November 3, 2021: When you digitize pathology, it creates all sorts of advantages. Pathologists and laboratory staff are able to send results anywhere near or far for diagnosis, consultation, and second opinion. What are the drivers for Digital Pathology? What are the use cases? What is the future of AI in this space? How does interoperability, scalability, performance and security play into it? Dr. Justin Collier, Chief Healthcare Advisor for World Wide Technology and Michael Valante, CTO Digital Pathology for Dell Technologies join us today to discuss.
00:00:00 - Intro
00:02:00 - Digital pathology is an area of pathology that focuses on the data management elements of digital imaging. So the acquisition, the transmission, the display and the storage of the information that's generated from digitizing glass slides.
00:03:00 - We have a major shortage of pathologists
00:05:00 - How do we improve the workflow for pathologists and laboratory services?
This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.
Today on This Week in Health IT.
Digital pathology is being adopted worldwide and there are lots of cases of large regional networks now looking at how they can improve their productivity, enhance the delivery of their services by just being able to leverage digital data and communicate effectively digitally.
This is a solution showcase. My name is Bill Russell, former healthcare CIO for a 16 hospital system and creator of This Week in ???? Health IT. A channel dedicated to keeping health IT staff current and engaged.
If you want to be a part of our mission, you can become a show sponsor as well. The first step is to send an email to [email protected]
Just a quick note, before we get to our show, we launched a new podcast Today in Health IT. We look at one story every weekday morning and we break it down from a health IT perspective. You can subscribe wherever you listen to podcasts. Apple, Google, Spotify, Stitcher, Overcast. You name it, we're out there. You can also go to todayinhealthit.com. And now onto today's show.
Today, we are going to talk digital pathology. The case, the driver and the solution. We have Dr. Justin Collier, Chief Healthcare Advisor and Industry Practice Manager of Healthcare and Life Sciences for World Wide Technology. And we also have Michael Valante, the Global healthcare Business leader for Dell Technologies and CTO for Digital Pathology. Gentlemen, welcome to the show.
Thanks for having us. Thanks bill.
I'm looking forward to this conversation. I've joked a lot of times that this is the Education of Bill Russell. You guys are going to give me the deep dive into digital pathology, and we're just going to start from where we usually start the very beginning. I'd like for you guys to just give me a definition. What is digital pathology and how prevalent is it today in healthcare?
Yeah, sure. So digital pathology is an area of pathology that focuses on the data management elements of digital imaging. So the acquisition, the transmission, the display, the storage of the information that's generated from digitizing glass slides.
Yeah or you could just simplify it and say it's the modernization of pathology. In the same way that radiology moved from being physical films on a whiteboard to being digital films on
Sometimes referred to as whole slided imaging or even virtual microscopy.
So prior to that, we literally were looking at big rooms full of microscopes, glass slides, samples stored in cabinets. Like we used to store paper documents. Is that essentially what we were looking at?
Yeah, and it's, it's still, probably went that way throughout much of the world.
So I, I assume just like everything else, when you digitize this, it creates all sorts of opportunities. So give me an idea of what are some of the drivers that are leading us to digital pathology and and then we're going to look at some of the use cases that are emerging as a result of
I'll jump in first if that's alright Mike and give sort of my top three and then turn it over to you. So I, I think three main drivers that we see behind this, One is of course that we have a shortage of pathologists and a shortage of pathologists with specific specialization within that pathology duty.
So the ability to read studies in a virtual environment creates a lot of ability, both to scale as well as to access critical resources. So that certainly is a huge driver for it. The ability to read those studies from anywhere. The second one that I would point out is that it's impossible to have a backup copy of a glass slide.
You just can't do it. So having a digital image provides that capability of having a backup copy. If there's a natural disaster, if something happens and that slide gets broken if it's your own copy, then you're kind of out of luck. So having a digital backup is a huge advantage. And then the third one is that you can collaborate so much more effectively.
Very often there are second opinions when it comes to pathology studies or a patient will change institutions where they're receiving care. So the ability to more effectively send that information in those situations is fantastic. And of course, the research angle, being able to collaborate on research and do more with the images because of they're digital.
