In this episode, join me as I speak with Dr. Greg Rose, a retired radiologist who played a key role in the digital transformation of radiology. His journey offers valuable insights and lessons for the digital pathology community.
Key Points Discussed:
Greg's insights into the digital transformation of radiology provide a valuable perspective for pathologists looking to embrace digital tools and techniques. His experience highlights the importance of managing change, leveraging AI, and improving diagnostic workflows.
THIS EPISODE'S RESOURCES
đź”— Dr. Greg Rose's website
Become a Digital Pathology Trailblazer get the "Digital Pathology 101" FREE E-book and join us!
In this episode, join me as I speak with Dr. Greg Rose, a retired radiologist who played a key role in the digital transformation of radiology. His journey offers valuable insights and lessons for the digital pathology community.
Key Points Discussed:
Greg's insights into the digital transformation of radiology provide a valuable perspective for pathologists looking to embrace digital tools and techniques. His experience highlights the importance of managing change, leveraging AI, and improving diagnostic workflows.
THIS EPISODE'S RESOURCES
đź”— Dr. Greg Rose's website
Become a Digital Pathology Trailblazer get the "Digital Pathology 101" FREE E-book and join us!
Greg: [00:00:00] One of the issues about going digital is it's a major change in how you do your medical services versus on site versus remotely. But once we started our teleradiology company, we had to be digital. So we had to make things digital that weren't digital, like plain film, and figure out resolutions and data and all of that.
But it was an exciting and difficult time with political and financial and technical hurdles. That we went through. I believe that the hardest part of this was not the technology. The hardest part of this was issues related to personalities, because some people respond to change very differently than others.
Some people like that there's something new every day. And there are others who want everything to be the same.
Aleks: Welcome my digital pathology trailblazers. In the digital pathology world, we always say that we are behind radiology, like maybe 25 [00:01;00] years. We should look what radiology has done and learn the lessons and not reinvent the wheel, but I don't see us doing this, or at least I don't know what radiology actually went through, how it was, how the process looked like.
So today I decided to invite a person who has been there all along.
Through this digitization journey on the radiology side, welcome Dr. Greg Rose to the show. How are you today?
Greg: I'm great. Good morning. And thank you for having me.
Aleks: I am so curious about this episode because like I said in the intro, we always say, Oh, we're like behind radiology.
There is so much we can learn. How much is there? We can learn. How much do we need to reinvent? But before we dive into it, let's start with you. So tell us about your background and what you're doing now. Just tell the listeners, the digital pathology trailblazers, who is Dr. Greg Rose?
Greg: Thank you very much, and thank you for having me.
I'm a retired [00:02:00] radiologist, and I'd like to say that I hope that any assumption I make about our similarities in practice, is accurate and, and helpful and forgive me if I, presume too much.
Aleks: Don't worry, I'm going to ask clarifying questions.
Greg: I wanted to say I hope that anything I say today, I hope is of some help to you.
I did a PhD in physics, in the 80s and then spent a little time, doing some research for the army and then saw all the interesting things happening in the 80s in radiology. Like ultrasound and CAD scans and MRI. So I went to med school at University of Texas, did residency at Baylor in Dallas, and then did a fellowship in MRI at the Mayo Clinic, then came down and practiced in a small and then large practice in Texas.
And I started a teleradiology company. In 2005, and we did that for about 11 years. And then after [00:03:00] that, I did a couple of other things. I was the CMO for strategic radiology and then after that, I did some consulting work with unified radiology, which brought together many of the groups. And then since then I've been doing kind of this or that, a little bit of consulting with the American College of Radiology and some other things.
Aleks: So you've definitely seen the digital transformation in radiology firsthand, both on the practice side and like your own company side.
How did it happen, the digital transformation? How was it to go from the analog to digital and how easy or difficult it was for you?
Greg: Well, just a little bit of background is that we did x rays since 1895 until about 1940.
That's all we had for radiology. And then we started doing nuclear medicine, but still everything was analog. And then we started getting into PET scans into the 60s. But then the [00:04:00] 70s and 80s, we started doing radiology that was inherently digital. Ultrasound MR and CT. So that helped us as we transitioned over.
And then into the 90s, we were still printing film. We were not doing things by, teleradiology until the late 90s. And then there was a transition and we all went through this and we started to do things more and more digitally. And we're going to talk about this a little bit later. But one of the issues about going digital is it's a major change in how you do your medical services versus on site versus remotely, and we'll get into some of those details, but once we started our tele radiology company, we had to be digital, so we had to make things digital that weren't digital, like playing film, and figure out resolutions and data and all of that, and we're going to talk about that in just a minute, but it was an exciting and difficult time with political and financial and technical [00:05:00] hurdles that we went through, and whatever I could do to share some of those, wins that we went through to get that to work might make things easier for pathology.
Aleks: Definitely. I want to, I want to hear about all the hurdles and all the hoops that you had to jump through. So question, you said that you had to print film, print film, Does it mean it was already captured digitally and you had to print it or was it still captured in an analog way on the film like the photography film?
Greg: Right, so we could clarify that. A plain film, which is just x rays, those were still being captured analog and so those films had to be digitized. We put them through a digitizer. Which looked at each pixel and created a digital image. And, but then still into the late nineties, we were printing those films directly without digitizing.
So we had a huge film library. These patients. With films as much as seven years [00:06:00] old. And then with kids, it was up to 21 that you would keep their stuff up. But the CTMR and Sona, were actually captured as digital, but we printed them on film and then still. Read them off of printed film until we went to the transition of digital.
Aleks: So these CAT scans, what time frame was that? So what I'm seeing, you had both like the fully analog workflow that is very similar to what we have in pathology right now. We have the analog glass and then we're digitizing the glass and then we're archiving the glass and the digital depends. Sometimes you have to archive it for longer or for less long, but the glass has to stay there, according to the, regulations, but you guys had, which I kind of see, we have a little bit of a parallel, but the, the parallel is still on the cutting edge, side on the pathology. So you guys had the analog. Then you had the digital, the CT scans that actually could be like fully digital, but you had to re analog them for, by printing.
How, how did that feel? And let me just tell you, like, what kind of analogy do I have? A semi analogy. There are, imaging modalities in pathology that are, that are direct to digital, but they have not been adopted yet by the mainstream pathology community, I would say. These are more like research, applications, but yeah, tell me how it is to have something analog that you need to digitize and then you have something digitalized, digital, that you have to bring back to analog and working with those two modalities at the same time.
Greg: Well, I think you've outlined it pretty well. We know even then, when we're looking at these cases, that there must be a better way. In fact, with our studies, when we're doing, for instance, a CAT [00:08:00] scan or an MRI or an ultrasound, it's looking at cross sections. through the body. And to be able to look, we had to look at an image of that and then a slightly lower image.
