How AI and Spatial Biology Are Transforming Cancer Treatment and Clinical Trials

Dusty Majumdar is the CEO of PredxBio, a precision pathology company leveraging AI and spatial biology to transform cancer diagnostics and therapeutics. With over two decades of leadership experience in precision healthcare, he has held senior roles at the American Society of Clinical Oncology, IBM, Exact Sciences, and GE Healthcare. Dusty has spoken at numerous forums, including The Economist War on Cancer and a MIT plenary lecture on AI in healthcare.

Here’s a glimpse of what you’ll learn:

• [02:36] Dusty Majumdar talks about growing up in Eastern India and receiving his PhD in the US
• [07:50] How taking business courses sparked Dusty’s interest in leadership
• [11:22] Dusty’s transition from R&D to leadership at GE Healthcare
• [15:08] Leading product development and innovation in CT imaging
• [21:16] How Dusty became involved in improving early cancer detection using multi-omic data approaches
• [26:02] Dusty describes his leadership role at IBM, integrating AI across healthcare sectors
• [33:12] The clinical development of AI in healthcare
• [41:02] How PredxBio leverages spatial biology and AI to predict immune therapy responses
• [51:52] Predictions for virtual clinical trials and the future of AI-driven personalized medicine

In this episode…

Despite recent advances in cancer research, early detection and effective treatment remain major challenges. Traditional diagnostic tools often fail to fully capture the complex biological interactions driving tumor growth and therapy resistance. How can emerging technologies like AI and spatial biology help personalize cancer care and improve patient outcomes?

Spatial biology and AI can decode the tumor microenvironment at a much deeper level than genomics alone. By analyzing how cells organize and communicate within tumors, this approach can predict therapy responses with over 90% accuracy. Oncology, genomics, and AI-driven healthcare leader Dusty Majumdar emphasizes the importance of integrating multi-omic data, such as proteomics, transcriptomics, and metabolomics, to personalize treatment strategies. He envisions a future where virtual clinical trials drastically reduce development time and costs while improving trial success rates.

Tune in to this episode of Transaction Healthcare as host Zak Eisenberg chats with Dusty Majumdar, CEO of PredxBio, about using AI and spatial biology to transform oncology care. Dusty talks about liquid biopsy limitations, the role of regulatory bodies like the FDA, and the evolution of AI in healthcare.

Resources mentioned in this episode:

• Zak Eisenberg on LinkedIn
• Merritt Healthcare Advisors
• Merritt Healthcare Advisors email: contactus@merrittadvisory.com
• Dusty Majumdar on LinkedIn 
• PredxBio

Quotable Moments:

• “You have to deliver. And, you know, there were consequences for not delivering at GE.”
• “I realized the value of communication and being able to stand up and think on your feet.”
• “Genomics gave you the next step, but spatial biology unravels the mechanism of action.”
• “Without AI, combining all these orthogonal data sets in any meaningful way is just impossible.”
• “You can shorten the time of clinical trials from, you know, a few years to days.”

Action Steps:

1. Leverage spatial biology to analyze tumor microenvironments: Understanding cell organization reveals critical patterns influencing cancer progression and treatment response.
2. Combine multi-omic data sets for better predictions: Integrating genomics, proteomics, and metabolomics improves early cancer detection and therapy accuracy.
3. Utilize AI-driven predictive analytics in clinical trials: Predicting patient response upfront reduces trial costs, duration, and failure rates dramatically.
4. Collaborate with regulatory bodies early: Engaging the FDA ensures compliance and smoother adoption of AI-based healthcare innovations.
5. Invest in virtual clinical trial development: Digital models can simulate treatment outcomes, accelerating drug development and improving patient-specific therapies.

Sponsor for this episode…

This episode is brought to you by Merritt Healthcare Advisors.

Merritt Healthcare Advisors is an investment bank with a unique focus on healthcare providers and their businesses.

Merritt leverages the healthcare industry expertise of its owner-operators, clinicians, investors, and advisors to develop surgical facilities that perform safe, efficient, and cost-effective procedures.

To learn more, visit https://merritthealthcare.com/.

Episode Transcript

Intro 0:04  

Hello and welcome to Transaction Healthcare. I’m Zak Eisenberg, Vice President at Merritt Healthcare Advisors. Merritt Healthcare Advisors is an investment bank with a unique focus on healthcare providers and their businesses. Transaction Healthcare is a podcast focused on addressing questions and concerns at the intersection of health care transactions and business.

Zak Eisenberg 0:24

I’m Zak Eisenberg, a partner at Merritt Healthcare Advisors, and you host for Transaction Healthcare, where we address questions and concerns at the intersection of transactions, healthcare, and business. This episode is brought to you by Merritt Healthcare Advisors. Merritt is a full service investment bank with a unique focus on health care Merritt leverages health care industry expertise of its owner operators, entrepreneurs, advisors, clinicians and investors to advise owners, doctors, innovators and entrepreneurs in the health care sector about maximizing growth and running successful transactions. To learn more, visit www.Merrittadvisory.com. I’m joined today by Dusty Majumdar, who assumed the role of CEO at PredxBio in 2022 spearheading the company’s mission to lead the field of spatial biology through delving into the mechanism of action of cancer drugs and predicting patient outcomes. Prior to joining PredxBio as CEO. Dusty held strategic leadership and executive roles at IBM as the chief strategy and marketing officer, as well as at GE Healthcare, exact sciences and 3m recognized as a global leader in life sciences. Dusty holds a PhD from the University of Texas at Austin and a bachelor’s degree from the Indian Institute of Technology Karakpur. With that intro, Dusty, welcome to the show and glad to have you here. How’s your day going? Thank you. Great to be here. Good. Well, Dusty, I like to always start these conversations by digging into your specific background. I think it’s always helpful for our listeners to hear about how the entrepreneurs and leaders shaping today’s healthcare started in their careers and really grew into the leaders and innovators they are today. So maybe you can bring us back even to your time at the institute, Indian Institute of Technology, or starting at your PhD program, or even before that, did you always know you wanted to be working in the life science industry? Maybe when you were a kid, wherever you’d like to start around that time would be, would be great,