And so many technologies allow you to do more with digital images today, computer vision, things like that is that the sky's the limit.
Yeah. So I completely agree with Justin and might only add to that, that, in addition to the shortage of pathologists and the enhanced, the ability to collaborate.
You have an increasing workload. The advancements that are being made in comprehensive cancer care are really driving increased demand for laboratory services. So while you have a shortage of skilled resources and an increasing subspecialization, you have an increasing workload and demand on them.
And then there's the whole element of wanting to leverage AI tools for this. And you need digital data in which to do that. So all the things are kind of all these movements are pointing towards how do we improve the workflow for pathologists and for laboratory services and using digital pathology or digital tools certainly is there's a way to do that.
Just hearing you guys describe the drivers, it almost sets up the use cases pretty well, but I'd love to hear some of the use cases. Obviously we just came through a pandemic sharing those slides, sharing them digitally, I think would be an obvious use case but provide some of the additional use cases that we're seeing emerge.
Well, I think as Justin kind of just described them really well. I mean, there's certainly just the element of having a digital workflow, which certainly enables the ability to kind of eliminate the inherent geographic challenges, barriers if you will, that exist with traditional slide based pathology services. You have digital tools now you can be remote from the laboratory and sign out cases. You certainly have an ability with digital data now to mine that data and combine that data with other digital health data at which to gain newer insights. So there's an element of clinical decision support. There's an element of workflow productivity and of course the elements of research and analytic capabilities that one can take advantage of.
From a health IT standpoint, you're talking another imaging solution. How big are these images? I mean, if we were sorta to rank them between radiology cardiology, multi slice, that kind of stuff. I mean, how, how large can these images get?
Yeah, they are big, they're big. They range, it depends a little bit on how much tissue is on the slide and at what a magnification you're digitizing the data at of course. But typically we talk about average slides being, one to three gigabytes of data per slide.
Per slide. Give me an idea of 200 bed hospital. How many slides are they generating a day or a week?
So you figure an average case is probably depending on the the case mix, maybe six to 10 slides of data per case. And a hospital that size is probably doing 60 to a hundred thousand cases a year. The numbers roll up to how hundreds of terabytes to petabytes of data per year.
That's where I'm going as the CIO. I'm just thinking, but the reason you go down this path is the same reason we went down this path with a radiology, cardiology. I mean, you can get consults at a distance. You can do research. You can store it. You can you can create disaster recovery plans around it and those kinds of things. So there's, there's a lot of benefits, a lot of use cases that come out of it. And, so now that we've established the benefit and the use case around digital pathology, let's talk about building it out. And I always like to start with this question of who's the driver in the health system? What's the role that is driving this in the health system?
So ultimately I would say your ideal scenario would be to have a multidisciplinary team. You have to have the clinicians and of course the physicians that are actually going to be doing the pathology reading.
But you have to have the It support and the operational support as well. You truly need that multidisciplinary team just as we did when we digitized the electronic health record. Or digitized PAC systems or I guess, moved to PAC systems. But having those multidisciplinary teams is absolutely crucial.
And it will vary who is really leading the charge or helping a health system to move in that direction, being the champion for that, depending on the health system. Because there are so many different reasons to want to move in that direction. And some cases that might be a head of innovation, that's really pushing things forward.
In other cases, it may be the pathologist themselves wanting to move to that different workflow, that better workflow, and be able to access collaboration more effectively. In other cases that may be other folks on the team that are really sort of leading that push to move in this direction.
Yeah. And clearly this is a change in workflow as well. So, like most digital transformations it affects a variety of stakeholders. And certainly it's helpful to have all those stakeholder groups to include the technologists that are doing the work to slide prep work and that in the slide creation work on the lab managers.
So it really gets done with a group of people. I think as been stated, clearly these projects tend to start in the clinical team. They get envisioned as, a change in the way service clinical services are being delivered. But we do like to encourage that the IT group get involved early given, as we just mentioned, the size of the data is large. The retention requirements are large.
