Aleks: [00:00:00] Welcome, my digital pathology trailblazers! I do have a very special guest today. I’m not going to tell you who that is, but if you type I’m going to tell you later, but if you type digital pathology in Wikipedia, you’re going to see his picture. He’s sitting there in front of his digital pathology cockpit with one screen looking at an image and the other one for typing or something like that.
If you don’t know how he looks, you will not know that it’s him because the picture is just a picture of him. It doesn’t say his name. But most of you know how he looks.
My guest today is Dr. Anil Parwani, the digital pathologist. Welcome to the podcast, Anil. How are you today?
Anil: I’m doing well. I’m doing well. And I’m looking forward to meeting everyone and interacting with everyone on the, online.
Aleks: It’s so great to have you. I have read many papers of yours and my digital pathology trailblazers as well, but I would love to hear your digital pathology story today. So let’s start with your [00:01:00] background. How did you get. Into this area. Let’s start like even before that. How did you get from where you come from into where you are right now?
Anil: Yeah. So my journey really started with research. I was doing research on viruses. I, in fact, work with covid coronaviruses. 25 years ago when I was in college, but I was really excited by technologies. I was really excited by what it would involve and bring people together, for collaboration.
So when I was starting medical school, I wanted to use technology to help connect with experts. So during my first year of residency, we started a web series, which was digital. It was. Not whole slide imaging to a static images, but we were able to connect interesting cases with colleagues around the world, and they were able to reach out and connect and ask questions and get gain insights into that disease.
The first scanners was starting [00:02:00] to come out. We had one of those at our institute and I started scanning slides and I was amazed by the possibilities that this technology can offer to a patient who is in Saudi Arabia to be able to connect to an expert in the, North America or Europe or wherever.
And those are the types of things that fascinated me. Those are, so I didn’t have any formal training. I didn’t go to a computer IT. I don’t have a background in it, but I was amazed by it. So I started learning about it. And as soon as I finished my residency and I took my pathology boards, I was offered an opportunity to be a pathologist in Pittsburgh, which is where pathology informatics really started.
And I was in, in the midst of everything and I remember my first talk they asked me to present at the API meeting, which is now called Pathology Informatics Summit. They asked me to present on digital pathology and predicted at this point, this is 2004. We’re all going to be digital in 2007 [00:03:00] and here we are in 2023.
We are not 100 percent digital, but amazing leaps have occurred in the field. More people have adopted and folks like you are leading this effort in many ways and getting the word out, getting this message out and getting people to follow this important field. So where I am today now is I worked in Pittsburgh for 11 years.
I worked on implementing the digital pathology solution, worked with startup company. We built a solution, which I think was an amazing solution. It’s still being used today, but the reason why I came to my, so this was the Omnics, which was a MGE collaboration. They started a company called Omnics and I was one of the pathologists involved in that.
There were several other pathologists, but we built a software, which is still being used today by many pathologists for digital pathology [00:04:00] slide review and workflow management, they built a scanner, which is still, which was sold at the time. But this opportunity helped me understand digital pathology from a clinical perspective.
Also, and it made me understand that when you interact with an image, you don’t just interact with an image, you’re interacting with a patient and you need all the critical information that’s needed to make a diagnosis. But it also made me realize that a glass slide alone is at the end of its life cycle.
You can do only so much of the glass slide. But once you digitize it, I can share that image. I can bring all the information associated with that image into a workstation. More importantly, I can start to apply computational pathology and AI on that image, which is not possible on a glass slide.
So digital pathology is not new to pathology. It’s in the future, it’s all going to be pathology. But what is important to understand is there’s only so much you can do with a glass slide and it [00:05:00] extends the capability of the information on that class slide to digitize. Imagine if you are in a library and you walk into a shelf and you pick up a book at that point in time, you are the only one who can read that book.