Dusty Majumdar  2:36  

sure. Thank you. Zak. Well, as a kid, I was always a curious kid. I grew up in the eastern part of India in a city called Calcutta, a city of probably 15 million back then, I don’t know exactly where the population is right now. And went to the Indian Institute of Technology in Kharagpur, where I was fortunate to brush shoulders with the best of the brightest that came from all parts of the country, and interestingly, a lot of the CEOs today in the tech world, like Sundar Pichai Arvind Sundar Pichai and Google Arvind Krishna at IBM and a variety of Other folks went to the Indian Institute of Technology, and I was contemporaneous with a few of those people at that point of time. So, you know, this is the 1980s the late 1980s when you applied to the United States, you’re not really sure you know exactly where you want to go for grad school. And so I applied blindly, and I got into a few of the universities, landed up at the University of Texas at Austin to do my PhD, further continue on the engineer engineering front. And again, UT Austin, back then, was a really famed University in chemical engineering and computer science. And I did my PhD at the intersection of physics and engineering, as well as there’s a lot of pattern recognition involved in the image analysis that I did for soft matter system, polymer, polymer systems and other systems that I delved into. Was fortunate to work for a world renowned scientist, Professor Don Paul at the University of Texas, who took me under his wings, and I was very, very fortunate to publish around 16 papers with him in the span of less than three and a half years that took me to finish my PhD, which very grateful that I spent seven years working here as a grad student, and after that, got into the industry, starting with 3m and then spent around 20 years in GE Healthcare, which was amazing when I started there, GE stock was going up 40% pretty much. Welch. Every year, Jack Welch was known as the CEO of the century. There’s a lot of revisionist ideas about GE and Jack Welch nowadays, but I can tell you the execution that I saw at GE and the quality of people that I encountered there I haven’t seen after that. Absolutely fantastic teamwork, absolutely fantastic execution. And the biggest thing that over there, that I noticed was the accountability, not only your manager holding you to accountability in terms of your do to say ratio, but also your peers holding you to accountability. If you say that you’re going to do something, you better do it, because you will not really get the cooperation of your peers if you consistently fail to deliver. So that’s kind of my journey at GE Healthcare. I started off in the famous Edison labs. Thomas Edison formed the labs in Schenectady next, actually, it’s in nescajuna, and not pretty close to Albert in New York. And that’s where a lot of inventions happened over the last 100 years before I joined, you know, including the X ray cube Langmuir. You know, there’s inventions over there, which won the Nobel Prize for him. There were a few Nobel Prize winners, even when I was there in the GE research labs back in the 1990s and it was really a famed Center of Excellence for science and technology.

Zak Eisenberg  6:33  

And were you working on the R and D side while you were at I was

Dusty Majumdar  6:36  

mostly in product product development initially, and after three or four years, I was called to the business side, and I moved on pretty much to the P and L and the business side in the late 90s, and pretty much have been in the business side, running businesses and PNLs and working in the commercial arena, lot of it, also in marketing over the last 20 plus years.

Zak Eisenberg  7:04  

I’d like to go back for a minute to your time in your PhD program. First of all, I have some appreciation, because my brother did a PhD of how difficult it is to publish 16 papers in three years. So I understand that that’s a lot. Did you know immediately that you didn’t want to stay in academia and you always wanted to go into industry, and when? When was that transition for you? If at all, maybe you were very focused when you got into your PhD program, you knew you were going to go into industry, and this was a step along the way, or was it a gradual learning experience for you of how you transitioned into into

Dusty Majumdar  7:50  

industry? You know, I was always interested on the business side, even when I was doing my PhD. I took several courses in the business school at UT, Austin. I also attended a lot of sessions with the business students on the Toastmasters International Club that they had internally. Got to know a lot of the folks over there. And, you know, I realized the value of communication and being able to stand up and think on your feet and deliver a pitch. This is almost 30 years ago, and I thought that’s where I really wanted to be in the business arena, being able to articulate a vision ultimately and lead businesses. So I purposely made the transition soon after I finished my PhD. You know, went into industry directly, did not really apply for r and d positions in academia. But the interesting thing was, in the first role I had in GE, it was in product development, and it was directly related to my PhD, which was developing engineering thermoplastic materials that could withstand the impact test comparable to metals in exterior body panels in cars, and it was directly related to my PhD, because I had looked at Polymer polymer interactions and how you could actually get the impact strength and the balance of all the other properties in polymeric systems, and ultimately, within a short period of time, GE actually launched the products that my team and I had developed in the exterior body panels of cars like the Volkswagen New Beetle. If you remember, in the mid 1990s the New Beetle was launched that used the material that we developed. It also was across many different OEMs, Toyota. It must have been exciting to see, unbelievably exciting. You know, it was 50 million business that came out of nothing within a couple of years. So it was a fantastic experience as my very first role in GE.