The size and the specs, I would assume that research drives us. So academic medical centers are probably leading the way in this. Is that pretty accurate?
That's what we're seeing. And it's certainly what we're seeing as they tend to lead in some other innovative areas. But there is a huge research driver. The need to collaborate across institutions more effectively than sending glass slides in a package, through UPS or FedEx, and hoping to god that they get somewhere intact.
The need to receive images from outside for patients that are referred for that tertiary or quaternary care. And of course, the ability to then further analyze with modern analytic techniques, things like computer vision, which is AI that analyzes images very similar to radiomics for radiology but being able to actually analyze the image in ways that the human eye can't actually even see.
To do research, to do discovery and compare that data to the genomic data, the other imaging data, the other information about the patient, bringing all of those types of digital currency or digital information about the patient to bear, to do accelerated discovery, to find new things to, to test things perhaps in a sandbox type environment and be able to become predictive.
Which patients are gonna respond to which treatments? What new treatments have we not tried for a particular disease that might be effective based on what analytics can be done. And now that you've got those images digitized, pretty fantastic stuff.
Yeah, for sure. And I think you raise a really interesting point in the sense that there is this crossover between, we've been talking a little bit about healthcare and hospital organizations and laboratory services, but the clear crossover to research and then ultimately pharma and drug discovery, digital pathology plays a role in that. It becomes somewhat of a gateway to precision medicine. If we could frame it that way.
So let's start with the end in mind. The pre-project elements that need to be thought out. You've mentioned advanced technologies, AI, machine learning, computer vision. You mentioned storage. You mentioned workflow, you mentioned a lot of things.
And I, I like to consider those things going into the project. What are some of the pre-project elements that need to be thought out? And what are some of those considerations?
So digital pathology adoption, I think you think of it as a digital transformation. And that's what it is. And as was mentioned earlier, whether it was EMR adoption or whether it was moving to radiology packs and the digitization of radiology workflows, all those same considerations are going to impact a digital pathology adoption.
So considering that really looking at the use cases. And what is it you're trying to achieve? Is this about creating environment for remote sign out of cases. Will that may construct a certain technology design. If this is about digital archiving of samples, in which to build a repository, to do analytics with, or develop AI algorithms, then that may have a different architecture that, that looks to be deployed.
I think what we're seeing, however, is that across all of these variety of use cases, clinical decision support, remote sign out, digital archives, IT is underpinning all of those. And has a strong need to be involved in designing to ensure that whatever the enterprise has guiding principles or standards that should be adhered to, or thought about in, in the construct, not only of the solution design, but the project deployment design.
Yeah, very much to Michael's point. You have to consider what else you're doing with data throughout the organization and how this fits into that governance, how this fits into that larger strategy and also consider what elements you may already have deployed in your technology estate that can continue to be leveraged as part of the design.
So how does this impact what you've already got as a baseline and how is this going to impact all of those other areas going forward? If that makes sense.
Yeah. How do these projects, where do they traditionally struggle? Is it is it in getting them off the ground? Is it in the development of those workflows? Is it in the communication? What areas does does a project like this struggle? And I'm pulling those really from my EHR experience and saying, yeah, we had to work workflows out across these departments. That was a significant lift. We had to work out the architecture.
That was a significant lift. We had security. That was a significant lift. So there's a lot of areas where this is a significant lift but this might be basic blocking and tackling today for health IT organization's. So I'm just wondering this specific type of project. where does it normally struggle?
So there are a couple of them that I will point out. And then, and then Michael, I'm sure you have many, many others. But two critical ones. One of course is failure to appropriately plan and test. That's a place where any technology solution will really struggle. The second one is whenever you've got a shadow IT situation. Where somebody's sort of going rogue and just doing this without consulting the rest of the organization.
So failing to think about all the things that we've already talked through. If that makes some sense. So if you have just a, a group that decided, Hey, we're just gonna do this. We're just gonna buy the digitizers and we'll let it figure out what they're going to do with the digitized images once they're produced. That can be a challenge.
