In our fifth Ask a Pathologist Podcast episode, we had the privilege of interviewing Dr. Hillel Kahane about his favorite game-changing tools for pathologists. From pre-analytical tools to digital pathology to diagnostic AI, Dr. Kahane shares what tools have been the most useful in his practice and why he recommends all pathologists give digital pathology a try.
Dr. Hillel Kahane is a board-certified uropathologist with over 30 years of experience in the commercial USA laboratory industry. He has personally signed out more than 1 million prostate core needle biopsies throughout the United States, serving over 6,000 urologists and radiation oncologists. He is also the co-founder of StarPath Inc., a physician-owned and physician-operated digital pathology company that operates under PathNet, a global network of next-generation pathology practices. StarPath provides world-class quality pathology interpretations while leveraging the power of artificial intelligence coupled with the 50+ years of experience and expertise of its two founders.
Listen here:
Watch on YouTube Here: https://youtu.be/Dd7BM_kBjyQ
Ask a Pathologist Podcast Episode 5 Transcript
James Thackeray: Well, welcome to Ask a Pathologist Podcast sponsored by Lumea and the Digital Diagnostic Summit.
The idea behind this podcast is to provide good, helpful, useful information, and resources for both current pathologists and those in training. I am James Thackeray and I sit on the board of the Digital Diagnostic Summit and the Chief Commercial Officer for Lumea. And I want to give a very warm welcome to our guest, Dr. Hillel Kahane, who I’ve had the privilege of knowing for some years now, but I want to give him the opportunity to introduce himself. So Halal, why don’t you give us some of your background and introduce yourself for us?
Hillel Kahane: Sure, I’m Hillel Kahane. I’m a uropathologist. I’ve been practicing uropathology now since 1992. I’ve got about 32 years of experience doing this. I’ve always been in the commercial sector and I’ve only been looking at urological specimens, primarily prostate needle biopsies. I’ve worked with both Dr. Epstein at Johns Hopkins and Dr. David Bostwick at Mayo Clinic over those 30 years and I’ve probably looked at an excess of a million cases in the last 30 years.
James Thackeray: So that’s crazy. I mean that number is staggering. That type of experience is, I mean, there were probably very few in the country that could make a statement like that, right? I mean, that’s how many cases you’ve seen.
Hillel Kahane: Well, I mean, I can say I only subspecialized in prostate needle biopsies and I’ve worked with and diagnosed cases for probably an excess of 6,000 urologists and radiation oncologists over all these years. But I’d say primarily in the commercial space, not in the academic world.
James Thackeray: That’s great. I remember I knew your name long before we ever met because I was also in the commercial space, but Dr. Kahane, your name had a lot of clout behind it because you had such vast experience in this world. So it’s always been a privilege for me to get to know you over these last few years. Well, let’s jump into this. You know, our format is a little different. We’re just a smaller or I should say a shorter podcast that we think would be interesting content to other pathologists and those that are going into the field.
Today’s topic is tools for pathologists and we thought we’d start with that.
Maybe we kind of navigate from there a little bit, but as you just start to think about different tools that you’ve used in the past and so maybe we start there. Any tools that you’ve used in the past that have helped you in your practice of pathology and then we can maybe go towards tools that are now becoming available and how that factors into your workflow. We’d love to hear.
Experience is Your Primary Tool
Hillel Kahane: Well, you know, the primary tool essentially is experience, to be honest with you. The more you do, the better you get at it. So when you’re looking at a biopsy or two a week, you’re not going to get very good very quickly. And when you’re looking at, you know, 30, 40, 50 cases, whatever the number is, in a week or month, you pick up the nuances extremely fast. And so like I said, working at Dianon Systems, my first job, which is primarily a EuroPath lab in the early 90s, all the way up to about 2007, all I did was prostate needle biopsy.
So you got real good at it. And you, just from, you know, eight hours a day, it’s all you looked at any difficult cases we had at the time, we sent off to Johns Hopkins to John Epstein. And he would kind of fine tune our accuracy, let’s say. And so we learned a lot from him. He used to come up every quarter and spend a day or two with us at Dianon and kind of teach us all the little nuances and little pitfalls not to make and not to overcall, undercall, proper Gleason grading, all the nuances that go into it. So basically, in a nutshell, experience, that’s kind of what you got. You know, you can’t really read this in a book. You’ve just got to sign out cases, essentially.