You’re the only one who can enjoy that book. You’re the only one who can benefit from that book. But imagine if it’s digitized, it’s on Kindle, it’s on Amazon, it’s on social media, hundreds of people can read the same book simultaneously. So that’s really, if you convert that analogy to digital pathology, that’s what that digitization brings to the field.
Patient care, research, education, collaborations, those are the things we could not do on a glass light as easily. We could still do some, but this is really. opened up a universe of possibilities for everyone. So that’s why I came to my current position when I was offered this role to implement digital pathology at Ohio State University and offered an opportunity to really focus on cancer [00:06:00] research, but it Really got expanded beyond that.
So I’ve been here for eight years and have wonderful colleagues and a wonderful leadership who has made me bring this technology, with with the help of a large team into a reality, which we couldn’t have otherwise done.
Aleks: So let’s talk about your day. So is the whole department, the whole pathology department, everybody’s looking at slides on the computer screen, or do you have some instances where you still look at glass?
Anil: So we have some instances where we still look at class, right? So we still have issues like I’m a GU pathologist. I signed out prostate, kidney, bladder. The only reason I need to go to digital from digital class slide is special slides. I will use whole mounds, big slides and we just acquired a scanner to help us with that.
But for the most part, most of my work is now digital. But there are occasions when I have a Congo right or where I have a cytology urine specimen, which [00:07:00] is not guys apply. But I would say 100 percent of the pathologist here use digital pathology. I would say 90 percent of them use it for primary diagnosis exclude and that.
Aleks: Lets talk about that so you can like bail out 10 percent don’t do it for primary diagnosis. How does that work?
Anil: Valid reasons, right? So they might have exceptions that they don’t have the right technology to help them. So areas which we’re lacking is renal bad. They have not completely converted to digital impact.
They have not completely converted to digital. We do digitize all their slides and inside of that side of pathology is another area we need to buy them appropriate scanners for these stacking appropriate standards for that resolution that they need beyond a 40 X up to 70 X 80 X. So those are the things that they are they need.
They’re not opposed to it. I can tell you this. We had a [00:08:00] downtime recently where our system went down. So all the images that were being captured in the histology lab, they were not available for review for our pathologist and it was the longest downtime we’ve ever had in the last seven years.
It was like 12 hours and pathologists were crazy. They became go mad, they became annoyed because seven years ago I told them when we go digital, you will not use your microscope and they did not believe me. But here we are in 2023 in April, May, whenever that happened, they were complaining that we don’t want to go back to the microscope.
So that was a transformation that I witnessed. I felt bad. I wanted to fix it and we had colleagues who were working to fix it and it was fixed. It took about 12 to 15 hours to get back to the ground state. And it was a silly, simple security issue, which could not be resolved between the vendor and the.
But [00:09:00] that episode, that event made me realize that how dependent have we become on the digital workflow on digital pathology. We, in our institute, this is a clinical system if it goes down, we call it the help desk pathologists call it the help desk they come. So we’ve come from a point where we had one scanner, no barcoding.
Most pathologists were not interested in using digital pathology to a point where I would lose my job if I didn’t bring digital pathology back up more than 12 hours.
Aleks: So that was seven years ago when you joined. You like started basically from scratch or like almost from scratch.
Anil: We did and but the good news is at that point we had support from the leadership. We had resources. They were committed to it and they were committed for research purposes in the beginning, but they soon realized that if we can have pathologists review cases digitally, those cases will become research data. [00:10:00] Those cases will go into a database where pathologists could. Review these cases or clinical oncologist or even researchers could review these cases in real time.
So it became a reality in 2018 where we started signing out digitally for the first time primary diagnosis and we chose systems which were already FDA approved. But what we have learned from that process is that the technology itself is no longer a barrier. We can get image quality, which is equivalent or even better in some magnifications to the glass side.
And pathologists are very tuned to it they are very comfortable with making those diagnosis and they can make that leap with without any issues. So we’ve come to parallel universe where we have in the beginning we had 90 percent pathologists on class slide review and only 10 percent on digital.
