May 31, 2023: Erik Nystrom, Principal, Enterprise Imaging at Pure Storage joins Bill for the news. How does cyber resiliency factor into the selection of an enterprise imaging platform? What are the challenges and considerations of using the cloud for enterprise imaging systems? How has the role of workflow evolved in the field of imaging, and what impact does it have on the adoption of enterprise imaging solutions? What are the implications of AI in transforming medical image analysis and interpretation, and how will it be accessed and integrated into enterprise imaging tools?
"The Future of Care Spaces" is an upcoming webinar that explores the latest healthcare technologies and solutions transforming care spaces in America. Hospitals, clinics, and at-home treatments are all affected by the advancements in healthcare technology, which can improve workflows, treatments, and patient outcomes. What are the latest remote monitoring tools, advanced telehealth solutions, and other innovative technologies transforming care spaces across America? Join us June 8th, 2023 1:00 PM ET and Register Here
Alex’s Lemonade Stand: Foundation for Childhood Cancer Donate
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 Health.
They have to work around expensive storage, right? So they built a solution where you move the data from an archived territory, short term storage tier, and now presented out. Now we're at a stage where that storage that is so fast, you no longer need this (Intro) 📍
Welcome to Newsday A this week Health Newsroom Show. My name is Bill Russell. I'm a former C I O for a 16 hospital system and creator of this week health, A set of channels dedicated to keeping health IT staff current and engaged. For five years we've been making podcasts that amplify great thinking to propel healthcare forward.
Special thanks to our Newsday show partners and we have a lot of 'em this year, which I am really excited about. Cedar Sinai Accelerator. Clear sense crowd strike. Digital scientists, optimum Healthcare it, pure Storage SureTest, TauSight, Lumion and VMware. We appreciate them investing in our mission to develop the next generation of health leaders.
Now onto the show.
(Main) 📍 all right. It's Newsday and today we are joined by Erik Nystrom, global Principal Enterprise Imaging for Pure Storage. And Eric, welcome to the show.
Thank you very much. Glad to be here.
Well, I'm looking forward to the conversation. We have been talking so much about generative ai.
I'm looking forward to. Maybe we will end up talking about it a little bit, but I'm like, you can't have a conversation anywhere without talking about generative ai. I went to a family weekend to celebrate my parents' 60th wedding anniversary, and like the nephews were cornering me, they wanted to talk about generative ai.
I'm like, you gotta be kidding. I can't go anywhere without talking about this these days. Hey are you experiencing the same thing? Well, it's
everywhere. And what's interesting about it is I'm eating, hearing conversations from my parents, right? They come up and they say, Hey, what's this AI thing like being integrated into medicine?
What do you think about this? Do you think it's a good idea?
And you told 'em, just go watch. Go watch Terminator. And then after they get done, they can come back and you can talk to 'em more about ai.
Yeah, well both my parents are retired physicians and it's quite interesting because for them to believe that you can publish a paper now essentially automated a medical paper or a scientific paper is quite fascinating.
It is. Yeah, we talked about notes. We've talked about a lot of things, but I'm violating my own principle here. I wanna talk to imaging with you for a little bit. We have about five or six different stories on imaging. We probably won't touch on too many of them, but I do wanna talk about the trends.
And so the first story is five major trends shaping the future. Of enterprise imaging. And if I look at this, I'm not gonna hit it too deep. I want to just pull the major things that they have in here and maybe have a discussion around it. So they have major trends, use of cloud AI workflow, cyber resiliency.
As a service procurement model. So this, and by the way, the next story we will pull has a couple of different things. So, I'm curious, of those five which one jumps out at you that you wanna start talking about First?
Let's do uh, cyber resiliency,
cyber resiliency for a hundred. Please. It talks about imaging data is required in almost every step of the patient care journey. Paired with the increases in sharing of images, both within and outside the physical boundaries of the hospital, the number and frequency of malicious attackers who seek.
To disrupt image access has also increased. Cyber resiliency is the capacity, and you know what cyber resiliency is, and it talks about that. I think it's interesting, Lehigh Valley Health Network was breached and they actually stole images and they were very sensitive images and they used that as part of the ransomware request.
Where does cyber resiliency fall in your thought process when you're looking at an enterprise imaging platform?