I completely agree with everything the said, and I would only add to that, there's this other element of interoperability and integration with other systems. So you have a laboratory information system, which is the primary tool that pathologists use in their, they're signing out and they're diagnosing cases and in ordering recuts or new stains.
And so that integration between the image management system and the laboratory information system also has to be thought out. And because now we're dealing with communicating between multiple systems. There's a planning element of, it's not just, Hey, well, let's get a system in here. We'll get some storage, we'll get a scanner, we'll get a workstation and we can start being digital pathology because there's these relationships with other systems that are really important and have to be planned.
It's really easy to plan kind of phase one. And it's really somewhat straightforward to envision, phase N. It's how do you get from one to N? That requires a lot of planning. What is the trajectory and the velocity of adoption really look at and for me anyway, I tend to think of coming back to what are the use cases and what are the steps that need to be taken to support those use cases. And then the technology planning of course becomes part of that, of that design and that ultimately that plan.
Actually, as you were talking, it reminded me that anytime we're talking about data, it doesn't matter what the kind of data is. You've got to also consider cybersecurity as part of that plan. And it's got to be considered as a forethought, not an afterthought.
Well, thank you for saying that since it is national cybersecurity awareness month and as a CIO, it's amazing how many, it sorta made me cringe. When you said that there are places where somebody could bring in this solution and then notify IT later and say, ah, we just brought this thing online. Given the storage requirements, given the security, given the integration that's required, I would think this is definitely something you want to definitely have that multi-disciplinary team. I want to know who I'm going to partner with.
So you're talking to an IT person at this point, I listened to this and I go, this, this sounds good. I would like to you know introduce this concept or talk to me about the team that I'm going to go out and talk to. I'm going to start I assume with pathologists and have conversations there.
Where else? How am I building that team out? What does it look like for me to garner interest within my health system?
Yeah, so I think that's an interesting question and I think one that I think you can get to an interested team very quickly. We've talked about the different types of stakeholders that should be involved in the planning. I think there's some clear benefits to different stakeholders in there's some clear responsibilities for other stakeholders that have to be sure to be addressed because we're dealing of elements of workflow change.
We're dealing with elements of high-performance computing, enterprise storage. We've talked about interoperability, we've talked about scalability, performance, security. So from the IT perspective, I think it's pretty clear of the role and the guidance that they play in the planning of these solutions.
I think the pathologist piece we've talked about from their signing out cases and how they might be able to increase the ability to have clinical decision support tools with digital data. We've talked a little bit about the lab manager and maybe some of the technologists who are who have interest in what is the workflow change in the lab, now that you're introducing a step that I now have to digitize a slide. What does that change in that regard? And then there's the archiving perspective. If I'm still having to maintain a slide archive, does that change at all? When are you going introduce digital data? Or how do I make those processes work in parallel?
Because there's a bunch of data to be stored for a long period of time. So I don't know that there's necessarily one champion, although I think it's often helpful to identify who that champion is, who then can engage these variety of stakeholder. To bring that team together to surface the goals and objectives of what is to be done as well as surfacing, what are the kind of critical issues that need to be addressed or the challenges that need to be identified as part of the overall adoption.
What are the success stories? What do the success stories look like? I assume there's, there's a bunch of people that are out in front on this. Are they much more efficient? What are they able to do once they get this place?
So I think there are many, and certainly WWT has many, Dell has many working together we've, we've been able to join together, to see a variety of successes, whether those are as was mentioned, digitizing data for research purposes in an individual facility, or whether it's scales to providing better subspecialist load balancing across a regional network. Digital pathology is being adopted worldwide and there are lots of cases of large regional networks now looking at how they can improve their productivity, enhance the delivery of their services by just being able to leverage digital data and communicate effectively digitally.
Yeah. And I would say many of the success stories that are leading the charge here, certainly Europe is ahead of us in the US when it comes to the digitization of pathology, potentially no surprise there. Different regulations, different regulatory environment makes some things like this more permissive throughout the EU.
So there has been very successful for owners in that space. And of course, pharma is another place where this has been leading and where they've been seeing success.
What's the future here. What's the future of digital pathology?
