Are you trying to identify the right tools, processes, and vision needed to fully adopt and scale digital pathology? Have you been looking for insight into how labs are already leveraging AI to drive quality and efficiency? Nathan Buchbinder, Chief Product Officer at Proscia, breaks it all down on the Digital Pathology Podcast.
Listen to the episode originally entitled “Proscia’s Tools & Vision for Modern Pathology with Nathan Buchbinder” that appeared on Digital Pathology Consulting, a resource created by Dr. Aleksandra Zuraw, DVM, Ph.D., Dipl. ACVP, to help pathologists design their image analysis-powered studies in the best way possible. Or, check out the transcription below.
Transcription
Intro: Learn about the newest digital pathology trends in science and industry. Meet the most interesting people in the niche and gain insights relevant to your own projects. Here is where pathology meets computer science. You are listening to the Digital Pathology Podcast with your host, Dr. Aleksandra Zuraw. (0:02 – 0:24)
Dr. Aleksandra Zuraw, Host: Hi everyone. This is Aleksandra Zuraw, and I want to welcome you to another episode of the Digital Pathology Podcast. I’m joined today by Nathan Buchbinder, Co-founder and Chief Product Officer of Proscia. Nathan started Proscia in 2014 alongside his fellow co-founders, David West and Coleman Stavish. In his role as Proscia’s Chief Product Officer, Nathan defines the company’s product strategy and roadmap and delivers market-ready technology to the growing digital and computational pathology market. Nathan obtained his Bachelor’s degree in Biomedical Engineering from Johns Hopkins University and subsequently received his Master’s in Biomedical Innovation Development from the Calder Department of Biomedical Engineering at the Georgia Institute of Technology. I am really excited to have him on the show today. (0:24 – 1:26)
Dr. Zuraw: Hi Nathan. It’s great to have you as my guest today. (1:27 – 1:31)
Nathan Buchbinder, Chief Product Officer, Proscia: Thank you for having me, Aleksandra. (1:32 – 1:33)
Dr. Zuraw: So, let’s start with telling our guests about yourself and your entrepreneurial journey. (1:34 – 1:39)
Nathan: Absolutely. My name is Nathan Buchbinder. I’m the Chief Product Officer and a Co-founder here at Proscia. Proscia is a digital pathology and computational pathology software company. My journey with Proscia began back in 2014, early 2015. I was doing research in a lab, as was my colleague David, who’s our CEO. And what David found and what I could attest to as well through my own experience, and I’m sure plenty of researchers in the pathology space could replicate, is that there’s really a big problem in the pathology lab with regards to technology adoption. There’s all this talk about big advances in computation and computational medicine, but it really hadn’t hit pathology. And that was where I think it was needed the most, where you saw big challenges with being consistent and efficient in the identification of various patterns and correlations to various outcomes and assessing tissue specimen. And so, you know, me, David Coleman saw this as an opportunity to really change the way that the world practices pathology. And so we took it from there. We built up the company. We’re now a 30-person company based here in Philadelphia. We have really grown it into a thriving business where we’re excited to continue to shape that vision. (1:40 – 3:12)
Dr. Zuraw: So that was, what, five years ago? (3:13 – 3:16)
Nathan: Yeah. A little less than five years ago. (3:17 – 3:19)
Dr. Zuraw: Exactly. And like you say, there is so much talk about going digital, advancing through technology. Pathology is not at the forefront of this trend. So it’s great that you’re trying to advance that. (3:20 – 3:37)
Nathan: Yeah, absolutely. And it’s interesting that it’s not at the forefront because for such a long period of time in the history of medicine — call it the late 1800’s through the mid-1900’s — pathology was the innovative space in medicine. It was the area of most advancement and discovery in terms of our understanding of disease. And for whatever reason that stopped, call it in the 1960’s or 70’s. Not that research stopped, but the position of pathology as being the key determiner of how we interpret diseases like cancer or other diseases as well. We see this new wave of computational pathology of innovation in AI, machine learning, cloud computing, and those other various technologies as another opportunity to take pathology. And really reinvigorate it. (3:38 – 4:33)
Dr. Zuraw: Okay. So tell me what is the mission of Proscia? (4:34 – 4:39)
Nathan: Yeah, so our mission is to change the way that the world practices pathology. We see this as a really transformative period in medicine as a whole. And so we’re on a mission to leverage computer intelligence to reshape how we’re thinking about diseases like cancer and leveraging that to ultimately drive clinical insights that impact the patient. (4:40 – 5:09)
Dr. Zuraw: And why is this important? Why is this important in the pathology space explicitly? (5:10 – 5:15)