James Thackeray: I love that. I love that. And I was going to ask, but you answered the question already: how, you know, as you get more and more experience, how do you standardize that? And that’s pretty informative. The fact that you used, you know, one of the world-renowned GU pathologists to help kind of standardize that process throughout your experience and fine tune it as I think you said. So that’s a great answer. That’s very insightful.
As you start to look towards the future, and I say look towards the future, but knowing you, you’re very much a part of new developments that are coming out. What type of tools do you foresee being helpful in the practice of pathology? Maybe helpful is an understatement, more of game changer tools that are actively coming out that you’ve been a part of that will change kind of the way pathology is practiced.
Hillel Kahane: We’re talking about prostate needle biopsies, correct?
James Thackeray: Correct.
Helpful Pre-Analytical Tools for Uropathologists
Hillel Kahane: Yes. Well, I mean, one of the biggest game changers for me in my career was when I met you and Dr. Matt Leavitt and you guys introduced me to the BxBoard™. That just blew me away, to be honest to you. That was about 2018 maybe, whatever the year was.
And I was like so impressed by this contraption, mechanism, whatever you want to call it, innovation, that I realized, you know, very quickly, this can be a game changer for the way you sign out prostate needle biopsies because you only have to look at one slide with six cores on it, as opposed to six slides with one core. I mean, I thought I was clever at Boston laboratories as the medical director by inking the cores in three different colors and putting three cores per slide. So I look at four, four slides, because four times three is 12, and repeating the colors.
That was my huge innovation until I ran into Dr. Leavitt, who showed me he can one up that, actually double that, and showed me the BxChip™ and it’s oriented too. That also was extremely impressive to see how the cores are oriented. So we know where the proximal end is, we know where the distal end is. Therefore, when we diagnose a tumor the urologist has a sense of where along the core the tumor lies. That’s also a game changer, to be honest with you.
Nobody else in the country does that reporting, they just tell you there’s tumor and they tell you the Gleason score and how much tumor, but we tell you where the tumor is in addition to that. And not only by length, but also by area, which is another amazing Lumea innovation to be honest with you.
James Thackeray: Well that’s a great, great point though. So let me ask you this, as a pathologist, though, so you talked about the BxChip and the ability to put all those cores on one slide, but how did that, that’s had to have been and even though you were doing more cores on a slide than maybe traditionally was done in your days and you were figuring out ways to ink it and do those things, how did that change from your perspective? Because that’s a different way of, I mean, if you’re looking at one core per slide and now all of a sudden you’ve got all these different cores.
Just from a navigational perspective, how did that change the way you read your cases?
Hillel Kahane: Well, it changes the speed immensely. Now you can become a hell of a lot more productive, to be honest with you, because now you can do the six cores, not like, you know, ultra lightning fast, but it’s a lot easier.
And the other key thing that I didn’t mention was the cores are all linear, which is also a game changer because many times you look at cores, we call it the spaghetti strand effect.
They’re all kind of rolling on top of each other, the cores, and there is a possibility, to be honest with you, where tumor is masked by a benign piece of prostate that kind of lies on top of it. And it’s not that you misdiagnose, it just wasn’t, the technology doesn’t let you see beyond that core. It’s kind of like, almost like a curtain effect, essentially, it’s covering up the tumor.
But with a BxChip, the cores are laid out horizontally, or vertically depending how you look at it. But you’re not going to miss anything. Let’s put it that way. And I’ve noticed that, to be honest with you, because I keep close track of my malignant detection rate with Lumea. And I’ve noticed it’s very high, to be honest with you, which is kind of a good thing, in a way, because then you know you’re not missing stuff. Because I’m like in the mid to high 60% rate on finding cancer on everyday cases, and we do lots of them on a daily basis.
I was never this high prior to being part of this Lumea and PathNet thing. So there’s a huge advantage to using the BxChip.
Artificial Intelligence for Pathology
James Thackeray: Got it. No, that’s, that’s, that’s really helpful. So let me ask you this because I know that you’ve been involved in a lot of – as these new AI algorithms come out from a diagnostic perspective, I think because of your vast experience, a lot of these different organizations that are developing new AI have come to you to help them standardize their AI products and to create, you know, who’s the ground truth for a lot of these new AI products?
Just in general, why don’t you just tell us: how do you feel these algorithms are doing and the role that they’ll play going into the future?
Even, you know, maybe give us a status update of where they may be now and how you see them evolving into the future.