The next one, it was clear to us that it would make more sense to stack these images and be able to scroll through them to stay in the same place. That's a little bit different, I think, in pathology, where the images are not stacked like that. It's an individual slice, and you're not looking at a set of images that are stacked through the same parts in general.
So what we discovered was some people were kind of sneaking over to the CAT scan area where the images were obtained and would read them off of there and then sneak back to where the images were displayed on film. And so there was a little bit of a funny transition during that. But then it turned out that with all of these, these images.
It was taking longer and longer to do the studies because we'd have to place [00:09:00] them all up. We'd have to print them. And then the prior study we need to hang all that prior study of films and look back and forth. And it was a good way to make a mistake because you weren't looking at the same images at the same time.
So it was, we were pressed to do this. And that's what started a lot of the, of the interest in trying to get this to where it was all viewed on monitors and digitized. The whole idea of storage was still just film room, but we just kept filling up film rooms. But our CT studies, a whole CC study, is about 50 megabytes, whereas one of your slides is something like 2 to 3 gigabytes of information to fully do and we'll get into that a little bit later too.
Aleks: So, okay. So we have a couple of similarities, a couple of differences. Definitely the, I was laughing when you were saying that people would sneak in into the CT room where the images were obtained. And then I see [00:10:00] somebody looking at like a bunch of films next to each other and then like how easy it is to confuse them the levels.
I mean, I know they're labeled but it's so easy to just misplace the, the film that the like sheet of film or however that was. So any other things that you think are similar or different. Where do we, where do we converge and where do we diverge other than that? Anything else?
Greg: Well, one of the places where I think pathology will be migrating to, is to understand what will be our resolution.
Does it really need to be on a nanometer scanning scale? Because the smaller you do it, of course, the bigger the data set, the harder it is to move it, the more expensive it is to to retain it. But, there has to be agreement about this because it needs to be that you don't condense it so much that somebody would say, well, you condensed it beyond being able to [00:11:00] make a diagnosis.
So you guys are in a little bit of a hard situation. And, but we've gone to where we play film at first, they said, we just can never digitize it. That was the theories. It just always has to be analog. And then we finally got over that to say that we can do it with five megapixel monitors. And then we finally got it down to three megapixel monitors.
Cause they're a lot less expensive, smaller storage, except manual has remained at five megapixel. But with you guys, you've got a slide that has all that information on it. And it at some point, somebody maybe legally will want to go back and look at that slide again and say, well, maybe you missed this or that.
So it has to be able to be findable if they see it. And it's blurry because it's low resolution. It could be argued. Well, you've made this so that we can't make a diagnosis in retrospect. And that's something legally, that you guys would need to go through to figure out what is going to be our acceptable way of doing things. And [00:12:00] I would imagine the, I imagine there's an American College of Pathology, which is your government body and they probably by working with them, they figure out what this is. so that's part of what you guys need to go through. We had to go through that in radiology.
Aleks: I think it's a unique struggle for you because we kind of have, and I don't know if we can ever go like lower to save any type of data or resolution than the microscope. We basically have the microscope and the objectives on the microscope that I, that are used for the, let's say anatomic pathology, that will be the 40X.
So, currently, for primary diagnosis, when I work on digital slides, we agreed in our institution that it's going to be 40X. Before it was 20X before because the digital zoom was good enough. This was good enough resolution. I don't know, but that's a good question I don't know if it's like across the board 40X [00:13:00] and I know that for cytopathology It's gonna be higher because they have higher magnification objectives.
They will work with 100x oil to see all the cellular details, so that's a, that's a good point. So You guys, how did you get to the agreement that, okay, we don't have to do five megapixels. We can do three now. How much data did you have to show? And like, what was the proof that you have to show everybody to say it's okay to go lower?
Greg: There were a number of white papers that came out that looked at various things, that were critical findings that tended to be associated finding at a high resolution that if you didn't have a high enough high enough resolution, you might miss them. So they looked at things like subtle rib fractures, early interstitial disease and pneumothorax.
And so that's what they looked at on the plain films. And once they did that, they realized we really can, we don't miss [00:14:00] diagnoses by dropping down to three megapixel and we can live. and the data storage for that was, as I mentioned, usually the JPEG lossless they would use and then they got into JPEG 2000.
So then we, we agreed on it by doing white papers on the most challenging pieces that might be, critical diagnoses. So the easy things don't matter. Question then would be, it's a great question. I think you're, you're trying to get me to tumble to the proper formulation of this, of this response. But I, I think that what you guys would do is look at what are our most subtle differential things? What is, what is a handful of these? And then at what resolutions do we find that we're still able to distinguish these and then you write a paper to show that at 40X and perhaps half a micron per millimeter per pixel might be a resolution that we can live with and not be missing diagnoses.
So it's a result of, [00:15:00] White Papers. Yes, so actually, for veterinary medicine in TOXPATH, we did a validation study, and like you guys did, we had a list of critical features, part of those critical features was, suggested by the European Society of Veterinary toxicologic pathology. I think it was toxicologic pathology.
So they had a list where people were, concerned that, oh, if it's not good, if the resolution is not good enough, we're not going to be able to see things. And this was like, cilia and the ciliated epithelium maybe inclusion bodies, where you have a viral disease, things that are really where you need to play with the micrometer screw to see them.
And there was a list. Then there was a survey, We surveyed all the pathologists working in my organization. This is Charles River Laboratories. Actually, I'm supposed to work on paper on that. So thanks for [00:16:00] this advice. And so we added these other lesions and we had a list of critical features and we still went for 40x.
I don't know if we're ready to embrace that, hey, actually we could see it at the lower resolution. That's going to be our next step. Once we actually go to digital fully and try to figure out, okay, can we save the space? Which brings me back to this question about data. So, do you have any advice of how pathologists could manage the digital data avalanche heading our way?
So for the resolution currently for the 40x, the most common one seem and, and agreed on one seems to be 0.25, microns per pixel. That's at 40x resolution. So at 20, sorry, at 40x magnification. 20x is going to be [00.17.00] 0.5. yeah, what, what's your advice? What did you do? How did you do it? Is there any ways we can, navigate this data storage dilemma that we have?
Greg: Well, part of this, so there's a number of pieces to this. one of the pieces is with the storage part. Another part of this is the transfer from wherever the images are to you, and then there's the, there's the uploading of your report, which is small. It's kilobytes. That's nothing. What you need is download.
So one of the things that's helpful is the download speeds of the various internet providers has been increasing substantially. which in some ways might match some of these increased datas that you guys had back when we were doing it with DSL And ISTN. We were very data conscious about compression because it took so long to get the images now that we're mostly on much [00:18:00] faster routes, including things like fiber.