Nathan: It’s critically important in the pathology space. You really don’t get diagnosed with cancer unless the pathologist says that you’re diagnosed with cancer. It’s fundamental to how we understand disease. And I feel that I’m maybe preaching to the choir a bit, but for anyone in the audience who doesn’t know, pathology really truly is that center point at which cancer is first diagnosed and at which it’s first understood. And where most clinical treatments and treatment plans get started. It’s where we understand prognosis and likelihood of survival, likelihood of metastasis. And so anything that we can do to tackle some of these incredibly impactful diseases that are increasingly prevalent in our society and around the world, anything we can do to tackle that early in the pathology lab really has the potential to dramatically impact millions of patients’ lives. (5:16 – 6:13)
Dr. Zuraw: Yes, I think in the era of artificial intelligence, we often kind of forget that it all starts with the pathologist and with the diagnosis, like you say. So having that in mind is very important, very important for the success of digital pathology companies, I think. (6:14 – 6:34)
Nathan: Yeah, absolutely. Cancer is, at its core, a disease of the tissue, and pathology at its core is a study of tissue. So anything we can do to tie those more closely together and drive additional insight really has the power to be very impactful. (6:35 – 6:51)
Dr. Zuraw: So tell me a little bit about your team. You mentioned yourself and David as the founders of the company. Who else is crucial to your team at the moment? (6:53 – 7:02)
Nathan: Yeah, absolutely. So there’s myself, David, and Coleman, who’s our Chief Technology Officer. We’re the three founders of Proscia. I say that knowing full well that our entire team is built of entrepreneurial minds and spirits. We really wouldn’t have been able to take an idea and execute on it and build on it and expand on it conceptually in terms of the product itself and the market that we’re serving without an incredible group of people around us. Right now, we’ve got products in the market. We’ve also got really innovative technology and a pretty solid group — a hardcore group of engineers, both on the platform infrastructure and development side as well as on the more scientific and artificial intelligence, deep learning, machine learning side of the house. And so our company would not be anywhere close to where it is today without the incredible people that powered the work that we do day-to-day. (7:03 – 8:00)
Dr. Zuraw: Mmhmm. I guess that’s a crucial component. So tell me about the products and services that you currently offer at Proscia. (8:01 – 8:10)
Nathan: Absolutely. So our vision has always been, like I said, to change the way that the world practices pathology to drive additional insight. And at its core is what we’ve recognized as the manifestation of that is computational pathology, is AI applications, is machine learning and deep learning and other powerful analytic tools that really shaped how we interpret tissue specimen. However, it’s nearly impossible from a practical standpoint to deliver those kinds of AI or analytic applications into the pathology lab without some kind of vehicle, without a way of having a conduit between the pathologists, the whole slide images, the data that you’re creating, the images that you’re capturing from tissue specimen, and the analytics that you’re actually looking to run. And so what we offer today are two broad product categories. The first is Concentriq®. Concentriq is our digital pathology platform. And it serves as an image and workflow management system for both academic and life science research organizations, as well as for clinical operational laboratories that are processing a patient specimen, to really deliver an experience that allows pathologists to say, “Hey, I can actually realistically see a world in which I might shift a big bulk of my work away from the microscope and onto the computer, onto the software platform for collaboration purposes, for analytic purposes, for just performance and operational improvements.” That’s one category of our products – Concentriq, our platform. On top of that and really leveraging Concentriq as a launchpad, we have a whole suite of AI modules, of disease-specific AI modules, the first of which we released in June of this year. It’s called DermAI™, and it leverages artificial intelligence in a workflow capacity to drive operational efficiency and performance improvements in a true pathology setting. And so we anticipate expanding both the platform and AI modules. But today we serve a wide range of users with a pretty robust and enterprise-oriented set of solutions. (8:11 – 10:45)
Dr. Zuraw: Okay. So Concentriq is for image and workflow management and then there are different modules that are integrated with Concentriq. And for now it’s the dermatopathology module, right? (10:46 – 10:59)
Nathan: Exactly. And with several soon to follow. (11:00 – 11:02)
Dr. Zuraw: Okay. So you already mentioned who your target audience is, but tell me a little bit more about your customers. (11:03 – 11:13)
