Welcome to the “Zoomed In” podcast, a series on embracing the shift to AI-enabled digital pathology.
In Part 1 of our inaugural episode, our CEO David West explores the pathology status quo – and why it must be broken. David explores the current state of pathology and why the transition to digital is inevitable before breaking down the data opportunity that’s emerging. He then shares his take on the status quo and why some organizations have been hesitant to break it.
Listen or read the full transcript below.
Transcript
Bruce Hall, Host (00:10): Welcome to Proscia’s “Zoomed In” podcast series, where you hear the latest trends in AI-enabled digital pathology.
BH (00:23): Welcome to the first episode of “Zoomed In,” a podcast series from Proscia on embracing the shift to AI-enabled digital pathology. I’m Bruce Hall, your host. Today, we’re going to be talking about breaking the pathology status quo. I’m joined by David West, the CEO of Proscia. How are you doing today, David?
David West, CEO, Proscia (00:40): I’m great. Excited to be here.
BH (:043): Thanks so much for being on the podcast. David, we’re going to talk about breaking the status quo, something we often hear about when it comes to innovation and transformation. Why is this so relevant when it comes to pathology?
DW: (00:54): When you look at pathology, it’s field of medicine that hasn’t really changed in the past 150 years. You still have a physical glass slide, a physical microscope, and pathologists that are inspecting that glass side, using the microscope. In shifting to digital, just like you’ve seen in virtually every other vertical, including within medicine, is a huge change. It’s necessary and probably inevitable for many reasons. When you look at the current practice of pathology, it’s very manual and subjective, and there are not enough pathologists to keep up with demand. We’re seeing a declining workforce. The physical glass slides have their own limitations. They’re very hard to share and move from point A to point B if you want to talk about subspecialty expertise or getting second opinions. Or in biopharma research in particular, a highly collaborative environment, perhaps most interestingly, there’s this data opportunity that we’re really not touching on. A lot of life sciences organizations have already started to realize this – to realize how much power is in the data – and have gone digital to a certain extent, to better leverage their pathology data. And increasingly we’re seeing developments around AI on the clinical and diagnostic side as well.
BH (02:07): I want to expand on your point about the data opportunity. Can you tell us more about the benefits available from digital pathology data?
DW (02:15): When you scan a glass slide into what we would call in this industry a whole slide image, or take a physical piece of tissue and turn it into pixels – over a billion pixels per image, a gigabyte per image. So there’s a lot of data. The data contained in these images is representing patterns, which you see in tissue, but with new technologies like deep learning and artificial intelligence more broadly, computers can read and analyze these pixels in ways that aren’t necessarily readily apparent to a human reviewer Very simply put, digital pathology is making it possible to unlock data and insights that are otherwise hidden in these images.
BH (02:56): Have any use cases resulting in specific applications emerged?
DW (03:00): We can group them in terms of a couple of different types of applications. You might see workflow applications or predictive applications or prognostic application. So in workflow, you’re going to see applications that help automate routine pattern recognition tasks in the pathology workflow, helping drive pathology productivity gains and efficiency gains. You’ll see predictive applications that might help predict for responses to therapies and any long-term correlations to future health events. And you’ll see prognostic applications that might indicate patient outcomes for a given disease.
BH (03:31): So David, of workflow, predictive, and prognostic applications, which are the most abundant?
DW (03:37): A lot of what we see today in pathology is workflow related. This is the low hanging fruit. You can think about the image analysis applications that are already available in market, driving efficiency. They might be automating routine tasks, ensuring that better quality and decision making. Or you can look at a solution like what Proscia offers in dermatopathology, DermAI, which initially can help triage and sorts images of skin biopsies to improve productivity and increase confidence. So it’s very easy to imagine how this type of solution down the road could support diagnostics and lead to improved accuracy and quality. And then once we get into predictive and prognostic applications, we’re talking about benefits that include improved patient outcomes and increasingly personalized medicine. We’re really just scratching the surface here with respect to what’s possible. And I don’t even think we can really imagine what applications we might see in the future.
BH (04:27): Those sound like some great benefits, David, and it will be very exciting to see how all this evolves. What other benefits can organizations expect to see, especially in the short term, by adopting a digital pathology approach?
DW (04:38): There’s a great white paper that was published recently from the Dark Intelligence Group, which as folks might know, is one of the leading analyst firms for diagnostic laboratories. The report featured some digital pathology pioneers. At a high level, labs that are implementing digital pathology are seeing somewhere between 13 to 21% productivity and efficiency gains. For labs that work on very narrow margins, this is material. Now, it doesn’t happen overnight, but with a little bit of investment, we do see these gains. So when you look at an institution like UCSF, one of our partners who has gone almost a hundred percent digital and fairly quickly, it’s been able to significantly reduce the time it takes to diagnose and to share frozen sections. Pathologists at Granada University Hospitals have been able to process 28% more units since going digital.
BH (05:27): Beyond the clinical setting, where else are we seeing benefits?