There's a good chance that you guys could be getting these much faster. One of the things about your data, our data, when it came in, we saw the whole image. It was a 50 megabyte study. What would happen is when we're looking at the study, when there's a particular patient, we would click on the patient and it would begin downloading it.
We also had, when we have a work list, I'm sure you guys have work lists, is that once something it needs to be done, it could begin downloading it while you're already looking at another study. So we had that going on also, but on the live process of having the images coming down to us, they would start to load and we could begin looking at images while the rest of the images were loading on our system.
With you guys, it what it might be is because a patient might have some number of slides that you need to look at, you would probably start looking at one. Hopefully you order them in a way that you could {00:19:00] begin looking at. one of the studies that can load faster. And then as you're looking at that, the others go, there's with, with the slide, we wouldn't with radiology, we look at the whole image on our screen, we can always zoom in and look at other areas, but for the most part, we just look at it as it is. With you guys you scan the slide and you're looking at different areas. I have to wonder whether some of the technology that you're going to begin to incorporate would model something like is done with Google Earth, where you have the whole Earth in your image, and then as you zoom in, it only goes and tries to capture the data, which is within that area.
And as you keep zooming in, it keeps only looking at the data, which is part of the zoom. I would imagine some of the pathology technology is going to be incorporating that.
Aleks: And I can tell you that I have heard of, this being done like that, especially because there are, kind of two ways [00:20:00] of doing this.
You can download all those images locally, or you can view them online. In the cloud, depending, obviously, the cloud viewing, you are limited by the speed of your internet. The downloading, you're kind of limited by the time it takes to download them. We do have an option to, have them download, One by one and also, we do have an option to have them to prioritize the download like, And, and depends a little bit on are you in a diagnostic workflow or are you in the research workflow?
I'm in a research workflow for drug development support So my study is going to be all the animals that I have in this particular study, meaning it's going to be let's say 80 animals and each animal has 35 tissues, so I will be looking, in the study at, I don't know, 1, 000 to 2, 000 of slides, but I do [00:21:00] have a specific order.
I want to look at my controls, animals first, and then I want to look at the high dose. So I can totally, tell my system, and I work with a system that downloads the slides, this CyFix Patholytics, and I can prioritize what to download. So I can basically download work for the day. And then overnight it's going to download work for my next day.
So that we do. And I know in the cloud based systems, they, when you zoom in, it loads only the tiles around. This is similar to Google Earth. I don't know how widespread that is. I only heard that from one guest from a pathology watch and they serve dermatologists with their dermatopathology services.
So I'm definitely linking to this episode here. And maybe people are going to comment that, how come you don't know we all do it? And I hope they do, but, maybe not. So that's a good piece [00:22:00] of information.
So switching gears a little bit, the switching to digital, it's not only the technical stuff, but it's also change management.
So, I bet it was a bit of a hurdle race, like, and you already mentioned technical, political, financial, this all encompasses change management. How did you deal with these things and what were those? hurdles that you had to overcome?
Greg: That is a great question because I believe that the hardest part of this was not the technology.
The hardest part of this was, issues related to, personalities, and certainly, and you said the word change a couple of times to emphasize it because some people respond to change very differently than others, if you hadn't noticed, and some people embrace, you [00:23:00] know, it's. There's some people like that.
There's something new every day, and there are others who want everything to be the same every day. And what happens is that there's also an issue with those who have been around for a long time. They've got, they've become very set in their ways, and I don't mean that in a bad way. I mean that become very familiar and good at how things work.
And then this thing comes along. And one of the crucial things to know about this and this coming from an old guy like me, I think I'm in a position to talk about my brethren and what can happen is that the, the young whippersnapper. In the practice brings in this very new idea that they need to embrace, and it could be viewed as this person suddenly acquiring more control and power because of this knowledge and knowledge is power and where this person wants to go with this.
Sometimes they feel that, well, let's get going with this. Let's get this all [00:24:00] rolling up. We're all going to do this. This is going to be great. And what can be not seeing is the concern of an older practice member. Who may feel in some ways marginalized from the talk, he's no longer as knowledgeable and power and powerful and familiar with what's going on.
This young guy is having talks with the CEO and these other people in the technology group. And, and there's this older guy who is kind of cut out of this. So I begin by mentioning that can be present. So that to at least know, be aware that that might be there. And then we can talk about how to deal with that.
But it's, it's very astute of you to bring that up because that's very important is that everyone feels involved. and I find that when we started going through this, instead of proclaiming things, the, my just as an advice a little bit on this, instead of coming forward and saying, this is [00:25:00] what we're going to do across the street where our competitors are, they're doing it, and this is what we're going to do, I think should begin with an interrogative of sitting down with the practice and say, we're aware of some of these changes.
What do you guys think of this? And then they're part of the talk, they're part of where we could be and what we need to retain. And that's a good thing, because the practice people who have been around for a long time are the ones who commonly are best, most well known by the remainder of the staff of the leadership, and those leadership may commonly be looking to the older, more seasoned senior staff to say, do you guys buy off on this?
Will this work? Is this going to be a good transition? Yeah. So allow them to be in that part in, in, in that, in that role, because then they can actually facilitate getting these things done. So I thought I'd just start off with that.
Aleks: Definitely. I think also, regardless of the specialty, that [00:26:00] medical community is still very hierarchical.
It's, it's like that because it just takes a long time to gain the expertise to be, seen as an expert and to actually be an expert. So, there is kind of a, inherent reason for that. But now, like you say, it's not, no longer only the expertise, the like, a domain specific expertise, how well can you diagnose things, but also the technology component.
And like you say, you cannot accelerate the, diagnostic expertise gaining, but you can totally be on the cutting edge as a, an entry level practitioner, when you're interested in those technologies, so that kind of like, you can jump over the, the classical medical hierarchy, and that, [00:27:00] that causes a stir, and sometimes can even cause rebellion, and we don't want that, we want to, collaborate and basically have everybody participate in this change that I think it's inevitable, and brings a lot of good, but having, people who've done that for a longer time express their concerns is only going to be beneficial.
But they're only going to do that, if they're included in the discussion. So, yes. What about the other changes? Like, what, what, was there any specific technical, political, financial, anything else that was specific to the time when, when radiology was going digital?
Greg: Well, during that time, and that's critical, is how does it migrate?
How does it change? One of the things that we, we knew ahead of time was to take it in [00:28:00] steps, trying to put down dates. We're going to do this by this date. And everybody is aware of what these changes are and to take it in different pieces. And what's most critical when you go to do this, I highly recommend that you have a failed back is that when something starts to go one way, You have the ability to fail back.