Nathan: Yeah, absolutely. So we’ve been in the market with the product for a bit over three years now, probably closer to four years. And in that time we’ve brought on some incredible partners and customers. The Concentriq platform actually has a couple of different versions of it that gear towards one target segment or another. So on the academic and educational research side of the pathology ecosystem, Concentriq currently serves some of the biggest academic centers in the country and even in the world. So, Johns Hopkins is an example of an education and research customer of ours. So is UPenn, so is Thomas Jefferson University Hospital. So around the country and around the world, it’s serving some of these very large research organizations that are generating massive amounts of data. The product Concentriq also serves as a platform for life sciences organizations. So groups like CROs and pharma companies, big and small. An example of a pretty large CRO that we serve is a group that’s based out on the East Coast that uses our product to really manage their day to day work. This group, NSA Labs — Neuroscience Associates — is kind of at the cutting edge of technology and its application in the delivery of CRO services throughout the life sciences spaces, not only for their own internal work, but also as a way to deliver results more expeditiously to their own clients. And then on the true workflow side, so not focused on life sciences or on academic research or education, but on the true workflow and clinical enterprise side, we’ve got some large organizations that are using Concentriq, that are using not just Concentric, but even our AI product as well. (11:14 – 13:49)
Dr. Zuraw: And I forgot to ask you, what platform does or what hardware does your product Concentriq run on? (13:50 – 13:55)
Nathan: We work with nearly every hardware company, so we can take images in almost any image format, including the most common image formats out there. So whether that’s the Lecia SVS format or any format that’s similar to a TIF, we can work with the Ventana Roche scanners. We can work with the Hamamatsu scanners. We can work with the 3DHISTECH scanners, which are distributed by what was formerly Thermo Fisher in the U.S., so a wide range of brightfield images, a very comprehensive list of those image formats, as well as fluorescent images, as well as Z-stack images. And increasingly we’re seeing requests and we’re fulfilling those requests for other imaging modalities. That’s not something you expect from software that’s, from one lens, a replacement of the microscope. (13:57 – 14:59)
Dr. Zuraw: So basically you’re format- or a scanner-agnostic with Concentriq. That’s great. And regarding the personal computer, does it have to be a more powerful one? Do you need some special computer to use your platform, or the normal lab computers are enough? (15:00 – 15:21)
Nathan: Normal lab computers are more than sufficient. We have a big focus on cloud. We believe that cloud affords any laboratory the opportunity to build out a robust and infinitely scalable digital pathology deployment, which is important given just how large some of these images are and just how powerful and computationally expensive some of these analytics solutions are. However, even for an on-premises deployment, even for something that’s being deployed locally, which we do on occasion perform, our software is lightweight. It’s something that you can operate from not just any computer, but really any device as well. (15:22 – 16:08)
Dr. Zuraw: Okay. I guess that’s important, because digitalization and digital pathology already requires a lot of investment, so making this part of it light is definitely a big advantage. (16:09 – 16:25)
Nathan: Exactly. That’s one side of it. It’s a way of protecting the investment and reducing the amount that needs to be spent up front. The other side of it is even a little bit more practical than that, which is that digital pathology is a change. Digital pathology is not a matter of doing something that you’ve done before, but slightly different. It really is a radical departure from how you’ve operated in a laboratory for the past 30 or 40 years. And so when we look to deploy and when we’re looking to build out a solution for a laboratory, we want to do as much as we possibly can to have the pathologists, the histotechs, the lab managers, operations, the medical team feel as comfortable as possible on Concentriq and as they’re using our AI applications. And that means minimizing the amount of disruption that there is to their day-to-day work. That’s something that’s central to how we deliver our solutions. (16:26 – 17:32)
Dr. Zuraw: So what would you say is the most valuable contribution to the field of digital pathology that you have already delivered or maybe that you’ve planned delivering? (17:33 – 17:46)
Nathan: The thing that’s at the front of my mind is that we delivered in June 2019 the first commercially available, generally available deep learning product into the pathology space with truly clinical applications. That’s something that is remarkable. It’s something that if you had talked to most pathologists two years ago or talked to even many pathologists today who might not have heard of DermAI and ask them how long it’s going to take for AI to start reshaping how we’re thinking about pathology, many of them would say it’s three years, four years, five years away. Some would say it’s longer than that. And we are demonstrating now and what we’ve demonstrated through this release is that this isn’t something that’s an eventuality. This is something that’s here, that’s to a certain degree inevitable. But without the negative connotation of radical negative change, instead it’s something that’s an opportunity for these laboratories to take advantage of major revolutions in computation and in technology and in pathology. And it starts with an application like DermAI. So I’d say that’s probably the biggest contribution to date, although there are several others. I’m sure if I was looking to be a bit more expansive, we could find something else. (17:47 – 19:11)
