Aaron Leibtag, CEO & Co-Founder at Pentavere: Unlocking the power of clinical information using advanced AI

How is it in healthcare that so much vital information is lost, buried, locked, or out of reach?

In this episode of Bite the Orange, we are excited to have Aaron Liebtag on the show. He is the CEO and Co-Founder of Pentavere, a digital health company that has developed an artificial intelligence engine that provides insights from buried information within clinical documentation to help improve health outcomes. A tragedy his co-founder suffered that could’ve been avoided if the relevant information hadn’t been missed by the corresponding care team made the duo realize the magnitude of the unsolved problem healthcare had. After thorough research, Aaron and his colleague founded Pentavere and developed the Darwen AI platform, which uses AI and ML to curate, organize, catalog, and analyze information buried in unstructured clinical documentation, and validate it against peer-reviewed publications. Throughout this conversation, Aaron talks not only about how this technology can solve the problem they first targeted but also be applied to many other areas like patient identification, clinical trials, and even commercial pharma.

Tune in to learn how Pentavere’s AI engine is using technology to navigate medical information easily instead of getting lost within it!

FULL EPISODE

BTO_Aaron Liebtag: Audio automatically transcribed by Sonix

BTO_Aaron Liebtag: this mp3 audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.

Emmanuel Fombu:
Welcome to Bite the Orange. Through our conversations, we create a roadmap for the future of health with the most impactful leaders in the space. This is your host, Dr. Manny Fombu. Let's make the future of healthcare a reality together.

Emmanuel Fombu:
Good morning, good afternoon, good evening, ladies and gentlemen, and welcome to another episode of Bite the Orange, the podcast about innovation. And today we have a great innovator and great friend Aaron Liebtag, that is going to join us today and share with us some of the great things that he's working on. And trust me, you'll find out this is truly transformational in our industry, and I'm very happy to meet with Aaron today. So with that much ado about nothing, let's go right back. We start off by getting to know Aaron. So, Aaron, tell us about yourself.

Aaron Liebtag:
Sure, my name is Aaron Liebtag. I'm the CEO and co-founder of Pentavere. Pentavere is a digital health company that has developed an artificial intelligence engine that accelerates knowledge and speed to insight of the huge amounts of information that's buried, that's locked in the clinical documentation that makes up the majority of the health information that we create, and the majority of the information that is created that's not used to help improve health outcomes. Before that, I spent about 15 years working for some of the largest retail organizations in the world, as well as private equity companies who would acquire consumer-facing assets in order to improve their operations.

Emmanuel Fombu:
Thanks for that background, Aaron. And in my experience working in digital health space, I know there's been a lot of innovation in the financial industry and even in the transportation industry, right? If you look at airplanes, for example, especially when it comes to AI. Autopilots have been existing for a very long time, right? So I think having this kind of backgrounds in other fields and bringing them into healthcare is truly transformational. But with that being said, please tell us what was the problem that you were trying to solve with Pentavere and what made you actually start this company.

Aaron Liebtag:
What you just described, Manny, in terms of other industries using technology and using artificial intelligence in a far greater way than what's currently being done in healthcare, and a tragedy that impacted my fellow co-founder and our chief technology officer is the reason why we started Pentavere. So just a little background on our chief technology officer. Like me, his background is not in healthcare. He spent 30 years working for some of the largest financial service companies in the world, architecting and designing systems so they can make billions of dollars in real-time. His mother went into the hospital with what should have been a routine procedure. The fact that based on some co-morbidity she had, she needed to be put on a certain medication post-operation, was missed. It was missed by her care team because it was buried in a clinical note, and as a result of that, she died. So the question that we asked ourselves is how is it other industries, technology, AI, data analytics is used to help these companies make huge amounts of money, yet in healthcare, so much of that information is tracked, it's locked and it's not used because it's in clinical documentation. So we started Pentavere to solve that problem.