Zak Eisenberg  9:58  

Right, right. And. So you did really start in product and R&D, but you said you always had an interest in the business side, Toastmasters and some other courses that you were pursuing. So was it, did you start pursuing this business passion once you were at GE or did the opportunity arise and you took the chance? How did this transition occur? Because it seems like to me, and we’ve known each other for a little while now, once you made this transition, while you were at GE all of your subsequent roles really were much more business focused, and you really haven’t looked back, just just curious how that when that moment came, if there was a period of intense introspection, is that the type of transition that you you wanted to have? And, and, yeah, how it, how it came about? I always think it’s interesting and I know our listeners feel this way, to help people end up in their current situation, if it’s, you know, pre planned or, or is it destiny or or opportunistic, or if, or if people are really, you know, setting up the the opportunity for themselves. So anyway, very, very curious about those.

Dusty Majumdar  11:22  

Yeah, so, you know, the thing about working in GE in the 1990s was you had to really prove yourself, to really move from one function to another, or to get a promotion, you know, from one level to another. So to first demonstrate that I could execute and I could actually lead the teams that were working for me, as well as articulate a vision. And again, very fortunate to have attended a lot of the courses in the famed G crotonville Arena back then that was Jack Welch’s Management Institute, where people who were performing at the top were sent to attend various business related courses. And I put in every year in my yearly review that I wanted to actually go and attend some of those. So that’s where the journey started. And then again, very fortunate to be close to Boston during those years, where I could go and attend business courses at Harvard University as well. So I prepared for it for probably a couple of years before I actually made the transition into the business side. And it was, it was a great transition, because my first job was not in healthcare, but in automotive, and I was in Detroit leading the product marketing function for a product line that was actually going into these exterior body panels. So I saw it through the actual discovery phase of these novel materials that could withstand the impact, the heat, the stresses in the exterior body panels of cars and replace metal completely, all the way to the commercialization of this to probably around 15 OEMs around the world. So that’s how I made the transition. And again, performance was the barometer to see if I could go to the next level, and then if I could go from one business, which was at that point, GE advanced materials, or GE plastics, all the way into GE Healthcare. And that’s where, in the early 2000s made the transition into.

Zak Eisenberg  13:24  

Yeah, which is also quite different for you, Dusty and lots of people in healthcare that you didn’t start out in the space. You fell into it through opportunity, and it’s really been an area that you’ve stuck with for a long time now. And I think the theme of performance is something that’s going to come up again as we start talking about what you’re working on now, but it’s a consistent, I think, issue, especially in the healthcare, life sciences space, where efficaciousness is so key, and this is true across science, but particularly when you’re talking about people’s lives. Efficaciousness is just so key, and in that, performance in the actual product really drives success much more than anything else. Okay, so you transitioned to healthcare GE and this was a fairly long stint after your PhD, that you were at GE, maybe 15 years, if I remember correctly.

Dusty Majumdar  14:29  

And that was for 20 years, and you know, over half of that in healthcare.

Zak Eisenberg  14:33  

Yeah. So from there you decided to leave GE, or did someone approach you with an opportunity that you couldn’t refuse. How did you leave this company, which in many ways it sounds like was a really great home for you for a long time, and a great community and supporting you and growing as a healthcare leader. How do you make that transition? And what. Happened after GE, between leaving GE and really where you are today, and maybe just walk us up, yeah, absolutely where you are now.

Dusty Majumdar  15:08  

Yeah, I think I enjoyed very much the years in GE, and so I worked on imaging the last, I would say, close to a decade in at GE Healthcare, and my last role was in the computer tomography business, leading the product side of the business for what we call the premium CT product line. And again, learned a lot about imaging. Learned a lot about the clinical — various clinical applications. Got to know pretty much all the KOLs around the world, and developed the value proposition for some of the smartest CDs today in the market. It’s the ones that use spectral imaging, where you can actually peer into the lesions and actually tell if the lesions are benign or malignant by flipping between two different energies, development of cardiac CT, we’re now in one rotation. You can capture the human heart or the brain. So that is, you know, is amazing, even compared to the time when I started in imaging to when I left. Now, you know, I found it a fascinating area, and G gave me everything that I needed to succeed with the resources and the teams. And the interesting thing with imaging, though, is that you learn a lot about the anatomy, but you don’t really understand a lot about the physiology or the biology or the mechanism of action of what is really going on, you know, inside and back in the late 2000s I began to look at genomics. And the applicant, as you know, the human genome project was finished around the early 2000s and then next generation sequencing came by. There are companies like Illumina who were making a lot of noise. And I was in Wisconsin in the last few years that I worked with GE and there was a company in Madison, Wisconsin, which is doing amazing work called exact sciences in detecting colorectal cancer early using some of the genomics and the methylation of these different genes that are there, that express themselves. And I had an opportunity over there to go and lead the strategy and the innovation functions, working directly for the CEO, Kevin Conroy. And I thought that was a fantastic opportunity to learn about a whole new space that I’ve pretty much been oblivious to. And it was just a fantastic way to delve into genomics and really get to the bottom of you know, what really happens when the path to cancer starts in the tissue? Cancer can take around 20 years to develop. In most cases, it does. And how do you catch cancer early, even before stage one, as soon as dysplasia, for example, sets in and and the fact that you can actually catch it in blood or in stool or in urine, fascinated me, because now you can catch some of these diseases early on.

