yeah. And they they're running the business really well, that the private equity firms excited and they're letting them go and it you know, yeah, there's some direction and stuff, but like, it's great and I've seen it happen the other way too. So it's a it's probably a better bet to And also going to your question, Keith. Peter Thiel would have called would characterize our company as a as a complex coordinated business. And that is, as he has even pointed out, that's one of the most valuable sorts of businesses. But venture capital people would never fund it because it takes years to mature and years to develop all the capabilities across the areas of the various parts of the business. So he uses Apple as another example of that. Right? Let me ask more question though I was curious about. I don't know when this happened, like you're starting out, you're figuring out the the testing. Right? There's gotta be this like, okay fine, you now know the genetic makeup of a specific type of cancer. How do you get the connection to the efficacy of the drugs to that? Like that's got to be a third party research, right? So are you tapping into like a database That's a really good question and AI is going to help with that. But what generally happens and historically has happened, and this is just going to accelerate with AI, is actually it's the pharma companies. They tap into there's just the cancer research conference that just completed here in San Diego last week. There are a lot of people, investigators as you know, are doing cancer research. Pharma is looking at all of the studies being done worldwide and they're looking at potential targets for intervention. So as you talked about, there's the genomic architecture of the cancer. And then you go inside and you realize, well that's a pathway and it's pushing proliferation or it's stopping apoptosis. There are ways we can intervene and they look for those targets for intervention. And for instance, FLT3 is a tyrosine kinase, which is when it's mutated, when it's mutated with the FLT3 ITD, it's a little problem with the with the protein. It turns it on and causes, AML to take off. So they're looking for the targets and then they design the drug to to bind to the tyrosine kinase and shut it off or somehow other intervene. There are other things like menin inhibitors in AML and other things but to your point it's exactly what you said. And if a doctor, you said you're in seven hundred labs, right? One of the other labs, is a doctor not doing this type of, what do you call precision diagnostics? Are they not? Is it like they're using another company to do it or they're not doing it? What's the alternative? One of the things that is most prevalent is there a thing called laboratory developed tests. That's tests brought up by each individual lab to do to look at the biomarker and test for that biomarker, that variant, right, mutation. So most test centers have some LDTs, they call them. So when I talk about seven hundred customers that use our tests, most of those customers not only get testing for their center, but they're a regional hub for tests, what they call test send outs, so that hospitals that don't have the test and realize the value proposition of the test, they send their sample to that one of those seven hundred clinical test sites and they run the test for them. And so I would imagine that the pharma companies that have, that are making the drugs that are more effective than chemo for a specific type of cancer with whatever specific biomarker, they have a vested interest in doctors utilizing this type of testing, right? And I'll point out where it's all going because that's it's kind of like an AI thing, right? There's a with using measurable residual disease testing, so you're looking for that canary in the coal mine, that one in a million sensitivity. The FDA and regulatory agencies worldwide are now getting ready to adopt what they call a surrogate endpoint for clinical trials. So it used to be that if you looked for a cancer treatment, you had a standard of care and you would put your cohorts of patients against the standard of care with your drug and you'd wait until a sufficient number of people had died in order to power the study to get your drug approved. And as a matter of fact, the first drug we helped get approved with one of our partners, they had started a decade long trial. But now with the surrogate endpoint, if you have a targeted therapy, the regulatory agencies are open to you using that test to show the diminution of that target in the body is linked to the efficacy of the drug, and now the drug could get approved in as little one or two years. So that's what's really exciting about using MRD testing now. It really cuts the cost of doing clinical trials, and it accelerates getting new treatments to patients who are in dire need of assistance and intervention. Fast forward to two thousand thirty, a few years from now, what does this, your world, look like? If you prescribe whether that is the testing and how that's all working, whether that's the the how the manufacture, the drug manufacturers are interacting with these, like, what do you think the world looks like at that point for you? The sizable change I see, we just are scaling. Everything is scaling. We get this question a lot. And what I think is going to be the big improvement is, as I pointed out, we spent the first three decades improving the tests and analytic tools to make the tests optimized. Right? Now, the biggest source of failure, the biggest source of of problems with even the optimized test is poor sample quality. So now with our prep quant system, we're intervening there and standardizing there and making the samples better and more, you know, consistent in quality. So now the tests are going to get better. So that whole pathway, we're we're really addressing now the entire workflow from sample receipt all the way through to test result. That's gonna be a huge, boom for the whole diagnostic industry, I think, and the whole community. And what this ultimately means is better identification of cancer, better identification of the treatment that will affect your cancer without needing to nuke everything else in your body. Right? And I would imagine less of a recurrence of cancer. And the dream is, I go back to the old example, when HIV was for, you know, going back twenty years with HIV, people died of AIDS. Once they found an intervention, people are now living a full lifetime with AIDS because they have targeted treatments. And they have MRD tests, they call it viral load tests, that can monitor how well their treatment regimen is working. When the cocktail no longer works, the physician changes the cocktail. And so people are living a normal life with HIV, whereas ten years ago they would have been dying twenty years ago. So the whole hope here is if we can't cure cancer, you die with cancer but you don't die of cancer. So you're ninety years old and you die of old age and you happen to have been diagnosed with cancer when you were sixty. So that's, you know, if we can't cure it, at least we can knock it down and keep it under control. That's the dream. It's very exciting. Thank you so much for being here. I your time and your team's time. Thank you.