Really it's the same future that we realized when we moved all of radiology off of physical film. It's that and more. So we've just gotten to the point where we've got the right technologies to be able to do these things well. And we've gotten the right technologies for the high-performance computing, for the analytics of imaging data on a pixel by pixel or boxville by block. So I would say the future is extremely bright and probably AI augmented.
I think that's right. There's a great future and it, it just expands if you look at it through the lens of an enterprise, right. Whether it's in the area of advancements in precision medicine or comprehensive cancer care or disease diagnostics, multidisciplinary teams we've talked about. These are all elements where digital pathology can have a positive impact in improving the way those services or those capabilities are undertaken and achieved.
Yeah. Michael, I want to talk a little bit about the architecture. And so as I think about this, we're going to be storing little probably on our network, on our physical network, we might be sharing that across a wide area network within the health system, but we're also going to be moving some of this information in and out. And I'm wondering what the typical architecture looks like. Are we utilizing cloud technologies? Are we utilizing mostly onsite on prem?
Is this a typical environment where we're pulling in those capabilities? The advanced capabilities that we're talking about in terms of AI and other tools, are we pulling those in from the cloud and just applying that on top of our internal architecture, what does it look like?
I think it somewhat mirrors what we've been seeing happening in the rest of medical imaging. Right. There's clearly a predominance of on-prem technology, asset storage, being utilized, given the nature of the size of this data. Given the opportunity to build enterprise repositories that maybe span multiple different types of image objects. And some of it really has to do with what is the rate of adoption. If there's kind of limited smaller use cases we're not doing everything. Maybe they're just going to focus on a particular tissue type. Those are relatively, smaller from a volume perspective, then sure let's manage these with existing infrastructure. Maybe scale it out. But the elements of the architecture I think are, are pretty consistent.
You need a highly scalable architecture, because your rate of adoption and your rate of growth is sometimes hard to anticipate. As I mentioned before, it's easy to kind of plan for phase one, but, and maybe envision phase N but the rate from one to N maybe unsure so an architecture that's easy to scale and how to scale was important.
Reliability is obviously key. Speed is always, always key. And given the world that we're living in the elements of security and being able to ensure that this data and the access to this data is protected is also key. So robust infrastructure is really important in, in driving this adoption digital cloud.
When it comes to cloud Bill, I would say that there are probably three real drivers that we see running up front with those that are leveraging cloud as part of this, and one is clearly collaboration just the ease at which you can collaborate either across entities or across across a system.
So that's a key one. Access to analytics resources is another key driver for the cloud piece of this architecture in some cases. Plenty of institutions that are leveraging all of the different flavors of public cloud for the analytics excellence that can be done there. And then the third is for archive. Just the ability for that cold storage archive. It reduces some of the costs involved in cost worries. When you think about egress fees of very large file sizes. So leveraging that more for cold storage, if you would.
The thing I like about this the most is you're talking about a set of skills, a set of technologies, and a set of practices that within it, at least we have, we've been doing this for years.
We've been doing it in the imaging space. And all we're going to be doing is changing the focus and really focusing in on pathology and standing up capabilities and technologies that we are already familiar with. We know how to store large images. We know how to archives large images.
But we just need from where I sit as a CIO, it's really the orchestration between the clinical side and the technology side, and really showcasing that to the to the board, making the case and those kinds of things. So I love the use case. How, we know there's a significant nursing shortage. How acute is the shortage of pathologists?umber of pathologists between:
So those that are perhaps an expert at reading since it's breast cancer awareness month reading breast cancer by slides that expertise may not exist in the same places in the same volume of experts brain tumors other types of tissue diagnoses. And then there's the infectious component. Certainly pathologists are involved plenty of the time when it's not a cancer diagnosis. It could be an infection. It could be another scenario where you need a different level perhaps of sub-specialist expertise. And those sub-specialists just aren't evenly distributed and available at any old hospital.
Will we see this as an opportunity for some health systems, some of the larger health systems to expand their reach and to partner with some of the smaller health systems
Oh, absolutely. And I think we'll probably also see the rise of the same types of maybe an overnight coverage service or those other things for times when your limited supply of pathologists may not be available. I think there will be quite a lot of both of those scenarios.