Emmanuel Fombu:
That's truly incredible. And I'm sorry to hear about the passing of your partner's mother, but definitely we have to make sure that this doesn't happen again, right, and to improve things going forward. And that's the point of you starting this company, Pentavere. And so to begin with, to boil all that down into simple language piece, I understand that you offer several solutions right through Pentavere. And tell us about, first of all, I know you have an engine called a Darwen AI. Maybe tell us about Darwen a little bit and then tell us about some of the solutions that you actually provide.

Aaron Liebtag:
Absolutely, so as a result of that tragedy, we naively thought that there would be systems or solutions out there in the market that can solve this problem. We embedded ourselves for two years in one of North America's leading hospitals to really understand the data problem from a privacy and a governance perspective, from all of the different silos and the fact that systems don't speak to each other as well as all of the information that exists in clinical texts. We were surprised that no solution could solve the problem by unlocking this type of knowledge from electronic health records, from clinicians' transcriptions. So we started to build our own technology, which is the Darwen platform. And one of the things that makes Pentavere unique, and likely why many of your listeners haven't heard about Pentavere before is, we haven't raised any money. Instead, we approached many of the largest pharmaceutical companies in the world and said, We are committed to working and solving this problem. And through collaborations with them over the last five years, we have built and validated our Darwen technology, which, better than any other technology in the world today, can scale the curation organization, cataloging, and analysis of information that's buried in unstructured clinical documentation and have another number of peer-reviewed publications in prestigious journals like the Journal of Thoracic Oncology or Journal of Medical and Informatics Research, attesting and validating our technology's ability to do this. And through that, we have developed and identified another number of high-impact business applications in clinical trial design, clinical trial operations, real-world evidence generation, as well as patient identification and insight identification that is of high value to the commercial teams within the pharmaceutical companies that we work with.

Emmanuel Fombu:
That's quite impressive, Aaron, to create as much evidence without raising any capital from venture capitalists, right? And a lot of times the big challenge in this particular industry is to show evidence, to show a proof of concept that something has worked and which you have done significantly. I've looked at several of the studies that you've published and which is available in your website. Anyone could log in on your website Pentavere.ai, and we will have it on the show notes as well for anyone that wants to actually go in and read some of these papers. So but what are your ideal clientele? So what do you sell your solutions to?

Aaron Liebtag:
So if you think about the scope and scale of the problem, which we are hearing is one of the largest problems in healthcare, is that we still cannot efficiently and cost-effectively get at the lifesaving information that we need when we need it. So in the early days of AI in healthcare, the system that we've created can benefit many stakeholders. However, what we're focused on specifically is around our life science companies and pharmaceutical companies who are spending a huge amount of money, taking up a huge amount of time to identify patients for clinical trials, to review data, and manage data as part of our, their clinical trial programs. Also to identify patients who should be on marketed therapy based on their diagnostic criteria but aren't. So our core customers today, as we are beginning to commercialize and scale the system that we have developed, are leaders within pharmaceutical companies who are looking for better solutions, real solutions, validated solutions that accelerate knowledge from real-world evidence that's buried in the huge amounts of clinical text that is not helping them get the right treatments to the right patients at the right time that is needed today.

Emmanuel Fombu:
And that's very interesting as well, because I see you basically offer end-to-end solutions, which is very unique in this particular space. So you offer solutions for clinical trials in terms of patient recruitment, patient identification, but the same model, the same concept can be used on a commercial side as well, correct?

Aaron Liebtag:
Exactly, what we demonstrated together with some of the largest pharmaceutical companies in the world as we developed and validated our AI engine is that the business applications are numerous. That same engine that can extract data from an electronic health record in order to generate real-world evidence can also scan the huge amounts of electronic health records and patient records to identify patients based on very, very specific inclusion or exclusion criteria as part of a clinical trial. It could also identify patients, like what we're doing right now in heart failure. Patients with confirmed heart failure who aren't on medication, who should be on medication, who are at imminent risk of a heart attack, which will cost the health system a huge amount of money. But from the commercial side of a pharmaceutical company, these patients should be on medication, they just need to be found. And our engine can add value to each of those applications, but what we're really excited about is the idea of patient identification. That is the core capability that we are really scaling as part of our solution right now.