Now we are in a parallel world, we have 90 percent digital, 10 percent non digital, [00:11:00] so we’re moving into that process. But we don’t want to push it, we were probably one of the first institute to go live with digital pathology here, but we didn’t push pathologists to do it, we made it grow organically and we want, if you are going to buy a new laptop or a new phone, if you know about it.
If you are excited about it, you’re more likely to use it to its fullest capability but if I just gave you a phone because it’s cheaper, if it’s, like something it’s convenient, you would like most likely not going to use it as much. If you’re not passionate about it, you’re not going to use it.
And that’s something about technology, which I’ve learned over the years. You can push technology into people. It’s all about change management and appropriately implementing the solutions. For people to enjoy what they do, love what they do. And be passionate about it. So I hear, okay. So then 10 percent gap is not really the pathologists not wanting to do it.
It’s just a use case that still [00:12:00] needs their equipment, their infrastructure, and like basically the capabilities that the classical. 40X, let’s call it the classical 40X pathology already has. And recently I’ve also heard the story where a big hospital, large hospital was transitioning to digital. They would provide the pathologists both glass and digital slides and let them make the diagnosis on the chosen medium for as long as they wanted.
And it took six months to have the last pathologist totally switch from glass. What was the case at Ohio State? So for us, how did you manage to change? Yeah. So for us, we were very early in this adoption cycle, not just at Ohio State, but nationally, I would say we were one of the first ones who took the chance, brought all this technology, spent almost 10 million to.
fully outfitted across the institute. We bought high end monitors for all [00:13:00] pathologists, we bought input devices, we bought storage. Our goal was to get to a point where pathologists would love this technology. It was, there was resistance from the financial people. Pathologists are already signing out on a microscope.
What added benefit does it bring you? You are producing a report which is being used by oncologists who treat patients. You are producing a report which is done in a timely manner. Why do you want to introduce a technology which will possibly slow you down? So we had a lot of myths that we had to overcome.
High cost, the myth of delays, the myth of even misdiagnosis. So we chose to do this organically. We didn’t choose to just go live with prostate pathology or sarcoma pathology. We chose to have five or six people who were really passionate about it, who were super users. And as they started using, they will sign out their neighbors, the pathologists [00:14:00] were walking out in the corridor, they would see them doing this and they were asking, where are your slides?
I don’t need them. That grew incrementally. But I think the biggest push for us was COVID. Yeah, I think. For the whole industry, there was a it was a pandemic, but it pushed so many technology advancement forward. And I’m glad that CMS kept the remote sign out and all the advantages that all the digital pathologists basically work.
Towards during the pandemic to build more proof that this technology is robust and we can rely on it. Yeah, tell me about your pandemic experience at Ohio State. No, so there was a time, I think it was April, March, April, 2020, where we were asked to shut down and I could not. We could not even sit down with one, one on one with the residents or fellows or collaborate.
And there were pathologists who were stuck at home because that child care issues. They were at risk. They were at risk to be exposure. [00:15:00] We were already had the digital pathology solution in place. So we saw a 22 percent bump during COVID in adoption, from a few pathologists to double digits to in the 20s and S.
Aleks: Yeah, but still I would think it would go even faster and more because you already had the solution that solved the problem in place. Why only 20%?
Anil: So 20 percent from where we were.
Aleks: Okay.
Anil: Not from zero, we already had 10, 12 people. So we saw, we started to see a bump every month, gradually 20 percent increase. So by 2021, we had 25 pathologists who had signed up.
Aleks: How big is your group? So we have 40 anatomical pathologists currently. So we saw pathologists were already starting to use this pathology. They were starting to review their cases remotely. They were able to, look at cases and interact with presidents and fellows [00:16:00] remotely.
Anil: So we saw a big shift in how you were training residents, how we were teaching residents, how we are triaging cases. So now we’re almost at a point where I would love to have the remote pathologist to come back to the office. That’s still me. I still think digital pathology should not be used as an excuse to not interact with, right?
