Artificial intelligence is the #1 go-to tech trend word in the health sector. AI, ChatGPT, and GenAI are the hottest technologies transforming the entire health ecosystem – from drug development to patient diagnosis to determining who is at risk for illness to sorting through virtual reams of real-world data waiting to be mined and applied. The conversation around the technology’s implications and use cases only accelerates as big gun speakers step onto mainstage platforms.
Last year, Medika Life published a piece titled The Top 20 AI Voices to Watch, and most of the thought leaders, influencers, and theorists featured continue to demonstrate why they lead the AI conversation. Among those key voices – Tom Lawry and John Nosta – have focused on why conversation must center around “not will be” rather than what is! What is happening with this technology right now to improve the system and human health? Nosta points to how AI is an extension of the health professionals’ cognitive abilities, and Lawry reinforces that health systems that adapt and implement remain ahead of the care curve.
Too much of the side conversation distracts from the reality that AI isn’t a future shock; it’s happening – it’s been happening – now and for decades and is impacting health professionals’ abilities to rise to higher levels of performance and contribution within their communities.
We had a chance to sit down with one of those market leaders whose ideas and organization are demonstrating the efficient value of AI to make medical diagnoses more timely, efficient, and effective. Joseph Mossel, Co-founder and CEO of Ibex Medical Analytics is an engineer, entrepreneur, and supply chain workflow expert and has been partnering with pathologists, health systems, and pharmaceutical companies to add speed and reduce the human stressors around accurate diagnosis.
Medika Life sat down with Mossel to explore how he and his team at Ibex bring real-time practical value to the pathologist’s workflow and world and, in doing so, bring an authoritative voice and application to their output.
Conversation with Joseph Mossel – Ibex Medical Analytics
Gil Bashe, Editor-in-Chief, Media Life: People often discuss augmented intelligence – AI – as “futuristic” or theoretical technology. Then, there is abundant conversation about AI’s pros and cons. Can it replace skilled medical professionals? But, as you and Ibex AI demonstrate, it accelerates critical analysis – sustains human life, and enhances research decisions. Your company’s customers embrace this “thinking” technology to improve workflow and reduce throughput stress. Perhaps most important is its ability to amplify critical thinking skills to provide deeper understanding.
People want to understand how leaders take this technology, manage vast amounts of information, and operationalize it. You turn a wealth of information into practical business and clinical solutions that sustain and save lives. Do you agree with that basic premise?
Joseph Mossel, CEO, Ibex: Yes, I do. When we looked at this field, we saw a huge opportunity to do something that had not been done before by applying AI to pathology. While it’s been done successfully in radiology, digitization in pathology is far more recent, presents unique challenges, and offers significant opportunities for the research and clinical care communities. For us, it was crucial to feel that we were doing something meaningful and helping improve people’s health globally.
What’s interesting about Ibex is that our founders come from a computer science background. We set out to harness the technology within a clinical environment. But we were true to the goal of helping patients and physicians. This singular focus guided us throughout our journey to do something meaningful and, as you note, practical.
Bashe: Pathology, traditionally involved in examining wet slides under a microscope, now embraces digitization. You were applying computer science and thinking through an idea that would transform an insight-driven medical profession that navigates within a larger physician community.
When you and your co-founder began this journey, did you study the entire process, from sample collection to the pathology lab’s inner workings to how patient-centered professionals use these data? What was your approach to understanding and improving this sector?
AI Isn’t Only About Data – It’s Also Work Flow Change
Mossel: More than that. We come from a background that requires understanding how to build products that improve throughput and output. It’s in our DNA. This isn’t just about analyzing data; it’s about having products running in the market, accomplishing an engineering feat and a successful user experience. A lot of sector knowledge is core to the company’s DNA.
As we looked at pathology, three questions drove what we were doing:
- The pathology lab resembles a sophisticated factory with advanced machinery and highly trained technicians. We wondered how to improve this (life-sustaining) factory—even considering how the experts’ experience improved.
- How do we capture mistakes early to reduce professional anxiety?
- How do we ensure the right cases get the right priority? We explored the steps, bottlenecks, challenges, and the IT status and infrastructure of the pathology lab.
The next set of questions is clinical. We needed to understand what happened after this sophisticated “factory” had completed its initial work. How is the information used in patient care or research? How can it be prepared and offered so that it continues to advance within the decision-making process?