Emmanuel Fombu:
And I think all this work is not going unnoticed. I'm also familiar with the idea that, I mean, I think you were nominated by the Galien Foundation for the 2022 Prix Galien USA, which is a great, great organization, and nomination. I've served on the board of the Galien Digital Health Committee before as a judge, and so I fully respect the companies that actually are nominated for this prestigious. And also, I noticed you also won an award from Takeda, Canada.

Aaron Liebtag:
That's correct. It's really a testament to the team, to the partnerships that we've been able to bridge, and the collaborations we've been able to do with companies like Takeda, like Novartis, like Amgen, like Johnson and Johnson. But it's interesting you mention our Prix Galien Award nomination because as I said, no, we haven't raised any significant capital. We don't invest in marketing or business development. So when we got the call saying, A, they've heard about Pentavere, but B, we were one of 12 companies globally nominated for the reward. I said, How did you hear about us? Like, how did a company like Pentavere that is developed our technology and the way that we have, get on your radar? And the answer we got back from the Galien Foundation was that for the digital health, one of the criterias is publications and peer-reviewed journals. And you're one of the few AI companies that met that criteria, and we believe in the impact that you're having for human health, and therefore we've nominated you. So it was an incredible honor just to be nominated, and hopefully, we will win, but we'll hear those results in October.

Emmanuel Fombu:
But congratulations on that. I think that's a great achievement by itself, and I wish you the best. Whether win or not, it's irrelevant to the point that you have great publications and we're making lives even much better. With that being said, I know sometimes when everyone looks at success, you think it comes easy. So what are some of the challenges that you faced so far?

Aaron Liebtag:
So I think any entrepreneur who's listening to this podcast, who's stuck with it and has gone through the ups and downs, knows how hard the journey is. And that's the challenge, is the journey. Two individuals with no healthcare experience coming off of a tragedy that just fundamentally impacted my partner's life and then making that choice to fully commit to solving a problem despite the probabilities of being successful, and that, the journey, is the challenge, but I must say we've been incredibly lucky along the way. By staying true to our North Star, we've been able to build friendships with incredible individuals like yourself. We've been able, through real projects, real delivery, real value, forge relationships with individuals in the life science sphere, prestigious researchers by breaking through barriers and building technology to have impact, believing in the problem and saying, if we can solve the problem to unlock the huge amounts of knowledge, the huge amounts of unknown insights that's buried in clinical text, electronic health records, pathology reports, clinician transcriptions, we will find those business applications that can solve huge addressable markets and change the trajectory of human health at the same time. And that North Star has enabled us to push through many of the barriers, barriers of not receiving funding, barriers of being a technology company not birthed in Silicon Valley or Boston or London, England. But despite that, we've achieved results that what we hear from our partners are unlike any results that many other companies have achieved.

Emmanuel Fombu:
And you mentioned the North Star, and talking about North, I mean New York, up in Canada, where are you located? And generally, in the AI space, I think Canada has some of the top AI engineers in the world, right? I know University of Toronto is pretty solid, especially around a particular ... of things. And several books I've read have come from professors out of Canada in Toronto. So the science from AI perspective is very solid up north for where we are, right? And I think we met the first time about three or four years ago at a time I think was pretty early stage where you didn't have all the studies that you have now. And so despite COVID and the pandemic and everything shutting down, I was quite, very impressed when I saw all the papers they came up with and all the partnerships you were able to do, even despite the hard times. So with that being said, you've accomplished quite a bit in a short amount of time, right, without even getting a lot of funding. So what does success look like for you? What would you like to go in the next six months or the next year?

Aaron Liebtag:
Yeah, so the inflection point that we're at is we built and validated and proved out a number of high-value business applications. Now it's all about scale, scaling commercially, scaling with healthcare providers, and having global impact by taking a technology that we've developed based on real data, real different types of data, and how can we take some of the solutions that you've described and demonstrate that can solve problems at the global level around evidence curation, patient identification. You mentioned the Takeda challenge that we solved. This is all around us being able to prove how our AI engine can help identify patients with rare diseases, but just as importantly, how can we build algorithms to identify them earlier so they can get diagnosed faster, because it could take up to a decade to diagnose patients with rare disease. And there's so few patients with rare diseases that you wouldn't think AI can help, but the way we built our technology demonstrates that it can identify patients with rare diseases. So how can we take that capability and really scale it globally and help these people?