Yeah. And that'll be looked at on a global scale because there are some countries where the pathologists short. There are several in the Asia Pacific region where it's really significant. And just don't have the resources to provide the quality care that they want to provide. Which, laboratory services are involved in such a high percentage of care protocols today.
I love this solution. Is there anything else we haven't discussed that maybe I've overlooked in the questions that you think the community would benefit from knowing or understanding about this?
Being at WWT of course I would be remiss if I didn't point out our advanced technology center which is the massive innovation lab that we have built out strictly for the use of testing training and proofs of concept. Things of that nature for customers for our partners. It's about a billion dollars worth of investment that the company has made. None of which is used as a production environment for us to run on, that's a separate system. But it's strictly as that innovation environment where we bring all of our OEM partners, gear, and software and tools and techniques.
I have it connected to the cloud and it's actually four physical data centers plus an equinix co-location facility that has those high speed connections to the public cloud. So the ability to test out which things in your existing environment could be leveraged going forward. How do you maintain your existing environment or those existing investments?
How are they going to play with the new technologies? Which things need to be replaced? What are the right things combinations of new technologies from Dell and some of our mutual partners. Perhaps to bring together for a solution like this, that will work best for you. Which parts of the cloud do you want to leverage and how? So being able to test that very rapidly. Cutting that cycle from months to weeks or even days in some cases and it's supported with engineers. You don't have to worry about your engineers. You have a busy day jobs having to find extra time to be able to do the testing on prem. If you can even get the things onsite set up racked and stacked, et cetera, to be able to do the testing.
So knowing that you have virtual access to this environment from anywhere in the world and you can do that testing in partnership with our engineers. I think it's a, it's a pretty powerful tool to help accelerate moving in the right direction and prevent the shortfalls and pitfalls.
Yeah. And I would only add from Dell Technologies' perspective that we've worked to develop a rather robust ecosystem of partners in doing this that provide all the components slide scanning and application software and doing software to do this and working with really strong partners like WWT.
We really bring solutions that are based on none of the experience in digital pathology, but leveraging the experiences that we did in the transformations that we've done within radiology for, for many years and EMR modernization even more recently than that. But I was really encouraged Bill by your statement that which I think we agree with that we understand the technology. We understand what these transformations look like. So we do understand how they are to be done. And while the image sizes and the file sizes of the digital side, they sound large and they seem like it's a lot, but the work to be done is very achievable and it does leverage the same kind of principles and behaviors that we have a lot of experience with in other imaging digitizations. So` there's a lot of history in which to benefit from as we help pathology departments with their digital transformation.
I would just say yeah, of the hard things in front of our iIT organizations today this one is certainly one of the most readily achievable even though it is large in scope
Fantastic gentlemen, I want to thank you for sharing this solution with our community. I also want to thank you for dressing up so nicely. Some people are going to hear this on the podcast. They're not going to know how well dressed you are. I just want them to know that the lapel boot near the the pink share for breast cancer awareness month, but both of you wearing suits which I really appreciate elevating the level of professionalism on the show.
During COVID, everyone's working out of their homes, not many people have been putting on their jackets. So I appreciate you appreciate you doing that. And I appreciate you sharing the solution with us.
Good to have a reason to dress nicely.
There you go.
What a great discussion. If you know someone that might benefit from our channel, from these kinds of discussions, please forward them a note, perhaps your team, your staff. I know if I were a CIO today, I would have every one of my team members listening to this show. It's conference level value every week. They can subscribe on our website thisweekhealth.com or they can go wherever you listen to podcasts, Apple, Google, Overcast, which is what I use, Spotify, Stitcher. You name it. We're out there. They can find us. Go ahead. Subscribe today. Send a note to someone and have them subscribe as well. We want to thank our channel sponsors who are investing in our mission to develop the next generation of health IT leaders. Those are VMware, Hill-Rom, StarBridge Advisers, Aruba and McAfee. Thanks for listening. That's all for now.