So So that’s where we are, where most pathologists are digital, very comfortable, they love signing out digitally. We’ve not seen a big difference in discrepancies on a class versus digital and some of the pathologists are mostly digitally signing out and they’re signing out remotely or the frozen sections and review of materials.
So we’ve seen a large shift in the demographics of who signs out digitally, where are they signing out? But. On the flip side, I still think we need to keep that engagement alive. We need to have pathologists [00:17:00] interact with one another in life, in person, post COVID. So that’s the challenge that I’m dealing with because we love the technology.
They don’t see a reason to come to the office and sign up. But we have many pathologists who are using the digital remote review as a way to triage cases. But at the end, they will come back to the office and sign out the cases, teach the residents. So we’re going through a transitional period right now.
Aleks: Do you see more interest in taking pathology as a specialty because of the advantages that digital pathology gives? Do you have like more applications because digital pathology is now an option or you have not seen any influence on that yet?
Anil: So what you’re asking is how does digital pathology promote more people to go into pathology?
Aleks: Yes, exactly. And if you have seen like a tangible change, positive change in that. Because…
Anil: Yeah.
Aleks: You know what?
Anil: What I’ve seen is in the United States today, if you look at [00:18:00] pathology residency statistics, I think we have about 560 presidency slots every year for pathology residents of life, medical students of life.
In the last five years, we’ve seen a decline in the number of U. S. Medical students going into pathology. So Of all the 500 plus positions, 40 percent of those positions are not filled by U. S. medical graduates.
Aleks: Okay, but they’re filled by foreign graduates. They’re filled by foreign graduates. It’s not that they there are any empty, right?
Anil: There is no empty slot.
Aleks: Okay.
Anil: Luckily, There is a competition at the end of the people scramble, they apply for those positions. But what we’ve seen in the last two years or so is more U. S. medical students are looking at pathology as a career. In my institute, we have several new electives and courses that we offer to trainees.
I have a student right now who is working with me. He’s from the School of Engineering [00:19:00] and he has been working and building algorithms with me all summer and he’s Brilliant. But then I asked him, why don’t you want to go into pathology? I had even, I had not even known about that as a possibility as a career and now he’s interested.
So there are many national efforts to improve the recognition of pathology as a career, which is cool, right? We think about pathology as a pathologist.
Aleks: I think we’re super cool. Being pathologist.
Anil: I think we are super cool too, and we are very tech savvy, more than people think we are. I can teach a few things to my kids about smartphones and databases that they don’t know, but they don’t believe that but at the end of the day, we have more medical students and graduate students and undergraduates who are intrigued by pathology. They’re interested in joining a career in pathology. We’re offering courses and digital pathology is the perfect solution to engage.
More and more people because I always tell [00:20:00] people if you’re looking at a slide, you’re looking at it around, you’re looking at an image. The possibilities are endless research, education, collaboration. It just opens up a new whole new world for you. So we have seen. Many more applications now, and this year for the first time in the last five years, we had all U. S. medical graduates, medical students match into our program and speaking to other programs around the country with their seeing this seeing the same track. So pathology, perception of pathology and digital pathology is changing. from a training perspective, from an education perspective. And I think it’s going to take us to a point where digital pathology will just be part of a modern pathology department.
This will drive clinical volumes, this will drive education, this will drive research. And we would be, for the first time, be able to engage at a very global level with pathologists around the [00:21:00] world. Looking at the same cases, sharing ideas, learning from each other, which would not be possible on glass slides, right?
When I was a resident, we had to go to a conference and somebody would take the Kodachrome slides with them and present this unusual case and today on Twitter, I can see it within seconds or on your website, on your post, this world has changed tremendously and, like I come from a world where when I was a resident, I only had one hour lecture on how to turn on a computer and turn it off.