Dr. Zuraw: I am so happy to say that the AI is actually creeping into the pathology labs. It’s going to be a process. Not everybody’s going to adopt it immediately, and pathologists have to get used to it. So I’m very happy that there is something that they can already start using. (19:12 – 19:32)
Nathan: Absolutely. You know, for any technology, take next-generation sequencing as an example, or even many genetic applications. For many of those technologies, it wasn’t that suddenly you flip the light switch and every single medical expert who was impacted by this technology was using it on a day-to-day basis. Something that took a little bit of time. But the introduction of the first applications in NGS really did serve as that point in time where if you were to put a pin on a timeline and ask “where did things change,” that was the moment that it changed. And we see ourselves in a very similar position. (19:33 – 20:14)
Dr. Zuraw: Definitely. Every technology requires some time to be adopted. I still remember when I didn’t want to buy a digital camera because I thought my analog one is so fantastic, and I’m going to lose something. And now I don’t own an analog anymore. (20:15 – 20:29)
Nathan: Exactly. And you would never think of going back. (20:30 – 20:33)
Dr. Zuraw: Exactly. (20:34)
Nathan: Yeah. And that transformation is the same thing that we expect to see in the pathology lab as well. It’s just a matter of getting everybody to the point where they take a look at their current analog cameras and say, “You know what? This really doesn’t work for us anymore.” Any resistance that we’ve had in the past just doesn’t bear any weight. Any of those old arguments just don’t make any sense, and we probably need to start switching over to digital. Whether it’s for cameras or for pathology. (20:35 – 21:04)
Dr. Zuraw: Pathology or any other technology. Right. So how do you work with pathologists for developing your products, your services, current and future ones? (21:05 – 21:15)
Nathan: We do have some pathologists on our team. However, when we’re looking to deliver products in the market, what we depend on, what we rely on is engagements with our future customers, with our partners and our prospects. And so, I think it would be fair to say that our customers, partners, and prospects are as much a part of the development process as the engineers who build the technology. In fact, we have members of our team who are dedicated specifically to going out to the market and understanding what problems have remained unaddressed, how best to solve them, how that manifests itself in terms of software. And we never release technology without having run it by many of our existing customers first. In our minds, the worst thing that could happen for digital pathology is the introduction of technology that doesn’t work or that works, but doesn’t solve a real problem that pathologists need to have addressed. And so we’re very cognizant of the role that the pathologist plays in shaping their own future and shaping what digital pathology looks like and in the years to come. (21:16 – 22:26)
Dr. Zuraw: I think that’s crucial, too. So you are an innovative company per definition already, but can you tell me what innovation means to you and how you innovate at Proscia? (22:27 – 22:41)
Nathan: Yeah, absolutely. I think there’s probably two ways of thinking about innovation. One way is thinking about innovation as taking something that exists already and doing some kind of manipulation to improve on it. The other way is to take a look at a problem that exists and develop a radically new way of solving it. At Proscia, I think we kind of blend the two, and I’ll explain a little bit more. So at our core, we truly are introducing something that has the potential to be disruptive. So it really is that second kind of innovation, the one that’s a radical change, a radical departure from anything that we’ve seen before. However, the laboratory space, and medicine as a whole, is relatively conservative for good reason. You know, we’re talking about patients’ lives. We’re talking about practices that have been developed and well studied over the course of the past several hundred years. We’re looking at a profession that’s foundationed, that’s principled on this core concept of providing consistency and providing certainty in areas that matter the most. And so when we look at innovation, what we’re trying to do is balance those two concepts, balance the need for massive change and the opportunity that it presents for labs with the concept that there are certain aspects of the medical profession that need to remain constant, that really drive home at the fact that there’s a patient at the end of the story who expects to have a person in front of them, that expects to have that person be their source of insight and knowledge. Innovation at Proscia is really founded on that concept of entrepreneurial-ism bearing the end user and the end recipient of this information in mind. (22:42 – 24:43)