Zak Eisenberg  18:20  

Even before anatomical presence.

Dusty Majumdar  18:23  

And that was the most fascinating part. I worked with Luis Diaz on some of these different topics. And the thing that fascinated me was, if you had a surgical excision in the colon. Once a patient has cancer, and on the next day, you look at the liquid biopsy and you still see some CtDNA floating in the blood. You can be 99% sure that even if the doctor tells you the famous words, I’ve got it all, cancer is going to come back within six months. If you don’t see any CtDNA in the blood the next day after you have a surgery, cancer typically doesn’t come back within five years. And the other fascinating thing was, when the cancer starts coming back, radiologically, you don’t really see anything, but you start seeing things in the blood two months before a CT can actually detect a lesion. And those are the kinds of things that fascinated me, and I wanted, you know, propelled me, to go beyond, you know, just diagnostic imaging, and get into this whole world of molecular and genomic transformations that get into the biology of cancer and other complex diseases. So that was exact sciences.

Zak Eisenberg  19:38  

And then, do you see that as your transition from more, your chemical physics background, to more being focused on biology?

Dusty Majumdar  19:47  

Oh, absolutely. Well, you know imaging, obviously you have to learn sufficient biology when you look at perfusion, for example, of the brain, sure you look at the flow in the heart and, you know, ischemic failure. Is that you notice, I mean, you know, you got to understand, you know, about what really happens when you’re looking at a malignant lesion versus a cancerous lesion in the liver, and how you can differentiate that with spectral energy. So there’s always biology involved, but genomics gave you the next step,

Zak Eisenberg  20:21  

right, right? Yeah, it’s a different dimension to all of them, different dimension and different level of knowledge. I would say I know plenty of people in the imaging space who, yes, they have, of course, biological knowledge, but not necessarily as specific or as in depth. And yeah. Anyway, that’s an interesting transition to me, given just what I know now about where you started, in your background and your initial focus really on material sciences. Okay, so, so exact sciences, you spent time there in on innovation and what happened after exact sciences?

Dusty Majumdar  21:16  

So one of the things you know, while I spent time in innovation, or in genomics and innovation, was you notice how difficult it is to catch cancer early in blood. And, you know, I started wondering, and even if you see some of the companies who are doing liquid biopsy for early cancer detection in blood, their sensitivities are still in the 20 or 30% range for stage one cancer, that’s where you want to catch it. And I often wondered if the combination of multi omic data, for example, looking at methylation, mutation, proteomics, maybe even metabolomics, and getting all this orthogonal data together could actually give you higher sensitivities and specificities in early cancer, and that was just an idea back then. So I was looking around to see who are the companies that are involved in that, and what are the tools that could actually get you there. So what about IBM Watson really delving into this arena some of the acquisitions they were making, they’re making a lot of noise in the marketplace, and obviously Ken Jennings was defeated by IBM Watson computer in the 2011, 12 time frame. And there was a lot of excitement about what AI can do in healthcare based on some of the promise back then. So an opportunity came up for me to move back to the east coast and work in AI and lead a team of over 200 strategy and marketing associates, because IBM was actually acquiring six companies at that time, spending close to $8 billion Truven Exploris merge. They had their own people developing oncology and genomics solutions. And then there were solutions for payers as well, and the opportunity to work across life sciences, radiology, oncology, genomics, the pair segment, interacting with various governments, all on AI and healthcare and life sciences arena, totally fascinated me. And that’s when, you know, I thought it was a great opportunity to come and lead the marketing, commercial and strategy function. So that’s what I made, the transition from Wisconsin to Massachusetts in the East Coast after spending almost 10 years in the Midwest.

Zak Eisenberg  23:58  

Yeah. Well, I know you, you spent your early days on the East Coast, in the northeast, so maybe it felt like coming home a little bit. Yeah, for you. So this was obviously a much different type of role than some of your earlier roles, but it sounds like it really combined a lot of your previous experience into what you were working on, then really product development, innovation, biology and some of your more tech background. So while you were at IBM and leading IBM Research. How strategy in the Watson, Watson space? How did you start, or when did you start to think maybe you are going to start working on your own and. And start a company, because that’s a very large departure from the types of roles that you’ve been at until this point, you’ve worked at very large companies and very, you know, highly structured corporate environments and leading very large teams with near infinite resources, right? And obviously, now, and we haven’t quite gotten here, you’re running a scrappy startup with, you know, potentially world training technology. But when did you start having this inkling that you wanted to take a step into entrepreneurship? Was it? Was it during your time at IBM, as you were seeing the promise of or some of the promise of how AI could impact the space you have been spending at this point, several decades in healthcare and diagnostics? What or did it come up after you left IBM? How did that transition happen for you?