I’m in a world now where I am personally learning how to code, how to do data visualization, how do I, I can actually write code now, which I.
Aleks: What language are you coding?
Anil: I’m looking at, I’ve been learning R, I’ve been learning Python. In my college, I learned some basic Fortran, Pascal, but now I’m learning the current languages.
But the point is, if you don’t think [00:22:00] about it, you will not know about it. If you don’t know about it, you’re not going to use it, you’re not going to implement it. So the perfect way to move forward is to understand the limitations of the technology, understand the benefits of the technology.
Aleks: That brings me to, to another question, because you mentioned at the beginning, you were heavily involved in the create, in the creation of a software solution for digital pathology.
Why did you choose to stay in academia rather than join the industry? Or how are you involved in the industry because you’re super involved everywhere. So tell me the story about that.
Anil: Yeah. So I had an opportunity, I have had several opportunities to join industry, be part of an industry which is growing incrementally, but I love to teach.
I love to interact with residents and fellows. I love to, to take an idea and implement. And if I was an industry, I see that barrier between the industry and academia is always there. It’s [00:23:00] getting better now. It’s clearly defined, but I saw that as a barrier to, I thought I was, if I was on this side of the fence, I would be more impactful.
If I joined the industry, I would probably work for a company. I would maybe become a chief medical officer for a company and advise that company and maybe make it successful. But I would be only focusing on one product for one company and the impact will be limited. Personally, I would probably gain a lot more financially working for the industry.
But if I’m on this side of the fence, I can still work with companies with a very well defined industry, academic engagement, disclose it and work in the confines of those boundaries. But impact it more broadly, like I can now compare three products, software products, which are making an algorithm for prostate cancer detection in the confines of a research environment.[00:24:00]
That’s the reason why I chose to, I love to teach, I love to interact with that. That was the reason why I stayed in the academic side.
Aleks: So a different question, do you use AI in a different than a research setting? Do you use it for your daily practice, pathology practice? And what AI are you using if you are?
Anil: Yeah, so I personally. I’m a GU pathologist, so the only AI that I’ve been using and testing is mostly right now prostate cancer. We have used AI, we’ve built algorithms at our University of Florida Cancer, and I’ve also worked with collaborators at different institutes to recognition and classification.
In the clinical setting, we are mostly using AI for ordinary tasks like detecting lymph nodes, cancer mats and lymph nodes, or finding H pylori in gastric biopsies, or quantitating [00:25:00] ER, VR, and hertoneum. So those are the things that we have implemented at our institute. Because I’m a chief pathologist, I don’t use many of those in my daily practice, but I’ve been involved in Implementing them, helping working with our other pathologists and colleagues to bring those solutions.
What I’m most excited about the possibilities of using broader algorithms, not just a disease specific algorithms. So I’ve been working with colleagues at Mayo Clinic and other places to build algorithms, which would classify an unknown image into a known disease or a known cancer. So if you’re almost like a Google approach where I Take an image, or the computer takes an image of a region of interest, feeds it into an algorithm, and gives you the differential diagnosis.
That’s how we learned as pathologists. Like, when I was a resident, I would look at a tumor under the microscope. I didn’t know what it was. I would start looking for pages and it was very cumbersome, took a long [00:26:00] time and I was almost always wrong. But today we have technologies which I can just press a button and compare that image and search for similar images in a database.
So the answer to your question is I don’t use it a whole lot today because I still feel with. my expertise in prostate pathology. I can do better than the algorithm in many ways, but I also like the fact that it can bring more efficiency into my work. Look at my, we talked about my office. I have three monitors and I have One, which I use for reporting, one used to for all the images that I review.
And the third one is for all these conversations that we have about digital pathology and emails and work. So many pathologists at our institutes are now using AI algorithms. And they’re using it with either solutions that they purchased, we purchased off the shelf or something we’ve been building here at OSU.