So with me being here at Pure, and definitely when it comes to cyber resiliency, it's to protect the data. So in case the data gets taken or taken hostage, we can restore that data and bring it back fast and bring the health system back up and running in any kind of cyber resiliency.
Plan, we plan to mitigate, right? There's detection and then there's mitigation. Detection is good. You're being attacked, you know what's happening. Mitigation, we can bring ourselves back from the incident without paying these attackers, so to speak. Right? Right. So with my job and what I look at doing is with Pure, we have a feature called Safe Mode, which allows you to essentially mitigate once your environment gets hit.
And so I think that it's a lot about mitigation, bringing it back and not allowing these actors, so to speak to take advantage of your data.
So resiliency is a design concept from the get-go now. I mean, it used to be that, that when a lot of these imaging solutions were selected long, long time ago.
We would select the imaging solution based on, workflow, usability, the radiologist, do they like it? Do they not like it? That kind of stuff. And then they'd look at it and say, oh yeah, and make sure it's secure. And now that's it. That's really becoming more of a bottom line architectural.
loud for a minute. So back in:One of the things we moved was our PACS system to the cloud. Now, when we moved our PACS system to the cloud, we decided to go with a essentially a hybrid environment. And so it was in a colo data center. It was cloud hosted, but we were doing our radiology imaging, our PAC system for 16 hospitals from a single data center.
to sign a BAA with us back in:They were looking at us like you're crazy. We're not gonna sign that kind of document. The second thing we were worried about was lock, lock in. That was the other thing. And then the other thing we were worried about was the the cost of storage in the cloud. It was kind of nascent at that time and we weren't sure.
What the pricing models were gonna be as we moved along. And so we decided to build a private cloud. That's what we felt like and we felt like if the public cloud ever emerged as an option, we could always just move there. What are you seeing, or what are you hearing with regard to the use of the cloud for enterprise imaging systems today?
So a lot of the ISVs today, they're pushing cloud solutions, right? And that's kind of the push that it's been. And we are going strong to go into the cloud. Most, I would say, pretty much every single I S V out there for packs, they have a solution that is cloud driven. However, one of the things I see that we don't discuss or talk about is kind of what you touched on right there.
What is the storage cost gonna be? After I've been in the cloud for a while, or you go to a PACS replacement, what is it gonna cost me when I pull out of the cloud? Because today, as we know it, if you move from PACS A to PACS B, there's a migration involved and there's a fee per exam that you're migrating.
There's no egress ingress fees because you run everything in your private environment or your private cloud. But once you go into that post cloud environment and you do go to APAC RFP or a v RFP for that matter, what does it cost to essentially move between these solutions running either in the same cloud or different clouds?
That is a concern to me that I often don't see as discussed, right? Because everyone is pushing, so, we're gonna reduce workforce because we're gonna go into the cloud. It also touches a little bit on the cybersecurity that you mentioned, right? So ISVs are saying, we're offering that by running this in our cloud, we handle.
The cybersecurity aspect of it. But there's an important question to ask there. If Azure or Amazon or GCP gets compromised, who is responsible for that? Would it be the cloud provider or would it be the I S V that guarantees the uptime? Yeah.
And yeah, that's where the BAAs come in and it's really interesting the language and how it gets stripped as you go in there.
It's interesting to me because back when we were designing our pack system, we had how many tiers of storage We had, I think three, at least three tiers of storage. You have your long-term storage, which costs you like pennies. I mean, it's like costs you nothing. And the reason you're moving images out there is potentially the patient died, in which case you're really, unless there's litigation, you're really not gonna need those images ever again.
And so you just move it into that sort of cold storage and it sits out there and it's, it's. Pennies on the dollar every year to just keep that image, which you have to, depending on your data retention policy and that kinda stuff. But then we had obviously the people who are coming in tomorrow and we were moving that into well we were using caching cuz we were centralized, we were using some caching things to move it down locally.
To the 16 hospitals so that when those patients came in the next day, the physicians could pull up the various images and whatnot. And that was expensive. That was fairly expensive storage, and that's what we were doing. And then we had the pennies on dollar. And I, so one of the I guess one of my questions is back then when we were looking at the cloud, It was less expensive than the storage we had at the time in our data center, but the long term cost of storage was higher.