The slides are already there, and a pathologist is looking through, historically, the microscope, now on a computer screen, and conducting the analysis. The key is understanding the clinical questions pathologists must answer and having confidence in their observation beyond a black-and-white diagnosis. That is much more complex than just saying whether cancer cells are present or not. It adds value to this professional’s presence in the care or research process.
Our goal is to identify tasks where AI excels, like counting cells, and tasks where human expertise shines, such as spotting very small, suspected regions. AI can ensure that everything is scanned and nothing is missed, highlighting interesting areas and ensuring the pathologist can confidently examine. Augmented Intelligence, in this case, is not replacing the pathologist; it’s helping them do their job – the function that makes them essential in clinical care – even better.
If the Benefits Are Clear Change Is Possible
Bashe: No matter how valuable the idea or technology is, sometimes the culture of a community can create roadblocks to change. As an engineer with computer science and product development expertise, you understand the pathologist’s role as an ally to allied health professionals and as your customer. You have the keen ability as an engineer who understands computer science and building products, which means you understand customer expectations.
There is often a feeling that physicians feel threatened by AI; what was your experience sitting down with your advisors? What comments or feedback were you getting from those people?
Mossel: My view might be biased as I primarily meet with people already inclined toward using AI. The main reason I hear from early adopters is that pathologists are struggling to cope with their field’s increasing complexity, and they are genuinely open to help.
One of the things that we hear from many of our users is that it reduces stress levels. It also just helps them get through the cases faster. There’s less buildup of what they physically have on their desk – slides waiting to be made, for example – which adds stress.
But there’s an even deeper element: they don’t need to go home with this feeling at the end of the day. ‘What did I miss today? Was there something I didn’t see?”
So there’s always these remaining questions, and I think we’re helping them sleep better at night. In general, we make the experience of being a pathologist meaningful – even crucial to the next part of the patient’s journey. We’re not at all considering replacing pathologists. We’re focused on making their experience and clinical voices more authoritative.
Bashe: Technology augments the skills of health professionals; it doesn’t replace them. The human deploys or calls upon the technology to provide a greater value to the system, the patient, and the referring physician, and it becomes a much more authoritative voice because they’re championing it. They’re channeling their wisdom through technology.
Ibex chose to focus specifically on cancer diagnosis. Was that a personal, strategic, or market-driven choice? Because of the ability of the technology? Can we say, ‘Wow, it’s not that they have this type of cancer, from the molecular structure in the staging, it’s this type of cancer, and therefore, we’re best to go with this type of clinical approach’?
Paving the Pathway for Precision Medicine
Mossel: Yes, we are paving the way for precision medicine. I think that’s the third driving question for what we’re doing. One revolves around efficiency and the other concerns clinical quality. We primarily do that with our partnerships; we partner with pharmaceutical companies to use this information.
Data coming out from pathology drives new insights that pathologists can’t do alone. We are finding these important clinical questions that guide treatment. The key here is getting your hands on the correct data sets that allow you to build what we like to call AI markers, like biomarkers, by analyzing these slides quantitatively, which is different from how a human pathologist does it, and putting this into a machine learning framework, generating novel predictions, creating novel tests.
Why cancer? First of all, you’re right. It doesn’t have to be about the technology. Pathology is not just about cancer; technology lends itself to anything. I think that the answer to your question is all of the above; some of it has to do with firsthand experiences with family members who had cancer. Cancer isn’t everything pathologists do, but it’s undoubtedly the most critical, meaningful, and at the heart of the profession.
If you tie it back to therapeutics, determining the right therapy is where the most significant need and opportunity resides. I see us expanding to other disease areas. Clinical diseases also vary. The diagnosis is a determinant of pathology, and we can certainly go into those domains. We are not limited just to cancer, but still, we think of ourselves as a cancer diagnostic.
Bashe: Cancer is the third major cause of death in the world. 72% of disease deaths relate to what they call noncommunicable diseases. Cancer has a substantial economic impact. If we could get to a correct diagnosis earlier, we could deploy treatments earlier; the fastest or quickest institutions to deploy technologies that make their people smarter, faster, more effective, and more efficient are often major academic medical centers.
Does this type of technology make academic medical centers more important to the biopharmaceutical industry because the quality of the data the department shares is more specific? Will it make the data more valuable to pharmaceutical companies when submitting new medications?