Hillel Kahane: Oh yeah, well right now they’re very good. They’re very robust algorithms. They detect the tumor almost every time, but not 100% yet. That’s why we’re constantly evaluating different algorithms so we don’t choose one over the other right now. We’re just test driving, so to speak, multiple company algorithms. They all work very similarly. And the beauty of it is, it really is almost like another game changer because it’s almost like providing the pathologist with a life, like a net underneath him, like you’re on a trapeze and you fall off, you’re not gonna die, so to speak.
And so the algorithms are extremely helpful, both for detecting the cancer, but they also serve its weight in gold when cases are benign.
And the algorithm tells you it’s benign, it takes a while to get used to the fact that, not that you’re not gonna look at these cases, but you have this feeling that the algorithm, it’s another set of eyes, essentially.
And so you’re getting, like, almost like two pathologists, well-trained expert pathologists, for the price of one. The algorithm immediately, you know, kind of highlights the tumor, even breaks it down into Gleason grades between three, four, and fives. In some, many of these algorithms detect perineural invasion. So there’s lots of things that these algorithms do.
We currently use it as a QA/QC tool. So after we’ve annotated our tumors, if there are any, or if it’s benign, then we turn the algorithm on and see how smart or how non-smart we were, so to speak.
And it’s a game changer because eventually when we get to the point where the algorithms are extremely accurate, it’s almost like an autopilot on a jetliner, essentially. It’ll basically fly you from A to B like it does today, but it’s like the analogy is even though the autopilot can actually take off and land the plane, I don’t think anyone’s gonna get on a plane when there’s nobody sitting in the pilot seat. So the same thing here, you need a pathologist, obviously, in case – because sometimes the algorithm does go askew and it’ll either overcall tumor or mistumor or tell you benign areas are tumor. So you always need the pathologist to kind of hit the kill switch, so to speak, in case something goes astray. But we’re getting very close to being almost 100% accurate. And that’s the goal here, of course, is you cannot be 95 or 90% accurate.
When it comes to prostate or any biopsies in this world, you can’t say, well, we’re right 95% of the time. That five percent’s a huge number, by the way, based on the number of prostate biopsies done around the world and around the U.S., so this thing has to be like a zero defect kind of a product, essentially. And these algorithms are getting very close to that. So it’ll really assist us as pathologists in the future. And I see one day that the algorithm becomes the report, which will be kind of another game changer.
James Thackeray: Yeah, and I love that qualifier. So you talked about the way you use it currently as more of a QC/QA check on the back end. And I like the thought of this second set of eyes. And as these algorithms get better and better, and to your point, they need to be 100% right. I mean, you know, I think that’s the process of the regulatory agencies that help us figure out when they can be used as primary diagnosis and that may be who knows how far away, but the use of algorithms as a QC check on the back end seems like a great way to at least start with these already fairly good algorithms.
So I think that’s great.
Hillel Kahane: I mean, I saw a study a couple of years ago out of France where they took a bunch of benign prostate biopsies that were a few years old, let’s say, and they ran the algorithm on them. And between 9% and 12% of the cases were not benign. That’s frightening, to be honest with you. The algorithm detected high-grade pains, atypical foci, and full-blown cancers in many cases. And that’s really, I think, irresponsible, you know, not doing any patient justice here, obviously, by missing cases. Not that they went out to miss there, I’m not blaming anyone, but that’s kind of the state of the art at the time. I don’t know who read these cases, what their training was, blah, blah, blah, but the fact that the algorithm’s able to pick up that big of a percentage of, quote unquote, “missed diagnoses”, it shows that the algorithm is really ready for prime time. And can save patients’ lives, obviously.
Would Dr. Kahane Still Choose Pathology if He Could Go Back in Time?
James Thackeray: Yeah, yeah, no, that’s a great point. So here’s maybe a little bit of a different question, and we can maybe end on this question. So if you were looking back, if you go back in time and you’re picking out your specialty right now in medical school, what do you think you would do? Would you keep with pathology, knowing all the advancements? I mean, would you embrace that? Would you look at it as an opportunity or do you think you would look at a different specialty based off of maybe the advancements?
I’d love to hear your feedback.
Hillel Kahane: I always liked pathology as a medical student. I just enjoyed looking in a microscope and so every case is a challenge to try to challenge you intellectually. What are you looking at? What are the processes, the disease processes? Is this malignant? Is this benign? I personally like that. The majority of medical students don’t go into pathology, just so you know that. So it’s not like a high in demand kind of a specialty. It’s actually towards the bottom of the list of specialties throughout the training. However, there are people like me who enjoy it and I enjoy working at a microscope, now on a digital tablet. I got rid of my microscope probably three, four years ago.