Than the low cost storage we could do in our own data center. So the really fast disk and that kind of stuff, they could do it maybe a little less expensive, but the long term storage was a little bit more, I mean, has the cloud provider sort of addressed that and given us, multiple tiers of storage that, that make, that, that cost model really not an issue.
So I believe the cloud providers today, they have provided abilities to spin up different types of storage. You can have really, slow glacier storage. You can have really fast storage, you can have interior storage. But one of the things with PACS and radiology workflows that you need to remember is there are mechanisms within PACS to move that data.
Of course, right, right. But the storage does not know when X, Y, and Z is gonna have a car accident, or the storage does not know when there would be an incident. So therefore, we always say that you should always have the PACS managed where that data lives. If it lives in a slower tier, a mid-size tier, or a faster now, One of the questions that keeps on coming up to me is, Eric, why after so many years working in PACS did you go work for a storage company?
And one of the reasons for that is what I saw, and what happens is when you run your PACS on an all flash solution, it becomes extremely fast and essentially takes it to the moon. So it doesn't matter. When you come talk about money in class, right? When we have coming to a point right now where flash is at the same price as regular disk, now you no longer need to tear your data within a PACS.
You can provide the same speed for the data, regardless if it's on the long-term archive. Or in the short term storage. And this also allows you to run AI across the entire workload without compromising the radiology workflow overall. So, I mean, I've segued a little bit into AI here, but yes, a
AI and workflow.
And I'm trying to figure out which direction I want to go. Let's, workflow is so important in imaging. One of the things I'm seeing more and more. Is we used to have a pathology system and actually we didn't do digital pathology at the time, but we had a pathology system.
We had a radiology system, a cardiology system, and you have all the different systems. But I'm seeing more and more people talk about enterprise imaging. That seems to be, hey, let's cut down on the number of systems. A single system that handles imaging for all the ologies seems to be the direction.
Is that the direction that you're seeing as well?
Yeah, I mean, we're seeing, I mean, overall it used to be, like you said, it was PACS V D a, cardiology and a little bit of pathology. Now we speak to it all as enterprise imaging, right? Under enterprise imaging. The last add-on that we're seeing added is pathology, and actually a little bit of genomics also is coming slowly with some of the ISVs, but to that point, What we're seeing with pathology is really interesting because pathology is what radiology was.
When I started my career, we were looking at a decision path. What do we have to do with the images? We have to keep the images and the digital images? We're seeing the same thing today with pathology. The pathology labs, they still have to keep the digital slides and they're digitizing the slides for diagnostics next to the radiology images.
But then the institutions are saying, Okay, since there's no true direction yet on this, some institutions, they're scanning the slides, keeping them for X amount of time, and then actually deleting them, and then relying on the raw glass storage versus some other institutions that are just scanning everything in, holding onto both.
But I, I strongly believe that pathology will fall under. Umbrella of enterprise imaging, just like we're starting to see with a little bit of genomics on the corners starting to do,
yeah. No, yeah I'm hearing that. I'm seeing that as well.
Alex's lemonade Stand was started by my daughter Alex, in her front yard. It By the time she was four, she knew there was more that could be done. And she told us she was gonna have a lemonade stand and she wanted to give the money to her doctor so they could help kids like her.
It was cute. Right? She's gonna cure cancer with a lemonade stand like only a four year old would.
But from day one, it just exceeded anything we could have imagined because people responded so generously to her.
We are working to give back and are excited to partner with Alex's Lemonade stand this year. Having a child with cancer is one of the most painful and difficult situations a family can face at Alex's Lemonade Stand Foundation, they understand the personal side of the diagnosis, the resources needed, and the impact that funded research can have for better treatments and more cures.
You can get more information about them at alex's lemonade.org.
We are asking you to join us. You can hit our website. There's a banner at the top and it says, Alex's lemonade stand there. You can click on that. And give money directly to the lemonade stand itself
now, back to the show.
All right. We, I've avoided it as long as I can. Let's talk AI a little bit.
Let me give you some of this paragraph. So, AI is transforming the way medical images are analyzed and interpreted. AI algorithms can assist organizations with workflow improvements. AI tools can reduce. The workload of radiologists by automating repetitive tasks, which means they can place more focus on complex cases by automating repetitive tasks such as image tagging and analysis, AI can improve workflow efficiency.