DW (05:30): Digital pathology is also benefiting life sciences as well. These organizations deal with massive amounts of data, which often needs to be shared across multiple sites, different CROs, partners, and stakeholders. As an example, we work with one leading pharma company that’s been able to organize this data and make it accessible across the entire organization as a result of going digital. And of course, with everything going on with the COVID pandemic right now, I’d be remiss not to highlight that pathology enables remote operations more broadly. And this has been very critical in enabling the community to continue working despite the need for social distancing. Labs can leverage digital pathology to read and share cases remotely. Researchers can collaborate and access the data they need. Even medical schools can use digital solutions to continue to educate their students. So, with all this said, going digital is really just the first step in realizing the power of computational pathology or AI. Organizations that don’t go digital now not only miss out on these near-term benefits, but they also fall behind from those that are able to leverage AI as those applications increasingly become available.
BH (6:32): So David, with all of the short-term benefits, long-term benefits, and new data opportunities, why have organizations been hesitant to make the full transition to digital pathology?
DW (6:41): To be clear, a lot of organizations have already started their digital journeys to some extent. For some labs or research organizations, it’s more about realizing the full potential of digitization. If we take a step back, the technology to enable all this really hasn’t existed until the last five years or so. Cloud storage wasn’t at a place where it could support the massive file sizes of whole side images. Compute power wasn’t strong enough to process them, especially when we think about what we want to do with very computationally intensive algorithms and deep learning. And we didn’t have affordable, high-throughput, clinical-grade scanners. But now we have companies like AWS that are delivering storage, or NVIDIA that are lowering the costs and improving the speed of GPUs, and scanner players like Leica or 3DHITECH or Hamamatsu that are building some very high-quality and high-throughput scanners at the right price point. Of course, organizations looking to go digital have to acquire this infrastructure. It’s an upfront investment. And there’ve been regulatory hurdles in the past that have been overcome in recent years. And the reality is that it’s a big transition. And once again, it’s a necessary, inevitable transition that pays off dividends, but it’s a transformation. And we see this in many industries. To break away from the status quo requires time, effort, and resources and the right mindset.
BH (8:00): So now we get to the topic of our podcast, which is breaking the status quo. But before we get into exactly what kind of changes are required, tell us exactly where you think the status quo is today.
DW (08:10): I think the status quo is probably the most important aspect when we think about adopting digital pathology, but it’s not something that you can touch or feel always. When I think about the status quo, it’s about the current state of things. It’s almost what no one really notices in their day-to-day because it’s been that way for long enough. Often, no one questions it because it seems good enough, but often it really isn’t good enough. This is what we’re seeing when it comes to digital pathology. This is what I was talking about earlier in terms of the inevitable shift to digital.
BH (8:40): What is it going to take to break the status quo?
DW (8:44): So breaking the status quo has to do with bringing on disruptive change. Think about big disruptors like Uber, Airbnb, or Netflix. They totally transformed their industries. And so someone might even resent you for breaking the status quo. There’s a cultural element. After all, you’re questioning and reinventing what feels comfortable. Doing this as the only way to drive meaningful progress forward.
BH (09:06): Makes perfect sense. I think we can all intuitively recognize that breaking the status quo is hard. There can be throes of change that go along with it. But in the field of digital pathology, what specifically do you think is holding us back, whether it’s getting started or moving along the journey to realize its full potential.
DW (09:23): That’s a good question. I think it’s really important to understand the different factors that come into play here. There’s several that we often see in our conversations with labs and research organizations. And it’s only after you understand them that you can work to address them. One of the most important aspects here that is often overlooked is a lack of dedicated resources. And we see some of the most successful organizations have a dedicated project manager or a team that’s helping to push digital adoption forward. Maybe it’s a focus group. It’s a process. It takes work. Otherwise, a focus on maintaining day-to-day operations often takes over. This can be a full-time job. So, if you’re an organization that’s serious about digital pathology, you need to dedicate the resources to accomplishing that. Otherwise you’re almost certainly will fail.
BH (10:08): So, a lack of dedicated resources certainly makes sense. Any other challenges?
DW (10:13): We often hear another challenge. Pathologist themselves, in their day-to-day operations, they love their microscopes. Sometimes they can even work faster on their microscopes. I totally get that. A lot of pathologists have been using the microscopes for 20 or 30 years, and there’s a deep ergonomic connection to the instrument. I think the reality is that the digital pathology is about everything you get beyond just image viewing – from seamless sharing to collaboration, and the potential of AI and these computational pathology applications. And a lot of these ergonomic challenges can be overcome with even just a few weeks of a little bit of practice.
BH (10:48): This concludes part one of our podcast. Check back for part two, where David continues discussing obstacles to breaking the status quo, explains downsides to not breaking the status quo, offers advice for moving forward with digital pathology and what to look for in a partner, shares where to find more info, and offers a glimpse into what’s ahead for the “Zoomed In” podcast series.