Dr. Zuraw: So AI is of course a part of the innovation in this space and you already have an AI application. So tell me, what problem did using AI solve that you couldn’t solve before? (24:44 – 25:03)
Nathan: So until we had AI, what we were able to do is solve problems where we already kind of knew what the answer was. We could solve problems where we knew that there was a visual similarity between the histology on a tissue, on any given tissue, and the end outcome that we’re trying to correlate it to. If you know exactly what that pattern is or if you have an idea of what that pattern is, you can start to build your own non-artificial intelligence, non-deep learning application to address that problem. So, maybe IHC quantification is a good example of that where you know that what you’re looking to measure is the expression levels of various proteins or various biomarkers in a given specimen. And usually it’s a pretty straightforward measurement that you’re looking to make as well. It’s intensity, for example, but where AI comes into play, where traditional computational techniques fall short, is where the patterns are a little bit more complex or they’re a little bit harder to describe as a single variable or where you may not even know exactly what the pattern is in the first place. But you do know that there is some correlation between what you’re looking at under the microscope and what the end outcome is for the patient. Whether that’s survival or response to therapy or any other metric that you’re interested in. That’s where AI is not just a nice application, not just a way of doing things. It’s really the way of doing things. It’s the only way to solve those kinds of big problems. (25:04 – 26:39)
Dr. Zuraw: Mmhmm. So the patterns that we may not be aware of, but which are probably there. (26:40 – 26:46)
Nathan: Exactly. Or even for patterns that we’re aware of, but from which we have a difficult time finding a way of effectively communicating that on a consistent basis. The other area, by the way, where AI has the potential to really be the only innovative solution is for patterns that exist that the pathologist does not typically encounter on a day-to-day basis. What I mean, for example, is the pathologist in most contexts does not do a 10-year follow-up with every single one of their patients. However, there are almost certainly trends in 10-year follow-up data between the histology, the original specimen, and maybe the likelihood of recurrence of a tumor sample after 10 years or their survival after 10 years or occurrence of other disease after 10 years. So that’s something where AI is really the only approach to take for those kinds of robust and complicated forms of pattern recognition. Something where it really is going beyond the current practice of pathology and therefore requires solutions that can intelligently identify these patterns and translate it into something that the pathologist can use. (26:47 – 28:07)
Dr. Zuraw: Wow. Yes. I guess that’s going to be the future that is already the present with your application being introduced to the market. Is there anything I should have asked you but I didn’t so far? Anything else that you would like the listeners to know about? (28:08 – 28:29)
Nathan: The only thing that I’ll say, and I think this is maybe a bit of a reiteration of something that I said a few minutes ago is that it’s wrong to think about this wave of innovation as being optional. And it’s wrong to think about this wave of innovation as being 10 years away. I think we’re looking at a space that’s changing rapidly and for which there’s going to be solutions that really drive the future of how we look at specific specialties. And so what I’d say to anyone who’s listening who hasn’t already taken a serious look at digital pathology in any capacity is that now is the time to start acting before it turns from something that you have the time to consider to something that’s an urgent necessity. (28:30 – 29:16)
Dr. Zuraw: Okay. Thank you very much. Before we finish, just one more thing. Where can the listeners find you online to find more about your offer? (29:17 – 29:27)
Nathan: Absolutely. You can visit us at our website. It’s www.proscia.com. That’s P-R-O-S-C-I-A dot com. (29:28 – 29:37)
Dr. Zuraw: We’re gonna include the link in the podcast notes. (29:38 – 29:41)
Nathan: Thank you very much, Aleksandra. (29:42)
Dr. Zuraw: Thank you so much for being my guest and making us aware of what you’re doing. Good luck with everything, and I am keeping my fingers crossed for more AI applications and for your success. (29:43 – 29:57)
Nathan: It was a real pleasure. Thank you so much. (29:58 – 29:59)
Dr. Zuraw: Thank you. Bye (30:00 – 30:01)
Outro: Thank you for listening. For more great digital pathology resources, visit the Digital Pathology Consulting website and subscribe to our newsletter on digitalpathologyconsulting.com. After subscribing, you will get access to the annotation guidelines, which will help you annotate slides consistently in all your digital pathology projects. Talk to you in the next episode. (30:02 – 30:21)