Dusty Majumdar  26:02  

Interestingly, I always had the desire to run a business out there, and whether it’s a private company or a public company, and it really came from the fact that GE was run like a startup, believe it or not, even back of the 90s and the 2000 we were on super lean, you know, we had to really prove ourselves every single day, and we had to establish, you know, the brand in the marketplace to stay ahead of fierce competitors like Siemens and others. Kind of similar to what you face in the startups arena, where you have startups mushrooming out of nowhere, and then you’re competing in a big way. So, believe it or not, it came during my GE days because I always wanted to, always have the desire of running a business where I could lead a group of really passionate and dedicated people who wanted to make a difference in the world. And interestingly, GE was run in a similar way. You know, we always felt like we were in a startup environment. The resources were lean. You had to deliver. And, you know, there were consequences for not delivering at GE, and you have to get along with the teams that you’re working with. And in some cases, you know you have to have both the competitive spirit as well as collaboration at the same time. You know, as you’re working with various teams around you, and I think that really was the crucible of my leadership and marketing training. Now, the other thing that I haven’t mentioned too much is I really cut my teeth in marketing. Every single year that I was in GE Healthcare, I worked with some of the fantastic leaders in marketing who would write the who really wrote the book on marketing back then, and worked with agencies like profit, profit in Chicago, where Scott Davis, who is a senior partner there, has been a mentor for 20 years in the arena as a branding and go to market strategy. You know, how do you really leverage some of the new tools like social media back then? You know, 2008, 910, how to use digital tools. So I was one of the pioneers, I would say, back then, inside GE Healthcare, going into these new tools that came up to propagate our value proposition, to really, kind of strengthen our brand, touch points back then, and then have the messaging that really would resonate, not only about the bits and bytes, but emotionally. How do you really have the messages resonating across a wide range of stakeholders? And I felt that a lot of those things that I picked up could be used later on, and obviously had the opportunity to use those at IBM. And as you said, with close to infinite resources and the desire to make a difference over there, and we had to position IBM Watson Health, which was not really known in the marketplace out there, in the marketplace, so that people would know what we do. And I think we did that with varying, very varying degrees of success in different arenas, position AI with clinicians who did not really believe that AI could make a huge difference, and also the whole value proposition of six different businesses. The I mentioned that IBM acquired six different businesses back then, like proven explorers and others. And what is the combined value proposition of all of that? What is the brand architect? How does the brand architecture look like? You know, once you acquire each of the businesses with that which had, you know, 150 different solutions, how do you get them together in. Again, working with pioneers in the field of branding and marketing like Scott Davis and others at profit. You know, we got that moving, and I would say that we did a great job in positioning IBM Watson held as a new company within IBM.

Zak Eisenberg  30:19  

So it’s amazing. It’s amazing to me as you fast forward just a decade, because I remember some of these commercials that Watson had at the time. They were really the leading market voice in the space at the time, kind of championing AI. And we had, you know, a transitional period of 10 years where the promise didn’t seem to land, and all of a sudden, of course, in the last two, three years, it’s just totally exploded into the public consciousness. Whereas, I think for people who have been paying attention to this space, this is a natural evolution, as you said, You’ve been in that space, really starting in the mid 20 teens. I, too, had been following the space for a long time. I was a little earlier in my life than you were at that time, but just coming out of college around the time that, you know, Facebook was bursting onto the scene, and some of this was transitioning. So it was definitely part of, I think, my generation’s consciousness of AI being not just a real potential, but a definitive part of our society, in our immediate future. Now, of course, I think everyone listening will have heard of AI in the consumer centric space. But what’s so interesting to me is how little people know about what’s happening, I think, in more scientific arenas and happening more on you could say B2B sides, but more more more of the scientific advancement. So, you know, often bring up Alpha fold, which I’m sure you’re familiar with this project by Google’s Deep Mind. Very few people really understand the implications of something like alpha fold. And I think there are lots of examples of this in the scientific realm now, and you know, maybe products bio is one of them that are really showcasing what the true potential is of AI, not just as a chat bot, but as a real way to drive innovation and change for the better, hopefully. Of course, there’s some people that might disagree with that, but yeah, so you’re, you’re at IBM, you are driving this team, bringing together these different products and businesses, and where do you transition? When does PredXBio burst onto the steam for you? Or was there another transition? Or any time between IBM and PredX? What happens?

Dusty Majumdar  33:12  

Yeah, absolutely. So, you know, I mean, you’re right about the fact that now AI is becoming more real. I think one of the good things that happened as a result of IBM Watson Health transitioning out of healthcare is the hype that was there at that time, not only from IBM, but also from some of the other companies, like Google, even Microsoft, was it was a very hype oriented, frothy atmosphere that was created back then with AI and healthcare and life sciences. Healthcare and life sciences are hard. You know, you really need clinical studies, clinical trials, in some cases, to make claims. And I think one of the things that was a realization with the, I would say, the IBM top leadership was that advertising and TV and telling doctors that this is a great way to do about your business is not a very good idea. So I think that I think it kind of solved. It solved. So, you know, I would say, you know, it resolved, you know, some of the unnecessary hype that was going on in the marketplace through the demise of some of these different organizations that were pushing it. And I think right now, we are in an arena where I think the hype has been mitigated to a great degree. There’s serious work going on in the last, I would say, five to seven years now, with AI and life sciences and healthcare, and we are beginning to see the results, you know, the low hanging fruits, like basic machine learning, deep learning, you know, pretty much commodity at this point, if you’re not. Using AI, then you know, you are essentially not going to be successful at this point of time in many different fields. So my transition from IBM was an interesting one. I went to work with ASCO, or the American Society of Clinical Oncology for a couple of years working with the leadership there to launch cancer link, and that was something that I enjoyed very much in the very different position, cancer Link was an AI based solution which would actually help doctors in community hospitals really get better guidance about their treatment, their quality metrics and other things that are critical, which takes a long time and resources and effort to do. And cancer link addressed some of those pain points the way cancel link did it is they sucked in the data from the oncology EMRs and then was able to curate that data and and then get back, you know, to the oncologists, whether it’s in academia or in community hospitals, with some guidelines about treatment, you know, how to manage their quality metrics, etc. So again, a very serious and deep in the oncology field, and really understanding either real world data and real world evidence, and that gave me an opportunity to be in the conversations back then with FDA and some of The companies like flatiron, concert, AI and others Tempest that were actually involved in this RW, D, R, W, E space. These are the early days. You’re talking about 2019, 2020, of all of this. And ultimately, you know, those, those, those, those, those different spaces became very relevant in terms of getting clinical evidence and reaching clinical endpoints, including in oncology clinical trials as well. So that was an interesting transition. And I also worked for Mitsubishi Gas chemicals, an arm of the Mitsubishi empire, who were getting into life sciences and healthcare back then in the COVID times, I was supposed to be in Tokyo almost every quarter, starting 2020, March and right then. So a lot of it was virtual, which is quite challenging to work virtually from from Boston, with a team in Tokyo who were also, you know —