Emmanuel Fombu:
And I think you mentioned something there that is very interesting about diagnosing patients with rare diseases earlier. That's been a big challenge even with rare diseases, even with chronic diseases, right? Things like heart failure, hypertension, ... Disease, ... Disease, the bunch of other diseases out there that we could talk about. But early diagnosis is very important for several reasons, as you know. One is, if you identify a patient early, they could get on therapy early and have better outcomes, right? Waiting for someone to get really, really sick before we diagnose them doesn't really help the outcomes that much, right? So early diagnosis is very critical and very, very important. With that being said, does this involve embedding a technology within EHR systems? And the second part of that question is do you work on Epic, Cerner? Is it like a particular EHR or are you EHR agnostic?

Aaron Liebtag:
So we're EHR agnostic, we're cloud agnostic, and we're format agnostic, meaning we can ingest data in almost any format, including PDFs, fire APIs, custom APIs, just giant data dumps, we can ingest data. Our engine can deploy on-prem, it can deploy on the cloud. We are one of the first commercial AI algorithms deployed in production at one of the largest pharmaceutical companies in the world, clinical trial workflow. So a lot of that design of our technology came out of, again, us embedding ourselves in a leading hospital for two years to really understand if we want to have impact, how can AI work within the realities of a healthcare system, within the realities of siloed data? So to that end, we've ingested and published on data from systems like Epic and Cerner. We've ingested and published on data that came from PDF diagnostic reports, physician correspondence, physician transcription, variable concepts, errors in data. We've been able to demonstrate our ability to ingest all of that information and be able to give high-quality curated data coming out the other side that can identify patients or is used downstream to accelerate AI and machine learning. Because, as you know, the more labeled and structured data that you have into patients, the more that empowers data scientists, both at hospitals within life science companies to build predictive models, to predict our various outcomes for various types of patients. And we're finding we're playing a big part of that workflow and that impact.

Emmanuel Fombu:
Which is quite impressive because if you look at like voice notes and you can look at PDFs and look at EHRs, so these are different kinds of data sets that you're talking about here, which opens up the entire globe, for example, right? Different parts of the world, in developing nations where you probably have data still on paper, right? In some parts, in the developed world, where that happens. Are you limited in your business to geographical location? For example, are you only working North America? Can you work in Europe? What is your geographical location? I'm not saying where you are now, where are your capabilities?

Aaron Liebtag:
To date, our applications have been in North America and in Europe. For the work we're ingesting in clinical trials, it's global. We're multilingual, so most of our use cases is in English. We've done work in French, but we have capabilities that can organize unstructured clinical text and find signal that scales across all languages where words can be tokenized. So there are certain symbolic languages that our technology is not geared towards. We don't do handwriting, but if any type of text that we're able to ingest it, we're able to organize it and add significant value to it, so that even if it's a person in the middle labeling data or identifying certain clinical concepts, we've proven to be able to accelerate that process with our engine and we've proven our ability to do that at a global scale in the applications and pilots that we've done.

Emmanuel Fombu:
That is fantastic. And as we wrap up this segment of the podcast, what I would like to know is, if someone is listening to our podcast right now and listen to all these great things you're working on, what is your immediate need right now, right? Are you looking for partnerships with remote pharmaceutical companies? Are you looking for more partnerships with providers, with payers? What is the next immediate step? What kind of people would you like to reach out to you?