Working with an iPad Pro has some huge advantages that we didn’t really get into. The fact that I could sign out literally from anywhere in the world, as long as I have a digital connection.
The fact that I can consult my colleagues or they can consult me in real time is huge when it comes to digital pathology. I’ll annotate something or somebody will annotate something on the BxChip. They’ll send me a text or an email, say, what do you think? Is this benign, malignant? And they get an answer back literally in minutes. I remember the old days, we used to send the slides, package them up and ship them to Baltimore and wait a week or two or longer, depending on John Epstein’s availability or whoever we were sending it to. And this is a common occurrence. And now the patient care is just so, so streamlined today and the second opinion comes back literally in real time.
Digital Pathology for Second Opinions
James Thackeray: Yeah, that’s such a valid point. We haven’t gotten into the digital side as much. We talked about AI, but the ability to improve clinical care through second opinions and consults, even peer review consults, or just your colleagues that you work with, and I love that. So it’s basically in your workflow, it’s as easy as just giving them access to that case, that de-identified information, but giving them access to the case, and then their availability to jump on wherever they may have a good reception, right?
Hillel Kahane: Well, my colleagues at PathNet already have access to the case.
Right, okay.
Hillel Kahane: I mean, we just say, hey, what do you think of case 1234? Look at core number C. You think that’s a Gleason six? You think it’s been, whatever. And then within minutes, literally, you’ll get an answer back.
And the fact that it’s been annotated, they can see exactly what you’re talking about.
That’s the beauty of it.
I’ll annotate something on the iPad Pro and my colleague doesn’t have to waste their time looking around. They know exactly where, or we can actually just jump on a Zoom call like we are now and we can actually look at the case together. That’s, like, extremely valuable. It’s almost like a virtual multi-head scope.
James Thackeray: Does it make you more apt to bringing your colleagues in because it’s so convenient? I mean, you think of the alternative of having the lab package up the slides and ship them to wherever. I mean, how often are you doing these types of reviews with your colleagues?
Hillel Kahane: Probably once a week. Well, it’s not like a scheduled thing. It’s like when you come across a difficult or challenging or interesting case and you want to share it with a colleague, because that’s how you learn from each other, and you immediately just send out an email saying, guys, take a look at this case, at this part. What do you think? You can invite as many people as you want or just a single person or you can invite everybody to look at the case. It’s that just, you know, that’s how pathologists learn from each other. When you’re in a hospital, it’s usually by the water cooler, you can kind of share and go into the room with a multi-head microscope and show each other cases casually. But here you can do it from the benefit of your backyard.
I mean, literally anywhere you are, you can share cases with colleagues. And that’s invaluable, to be honest with you.
What Would Dr. Kahane Say to Other Pathologists About Digital Pathology?
James Thackeray: I love it. Okay, this is truly my last question because you and I could speak about this forever, I think. So, you’re one of the more experienced GU pathologists that I’ve certainly ever met. I think you’re one of the more experienced ones in the country. But your adaptation to digital pathology, because you adapted later in your career.
For those that are maybe later in their careers and are looking at digital pathology and the workflow for the first time, I guess what input or feedback would you give that transition from going from the microscope to whatever device that you’re using to look at it? How was that for you? Was that a difficult transition or what was that like?
Hillel Kahane: No, for me personally it was very simple. At first you’re kind of hesitant because is the clarity, the quality, the image equal to what you see under the glass slide? And it is with the current system I’ve got with Lumea and PathNet and so forth.
It’s just a matter of taking the leap. I would encourage people, if you have the interest, to take the leap, jump into it, see what it’s like. I mean, you can always go back to the microscope if you really hate it, to be honest with you. Some people probably have, but there’s so much upside potential to going digital, that it’d be foolish not to at least give it a shot. Try it, because the world is moving forward, not backwards.
And AI, you can’t do AI on a glass slide, obviously. There’s no digital image for that. So that’s useless to you.
I mean, if you want to be part of the 21st, 22nd century you’ve got to go digital because there’s so much information on these slides that we don’t even fathom today, to be honest with you. It’ll be available down the road, somebody will come up with something else, say, oh, it’s all digitized.
So all these images stored up on the cloud there, there’s so much information we haven’t even tapped into yet, to be honest with you.
James Thackeray: Yeah, that’s a great point. Well, this has been great, Dr. Kahane. It’s been such a privilege to have you on our modest little podcast here and get your input. We’re really appreciative of it. And yeah, look forward to speaking to you at another time.
Hillel Kahane: Sure. Anytime. Appreciate it.