You probably won't have to fail back very often, but there's, there's something about in the psyche about the knowledge that if this isn't going to work, at least we can fail back and then they're not distracted with what could go wrong. They know that we could always fail back. So I highly recommend that to any group that decides to do this.
The other piece is that there will be comments that will come up about. The fact that if you go digital like this, this means it can be done remotely and when things go remotely, then the politics change significantly because one of the concerns that will be, [00:29:00] stated mostly by your senior staff, but your mid and younger staff will note this.
Also, once it can be done anywhere. It means it could be done by anyone. And then wait a minute, what happened to our contract? Because we have a contract here and now we've got some group out of some other state that says, well, we could do it for 10 percent less. Now that it's just out there. So with these discussions, as you begin to make it so that anyone could do it, it's critical, especially your senior staff.
Established relationships with the hospital to show that it's more than just a dictation of a slide. It has to be more than just a simple transactional state that I give you some coin and you give me some piece of vegetable. There has to be something beyond that, which has to do with the custodianship of the practice of working with with the tanks being available by phone to discuss things.
Having a technology [00:30:00] to show that you're kind of still there. I just did a comment because we did this too, is it's not a bad idea. As you go remote. That you have the ability for the clinician to very easily get you on video, not just a phone to talk to you, but there's a personal connection of video.
It's much harder for him to, to change over to another pathologist. If he, if he sees you and he knows Bob and he knows about Bob's ishing, and they connect on that, then if you're just from an unknown IP address. So I recommend that you begin to build what those connections are. Before you decide to make it all remote because the moment it can't be remote from your hospital, the, the sharks will notice that and then come to try to bid.
And then you find yourself in a, in an RFP with a bunch of competitors when you've had that contract for 25 years. [00:31:00]
Aleks: Super important thing to think about. And I think it's a lot of this change is It's psychology and so I think even this relationship building piece was not that crucial when everybody knew everything is on site and this is the institution we work with and here are these pathologists as the institution's pathologists.
And then you have to think of something you really didn't have to think about before and leverage new technology to actually achieve the same result, or, or, like a new result because, right, the, the pathologist would, I don't know how often they, like in veterinary pathology, you just write the report and that's it.
Sometimes you contact the clinician, but it's definitely not a patient facing profession. You have subspecialties like cytopathology and you can get more involved, but basically in general, pathology [00:32:00] is not the most patient facing, profession, right? So then Establishing this relationship with the treating physician.
Iinteresting. Anything else? Any, like, any advice on change management in any other capacity?
Greg: Well, this is a, and this is a very interesting topic. I'll just touch on a few things and, perhaps in the future, your, your audience may be interested in diving deeper into each of these, but another one to consider is that whenever you.
Move over to this technology. Remember that you're now in the game of other competitors, and that means you have to bring more to the table. Be prepared to do that. The key phrase you can ask yourself is that when another competitor, because competitors will come after your contracts and it's your livelihood when they come after it.
You want the CEO of the [00:33:00] hospital to say, well, we can't really change over to you guys, even though you're less expensive, because the current service also provides one, two, three, four, five things. And then you have to find those things. So that leads to this question. It's an interesting question is, is for your leadership, who get along with the CEO and, and, other people who make these decisions?
Need to go and meet with these people. And this is subtle. Commonly, when you ask a person, what can we do better? Do you have any concerns? Are there any complaints? What can we do to be the best that we can be? Commonly, they will say nothing. They will say, Hey, there's nothing. You go away. And in the mind of the CEO, it was, We have a group that's interested in always improving even though he didn't give any suggestions, but he knows that's who you are So with this other group, it's an unknown So the devil he knows is the group that says listen if there's anything we can do to [00:34:00] improve, let us know and he may not know anything.
So then when he asks be ready with a few things. And be ready to say, is there anything you, we can do differently? How can we improve? What can we do to be the very best? He might say, well, isn't that your job to figure that out? And say, yeah, so we have a couple of programs that we're adding on to grow with, with this.
And one is that we're going to be able to have faster turnaround times because we're going to be able to get this slide to maybe a subspecialist. They'd be able to shop it with different pathologists now that that's available and then have a handful of these things you can already say. As far as the technology, once you go off site, there'll be an expectation from the clinicians and the CEO that things are going to be more efficient.
They're going to be faster. So then you have to figure out, well, if they're going to be faster, are we going to provide 24 seven? You must do it for [00:35:00] surgical paths because the people are in the middle of surgery, but there's not a lot of surgeries after hours, unless it's trauma. Generally, that's not pathology involved, I think, but they may want faster turnaround times.
So now what you need to provide with the technology, now that it's digital, you have a work list, you're able to give turnaround times. and do QA on all of these. It's highly recommend. We did this with with radiology is you provide a report, start to provide a nice digitized report that gives all the critical information that the that the CEO want to know.
How long did it take for you to get the study? What was your timeline? How fast did you report it back? What fraction of these by your internal QA system, where they changed, do you give a little histogram of how fast things were done? What is your discrepancy rate? If he gets a report with, colored graphs and pie charts and of what you're doing, [00:36:00] that's going to be hard to compete against.
That's giving the CEO that CEOs love a dashboard of what you're doing now that you're digital is it's just another element of upping your game.
Aleks: Okay, there's a lot of thought to manage the change and at different levels at the level of colleagues at the level of decision makers, and, and I bet there is.
So there's another possibility that we now have now that everything is digitized.
AI. Artificial intelligence. So, and there is discussion about this. Is it our friend? Is it going to take away our jobs? Should we be afraid of it? My question to you is how is it currently used in radiology? if there are any, application examples that you could give us, and how do you see it playing out in pathology?
Again, any potential [00:37:00] synergies or differences, synergies and relationships between humans and machines? What's your take on that?
Greg: Great question. I think the best thing to do is to first, stratify the response. So there's what happens, once the patient, a slide is created, let's say between the slide being created and it being presented to you, there's some time there where things can be happening.
For the pathologist thing where they would like those to be as efficient as possible and to be not as involved during that time, then there is the actual interpretation when you open the study and you're looking at it and you're looking at the slide and making a diagnosis. which, and I, I won't get into that too much because I think that would be presuming your process a little bit, but it's kind of like what we do.
Aleks: But you can give examples what you did in radiology. I would be super curious about that.
Greg: Sure. So one of the examples is that they've been trying to get AI to, to make a diagnosis. [00:38:00] So, during the reading part of it, they've been, they put in case after case after case, knowing the diagnosis and, and working on a neural net to give a diagnosis.
There's been some early work on chest films where that's a, a very low paying case, but it's been able to at least detect stuff that you ought to look at. So this is the issue is, and I'm glad you brought this up, the big issue is AI can look at something. And is better at determining something unusual and determining if nothing is there rather than characterizing something.