Aaron Liebtag:
Yeah, so our immediate next step and we've started a process, is looking for partners who have the existing infrastructure, the existing sales, and marketing infrastructure and spend, who have existing customers who are also shared customers, whereby one and one doesn't equal three, but one on one may equal 100. Where we can become part of that workflow and really scale clinical discovery using our technology and using strategic partners, existing infrastructure. We're a small organization that's made up of engineers, scientists, clinicians, linguists, AI engineers. What we don't have is any sales or marketing, no business development infrastructure, and we're at the inflection point saying, do we raise capital to scale that on our own? Which is a strong option and we get those calls quite often, or because of why we started the company, which is to have impact and to have impact as quickly as possible. And the way we went about doing that is to build and validate a real technology that delivers real results, solving the problem. How do we roll into existing companies who may be two chairs short of the full dining room set? And together we can really change the trajectory of healthcare. We've got a number of business applications that speak to different targets, but the common denominator is about unlocking knowledge that's buried in huge amounts of clinical text, huge amounts of documentation that the current way that we get at that information today is we hire thousands of people to manually review information. It's not scalable, it's not reproducible, it's incredibly expensive, and it's holding back the promise of AI in healthcare. So we're looking for those partners where together we can accelerate the impact of AI in care, where we become that data curation engine.

Emmanuel Fombu:
Thanks a lot, Aaron. Dear listeners, if you listen to what Aaron just said and what the work he's doing at Pentavere, you believe that he's a great fit for Bite the Orange. And if you believe in the Bite the Orange movement, as we discussed, I believe that what Aaron is doing supports that philosophy, and the whole idea of changing the way we've done things in the past, which fits perfectly. Remember the orange? The reason we're called Bite the orange is because the skin of the orange is bitter, but the inside is sweet, right? But all the nutrients are on that skin, right? We believe in what Aaron is doing, please make sure you bite that orange and support Aaron on this. And Aaron, we expect you to bite the orange as well, and you know, and share your picture. Hashtag Bite the Orange and hashtag Pentavere to support the movement. And once again, thanks a lot, Aaron. We'll have the contact information for Aaron and Pentavere in the show notes when the podcast is live. So you could definitely reach out to Aaron at any given time and post questions for Aaron. Any support or any business opportunities you have, Aaron, you want to support the movement, please feel free to reach out. So once again, thanks a lot, Aaron, for joining us on this episode of Bite the Orange.

Aaron Liebtag:
Thank you so much, Manny. Thank you for the incredible work that you and your movement is doing.

Emmanuel Fombu:
Thank you. Together, we can make the future healthcare reality. Thanks, everyone, and see you in the next episode.

Emmanuel Fombu:
Thank you for listening to Bite the Orange. If you want to change healthcare with us, please contact us at info@emmanuelfombu.com, or you can visit us at EmmanuelFombu.com or BitetheOrange.com. If you like this episode and want more information about us, you can also visit us at EmmanuelFombu.com.

Sonix is the world’s most advanced automated transcription, translation, and subtitling platform. Fast, accurate, and affordable.

Automatically convert your mp3 files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.

Sonix has many features that you'd love including enterprise-grade admin tools, world-class support, advanced search, automated translation, and easily transcribe your Zoom meetings. Try Sonix for free today.

About aaron leiBTag:

Aaron has fostered successful partnerships with, and secured funding from, many of the largest global pharmaceutical companies and leading hospitals to develop DARWEN™ AI, which has achieved industry-leading accuracy published in prestigious journals, and presented at influential global conferences. Aaron has a proven leadership track record leading organizational transformation resulting in liquidity transactions for multiple private equity-backed consumer enterprises. Aaron is the former board director at the Museum of Contemporary Art Toronto Canada.

Things You’ll Learn:

  • The Darwen AI platform can also be used for clinical trial design, clinical trial operations, real-world evidence generation, as well as insight identification.

  • Darwen AI can identify patients with rare or chronic diseases and help them get the right treatments at the right time.

  • One of the criteria for the Galien Foundation’s Prix Galien award is publications and peer-reviewed journals.

  • Darwen AI is EHR, cloud agnostic, and format agnostic, which means it can ingest data in almost any way.

  • The more labeled and structured patient data is, the better for data scientists who are building models to predict various outcomes for various types of patients.

Resources: