Aleks: [00:00:00] Why do we have three hosts in this episode? So David Tulman, the co-founder of Instapath was a guest of mine on one of my podcast episodes. And when we were recording this, he said, oh, I am the host of the new, Beyond the Scope Podcast. This is the DPA Digital Pathology Association podcast. And I’m gonna invite you to an episode as well.
And then time passed. Nobody ever invited me. Till this week or two weeks ago. Anyway, recently David sent me an email and they invited me to a crossover podcast episode. So it’s gonna be David, it’s gonna be Giovanni Lujan, who is the cohost of Beyond the Scope and me.
[00:01:07] David: All right. So here we are. This is the first time that we’re doing this. This is a crossover podcast of some sorts. We have Aleks and Giovanni here, so we have hello? Hello. Hello. Hello. And the purpose of this is we wanted to get the, we wanted to get some of the podcasters together just to talk about.
The hot topics in digital pathology. Talk about some of the news items we’ve been following. We’re always like interviewing guests. So we never give a chance to give our own opinions on what’s going on. So this is a fun opportunity to do that. And so since we probably have all sorts of different audiences here, this is gonna be published on a couple different podcast feeds.
Aleks and Giovanni you two are meeting for the first time. So why don’t you guys we’re meeting
Aleks: for the first time? Indeed. So yeah, so for, yeah, we are all guests and all hosts today, right? Cause this is Beyond the Scope and Digital Pathology Podcast. So Giovanni, welcome to my podcast and feel free [00:02:07] to welcome me to your podcast guys.
Giovanni: Very happy to have you I’m a fan.
Aleks: Yeah, I’m a fan of yours guys. I have been looking at the, your podcast and listening to your podcast and I stole some of your guests. So for example, I love the podcast with David Clunie and I totally, after listening to your podcast, I was like, oh, he has to come to mind as well.
David: So I think I think if we’re not stealing guests, then we’re not doing it right. Cause our goal our goal for both of our podcasts is to get the best guests and. Naturally we have to share.
Aleks: Exactly. And also this is not such a huge niche. I would say probably everybody heard about everybody and it’s good to cover topics from different angles.
So yeah, I’m honored to be here guys, and nice to meet you, Giovanni, because David, I already know because he was already guest on my podcast before
Giovanni: Yeah. He told me about that and I think we started at the same time because I remember [00:03:07] I started following you, I think in LinkedIn.
And we were just because I like what you were posting there. And then I noticed that you were also coming up with a podcast. And so there were several similarities and yeah, and as you said, we get it from different angles, different perspectives. And perhaps we share guests, but we all have different perspectives, different point of views that we elicit on the guest.
So I think it’s good that the guests go around sharing their opinions with us because they always have something different new to say.
Aleks: So I remember also Chen Sagiv, was your guest once and she was my guest as well. So I’m gonna be linking to all those double podcasts in the description and in the show notes.
But one question my question to you guys, before we dive into hot topics and everything. Yeah. What is your mission with your podcast and like what’s your [00:04:07] main angle, and then I’m gonna tell your mine.
David: Perfect. Giovanni. Do you wanna do it for us?
Giovanni: Yeah. It was an idea of one of the committees of the Digital Pathology Association to put some faces and voices out there to try to broadcast the news, to put some of the projects we’re working on out there, get people to know about what the DPA does. And that evolves into also presenting people who have something to do in the digital pathology world involving everybody, not only academic, but also the industry research in all these different characters or professions that have come together. To be this wide world of digital pathology, which is not a physician world [00:05:07] is not an engineer’s world is not a business world.
This kind of a combined world where we all came with different backgrounds trying to promote this. So I think that’s the main goal of us, to get everybody together, learn from each other and let people to know. Digital Pathology Association, anything I missed David,
David: The only thing that I would add is I came in a little bit late in the process.
This was not my original idea, but the first time that I was on this, that I was on this committee meeting. I remember Giovanni like super excited talking about how he was really into doing it. He just didn’t there just wasn’t like a plan to actually do it. So I had done, I had you
Aleks: had a podcast
before, right?
David: Yeah. I still do it.
Aleks: The perfect candidate.
David: Just for fun. I have, and I still do this, college football podcast with my friends. It’s more of an excuse just to get together every week and talk, cuz we’re all in different [00:06:07] places. But I actually know, I actually of knew what I was doing in terms of producing the podcast.
So I happily
Aleks: I see your headphones and I see your microphone and I like. I okay. Full disclosure. Okay. I’m in Poland now. My internet is so bad and I am recording on my phone today, so totally like a zero professionalism of a podcaster. Whereas David, if you’re looking on video, you are gonna see his mic and everything.
And if you’re listening to audio, I hope my audio is not that bad today.
David: Sounds really good right now.
Aleks: You okay? It looks very good. Tell you how great we can do some post processing. Okay guys, now we’ll closure with the light. This is my light. nice.
David: Yeah, I think we figured it out. And I think the last thing I wanted to say about our podcast is just to echo Giovanni’s thought on how just digital pathology is such a collaborative community.
I’m just really happy that we. Doing the cohosting. We have [00:07:07] the voice of the voice of a pathologist. That’s actually signing out cases with digital pathology represented. And then we have a voice of someone like me. Who’s innovating in the space, knows a little bit about this business engineer by and knows engineering.
So I’m really happy that we have the two different perspectives.
Aleks: Yes. So guys now the three of us cover even more, cuz I’m a veterinary pathologist exactly. Working in pharmaceutical industry in a CRO supporting pharma. So totally all the angles. We do not have a computer scientist. Do we?
David: No not as a host, but we get computer scientists for 30, for a third of our guests.
So by nature. So it works
Aleks: out that way. No. So we cover quite some spectrum of this collaborative discipline that we are engaged in digital pathology. And I totally second you guys, it is multi- profession, multidisciplinary area. And the goal of my podcast is, what I have seen that there was a big gap between [00:08:07] pathologists and computer scientists.
And then basically, because we have so many different stakeholders I was in a situation where it was pathologists and computer scientists, but you have pathologists engineers, you have pathologists, regulatory, and like all of them together. And you need to build those bridges. So I think podcasts are a fantastic way to build those bridges for an area that needs bridges to function well.
Yeah, that’s my mission. And I guess we have a bit of a similar mission from a different angel, yeah,
David: let’s ask. I’m gonna, I wanna ask Giovanni, because you, a lot of your projects are rooted in computer science, right? So you’re working with a lot of computer scientists.
Giovanni: Yes, definitely. And that’s our second, like what does stage,
Aleks: how’s that going? Giovanni?
Giovanni: interesting because I have interesting, right? Not much of a computer background or an engineer background and When I started all this, not too long ago, I started [00:09:07] with the pandemics. I that’s the time when I came and moved to Ohio Columbus where I am right now from Dallas, Texas.
So I came here with the purpose of helping with the implementation of the digital pathology workflow. And at the beginning it was, Or daily routine work, just moving from glass to, to, to digital. So that was the first part of the transition. Now we became fully digital in January. This year, we’re all over the slides get digitized after production.
And now we are on the verge of deploying our first sets of algorithm. In, in or daily use. And in that part is where I started interacting with all these different engineers, computer technicians, different engineers, I know a complete different lingo, a complete different set of expectations. And [00:10:07] then is when you realize that in order for a meeting to be completing, to get somewhere,
you have to start including all these different representations, because there are only pathologies. We do nothing. We just can’t wish. And so we need a, an it person who tells us who has the what’s doable, and what’s not based on what we have or what we need to acquire. And then we’re working with the ice.
So we need somebody who knows programming. Who understand what that’s about and who can judge and qualify an algorithm, because there are so many vendors right now and they can come and tell us everything. And the only thing that we do is open our mouth, right? Oh, but we need somebody to ask this substance, our questions, that, how do you assess an algorithm?
And I’ve been learning an immense deal of things. And I’m hoping that my colleagues, the [00:11:07] engineers and the computer scientists are also learning from me pathologists. So everybody of us has a different component. And again, if we are all not in the room or in the phone call, the meeting goes nowhere because that part is missing very.
Aleks: I hear a shout out to all those engineers, all those computer scientists and all those vendors. You go and drop comments below this this podcast, wherever you’re listening to it, we need you. And it was just a joke. We love you and we need you.
Giovanni: totally, absolutely. They are. They’re in the same ball game too, because everything is new.
They’re trying to come up with solutions every day from hardware software. Algorithms, what have you, everything is new needs to be tested, needs to be, proven not only that it works empirically, but also that it delivers. And when this implemented clinically [00:12:07] and so if not, if it wasn’t for them, we wouldn’t go anywhere either because that’s basically that was become sort topic of conversation.
Another thing is. And this is what vendors are learning. I think from the experiences is that they cannot do everything. Digital pathology, computation and pathology, I think initially was thought to be, oh, this company is gonna go into that and it’s gonna do everything. Hardware, software they.
The, whatever the steps,
Aleks: It didn’t happen. Did it?
David: I can speak to that. I can really speak to that because when we started when my company spun off from Tula in 2017, We really genuinely thought we were building like a platform microscope that was gonna be able to do literally everything in pathology.
as we got out there, we did our pilots and we did our beta testing. We were so wrong. We just, we learned that we had to, we really had to, in order to at least get some initial traction we had to [00:13:07] significantly narrow our focus, pick one lane to go down and to pick a very narrow number of use cases and applications and truly validate those and the whole one company or one product to, to rule them all was just never gonna be possible.
Aleks: This idea, died quickly. I think it’s similar to the AI, like the concept of AI. Everybody’s thinking, oh, it’s gonna do everything. No, it’s doing one thing at a time. And you can have like hundreds of them that do one thing at a time, but they still do one thing at a time and they still have to do this one thing good to be accepted in the community.
David: I think that’s the perfect transition, cuz we wanted to talk about some of the hot topics and some of the news items that we’ve been following.
Aleks: Yeah. Agenda do we might we have, you have a very like ambitious one and we might need to divide into two episodes, but that’s fine.
Let’s go over it.
David: No, I think the first, the relevant one [00:14:07] is. In the last year we’ve seen CAP produce guidelines about how institutions should be doing internal validation studies. If they are planning on go going digital, what should they be doing with the whole slide scanners that they’ve just purchased to validate?
And I think like the highlights are. 60 cases seems to do it looking for a 95% concordance rate and then some kind of washout period to, to re-review those images afterwards. And so that’s now a guideline that’s published, Giovanni had, is that something that you guys followed at Ohio State?
And I think we did a podcast on this, but maybe you guys came up with some analogous guidelines as well that you followed when you did your validation.
Giovanni: We’ve been using everything that the CAP is putting out and I happen to be part of the digital and computational pathology committee.
And this committee is not supposed to put out guidelines, but we are recommending the guidelines for the committee who [00:15:07] put out the guidelines. And I am playing an important role there because not too many people are actually doing the digital workflow in this committee. So one of the ones who bring that kind of experience and that’s actually my, my, because we are very fortunate in this committee to have
people with expertise in all the different areas of computational pathology, which I’m lacking in terms of training. Because again, I don’t have a computer background or a scientist background to understand all the nuance of the decision process of AI participation, but we do have all this in this committee.
And my claim to find in the committee is that I’m actually doing the labor . So they can, I can provide that feedback if it is working or not. So we are using the CAP guidelines and providing also feedback for what, is not in the guidelines yet, for [00:16:07] instance, for validation of, for our systems at home. We started doing remote digital pathology and there were no guidelines for that. So we develop our own internal validation and guidelines for that. We giving that to the CAP and they are collecting information from other centers to, and eventually with this recommendations a set of recommendations will be put out to, for people who wants to validate the pathology from home as.
Oh, everything from the works, everything is in the works and not much is out yet just the tip of the iceberg, but a lot of things will be starting coming out slowly from the CAP,
David: I was just gonna to say that from on the vendor side, that’s certainly what we’ve started looking to anytime we set up a pilot with an institution, we just go to the guidelines and say, look, CAP is recommending 60 patients. This is the concordance metric that we’re hoping to hit, and we’re gonna do the washout period.
And we [00:17:07] find that seems to be, provide enough evidence for institutions to make a decision if they really wanna do this workflow or not.
Giovanni: And the FDA also, it’s been interested in collaborating with us. With the C P and the digital pathology association, because they see.
This is just the beginning of a wave of algorithms and techniques and new tools that they will have to validate. So they’re trying to set up initial guidelines and processes set up these processes that people, vendors can follow to have their tools validated in a relatively short period of time.
Aleks: Yeah, so guys, we have guidelines as well. We meaning there has been an FDA guideline document published it. I think the comments have been submitted I don’t know, last month for use of digital pathology in [00:18:07] non-clinical work, GLP compliance. So good laboratory practice compliant, which in theory… And in those guidelines, you basically, when you do digital pathology, you read those guidelines and it’s like common sense. So there is, so we have a process called peer review. So when I would evaluate a study for a pharmaceutical company, the pathologist from the pharmaceutical company is checking some percentage of what I did.
And if if we agree. So for this peer review this was already in the works. And now me as the one who is evaluating those studies as the primary pathologist can do it digitally, it, it could, it was already possible to do all the time because the GLP guidelines have a GLP compliance defines how to validate computerized systems and digital pathology system is a computerized system, but nobody has done it so [00:19:07] far.
And nobody was officially talking about digital pathology as like a special type of computerized system. So now we have guidelines pharma industry, and CROs submitted comments. And there was a first instance of validation for primary read at my day job at Charles River Laboratory. So I actually started reading studies digitally, but we also have official guidelines and I have to do a podcast episode about that cuz I think people are.
I don’t know, think the guidelines are more complicated than they are. Yeah. Do you have that? Do you have that impression, David, because you said you were reaching out to vendors and they’re like not really knowing what to do and you show the guidelines and it’s clear.
David: Yeah. And I think the guidelines, at least for the clinical validation make it clear and they make sense. And I felt like it was. You said the word common sense? I felt like it was just common sense what people were doing before it was officially published. And I was gonna ask you now that your guidelines are published,
were there any [00:20:07] surprises or is it just writing down what you guys were like publishing officially what you guys were doing anyways,
Aleks: most of the things is just writing it down. There is a lot about archiving that there was a little bit of discussion whether like everything always has to be archived.
For the peer review. So the second pathologist of just reviewing my work, you don’t really archive that you just say it was done, but now when it’s done on digital slides the guidelines say, oh, you should archive that. So there was a little bit of pushback okay, when we are, why should we do. It’s just like input the second pathologist.
They’re not the primary person evaluating the study. It’s the study pathologist. So the main pathologist and every like all the data comes from the main pathologist. So why should we archive that other part we’ll see, with the many institutions submitted the comments. So we’ll see if we’re gonna get [00:21:07] a waiver on that or not.
When we are actually doing the primary digitally, it doesn’t really matter because you have those slides and You, you have them digitally anyway, and then you send the digital slides to your peer reviewer, but sometimes the first pathologist is reading on glass and then we only scan and give it to the peer reviewer.
And that was also COVID not invention, but COVID implementation because usually for peer review people traveled. So I would read the study in my day job. And somebody from a pharma company would come to my place and meet with me, look at the slides on glass and talk to me about, okay, do we agree here?
Do we disagree? And now of course they couldn’t come. So we started scanning. And so there’s a little bit of discussion. Oh, shall we archive everything? Because it’s so much data, huge data and all that jazz, as we know very well from digital pathology. But other than that, mostly common sense what people were already doing and what was [00:22:07] GLP compliant anyway,
Giovanni: do you, when, because pharma has been using digital longer than we’ve been using it clinically.
When is there still how is the adoption of digital informa is still a lot of laboratories or pharma companies using glass, or how is the transition going there?
Aleks: Yeah, it depends. Like what part of the drug discovery and drug development pipeline. There was a lot of image analysis being done across the board.
So for this, like you do digital pathology period. There is a lot of discovery, so discovery is What I’m doing is safety, toxicological, safety, tox safety we call it, which is when you already discovered your compound. You checked on animal models. And this discovery part is already done and you’re checking, okay, this is [00:23:07] gonna be safe.
And you’re check first, the safety in animals. And then if it passes a serious, like a row of different studies, then you can take it to a clinical trial. So in tox safety area, the adoption is not that great. I have to say, this is so the pre safety discovery. This is non GLP everything that’s non GLP you can basically not bother about validating your systems according to the GLP guidelines.
People do it anyway, because usually you have a mixture of of work For what I’m doing, the adoption is slow. I think the fact that we validated it and that we started doing it, and we work with so many pharma companies, my CRO that we’re a hub that then spreads it to other pharma companies.
I hope I, maybe it’s just wishful [00:24:07] thinking, but basically there it’s patchy. It’s like a puzzle. Some parts are strong and built together, but there’s the it’s not fully digitized yet.
Giovanni: It doesn’t surprise me because that’s exactly what we had experienced. Why do you think it’s slow?
Is the pathologist theself you think the main, the limiting factor for adoption or the cost? I don’t think it’s the cost because pharma companies, usually that’s not a factor.
Aleks: yeah. Everybody thinks that, but then I think that too, they do actually, I don’t think in general, when I learned from David Clunie how much an MRI costs and that this is like millions of dollars. It then a scanner is a cheap piece of equipment. So I think the problem is the interoperability and of systems and how many teams [00:25:07] it takes to make this thing happen. Let’s say you have a company, I don’t know, Giovanni in your place.
Now everybody likes digital pathology. Like all the pathologists, right? You’re fully digital.
Giovanni: I always do we still have some naysayers?
Aleks: They, do. Okay. Yeah. Yeah. Everybody has, everybody has
Giovanni: Everybody said to some degree, but still some say that they, even though they get all the images first, they still get the glass for one reason or another.
Aleks: Yeah. I think if it was just, dependent on one group of professionals, be it pathologists or I don’t know IT, , whatever, it would be easier to implement because cross-institutional work is difficult because it’s difficult because it’s cross institutional and you have to talk to people with different priorities in their groups than like bringing everybody on the same page and bringing the corporate leadership on the same [00:26:07] page to do this, it was just challenging, like with anything else? I don’t think it’s digital pathology specific. I think it’s new technology specific.
Giovanni: New technology,
Aleks: Characteristic for new technology,
Giovanni: people are very resistant to, to change. We always compare it with emergent technologies and how they are, take this period that you have early adopters, then the slow adoption.
And finally people. Years later, they still don’t have an iPhone or a smartphone. They still use some sort of. Because
Aleks: doesn’t have just one person. I know that doesn’t
Giovanni: do this. Everybody knows somebody. At least
David: that’s what I, side note. That’s like what I aspire. hope I can get to the point in my career where I don’t need a smartphone where I can I’ve done.
I’ve done enough in everything that I direct goes so well that I can be off the grid a little bit because you’re you get married to it sometimes. And it’s such a, can be a distraction, but that’s. Not there yet. [00:27:07] Not there yet. Not even close .
Aleks: Yeah. But what do you think, David? Why is it difficult?
David: , I’m gonna disagree a little bit because I think I, I do think cost, oh,
Aleks: Go ahead and disagree.
David: I think cost is playing a big role in it. You mentioned earlier that, relative to MRI scanners, that a slide scanner is cheap and cheap relatively is certainly true. But.
It seems to me like pathology departments are the departments that are purchasing this system. And the pathology budget. My understanding is that the pathology budget at a lot of institutions is just not even close to what the radiology budget is or what the surgery budget is. So while these systems may be cheap relative to the, the expensive MRI systems or CT scanners, those really big.
Things that are used surgery and radiology,
Aleks: bigger, so much bigger budgets and yeah, a surgeons and radiologists, go ahead and comment. Why do you guys have more money than pathologists?
David: That’s a [00:28:07] fair question. And I, I’m wondering if if we see digital pathology adoption increase, that’s going to increase the budget for pathology departments and that’s gonna help close this adoption gap.
But where I see if we’re talking about the adoption curve, We’re talking about the bell curve. If we’re looking at my whiteboard, we are on the very left side of the bell curve and the earliest of adopters phase right now.
Aleks: Good. Then that’s good for us because we’re early and then there’s gonna be a boom, right?
Soon. Hopefully
David: that’s what the, that’s what the bell curve tells us is that there should be a boom coming. But I think the bus, I think institutions need to work on the business case a little bit in order for this to become more mainstream in the us.
Aleks: David. So you guys were doing the fluorescent way of imaging, right?
Without , making slides so that’s like their roadblock that everybody says, okay, it’s different from radiology because they just went straight [00:29:07] to digital media and we still have the analog media. How’s that going? At your company?
David: It’s challenging. I think, like I mentioned earlier, we’ve really had to narrow our focus and pick a lane and we decided to just take a baby step and make our first clinical application just a tissue adequacy evaluation.
So you can use our system for a like a touch prep or a frozen section type evaluation. And that’s just, we felt like that was the best starting point. It’s a low hanging fruit from a validation standpoint. And it’s also not a primary diagnosis. That’s just like the baby step that we feel we felt was needed to get pathologists, more comfortable with the workflow and the image type that, we’re we produce.
Aleks: Yeah, pathologist comfortable with the workflow is in our challenge in itself. But yeah. Yeah. So I think what you guys doing and couple of other companies start doing this to think how to eliminate the glass. Yeah. And I think it’s a fantastic idea. And I also think this is gonna face such a [00:30:07] huge resistance.
David: Yeah it, from my experience, it definitely. I face a huge resistance, but it’s okay. Nothing great ever happens without resistance, without overcoming dis resistance. And I, but I also I accept the resistance and I understand, I do understand the resistance and that understanding just helps me figure out something that is going to work.
I think we’re quite a ways off from being able to have glass free pathology. I know that we have to, I know that we don’t have endless time here. We have to wrap up a little bit and I had this grand idea for all these hot topics we covered. We covered a couple of them really well. So I think we’re gonna have,
Aleks: because we have to meet again. That’s okay.
David: Yeah. I think we’re gonna have to do this. I think we are gonna have to make this kind of like more of a regular thing. I was thinking, we do this once a quarter or a couple times a year. We get together.
Aleks: Totally. We wanna do this? If everybody who’s listening to this comment below, do you want us to meet or was it [00:31:07] pointless rambling. Let us know.
Giovanni: yeah.
David: Suggestions for topics we should cover.
Giovanni: I vote. I vote. Yes, let’s do it. You vote? Yes. So we
David: have to do, we have to come up with some kind of like catchy name for what is this group called? Like we’re some like. Pathology
Aleks: let’s do competition on LinkedIn, like for our regular meetings.
David: Yeah. I think that would be great. And I think let’s just wrap let’s just go ahead and wrap this up then. Aleks do you wanna plug what’s what are the next big things for your podcast and your website? Giovanni guitars?
Aleks: I have a big thing coming.
Cool. Yeah, I do. Have I prepared for this question of yours, David? Okay, good. I do have a big thing coming. Hopefully already September. Maybe I’m gonna start in October. So I reached out to several computer scientists active in the In our digital pathology space and I wanna invite them for webinars.
They mostly work in the open source software space, at [00:32:07] least my first three guests and anybody who follows that can guess who they can be, but you will know soon. So I wanna have a regular webinar serious with scientists active in this space and help them promote and learn from them.
Those people are like they have a huge knowledge that I wanna share with the community and they’re not vendors. So they’re, not marketing themselves. They publish papers, who reads those papers.
Maybe we read them if we have spare time, which we don’t have. But sometimes we read them, but the rest of the community and the, non pathologists and people who are involved in the digital pathology effort, but do not have this as their interests or whatever, have other stuff to do.
I wanna deliver this as a webinar. That’s gonna be also on my website and parts of it are gonna be on YouTube as well. So that’s my new thing. How about you guys? [00:33:07] Awesome. We’re gonna
David: post that’s great. Yeah, that sounds awesome. I can’t wait to see that we’ll post those links for us. Our next big thing is we’re gonna be Giovanni and I are gonna be on site at the Digital Pathology Association Visions conference and in October in Las Vegas, we are gonna have a podcast booth set up. We’re gonna be doing on person interviews, like radio row at super bowl. And this is so cool. It’ll be, we’ve done. We’ve done all these remotely and now we’re actually gonna be pulling some people in at the conference and those episodes should drop end of October and through November.
Aleks: Oh my goodness. Fantastic. This is so cool guys. I. I did not plan to go, but sometimes plans change and that would be so cool. The last time I was, if
Giovanni: youre there, listen, if you are there, we’ll add you to the table.
Aleks: That’s to
David: you find to register at the link that we’re gonna post on the episode. All right.
Okay. This is really fun. [00:34:07] We’re gonna do this again. We’ll come up with a great name for this until then,
Aleks: and I’m gonna have my full setup then, and I’m not gonna be doing this on the phone.
David: Awesome. All right, can’t wait to do it again. Thanks. This was really fun.
Aleks: Thanks so much.