It has a little bit more of a challenge. But the problem is that when and they use AI on a lot of these studies. AI gives a recommendation and it will say, I think that this is abnormal, but your clinical acumen may look at it and say, well, I think it's normal though, even though you say it's abnormal.
Well, what happens if it comes back abnormal? Then they could say, well, you ignore the [00:39:00] AI. It was trying to help you, but are we supposed to just agree with the AI all the time? Or the AI says it's negative, but you say it's positive. And so they do a biopsy. It ends up being negative. And then they say, well, you ignore the AI, unnecessary procedure, plus there's the time.
If the AI does what it does, and I have to go and double check and second guess the AI, your times will increase. So the legal aspect of this needs to catch up to this, where we have to decide, are we going to use our ideas or the AI?
Because it's not fair to put us in between the two and then just come after us when the two of us don't agree.
So it may be a little bit of an opponent in the early stages until the legal part catches up. But they have tried to do with some wrist film, plate films. They've done it with some of the cast scans for screening for cancer. And so there's been some early work on that, but we're still caught with the legal aspect.
Aleks: So how are you, do you have any suggestions? I think this is because we only have one, so the algorithms and the models being built for pathology at the moment are very narrow, just for one particular, entity, let's say the one that this FDA cleared is for prostate cancer and it highlights the areas of malignancy on the prostate slide.
And then the pathologist goes in and says, yes, no, and does some other stuff. But basically it's like highlighting of something that should be looked at that is being suspected of malignancy. And it's a very narrow application.
I am not aware of, legal discussions other than in the context of regulatory clearances and lab regulations.
How does legal play into this? Like, how are you guys discussing this in radiology for, do you have a specific application that is like being [00:41:00] under review right now or something that you have in mind?
Greg: The AI being used to, and the diagnostic aspect of looking whether something is normal or abnormal and then highlighting an area that you need to look at.
This is an early stages. I think there hasn't been enough legal precedent. That I've been aware of where they're indicating whether the AI was actually our friend or whether it was like a It was really like a lawyer sitting next to you on every case and that's not much help if it slows you down And then it's an opportunity to say well this other computerized thing said this versus that I haven't seen enough legal precedence of I think it's normal he thinks it's abnormal and vice versa. To make a good comment, radiology is too early in that stage, but it would be nice in pathology for you guys to try to get together and say, we have to come up with something that's accepted behavior by the college rate college of pathology that indicates that you still have to make your call.
Otherwise, [00:42:00] the computer just doesn't and you got to whatever company does the AI, but if you want to have the, the pathologist be responsible, then he has to be allowed to make clinical decisions. Based on what he sees. And I know you, you will be stepping into this and any, bright, not necessarily ethical lawyer would look at this and say, every time something gets missed, I'm going to go back and scour the AI and see if they disagree.
And I'll bring that up in court. And I think that the American College of Pathology has to have a position on that. I would recommend before you get into that.
Aleks: Currently, the current, solutions are, so called computer aided diagnostics. So definitely the pathologist is in charge and has the last word, but it hasn't been deployed at scale at the moment.
So we might just not have a case like that and we might just not have a precedent. So, definitely an [00:43:00] interesting question and a specific question.
The AI that is being used on chest films, what, what is it supposed to detect? Like what kind of abnormalities is it trained on? Like any abnormalities or specific things that it's looking for?
Greg: The different AI, you put it through an AI analyzer and they generally have a specific thing that they're looking for. It can go through multiple AIs to look for various special things. One of them is just a neural net is that, as though, the film comes across and there's something abnormal. It doesn't fit into what's usual.
And it doesn't even tell you what it is. It just, it lies outside the neural net of acceptable normal. And then you have to go look at it, figure it out. Commonly, it's things like that you might miss otherwise. like nodules, pneumothorax, cellular fracture, which kind of echo what we were trying to do back with the resolution issue.
And, but more, more so AI is trying to look at [00:44:00] cross sectional imaging, which is higher contrast resolution, things like CAT scans. A little bit on MRI, they're also trying to get into, PET scans and those things they're not doing as much with. But it's CAT scans, mostly lungs, looking for nodules, looking for nodules that need to be followed up or biopsy or left alone.
Aleks: Okay, interesting.
So, How did going digital change, the reimbursement and the money situation in the radiology space? How did it influence the way doctors are paid? And we are just in pathology at the very beginning of figuring out what the reimbursement could be, even for reading slides digitally.
And so there is implementation of some like test codes that don't really get reimbursement yet. How did you do that? What was your path?
Greg: [00:45:00] Largely at the mercy of CMS. And as you can imagine, it's, I won't, this will inform anyone on this blog, but the job of the government is to not pay. That's their job.
Find ways to not pay not to find ways to pay. So whenever we would even do protocols there where it was a faster protocol to do something, they felt like, well, you're more efficient now, so you ought to make less money. So they even did that on the technical part. So watch out when you guys go to do this.
If your technology of scanning the slides and doing whatever happens, Once that's detected that you're doing something that is actually more efficient, then the reimbursable drop. We saw that with CT abdomen and pelvis when they put those studies together. We also saw it when we did MRI protocols to scan the patient faster.
And since 2000, the reimbursement for radiology has dramatically dropped, [00:46:00] partly because of just the efficiency went down by 38%. And that doesn't even account for inflation. So, when you look at all the different specialties that got, that got burned by this, it didn't even keep up with inflation, it went down.
Can we link that to the efficiency of radiology with PACS systems and teleradiology? Maybe, so that's something to watch out for now. You'll get a little bit of time before that's kind of detected and then enacted but I would suspect that as you become more efficient with what you're doing CMS will will identify that and will respond in kind.
Aleks: So how did you like counter that? Did you have to then just basically do more cases because you were more efficient? Or like, how did you not lose money as a, as you, your salary or doctor's salary? Like, was there anything, is there anything you can do? Is there any way to [00:47:00] react to this? Because it seems like it's going to happen whenever you increase efficiency. What do we then do? We just earn less or…
Greg: Well, you're, you're elected officials of Merritt College of Pathology can always, approach, approach Congress like we did and approach and say, this is pointed, you know, the more that you try to reduce reimbursements, The more you're going to discourage people from becoming doctors at all, which has happened.
So in light of that, Congress responded a little bit to this. The other issue is that directly answering your question is the way people made up for it was in volume, isn't it? You would just be doing more studies at a lower reimbursement to try to routine what that is. Now, what the natural issue with that is that now you're spending less time per study, which can be, cause you to miss things or mischaracterize things.
And then also now that you're doing more studies, your, your liability [00:48:00] exposure increases linearly with what you're doing added on to the fact that you're not spending as much time per study.
Aleks: Is there anything you can do with that? Can AI kind of help with that? I assume it could if you kind of like have an AI body that is helping you while you have to work faster, but I'm generally like, I don't feel great if somebody, makes me work faster.