It can also reduce subjectivity, thus improving the consistency of the diagnosis. This gets back to the conversation you're having with your parents. I mean, the AI is becoming an assistant to. The to the imaging professionals around there around the health system. And I guess the question becomes how do we access ai?
Will AI be built specifically for imaging? And will that AI be embedded in those tools that we're using for enterprise imaging? Or will there be more? General tools that are available that we need to tap into as well. I mean, what are your thoughts on that?
Well, I mean, it's really interesting when you think about it, right?
Because CAD for radiology has actually existed for quite some time, right? So when it originally existed, it would be a dedicated CAD server, and it would be run in parallel with your PACS You would send data over to it and then would process the data and send it back, and it would actually mark items of interest.
And then when the radiologist read, they would. Pay more, more extra attention to those areas now, over the last couple years, not only with ai, but also in conjunction with machine learning within radiology. And I think that it's will actually bring diagnosis to a second level, right? Because the amount of data that you can process with AI and go through and look at anomalies or regions of interest, that should be overread by a human.
Is immense. Right? And if we look at what I kind of talked about before, using storage that is capable of sustaining the performance that you need to run AI algorithms to gather this data across petabyte size archives. Right. This will overall, I think just like what we saw with Covid, right? Covid accelerated the remote reading, the remote access from physicians.
This will probably accelerate diagnosis overall and faster time to diagnosis. Yeah
it's gonna be interesting to see, what do you think the uptake's gonna be on this? I mean, how quickly, I mean, I realize AI's been around and in imaging is, it's been applied more than other areas within healthcare, but how quickly do you think we'll see the uptake of AI in imaging?
Just across the board that it just becomes standard practice to utilize ai.
I think a lot of. That ties into it, right, is how many algorithms actually get approved for insurance billing, right? So unfortunately we live in a society that is driven by that in a sense, right? So as more and more of these algorithms get approved by the F D A and they get approved for usage, for billing, in conjunction with that read for that exam, I think that we will definitely see a very strong uptick.
In the utilization of more AI within the radiology workflow. Now to the same point there might be tools out there that cannot be billed for like being used, but are sold and developed for the purpose of feeding up the workflow. And I think we'll see more and more of that. I remember when I went to sim about three or four years ago there was about 300 AI vendors there.
And now there is less, but there is more vendors that have more solid features and functionality. Right? More of them have integrated with ISVs to be able to launch within their hanging protocols. So I think we're getting more valuable, more refined, and more targeted tools.
Talk to me a little bit about storage.
I mean, since you live in the. Enterprise imaging space. You live in the storage space. I do remember when we saw the first all flash storage arrays designed for the data center, and we thought it was nuts, to be honest. At the time we were like, oh my gosh, this stuff's so expensive and and will we ever really need this kind of speed?
We need that kind of speed in a lot of areas
in healthcare now, don't we? Well, I mean, it's quite interesting. So the first time I heard about Pure Storage was when I was with my former employer and the customer actually come to me and he said, Hey, we're thinking about buying this an off flash solution.
We're gonna run our packs on it. And I said, are you joking? I said, you should buy something cheaper, right? Because you don't need all that performance. That's what we were thinking. Right. Right. And Then he said, well go take a look and, go see what it does for us. And I started getting really involved and I'm a generalist it from a background.
I wasn't super focused on storage, but I spoke to my customer and I started to see what this solution actually did for him. And what it did simplicity wise, and essentially it came down to never buy your same gigabyte again. Right? So if you're a pure customer, I gotta give you a small picture, but if you're a pure customer, right, you don't buy storage for that same exam ever again.
So as a health system, we talked about cloud on this call today, right? And to me it makes. Zero sense that you would put all your healthcare exams in a cloud and now you're gonna pay monthly for that exam until you're willing to push that delete button. And outta my 15, 20 years of doing this, I've had multiple customers ask me to write scripts to delete data out of their archive, but I've never had one willing to push that button to purch the data.
Right? So if they're not willing to do it on-prem, why are they willing to do it in the cloud? So for me to work for Pure, where we provide a solution, a storage solution, that not only shrinks in size over the years, but. We also make it so you as a healthcare institution, you don't ever have to pay for storing that exam ever again, right?
Yeah.