Zak Eisenberg  37:43  

Just the time zone difference itself must have been —

Dusty Majumdar  37:47  

Yeah, exactly. And the focus was an organ on chips. And FDA just came up with a regulation just a couple of weeks ago, or maybe even last week, that they want to basically stop all animal trials in oncology, right? And this was the time when combined, there was an effort between MGC, Mitsubishi and University of Arizona to do exactly that, using organ on chips to see if you can actually get much better way to predict how humans would interact with different therapies, so growing the tumor and looking at it. And the amazing thing is, you know, we are again looking at all of that with spatial biology. You know, can you actually take tissues from organ on chip? And now, you know, with the development of spatial biology in the last four or five years, can you actually look at the interactions of cells and transcriptomics and proteomics within the tumor in a growing the tumor outside the patient’s body, and then predict what’s going to happen in the patient? So it’s all, all of these things are converging, which is amazing right now. And then again, the area of real world evidence and real world data, again, converging with AI in terms of how you run clinical trials, and it’s just an amazing convergence that you see all across at this point of time.

Zak Eisenberg  39:14  

Yeah, and Dusty, it sounds like it’s also converging with your background in many ways, and transitioning now to what you’re working on today, maybe just you can give a brief, brief overview to the audience, but I’d love to hear your thoughts about where you think this space is going to Go over the next five years, because I think you’ve given a great education, and you’ve really had a front row seat to how AI and big data and some of the advancements in particularly in oncology research. Because this has been a theme. For you with exact sciences, as well as obviously with imaging, it’s a constant presence. And then ASCO, you went back to this after IBM Watson, how, how this not just products, bios, technology, but just broadly how, how how you think AI is going to impact the oncology space. And again, we can, we can talk specifically about what PredxBio and what you’re working on today is doing in that space. But it’s just, it’s such an interesting question to me about personalized medicine, really, in terms of eventual development for where this eventually, I think, could go really predicting therapies on a very individualized level, but, yeah, if you could comment on, just paint a picture for the audience, a Bit from your perspective. Sure.

Dusty Majumdar  41:02  

You know, a couple of months ago, I had a webcast with Dr Eric Topol and Zak I think you may have seen that on LinkedIn, and this is exactly on what we are talking about. I knew Dr Topol when he was a young cardiovascular specialist at Cleveland Clinic, almost 30 years ago, and nobody knew about AI back then. And just reflecting on, you know, how AI has pervaded all arenas of medicine in the last 25 years is just amazing. Just think about radiology, something where I spent, you know, over a decade and a half of my life. Now you can tell with a lot more accuracy when you look at a lung nodule, if the nodule is going to be cancerous or benign, even just from a CT image. Right? Previously, you used to do a biopsy of anything that’s suspicious, and in more than 90% of the cases, you’d collapse the lung of a 85 year old smoker who’s been smoking for 5060, years, whatever, and then find out that there was no cancer in the nodule over 90% of the time. So AI can now, in a lot more definitive way, in combination with liquid biopsy, tell you if the patient’s lung audio has cancer or not, right? So amazing leaps in radiology. And then the problem that we talked about early cancer detection in liquid biopsy AI is actually bringing together multimodal data now, from proteomics, genomics, methylation signatures. You know, metabolomics and without AI, you know, combining all these orthogonal data sets in any meaningful way is just impossible. So again, AI playing a huge role there. And then, you know, in my IBM Watson Health days, you know, there were products for genomics and oncology, for clinical decision making. And now we have gone from there, not just, you know, linking what the AI algorithm is telling you to NCCN guidelines and so on and so forth, to now, you know, in spatial biology, using AI, unraveling the mechanism of action. So not only predicting if a patient or patient cohort would respond with over 90-95% accuracy, as we do in PredxBio, but also answering the question, why is the patient responding? Or, more importantly, why are these cohorts of patients not responding to this particular therapy, and so huge leaps in the last 20 years with AI and I’ve been pretty much in the front row seat through this old saga, starting with the GE and then exact, and then IBM and ASCO, and then and then with PredxBio and several other startups that I worked in as as a consultant or advisor in the interim. So it’s an amazing field, and specifically what we’re doing in PredxBio, we work with pharma companies. As you know, pharma companies run clinical trials. They start with phase one, they have around 30 patients there. Then they go to phase two, probably around 75 to 100 patients. And phase three could be hundreds of patients. And what we do in phase one with pharma companies is we look at these biopsies that they get from the patients in phase one of the clinical trial. Then, in collaboration with the pharma companies, label them with certain markers. It could be proteomic markers, could be transcriptomic markers, and then once the image is uploaded in the cloud, we look at that with the. Computer Vision, and the AI algorithm that PredxBio has developed over the last 20 years, actually, and we try to unravel cell cell communication, as well as the organization of cells into certain patterns at PredxBio, we call those micro domains, which are distinct spatial functional cellular patterns that emerge from peering into the tumor microenvironment, you know, with our space IQ platform. And what the amazing thing that we have noticed over the last few years is that in solid cancers, as well as when you get tissues from liquid cancers, like lymphoma, you come across these patterns, which are constellations of cells coming together in a specific configuration, which are repeatable across the tumor microenvironment. And once you’re able to in a hypothesis freeway, and this is very important and an organic way to identify these patterns, you can pretty much predict if the patient is going to respond to therapy or not with over 90% confidence, in some cases, over 95% confidence. And again, this was different from anything I had done in the past. I mean, who would have imagined that you know, you can peer into a tissue and you see repeatable patterns. Cells are organizing themselves in repeatable patterns. So what this, again, taught me is that you can look deep with genomics, and you can look go real deep with single cell genomics, but if you don’t understand the spatial organization of these different cells and how they’re communicating with each other, and frankly, how they’re conspiring with each other to move the disease forward, you are probably unraveling less than 2% of that hidden circuitry of cancer that you can just do with genomics. And as you know, there are many drugs based on genomic mutations, right, EGFR, or whatever you know, ALK. And for a very specific number of patients, you know, 8% of patients who get non-small cell lung cancer, yeah, if you just have a drug that’s based on EGFR mutations, you were successful. But for the rest of 92% you’re firmly in the dark unless you really understand the organization of what is going on inside the tumor.

Zak Eisenberg  47:27  

Micro organization, anatomically, yes, yeah, anatomically, spatially,

Dusty Majumdar  47:31  

You know of the cells and their communication patterns, then you can catch with the proteomic or the transcriptomic biomarkers. So there.

Zak Eisenberg  47:41  

So obviously, I think at least where my head goes with some of what I hear in AI is, okay, this is where we are today. But what’s next? And I know we’ve had some conversations about this, about, you know, truly, because at the moment, you’re taking real world data from patients from a clinical trial, and your system is ingesting that data and responding with predictive analytics about if a particular therapeutic based on based On the treatment that had been delivered because those patients had received, some of them have received this therapeutic will have an impact across a larger cohort. And how efficacious will it be? Will it be efficacious and why? In particular, one thing that we’ve talked about is, what if you can do this digitally over overall, with truly predictive digital models of these micro environments, these, as You call them, micro domains, across a digital patient and and different types of digital patients with different, you know, genomic histories or biologies. And I think we were talking about this briefly earlier, but I’m curious your thoughts, if, if that future is where we will eventually get to truly personalized medicine, where a company is able to, in a rapid and affordable way, potentially even Design a therapeutic specifically for for you, because there could be a digital model which would allow repetitive testing of therapies at this micro domain level to understand efficaciousness across different therapeutic combinations, etc. Laura, I’m curious how real this is, because I’ve heard this from multiple people. I’d say in the AI life science space that this is more or less the holy grail of potentially where AI in this domain, in this space could eventually go, which sounds both fantastical, but also amazing for humanity, if we could truly achieve this for ourselves. So I’m just, I’m just curious, and I know products, bio is, you know, a, I think, a critical step along this way. First, you have to be able to predict it in real life patients. But it’s similar to me, and you brought this up with Mitsubishi of organs on a chip, and I’ve come across other companies, let’s say, you know, eight to 10 years ago, when I was working in venture capital, that were growing neurons on a chip. And neurons are very difficult to grow and grow in the lab. Structurally, they’re very, very challenging. And the possibilities that that opened up, which is really only starting to come into realization today. So even some of those technologies are very, very nation still in terms of usage in trials. But to me, it’s all about repetition and being able to move the cost curve down scale. You know, patients are most expensive, organs on chip, very expensive animals, etc., if you can do it digitally, obviously, that’s the best case scenario for everyone. But yeah, curious  to hear your thoughts on that.

Dusty Majumdar  51:52  

Yeah, I think we are well on our way to virtual clinical trials or in silico clinical trials. So if you think about what we are doing at PredxBio is to start with. Once we discover those micro domains organically, we can essentially extract the emergent biology from those micro domains based on who’s near to whom, in terms of cellular phenotypes, who’s talking to whom. How are they organized in space, you know, we can get to the network biology of that particular responder or non responder. And once we get that, that’s really kind of the, you know, the virtual network of what’s really going on inside the tumor microenvironment. And then you could perturb that network to test for various interventions, therapy A versus therapy B versus C, or combination of A and B, you know, with C, that’s where we want to go. And as we’re building our micro domain library, working with the data sets from NIH NCI, which we’ve been doing for almost 10 years now, that is a distinct possibility, because then you could shorten the time of clinical trials from, you know, a few years to days, virtually, right? And the cost can be slashed, probably by 98% it still will take some time. And obviously, FDA, I was just in a talk with Dr Makari from the newly appointed FDA commissioner two days ago in New York City. And I think his forward thinking in some of these areas will also help. Scott Gottlieb was very interested in doing something like this when we talked to him when he was the commissioner, just, I think, right before COVID. And then things kind of changed, right and now, I think, you know, with Macari in place, we see that kind of forward thinking again coming, you know, from him. So the FDA is absolutely interested in working with companies that can provide them the confidence that virtual clinical trials and the data, of course, that virtual clinical trials could work and taking it to the next step, many pharma companies in as you know, the cost is unbearable, $2 billion to run a cancer clinical trial takes around 10 years, and it fails 90% of the time. So what if you could actually simulate a clinical trial before you actually run it, and you predict, you know, what’s going to happen, and you design your clinical trial in a way where you know you have a high probability of success, you know, before you go totally virtual, you have to kind of show that in the virtual actually predicts.

Zak Eisenberg  54:38  

Yeah, there’s going to be a transitionary period where maybe, and again, I’m glad you brought up FDA, because one of one of my questions was, how, of course, whenever you’re talking about healthcare, you have to be talking about governments, not just FDA, but regulatory regimes across the world, because ultimately, healthcare is about people and the government. And is really the advocate of the people. And when you’re in healthcare, you are in business with the government, regardless of what you’re working on. At some level, you are in business with the government. And it seems to me, from what we’ve seen in the communication realm of AI, that the government is not reacting very quickly to AI, and maybe that’s a good thing in some spaces, but in healthcare, of course, we want to be cautious. I think conservatism is good when you’re speaking about impacting people’s biologies. But I’m curious your thoughts about how what role you think government will play, and what your hopes are for government over the next, let’s say, five to 10 years, there’s obviously, I think, much of this vision that we’ve been talking about, I think I share your excitement. Much of this seems to me, will be impossible to achieve without government playing a critical role and and being on board with the this new technology.

Dusty Majumdar  56:10  

I think the global situation is very interesting to look at. I know you just came back from the Middle East, and over there in Dubai or in Saudi Arabia, you probably saw the amount of dollars that are flowing into AI right now and healthcare together, right the Ministry of Health, as they call it over there, MOH, is increasingly interested in automation and having the most cutting edge AI systems in place, you know. So when I visit Dubai and Saudi Arabia that tell me that we want to do it tomorrow, if you can show us that it’s going to help our patients, right? So in some parts of the world, things are going to move fast. I think in other parts of the world, where you have a huge density of population and you don’t have the doctors really in place, like India, in remote areas, I think AI to analyze retinopathy, which is very common with the population that’s increasingly prone to diabetes and other diseases, as well as read scans in an ER room remotely using AI. I think those are things that are going to take hold, and the government would be highly supportive in the Western world, as I said, you know, with the new FDA commissioner, I’m hopeful that things you know, would start moving fairly quickly in that direction. And you know, Europe’s been a little slow in adoption compared to the US so far, but I see that picking up as well, you know, with companies like Auken out of Europe, you know, who are really trying to evangelize AI and hospitals, I think that could, that could really move very fast, especially in some countries like France, Germany and the UK. Those are really the three leaders and deep mind being situated in the UK, I think helps the ecosystem there, you know, for several different AI solutions, from alphafold to understanding the chemistry of drug discovery as well. So I see, you know, from a pharma company perspective, you know, beyond the government, with the kind of layoffs that we have seen in pharma over the last six to nine months. They need more and more AI based solutions to run their clinical trials efficiently, whether it’s patient selection, whether it’s going into discovery and accelerating discovery to yield new drugs, whether it’s stratifying and selecting patients, as I said, in clinical trials for phase one, phase two, and then ultimately predicting what’s going to happen in phase three, and selecting the right patients for phase three. I think it’s all going to be critical. I just see in, you know, this is and so many of these solutions that we talked about Zak are converging now. You know, you’re using AI to really understand what’s happening, maybe in an organ on chip situation. So one is feeding the other. You’re developing AI models, and you’re looking at what’s really happening in the biology and organ on chips, they’re feeding each other, and ultimately, you can go fully virtual. Same thing with clinical trials. You start with predicting what’s going to happen using in silico tools that we are developing, and then you figure out, if you can actually eliminate the clinical the actual real world clinical trial, it’s going to take time, but things are moving in that direction, and I feel that many of the regulatory agencies around the world understand the need for those developments.

Zak Eisenberg  59:38  

I think that’s a great place to wrap. I’m very pumped and upbeat about this Dusty. I know we’ve talked about this many times before, but, yeah, it’s an extremely exciting future, I think, for healthcare. And I you know, again, I said this earlier in the podcast. I think. So much of the public discourse around AI has been fairly negative or benign about job displacement or chat bots or deep fakes, and these are all things to be worried about. We should be concerned about this, but I don’t think we really as a species, because this isn’t just about what the US is doing. And of course, AI is centered on the US at the moment, though, of course, China and Europe have a role to play here, and they’re also pursuing this very aggressively. It really the impact potential for humanity is just massive, and it is absolutely exciting for me and to know you and know this journey you’re on. And I think for our listeners, this probably was a really interesting story for them to hear about how your life transitioned into the space and really how you’re converging. You said the technology is converging, but also your background is converging at this time perfectly for what you’re working on. So with that, appreciate you coming on the show and glad to have you today.

Dusty Majumdar  1:01:21  

Thank you very much, Zak.

Outro 1:01:23  

And that wraps up another episode of Transaction Healthcare. Hit the subscribe button to get notified when we release new conversations. And if you are someone interested in learning more about these topics, visit us@merrittadvisory.com or send us an email contact us at merrittadvisory.com.