I like, was in situation where I had to work faster and I missed things. And it just, like, wasn't worth the little acceleration in the evaluation time, because I had to go back, had to re read stuff, had to re write my report, and, and, you know, in, in, We have like kind of gateways in toxicologic pathology.
How I work, I always have a scientific review that is like kind of inter, intra institutional review by, a more experienced pathologist. [00:49:00] And then often I have a peer review by an external pathologist, which I think is a, in general, like this way of, of checking the quality. And helps, but this is not the case in a diagnostic workflow.
You don't have somebody looking over your shoulder and then suddenly you have to work faster and, oh my goodness, I wouldn't want to do that. No, I hope AI can help. it kind of is like, what happens when technology enters and you have to figure out like what, what the risks are and how to mitigate the risks.
And speaking of how doctors work or how they would have to work, to make up for like lowering reimbursement, in general, I was very enthusiastic about digitization being something that can improve doctor's wellbeing. But I guess there are two sides to this coin, [00:50:00] because if you're made to, work faster, there is no increased connection with the, treating physician or anything like that. But, in general, especially in the context of your teleradiology company, can you tell me a story about the time when, where digital radiology had direct impact on your well being and on patients access to care?
Greg: Yes. I, I want to say a quick comment about what you brought up before just very quickly Is that I have found that generally trying to get a doctor to go faster.
It doesn't work well, it's very hard to do and it's also hard if you have a doctor who's making mistakes It's very hard to get them to go slower that those are just two that was something. I just wanted to quickly mention so when digital has helped us when we went over to this company where everything was digital And we're able to share cases on the internet [00:51:00] very quickly with experts.
This has helped us a tremendous amount. We've had cases where somebody was looking at a case and they wanted immediately to check with another radiologist. For instance, one of them was a Mal rotation case on a baby. And one of our radiologists wasn't quite sure of it. Without digital, he wouldn't have been able to just immediately click a button and have a pediatric radiologist take a look at it.
And then she diagnosed it right away and then can help the patient. I mean, every night we would have subspecialist radiologists who are able to share cases across the board. And somebody may be in pathology, I imagine you have subspecialists.
Aleks: We do, especially in the U.S. It's not like that across the world, but in the U.S., yes, you have very subspecialized subspecialties.
Greg: Oh, yeah. And now you can, you enable that. And by going to that, you enable it. And interesting by bringing that up is when you decide to [00:52:00] move a particular practice from analog to digital, there'll be some expectation. There'll be some expectation. You're able to shop your cases between.
But it was almost every night we would have some case where I don't know if there's something wrong here. And then that brings up something the conspicuity of something being wrong Is that somebody can look at something and they they there's something positive, but they don't even see it. So never even had a chance to be shared with a subspecialist. But something doesn't quite look right and they can show it to somebody else. And regarding your AI question before just a quick comment is that even if AI tried to help us with our workflow and it could identify studies, it says it's basically normal.
The problem is that when you walk into a study and you already hear that basically it's normal. Do you give it the same kind of initial attention that you would another study…
Aleks: You have confirmation bias that it's negative and you like, [00:53:00] that's psychology as well. Again, like you kind of know the, diagnosis, whether normal or abnormal and you look corroboratory evidence to confirm that.
Yeah. That's another thing that I don't have an answer for because it is real. It's basically how our brains are wired. Okay, so this, yeah, I, I definitely take advantage of, being able to consult colleagues. Now that I work with digital slides, I basically can send them screenshots of the places where I am struggling and I can get an instant response because we're all connected in the same network.
I can just send a screenshot on Teams or I can share the slide. and get a second opinion that's informing my judgment. This is amazing. what I have noticed in pathology world, and you tell me if you guys, Have that as well. I assume yes, but basically now [00:54:00] institutions that do not have pathologists can tap into the resources of institutions that have pathology labs and pathologists Without having to employ them so community hospitals can work with independent practices or with like central histology labs and can give access to pathology diagnosis, to their patients where before it was a hassle and maybe they would avoid it because of, I don't know if they would avoid it, but let's say it would be a lot more to overcome to actually provide this. So then if it's difficult, you always think twice whether you want to do it or not.
Greg: As you do this, I think you will be able to connect to these databases, connect to academic institutions. there may be places that are, for instance, in radiology there, we would do what we do. But then we would be able to potentially connect with a place that was very good at like muscular [00:55:00] skeletal radiology and be able to connect with even the institutions.
But as your data builds up, I imagine, you're going to be getting into AI somehow. The better the AI can get the better it can at least tell you whether something is basically normal. An example is that there's when mammograms done in the United States. We have very good care here. They're told sometimes that while they're in the waiting room, whether the mammogram is naked, but there are countries where their their mammograms are backed up as much as 1 or 2 years.
Well, what good is a mammogram to identify..
Aleks: Two years, not really.
Greg: So, some of the implications that can be done with, and I would imagine, Pathology could do this is some of these poorly served areas. If they can at least digitize these slides up and have AI come over there and see if something's basically normal versus abnormal and then that smaller fraction of those could be sent to you for more analysis [00:56:00] and AI could have a good place.
So I'm just kind of mixing an AI with the database, but that's a place for you guys, potentially.
Aleks: Definitely. And, it's, interesting that you bring the mammograms because, I, mammograms is like, is like a population screening test, right? And we actually, I thought we didn't really have to, any of this, but we have the cytology, the pap smear, which is also a screening test.
And there is AI being developed to screen the cytology. I know that it's being deployed in several countries, including U. S. and it's also being deployed in China at scale, where like you say, the AI is picking up the cases that need further, work. And that's a similarity I found between radiology. I don't know.
Do you guys have any other screening tests? Like, is there a screen test for lung cancer or anything like that?
Greg: Yes.
Aleks: Yeah, you do?
Greg: With smokers, we do [00:57:00] CAT scan screening for, for the lungs. we, we certainly do that. We do some screening prostate exams. They, and of course the mammogram screen is the most common one, there's other screen exams we'll probably add on to as AI becomes more reliable, but we're still stuck with that idea that is AI going to read it or us?
And it's still, it's not so far. AI doesn't really help us save time. If we still have to go in, and read the study and do the work. So there's not really a time savings on it until we can trust the AI read.
Aleks: Okay. I understand. So what's on the horizon for digital health that gets your gears turning, especially for our specialties, for the image based specialties, pathology and radiology? Do you have anything that you think is going to be the next cool thing or something that you're looking forward to?
Greg: I think that [00:58:00] it'd be interesting for us to be able to develop these tappable databases. To be building these databases, so that we can, we can actually assist the doctors get the AI to be good enough to be sure of what is, what's negative and what is positive and then we do the rest of that work and then that raises the question of how do we then evolve as doctors because once we have AI doing a lot of this interpretive work that has been traditionally ours, what happens to our, our rules?
Do we spend more time instead of just cranking through. A lot of these cases where maybe 70 percent of them are negative. And that's where we make our money. We make our money on negative because that's what's done most quickly has the lowest liability and we're able to charge for what that is. We actually don't really make money on difficult studies.
We spend more time on it, higher liability. We're talking to the ordering clinician going over the studies. [00:59:00] So we don't really make our earning there. But if much of this negative has taken over, it'll be important for American College of X to get in there and state that we're going to have to dramatically change how we reimburse this because if a doctor's only cases, 10% Of, of the cases are patho… pathologic, let's say, and he could only get through maybe, 10 or 15 cases a day.
He's not going to do it because he can't. So we'll have to, as AI takes that over, that'll be a crossover. But what's interesting about this, what's interesting is to build the databases so we can be more accurate about this. When you go and read a slide and you say, you know, I think this is an oligodendroglioma, and then for you to be able to say that and have AI, look at this lesion and look at what you're seeing there and had come up with a differential with percentages, then you might say, you know, maybe it really is the second thing. Maybe it's an astrocytoma and I need to now reconsider. I'd like to see a real [01:00:00] strong database of those probabilities, based on the other demographics you capture on the case. That'll be very interesting.
Aleks: Okay. Yeah. I think there are, I mean, I know there are, slide aggregation. efforts going on. but they're more at the moment when, what I'm seeing the slide aggregation is more for development of new models rather than as a reference library, which I think is a great idea to just have a reference library to increase the quality or increase the certainty of the diagnosis.
Let's call it that way. that's interesting.
So switching back gears a little bit, I would like to get a how to prescription or, for somebody for pathology departments or practices that are just starting to think about digitization. Any practical first [01:01:00] step or major do's or don'ts. Like more high level, more like, think about this, think about that.
Think about like how to manage change. Is there like a, blueprint that you could recommend or like a sequence of, things to take care of before you go and choose your equipment and decide how to connect everything and make it interoperable?
Greg: There are some things. And, with your experience in dealing with so many different pathologists.
and all of their different departments. I think you would agree that if you've seen one pathology group, you've seen one, they're all so different. And just as much as they're disparate in radiology and different players, even within the practice, there's such a constellation of personalities and how they deal with things.
A couple of just, minor tips, I would, I would say, is that for those who are very knowledgeable, how things are happening and, [01:02:00] showing people, this is what this is, and this is how you do things. What I would recommend is just a couple of things. One thing is instead of a younger pathologist approaching an older pathologist and saying, “okay, let me show you how to do this”, which can be very dismissive and disrespectful, although you don't may not even mean it.
One way to change that is to say, let me show you how I do it. Because that opens up, it's just a very different way to say something that says something differently, but it's really kind of small.
Aleks: It's 99 percent psychology, Greg.
Greg: It's a lot of psychology. and then another one is that when people are told something that they're aware of, commonly, it's a common thing to say, I know, I know.
But that, that suggests the person isn't smart enough to know what your knowledge base is. And it kind of insults them when somebody says. We're going to be moving to digital in two weeks. I know it's interesting. I know, even in your [01:03:00] personal relationships substitute your right. It means something very different as far as the impact, but the overall meaning is the same.
It's just much less dismissive. And there's a whole bag of these things that you can learn as you go along. But to answer your question, specifically, how do we start? The blueprint is very rough. And I think it would, it should begin with is observation. Is for a person, a couple of people in the practice to have at least observed a few other practices.
It may have begun, begun this journey and bring it up at a, at a monthly meeting. So, you know, we've noticed this. This is going on. We're seeing some early wins and we see we see some early failures about this And then really couldn't cut and then stop short and say, “What have you what have the rest of you seen?”
And then people will start to vent their concerns and issues and well This will be the death of pathology [01:04:00] and other people say yeah, let's get going on this right away It'll start the conversations. I would recommend when you get operations going You look at a lot of different companies that do this reach out to some of these other groups Especially if they're not a competitor, maybe if they're in another state and ask them what's worked out.
You may say, you know, we went with company A, but, you know, looking back, and I wish we want company B, and then find some data, start to work on slowly like that, and then transition over a period of time. One thing that helps is not to transition the whole company. the whole group at the same time, same date, have a couple of people do it.
We talked about failed back in case there's an issue and just start taking it nice and slow. that you'll also need to have to get the turnarounds to have voice recognition. Is, would it be fair to say that pathology is completely voice recognition for dictations or are you still doing analog then somebody [01:05:00] transcribes?
Aleks: There's still analog, enough analog, but I've seen the integrations with voice recognition, at the recent USCAP conference and I was super excited about this. So yeah, we're going that direction. It's not fully, voice recognition yet.
Greg: I think voice recognition my, my personal feeling about it, and I don't mean to be somebody who's encounters where somebody says, well, Dr Rose said you had to do this.
What we found in radiology was that when you dictate 100 reports, and then you go back in the transcriptionist to transcribe those commonly radiologists will just sign all of them off. And they can't remember somebody accidentally said a right nodule left nodule after 100 cases, who knows. So I think even for patient care, there's an argument that as you're doing the case, you check it out, make sure it's right.
Our voice recognition would type out words as you speak. So I could finish a sentence and look at it. but I would do to hold it to the whole dictation, look at it and then sign it off [01:06:00] also. That improves your turnaround time, because then the clock already stops when you dictate it. So, in general, you want to go to telepathology. I would recommend, migrating over to, the digital voice recognition as you do that.
Aleks: I love that. I love the advice to go and see somebody who already done that and learn from their mistakes. I think this is super powerful. I don't think it's So it's done. Indirectly, because those who are have already done and have been doing it for several years, they're they're already in a position to talk about their mistakes.
When people are starting the journey, they want to have a success story for all the stakeholders to keep buying in to this, not to cut off the funding. But there are enough institutions that come back and talk about the mistakes and the, successes, but I don't think there is a, like a general culture like, [01:07:00] Oh, let's check somebody else before we start.
So thank you so much for this particular tip. And, one more thing.
So what did the regulators think about digital radiology? What was the situation for radiology and how did you deal with regulators? And, and I'm going to tell you how, pathology deals with this. So, in one way or another. So, so there are two kind of things you can do for the digital pathology devices.
And you can go to the FDA and have your device cleared as a medical device. This is for software. This is for the scanners. But then when you use those devices in a, CLIA lab, you still need to do your validation and then, you know, you have other, regulatory bodies that you, respond to and show the validation data to.
How did you deal with that in [01:08:00] radiology? What were the regulatory concerns when you started going digital? Were there any, or did they say this is the logical thing to do, go for it?
Greg: Okay, great question. So that has a couple of different pieces in it. One of them is money, and then another one has to do with clinical connection.
So there's money connection. There's clinical connection. Technically, there's also credentialing licensing and working those things. Let's start off with the with the medical. I'm not sure how you guys were operate with FDA approval or FDA clearance, those…
Aleks: Firstly, we have the devices that are on the market. They are class two devices. So we need the clearance or the sometimes if it was the first one, I think you need the pre market approval, but currently it is class two. So it goes under the five 10K clearance.
Greg: Okay.
Aleks: It used to [01:09:00] be classified as three and then we only got the first device, cleared when it, when the class was changed. So only when it was changed to class two, lower risk, where the comp, was the company able to provide the data and actually have it cleared by the FDA.
Greg: Okay, we, most of the FDA clearance or versus, you know, if the approval, if the FDA clearance you in general just need to show that it's similar to what others are doing FDA approval.
It's different kind of process. Most of that was done early with us. They had to do with monitors. And making sure that they're the right proper illumination and it was mostly a box checking kind of thing. I didn't, I don't think it was all that really clinically important.
Aleks: So, question. Do you guys need a medical device monitor like a medical grade monitor to look at, [01:10:00] your film or your images?
Greg: The way that worked out.
Aleks: Yes, tell me, because this is a discussion that's happening in pathology right now.
Greg: What was considered to be, well, it was, it was assigned, it was called medical grade. If the resolution of the screen was adequate for the three megapixel for all radiology, except mammo, Which is five megapixel and then that was kind of called medical gray.
Aleks: Okay, so basically certain specifications that needed to be met for the images to be viewed at the required resolution. That's kind of how we operate right now. And that, and, and you know what, the gaming monitors are pretty good right now. You can get them for not too much money on Amazon and they're perfectly fine.
Greg: And then the other pieces that come up, that you'll start to run into a little bit with your, with your system. And it depends [01:11:00] a little bit on how, I don't want to use the word progressive. It depends how forward thinking perhaps. that your CEO is, that there are those who feel that, no, we're not going to do any telemedicine of any kind.
It's bad medicine. They still exist. They're out there and they won't let any images, breach their firewall outside the hospital. And they insist that everything has to be done on site.
Aleks: Even in radiology?
Greg: Yes.
Aleks: Oh my goodness. Okay, interesting.
Greg: They're becoming fewer and far between because now the clinicians are now not referring to the hospital anymore.
And that is the food. That's the food of the system is the referral is if you start having surgeons going down the street because they can't get their pathology read outside or quickly or whatever that service is, if it has to be done on site, they can lose enough business that then that talks [01:12:00] to the CEOs.
Now, I don't mean to dismiss CEOs, but many of them have come to feel like I need everything on site because I want the responsibility of this. I don't want us to get hacks, although you can still hack a system through a firewall. So with you guys, what's going to happen is you need to overcome that. And the way to overcome that is nice and easy, nice and slow, approach them kind of sideways because you don't want to get blocked.
We've been blocked where everything worked. The clinicians were excited about it. They couldn't wait to get our radiology. And then the technical people said, we're ready to go. We're going to set up a VPN connection to connect with all these images. And then somebody in the upper level said, we don't, we are the custodians of our images.
We don't want anything getting outside the hospital. End of story. So the best way to approach that, that we have found is to try to find competitor hospitals who allow what this is. And then they will see I could [01:13:00] be losing my, my staff and a reference to those hospitals if I don't do this. And then you come with this very gingerly approached idea of this is how the images will be protected with a VPN.
It's the same as a firewall that moves outside. We've been doing this for years. Radiology has already gone through all of these situations to have secure images. We don't have problems with, but that could be a significant issue.
Aleks: Interesting. I love your perspective. I love your outsider perspective of how this process happened and how radiology succeeded with this and how to approach this change.
Is there anything else that you would like to share with my digital pathology trailblazers? Anything that I was supposed to ask but forgot?
Greg: Well, I don't know whether down the road when once you guys transition through [01:14:00] it now, here and in 2024, 25 up to 2030, that there will be others down the road who will then look to you and say, how did you guys get through it as a second generation of digitizing, modalities, because with, telesurgery and some of the other pieces that are coming into this, you may learn things that you are then teaching others going down the road.
But I think one of them is, is to step softly. And be highly conscious that the end results of most of these processes are somebody's sensitivities. There's a person behind all of this. And be very careful not to step on or step over anyone. Bring everybody into the talk. And almost always that will more satisfactorily grease the skids to success.
Then if you find an obstruction and you decide to go through it or around it, [01:15:00] that, that person might take it on as a quest to stop you. So we, sometimes we wouldn't even notice it. We wouldn't notice that we actually went around Bob. We didn't need to, we didn't even know that he was part of this, but then, he put his hands on his hips and Bob said, we have to stop this now because it was a usurping of power or ego.
That's an issue. That's a critical issue. just be aware of that. And I think that will help you in every aspect of financial, technical, and political.
Aleks: I think it's going to help me with life as well. Thank you.
Greg: I would agree.
Aleks: So Greg, do you still consult? Are you an active consultant? Where can people find you if they would like your advice on any of this?
Greg: There's, well, I have a website. which I'm just looking up the, it's at, if you, if you look up rosemachinewixsite.com, that's my website. It's at, [01:16:00] so rose machine dot W I X S I T E dotcom. And then that has a bunch of the things that I consult on and, what I do, but I'm, I'm happy to discuss anything about this transition and be part of any kind of a discussion, if you're in a monthly meeting. I'm happy to come in and answer questions and and talk about things. And sometimes it's funny how sometimes a third party person, even if you aren't that smart, like me, at least your third party, and you're outside, it can, it can bring you to a perspective that's very welcome. Rather than the usual game.
Aleks: Definitely. Definitely. I'm going to link to your website in the show notes. Thank you so much for joining us. And I hope you have a beautiful day.
Greg: Well, it was an honor to be here. I hope I helped a little bit, maybe one or two things that people can take away from this. A real honor talking to you.
What great questions. I think you really covered this very well.
Aleks: Thank you so much. [01:17:00] Bye.
Greg: Bye.
Aleks: Thank you so much for staying till the end. I think comparing pathology and radiology gives us a different perspective from somebody who's already been there, somebody who has gone through this digital transformation.
So if you're at the beginning of this transition, your institution is talking about it, or you just want to learn more about digital pathology, I would love to go get my book, I wrote the book, Digital Pathology 101 it's available in the PDF format for free. And I'm going to link to this in the description of this video and the show notes of this podcast.
So go ahead, grab the book and just start exploring what digital pathology is and how you can start and continue your digital pathology journey right now.
And I talk to you in the next episode.