But I would say we are not to the point where we [00:27:00] are using AI routinely. I would say it will take us about a year or two to get there and to incorporate it fully into our workflow and then it’s not something that’s under our control. It’s more under the control of the interfaces and the user experience and the consoles or the workstations that we’re looking at.
All the vendors who build scanners, they’re acquiring images. All the vendors who build AI solutions, they’re building great solutions. But they’re not talking to each other every day and that’s where you come in. You need to bring them together.
Aleks: I need to do that. Anil, before we finish, if there was not like, let’s do one, one piece of advice for somebody who wants to start in digital pathology, one for an individual and one for institution what would that be?
Anil: I think the biggest thing to do is think about your workflow today. What are you doing today? How can digital pathology help you? You might be a dermatopathologist and you are [00:28:00] majority of your cases are basal cell carcinoma. And you might want to buy a solution, which is not dependent on high throughput scanning or buying a particular AI, but focusing on that one problem, which will save 20 percent of your time every day.
So my rule is 20%. Think about a solution which will help you save 20 percent of your time. And it could be scanning, it could be remote sign out, it could be doing frozens remotely, where you have to travel somewhere for 20 minutes. Focus on one thing, focus on the low hanging fruit, focus on the biggest impact that 1 that you have to spend will have.
And I see large digital pathology projects across many institutes and they’re focused on going live on July 1 with everything and they get delayed they get delayed and delayed and that impedes the impact. The momentum is lost. So pick one thing. Which [00:29:00] you are most excited about as an individual.
Convey it to your institute. Buy people’s buy in. Get them to be excited about the same one thing. It could be I want to review all my amino stains remotely. I need to buy this scanner. It will take me this time. Make it a finite problem. Make it a smart goal. Make it something which is achievable, measurable in real time, and it produces a result.
Set expectations, which are realistic and I see that why digital pathology projects fail, it’s because we try to overachieve. Your goal should be underachieved. Your goal should not be, I’m going to be this institute who is completely digital. Your goal should be, I’m going to be this institute, this hospital, this lab, which will solve this problem and digital pathology will help me get there.
And then you can incrementally build on it. We are the perfect example. We didn’t start with 100 percent sign out. We didn’t start with the goal to, achieve this in two years. Our goal was to incrementally build a solution, which [00:30:00] people are excited about, because if people are not excited, if I’m going to just say July one, everybody’s digital, you’re my way or highway nobody will buy it. Everybody will. Leave Ohio State and go somewhere else.
Aleks: I love the advice that it should be problem like solution oriented rather than some achievement based. I think often, like you say, we try to overachieve and fail and then the perception is, oh, the whole digital pathology initiative fails.
And maybe it was just, one problem or 50 percent of whatever you set out to achieve was achieved. But from the outside, from people who are not involved in this, and often these are decision makers, this looks like a failure and they don’t want to embark on this journey anymore. So framing it as, okay, we’re solving this one problem that’s going to give us 20 percent more of our time, revenue, or whatever.
It’s a very sustainable way. Gives people the feeling of accomplishment and the motivation [00:31:00] to try the next 20%, the next 20%. So thank you for that and thank you so much for joining me today. It was an honor and a pleasure to have that digital pathologist on the podcast and so if somebody wants to ask you questions, what’s the best way to contact you?
Anil: I love. Emails. So if you want to send me an email, I’m very good and responding emails and be persistent. So if I don’t answer the first email, follow it by a second email, and I promise you I will answer.
Aleks: Okay, so rather email than social media.
Anil: Social media works well. LinkedIn is my preferred.
Aleks: Okay.
Anil: Social media, but I can go, I haven’t joined threads yet, but maybe we’ll start a threads email. You probably have already started it.
Aleks: I started, yes. And TikTok. So if somebody wants to find me there, you can and I’m definitely going to link to your LinkedIn profile and whoever wants to start a conversation can reach out there. Thank you so much and have a great day.
Anil: [00:32:00] Thank you so much as well. Thank you.