I it's it was an amazing transformation for us. I remember at the time cuz there was a bunch of applications that the bottleneck was the storage. Right, and we had really expensive high-end storage throwing it at it, doing all sorts of things to try to speed it up and whatnot.
And then these sand San arrays start popping up and we're like, wow. If it can do this. And then obviously, pure came out with essentially a cost effective and incredibly powerful solution. I mean, not to make, turn this into a. To an ad. But I mean, that was one of the transformation transformative moments in the data center was san storage arrays because it took those really intensive storage systems that were causing bottlenecks.
And it, it really eliminated a lot of those bottlenecks. Didn't actually it's whack-a-mole, right? So you eliminate that bottleneck and you realize you have another bottleneck over here in computer somewhere else, but,
right, right. It's interesting though, if you look at architectural design of pack software over the years, right?
They have to work around expensive storage, right? So they built a solution where you move the data from an archived territory, short term storage tier, and now presented out. Now we're at a stage where that storage that is so fast, you no longer need this. Two-tiered system. So a lot of ISVs out there.
Believe it or not, they've actually eliminated the two-tier environment. They have one flat file system and it's mostly running on smb. It's a little bit adoption out there for s3, but not much. But it gives the ability to give any image the same access time across from the health system regardless if you worked in, walked in today or four years, six years ago.
Right. I. And you're no longer limited by a technology.
So, Eric, I, we're almost at time here and I just wanted to give you the opportunity to talk about generative AI one more time, if you want to talk about it at all.
No, I think we're good. I mean, nailed a little bit sprinkled here and there throughout the
conversation I'll tell you what I am looking forward to.
What I'm looking forward to is, are you a Star Trek fan or have you watched Star Trek at all?
Sci-fi in general is usually a hit in my household, so it's a good thing. Yeah. So
Star Trek for years, I mean, back when it was, Nemoy and captain Kirk was Shatner, William Shatner back when it was them.
They would walk in and say, computer do this. And I'm like, that's it. That's where we want to get to when generative AI is literally pervasive. And I walk into my office and say gimme the the 10 latest news articles on digital pathology and it pops 'em up. Then that will, to me be the the moment at which I know that generative AI has sort of arrived.
I'm interacting with my computer with voice and it's responding like the assistant I want it to
be. Right, but the key point to that is that you need to have the underlying infrastructure and hardware that can support that workflow, right?
Yeah, absolutely. Absolutely. You keep coming back to that, like, like, and I've said on the show a couple times, architecture matters and people are like, what?
What do you mean by architecture matters? I'm like, are you sitting in a building right now? And they're like, yeah. I'm like, aren't you glad that there's like principles of design that there's things that the architect adhered to that the architect actually, understands load structure and all that other stuff, and you can sit in that building knowing.
It's not gonna collapse on you. I'm like, the same thing's true in software, hardware, data center, design, storage, all the, it, all, it all matters. Otherwise that's why we have breaches. It's why we have outages.
I'll leave with those. If the number one important piece when you take your PACS to the cloud is that today you're connected to a 10 gig switch and your data center back.
Is it still Is it still a 10 gig switch? I guess it is a 10
gig switch. I'm sorry, that is probably a hundred gig by an hour or a thousand gig. But the point is that essentially you're right. When you go from there and you connect to here, you set the expectations, right? Yep.
Absolutely. Eric, I wanna thank you for your time. Thanks for this discussion on on imaging and storage and architecture. Really appreciate it.
Anytime.
And that is the news. If I were a CIO today, I think what I would do is I'd have every team member listening to a show just like this one, and trying to have conversations with them after the show about what they've learned.
and what we can apply to our health system. If you wanna support this week Health, one of the ways you can do that is you can recommend our channels to a peer or to one of your staff members. We have two channels this week, health Newsroom, and this week Health Conference. You can check them out anywhere you listen to podcasts, which is a lot of places apple, Google, , overcast, Spotify, you name it, you could find it there. You could also find us on. And of course you could go to our website this week, health.com, and we want to thank our new state partners again, a lot of 'em, and we appreciate their participation in this show.
Cedar Sinai Accelerator Clear Sense CrowdStrike, digital Scientists, optimum Pure Storage. Sure. Test Tao, site Lumion and VMware who have 📍 invested in our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.