Mossel: The way to think of it is that AI is, in a way, creating a new modality of data: the ability to extract quantitative objective features from vast information. When you work with human pathologists in a clinical trial, they’re much more limited in the amount and objectivity of the data they can get on the site. So yes, you create this new, much richer, more objective information set. You can do more sophisticated machine learning AI on this data and introduce new insights.
Interestingly, the most significant successes are from academic centers. We work with many, though much of our work is with community commercial pathology labs. These are the backbone centers that oversee high volumes. People in the community commercial pathology labs feel the most pain and stress. Most of our success is in this domain.
AI Can Advance Health Equity
One of the biggest motivators for our actions is how we help promote health equity. If you’re a pathologist at one of these academic centers, you are well-trained with particularly good colleagues surrounding you. You have a fellow doing all your cases in advance of your searches. There’s a strong support network around you. If you’re a pathologist in a commercial lab, you work alone and don’t have the same level of support. We bring this lab our technology, which was trained by some of the world’s best pathologists and help drive up the quality. Most pathology cases go through these labs, and we are helping them—most people who don’t get their pathology sent out to Brigham and Women’s here in Boston.
Bashe: This is a two-pronged health equity front. Ibex enables the community-based or private lab pathologist to share equality with the state-of-the-art academic medical center’s pathology department. So, David equals Goliath when it comes to clinical output or throughput.
There is also the potential that people who often receive sub-optimal care, the Black, Indigenous Americans diagnostic, can get the same quality oversight as someone who has the private resources to go to Mass General or New England Deaconess or Dana Farber Cancer Center; they’re going to get the same quality review of their pathology data. This type of technology, deployed widely, can be a great equalizer in terms of both the ability of the pathologist to perform and feel comfortable about that, but also the ability of people who may have significant health risks to think that their data had the same oversight as someone from a different zip code. Is that accurate?
Mossel: Yes. This works within a specific country where you have a huge demand. This is not theoretical; we have actual numbers backing it up. But you see it within a specific country, the difference between the top academic centers and the periphery. It’s a dichotomy that exists both on a domestic and international level. If you go to the developing world, there is a massive shortage of pathologists, and there is opportunity there. That’s another part of our vision.
Our mission is to ensure that every patient gets the correct diagnosis. Everyone takes it as a given that cancer diagnosis is often wrong, and we do the best we can do. But we are now in a position where technology allows us to go beyond that, overcome these barriers, and make diagnostic errors exceedingly rare. But it’s not just a technological question, right? It’s a question for the healthcare system. For the technology companies, the providers, the payers, and the regulators – everyone needs to get together and solve this problem.
Bashe: You are a CEO, a founder. You’re an operational leader. You’re listening to the customer’s needs; you’re listening to the response. You talked about how you and your colleagues have created products before, and you understand that it has to be packaged, deployed, and used to assess market success. Could you share two or three measures of success from where you sit? What do you think the success measure should be?
Accuracy Decreases Workplace Stress
Mossel: We have one straightforward measure of success for Ibex that can be extended for the entire industry: the number of biopsies that go through our system. We have hard data showing that Ibex knows what we are doing and has proven that the more cases that go through our system, the fewer the diagnostic errors and the greater the efficiency for pathologists. And for us as a company, that’s what we rally around: driving up the usage. Everything will cascade from both the clinical and economic impacts and the success of Ibex as a company.
Bashe: Joseph, thank you for your insights – practical guidance on how AI accelerates decisions that save lives. You and your colleagues are drilling down into an area that is one of the most expensive areas of health: managing people with cancer. It’s one of the most terrifying diseases, where people want to know that they have an authoritative diagnosis and where pathologists and oncologists don’t want to screw it up. You are enabling people to be much more confident that the information they’re sharing with their patients and their clinical decisions is on point, not because of their own life experience or clinical experience. But the data is there to back them up. It doesn’t get much more practical than that.
Futurists have the advantage of rarely being wrong. Long after they’ve shared their predictions, the world moves along. People forget, and the edgy on-stage comment is forgotten. Pathologists can never be wrong. When they are, it’s considered an error – sometimes with dire consequences. The possibility of combining knowledge – trained and honed – with the aggregated intelligence of pattern recognition offers pathologists and their health professional community an added advantage. Ibex AI demonstrates not what is possible – what is happening now.
The concept of the future is always inviting and exciting. The difference between invention and innovation is scaled application. AI, ChatGPT and GenAI applied play to the strength of the curious and bold. Using these technologies can democratize information for humanity’s benefit.
