Jeff Elton (00:01):
Hi. This is Jeff Elton, CEO of ConcertAI. Welcome to our podcast series. Today, I have ConcertAI’s head of partnering (recently promoted to CRO), Claudio D’Ambrosio, who will talk a little bit more about himself, his own career, multimodal data, which is one of the most rapidly changing parts of the real world data industry, and how ConcertAI views partnerships, the importance of partnerships for advancing biomedical innovations. So, Claudio D’Ambrosio, welcome to the podcast. Let me start with, really, some questions about yourself. I know you had an academic career, really in cancer research. So what I’d really like to understand is what compelled you to the field, what kept you in the field, because you’re still in the field, and as you thought about making a transition into a company and a context like ConcertAI, what was it about what we’re doing that connected you back to the beginning of your professional career?
Claudio D’Ambrosio (01:03):
Thanks for having me here. Yes. I was a cancer research, a few professional lifetimes at this point, before I took on a more entrepreneurial and commercial identity. I was privileged to spend several years in a deep, kind of high-profile cancer research setting, studying basic mechanism of cancer progression, genomic stability. I don’t miss being in a wet lab.
Jeff Elton (01:28):
Where was that?
Claudio D’Ambrosio (01:29):
At Francis Crick Institute in London.
Jeff Elton (01:31):
Got it. Right. London. Yeah.
Claudio D’Ambrosio (01:33):
Yeah. While I don’t miss being in a wet lab anymore, I feel lucky to have touched, with my own pipette, several cancer pathways and that, actually, today, to answer your second question, are translating into new therapies, to really going from those kind of biological hypothesis a decade or two decades ago into real options right at the patient bedside.
Jeff Elton (01:58):
What was your focus when you were doing work back at Crick Institute?
Claudio D’Ambrosio (02:02):
Yeah. Really studying mechanism that leads to cancer, especially on cell division, mapping checkpoints, cell cycle checkpoints. These days, we talk a lot about immune checkpoints.
Jeff Elton (02:15):
Exactly.
Claudio D’Ambrosio (02:15):
But in that respect, I was maybe in the wrong lab. I spend most time in cell cycle checkpoints, DNA damage pathways.
Jeff Elton (02:25):
So now that you’re in an organization like ConcertAI, you have the opportunity to see the biomedical research activities of 35 pharma and biopharma. You have the flow of the largest collection of kind of research-grade data of anybody in the industry. What do you find motivating now being in this context?
Claudio D’Ambrosio (02:49):
Absolutely. That’s a great question. It really is a passion. I love to contribute to a world without a fear of cancer and other devastating disease, but I also love being at the intersection of so many innovations coming together in healthcare in general, but also through the work that we do, innovations in medical data,
Jeff Elton (03:32):
I want to pick up on this a little bit. This term, multimodal and multimodalities, has actually kind of become more in common usage, and in part, the accessibility of whether it’s genomic, transcriptomic, exomic, pathomics, digital pathology, clinical data, claims data is more accessible today than, certainly, was five years ago, and where we’re doing that. So when you think about that, give it some definition. So if we’ve got people that aren’t as familiar with that term as we may be, what does that term really mean, and why should somebody care about that?
Claudio D’Ambrosio (04:12):
That’s a great topic, and yes. The term multimodal is growing, and, as we’ll discuss, that’s a good thing. But let me first take a step back before I answer the question directly. First, I think we are at the point of reaping the benefits of the great health data digitization process, which really started this century. Every medical data at this point sits there in electronic medical records. Companies like us are really in the job of taking the collective information and use it to inform how to develop the next therapy and how to best treat the next patient that we meet.
This data now comes in different flavors or multiple modes, radiology data, pathology data, clinical data, socioeconomic data, and molecular data. Molecular data alone can be broken into genomic, transcriptomics, proteomics. So these are the uni mode of the multiomic construct. Now, we made enormous progress in extracting as much value and insights as possible from unimodal data or maybe dual-modal data like electronic medical record combined with claims. In fact, ConcertAI, I think, has contributed tremendously already in this field alongside a few others.
But the application of RWD from EMR data, imaging, molecular data, today, the progress, put simply, is already pretty amazing. So we have like VASARI features in glioblastoma developed from radiology scans. We have EGFR mutational signature from CT scans. We have BRCA signature from field digital mammography. We have molecular diagnosis. We have MSI-high. But there is so much more opportunity in the insight captured in the multimodal dataset.
Multimodal data can unlock the ability to discover actionable prognostic and predictive associations that, today, are invisible to us by dealing with unimodal data sets, so because unimodal data suffers from integrator variability, and databases are still not large enough to be used as really solid training data sets for machine learning applications. Like someone said this already, but in healthcare, big data gets pretty small pretty quickly once you apply the filter that we need to be precise, hence the world precision medicine, by the way. So I think that the opportunities by multimodal data are going to be unique. They’re going to compliment each other. They’re going to unlock new discoveries for research, for clinical development, and so forth.
Jeff Elton (06:47):
So I want to check a concept with you, because as I’m listening to you, a couple things are coming forward for me. So for one, we used to say, “Here’s a data set. These are the data we have. I hope it gives some utility to the questions you have,” but you’re bounded by that unimodal, maybe bimodal kind of data that’s there. But what I was hearing from you is I now have a richness to more physiological features and characteristics that actually can be captured in a myriad number of ways in different digital form, but all digital forms, all machine-readable.
So now, my question might be able to say, “Here’s my scientific, or here’s my research question. What are the appropriate data types that might begin to inform that question?” which is a very different way of approaching the field with actually a lot more rigor, really, and a lot more flexibility. But the other part I heard, maybe just to check this out, is methodologies here, because you used the term machine learning, et cetera. I might say, “I’m not bounded where the published evidence is, because I have these multimodal. Because of the richness, I can have high confidence because of that richness, but I can actually bring new methodological approaches to that.” Is that consistent with how you’re thinking about this?
Claudio D’Ambrosio (08:06):
That makes complete sense. I think, basically, it opens up to a much more hypothesis-driven approach to consume data.
Jeff Elton (08:13):
Hence more scientifically rigorous.
Claudio D’Ambrosio (08:15):
More scientific, rigorous, and multimodal programs and analysis are going to be different. They are more extensive. The combinatorial permutation of, say, even genomic and transcriptomic association, for instance, become pretty large pretty quickly. So we almost don’t know what we’re going to discover until you put some analysis into place. This is why I think bioinformaticians will become a scarce and valuable competencies as genetic counselor have been for the last decade. So by the way, kids in high school tune in, this is …
Jeff Elton (08:47):
Ah. I would.
Claudio D’Ambrosio (08:47):
I’m offering some free career advice.
Jeff Elton (08:49):
I was just thinking the same thing for my college-aged kids, that maybe that’s where I should be sending them for the next phase of their career. But let me ask you another question now. Let’s take some of the critical functions in pharma, biopharma, and let’s take this multimodal. So now, I’m in translational medicine. What does the configuration of a translational medicine function look like today, knowing that multimodal data and research methodologies are available to them?
Claudio D’Ambrosio (09:16):
Yeah. So it really closes the gap between hypothesis, development of, say, a therapeutic or new diagnostics, and the feedback that comes from the patient. It really shorten that cycle. It shrinks it. I mean, for a translational scientist, I think there’s an explosion of use cases in the molecular space, in multiomic, in discovery, in informing target selection or drug selection, in combination and sequencing once the therapeutic has passed IND and gets into human. This makes sense. Once you put a new drug into the complex human body, the more data we have about those molecular interaction, the better we’ll be able to accelerate.
Jeff Elton (09:55):
So you see trial data collection also now changes to take advantage of multimodal data.
Claudio D’Ambrosio (10:01):
Absolutely. Absolutely.
Jeff Elton (10:01):
It’s clear you have a lot of energy on the topic, and so what I’d really love to hear is, what’s one of your favorite examples of where multimodal data may have kind of come together and either provided a benefit that was prior unachievable or where there’s just a remarkable insight that became actionable?
Claudio D’Ambrosio (10:20):
So take a very exciting field like PROTAC. This is a technology that hijack millions of years of evolution to order a cell to degrade a specific protein. So you may have some genomic driver of a cancer, but if we are effective, a PROTAC, we may be able to shortcut and degrade a protein of interest that drive the disease. So, for instance, now, you have in development, PROTAC assets against, say, women estrogen receptors or ER. So we can use EMR data and transcriptomics together to test the effectiveness of these approaches before we ask thousands of patients to sit in a trial.
Jeff Elton (11:27):
That’s remarkable. Thanks for the depth, and thanks for sharing the example. So I want to take you a slightly different direction. You have an unusual background, as we kind of established, right, and clearly, still a passion for the field. But your role is leading partnerships. So when you think in your own context of a data-technology-rich company interacting with biomedical innovators, what our partnerships mean, and how do you differentiate that from more classic commercial and kind of sales process?
Claudio D’Ambrosio (12:01):
Ah. That’s a great question. Yes. Partnerships, to me, means we are interested in something bigger, more audacious, a goal that is meaningful to both of us through a collaboration that is more than a single transaction. It typically means higher stakes. It typically means programmatic alignment with a certain scale.
Jeff Elton (12:22):
For sure, on the higher stakes. Yes. For sure.
Claudio D’Ambrosio (12:25):
It often means this partnership are unfolding over multiple years where we codify the partners, the life science partner goal, into our technology and data roadmaps very deliberately. So it’s actually a great way to be an entrepreneur.
Jeff Elton (12:39):
So, without mentioning names, of course, just because the nature of some of what we do in the partnering area, just as you described is confidential, but if you had to kind of say, what are the hallmarks of the ideal partnership in your mind, where actually, the two parties really do make more than one plus one equals four or five?
Claudio D’Ambrosio (13:00):
Well, simply put, is when we have an impact is when our life science partner, armed with this new evidence or technology, can actually scale or achieve that something that wasn’t, not possible before this encounter took place. So when we are very programmatically-aligned, we ask a lot from each other. There’s executive sponsorship, but we also accept that innovation is art and create a forum and a cadence for continual revision. It’s almost like being part of the same company.
Jeff Elton (13:32):
I love that idea that there’s functioning as one, that it is so integral, it is as though they’re part of the same work units and the same programmatic, so that everything truly is synergistic, and it’s common set of interests, common set of outcomes we’re trying to realize. I want to kind of build on that theme a little bit, and I know that you personally are a very patient-centric individual and have chose to place yourself at the forefront of patient interests. Even as you’ve talked to me, and even as you decided that we were going to be working together, that patient’s interest, it really kind of emerged out of every aspect of your conversation. So as you look forward over the next three years, four years, no more than five years, and if you kind of actually thought through what you’re most excited about that you think could really bring value to patients, perhaps against some of their still currently-high unmet medical needs, what are you most enthusiastic about?
Claudio D’Ambrosio (14:27):
Yeah. It’s a great question. I can answer in two ways. First, I’m very excited by the opportunities of real-world evidence to really improve R&D, infuse new data, say, how to design a control arm. None of us will want any of our friends or families to be on a control arm with the least effective therapy. So how can we advance the field to make clinical development as humans-less as possible? Right? I’m also very excited about the opportunity to bring research to community to improve patient and trial matching ratios. Those are enormous use cases. If we can move the needle in decimal percentage point, would be great. If we could do that in full percentage point, that would amount to technology miracles.
To patients, the second message, I would say to humbly not assume your doctors, caregivers can know it all. There is more data out there than anyone can absorb. Every day, hundreds of new publications hit the field in probably areas that are relevant to your care pathway. So the field is moving fast, and you need to have an active role in your care. There are new options coming through trials and new approvals of new medicine regularly. So our genome changes, and so you also need to change and has to be involved. That will immediately raise the quality of your care.
Jeff Elton (15:47):
So Claudio, it’s a marvelous vision, and I think this notion, and I think your point about bringing clinical and the latest clinical research programs and capabilities to that community level, where 70, 80% of the patients get their care, and where trial participation rate can be one to three, and best 4%, and totally counter to the leading comprehensive cancer centers, that’s incredibly important. That’s a very, very powerful vision. So I want to thank you, Claudio-
Claudio D’Ambrosio (16:17):
Thank you.
Jeff Elton (16:18):
… for actually being part of the podcast here in the ConcertAI Cambridge facility, and I look forward to doing this again in the near future.
Claudio D’Ambrosio (16:26):
Thank you very much. Thanks for having me.
Jeff Elton (16:29):
So again, I’d like to thank Claudio D’Ambrosio for being part of this podcast. We covered a few real critical themes. Multimodal data now can allow some of the deepest, most advanced research questions to be answered. We start with the questions, and then we bring together layers of different data of different origin that surround the patient in service of that particular question. This will change and transform how we develop and test hypotheses and will accelerate early phase trial activities.
In the area of partnering, Claudio actually developed that biomedical innovations really is a team sport. It actually requires the inputs of multiple types of expertise, and it’s in the dynamics, operating as a single company or single entity, that we can really transform the performance of research activities. To all of those who joined us, I want to thank you. If you’d like more information about this or any of our other topics, please contact us through concertai.com. Till next time, good morning, good afternoon, good night.