Deep Dive into Gene Expression Profiling

[music playing] Welcome to Derms on Drugs. A video podcast brought to you by scholars and medicine, the best educational platform in dermatology, and totally free. Derms on Drugs is where cutting edge derm meets-- eh, meh, dermis comedy. I'm Matt Zyres, each week. I'm joined by my residency buddies, Dr. Laura Ferris and Dr. Tim Batten. And we use our 60 years of combined derm experience to discuss, debate, and dissect the hottest topics in dermatology. It is everything you need to know to be on the cutting edge of derm. And it'll be the most fun you've ever had while really learning something useful. New episodes drop every Friday on scholars and medicine, Apple Podcasts, Spotify, and any other weird place where you might get your podcasts. So let's go ahead and get into it. This week we've got a fantastic episode with a special focus on gene expression profiling in the management of skin cancer. So we're going to start off with three recent articles on this topic. And then we've got a phenomenal guest, my favorite derm path from the podium, Dr. Whitney High, from the University of Colorado, who really has some deep expertise in this gene expression profiling and how he thinks it ought to be used. But first, let's go ahead and get into our articles. Dr. Ferris, why don't you start us off? Sounds good. So I'm going to start with actually a recent consensus statement that was published by the Society of Surgical Oncology, assessing the evidence for in utility of gene expression profiling of primary ketaneous melanoma. This is Bartlett at all in the annals of surgical oncology. So this is a panel of melanoma experts, about 20 voting members who were chosen by the SSO. And they were convened to develop recommendations for using GEP. So they really looked at three different tests. So decision DX, which is the Castle Biosciences 1, 31 GEP, Melogenics, which is an 11 GEP, and then Merlin, which is CPGEP. OK, so they-- Were these were any of them dermatologists or were they all real doctors? I would argue dermatologists are real doctors, but these were surgical oncologists. I think maybe one consideration is that this is the Society of Surgical Oncologists. So these are the people who, I would say, have a dog in the fight of sentinel node biopsy, versus dermatologist. So I'll put that out there. OK, let's fair. So they wanted to look at this at GEP in terms of three different things. Should a guide sentinel node biopsy? Should a guide surveillance? And then should a guide adjuvant therapy? OK, so they were not looking at it. Because I think the perspective we often look at it as germs is counseling the patient. Like, can we give you more prognostic information? They were looking at it in a more practical, like, should we use it in making these decisions? Correct. OK. All right. OK, so basically, like spoiler alert, they said we shouldn't really use it for any of these things. This is really more of an experimental use in a research setting. But I think some sort of key information to pull out of that. So one, they did not think it should be used across for sentinel node biopsy across any thickness category from T1A to T4B. Area is where they said, well, maybe there's something promising. So they talked about CPGEP. So just this is Merlin. So this is the test that uses CP's clinical pathologic and gene expression profiling. And so they said, well, you said T2 melanomas may be promising. Negative predictive value was pretty good, 89 to 97%. In cohorts where sentinel node positivity actually exceeded 20%. So it's easy to have a negative predictive value if basically nobody has a positive sentinel node. So they did point that out. And they said, according to those data, if you used it for T2 melanoma, you could potentially reduce sentinel node biopsy by about 30%. They talked about 31 GEP. And so there's a study that I think is by VETA with all that is cited a lot saying, if you have a Class I A lesion and a patient who's over 65, this had a good negative predictive value. So you could maybe avoid sentinel node in that population if their risk of being positive was below 5%. But what they pointed out was that actually in the cohort that they looked at the sentinel node positivity rate regardless of 31 GEP result was actually less than 5%. And then they looked at another study which was the integrated I31 GEP. So this is sort of like the Castle Biosciences version of using clinical pathologic features, right? So not just what does the gene profile say. It's a continuous result. It's not broken into two or four categories. And it sort of gives risk. And so they noted that the positive, that the sensitivity for T1A, so high risk T1A patients, they said the sensitivity was low, 43%, positive predictive value was less than 5%. They did acknowledge that it was a little better in patients who had T1B within a sensitivity of 83% and a negative predictive value of 98%. And that that could potentially reduce procedures by 41%. And that it also the performance sensitivity and specificity negative predictive value were both 96% in patients with T2A melanoma. So now we've kind of got two different tests that they're saying maybe T2 lesions might have some value. But they said it's not been regulated. It's not been replicated in external validation cohort and also notably 25% of the validation cohort in this Castle Biosciences study did not actually undergo sentinel node biopsy, but they did never occur. And so they were presumed to be sentinel node negative. They did note that there's an ongoing study for a prospective study for CPGEP for sentinel node. Surveillance planning. Yeah. Yes. OK. So they said not recommended. And so they kind of specifically noted, don't use GEP in patients who have T1A melanomas. They already have an excellent prognosis. They do not recommend it. They did recommend, again, another study that's talked about a lot, which was a SEER database study where they just looked at patients who had a 31 GEP result. Not was it high risk, was it low risk, but did they have a result? And then they did, you know, it was retrospective propensity score matching. So they found, you know, similar people-- so here's your group and SEER who had the test. Here's a group, a similar group that was matched on a bunch of other criteria. And, you know, they said, if you had the GEP test, you were-- you're overall survival, a melanoma specific survival, were better than the group who didn't have the test. So just having the test made your survival better. So, you know, you could debate what does that mean? It was not prospective. So they pointed out that, you know, while they did do matching, there were some things like socioeconomic status, recede of adjuvant therapy and B-Rough status, and recede of treatment for recurrence that really weren't adjusted for. So that was discussed in the paper. And then for adjuvant therapy, you know, they said, there's really not data where we can say, this is going to define the patient who is going to benefit from adjuvant therapy. So you've got fully-resetced stage to be, to see melanoma, you know, probably the big population that you'd be looking at to say, you know, who should get adjuvant PD1 inhibitors? And there's really no data. They did point out that there is a study that's ongoing called NIVO-MELA. So this is a randomized study where patients with stage two disease who had surgery, they are basically being screened with the 11GEP assay. And then the high-risk group is going to get treated with adjuvant nivolumab versus placebo. So they'll pull out that high-risk patient group using the 11GEP. And then they'll randomize them to NIVO or placebo. So this is, you know, they point out a promising, like, prospective study, but we won't really have results until 2028. So, you know, they're cost cost of a, what's that? That's any stage two, not like 2B2C. I believe that it is going to be 2B2, I believe it's 2B2C. I think because that's the group that would be eligible for adjuvant nivolumab. Right. So you went through a lot of stuff. What is the takeaway here? Like, does it? Because it-- The takeaway is they think this is an investigational tool and what we really-- that what should be done are prospective randomized clinical trials. So, you know, prospective trials, like, you get the test result, and then you randomize them to get Sentinel node or not. Prospective, like you get the test result, and then you randomize them to get neballium app or not, or imaging or not, that really what they've primarily done has been, you know, sort of retrospective. That's the takeaway from this. - So, two, well, okay, I've got more to ask about this, but I think it'll be better whenever we get into it with Dr. High, because right, they were answering slightly different questions than what I think is most relevant to me as a Durham, because what I want to be able to say to what patient is, hey, there's, you know, this, and then that was negative, so now you've got a 90, now you had a 95% chance, now you've got a 97% chance, or 99% chance, which is meaningful, you know, or you went from 95 to 90, like. - Is that meaningful? I mean, I think that's what I'm looking for, to what the discussion, can you? - Yeah, that is not meaningful. - Yeah. - I don't see it as to me, huh? See, maybe this is one of the bad game. - You're doing overall, you're taking group prognosis and trying to tell that to one individual person, that's not how that works. - It's not how it works, but it's how it intuitively makes sense, and it's what I would want to know, is if you had a thousand people who had these results, you know, how many people did it come back in? Right, does that test say it's gonna come back in me, or not know, but it's a gamble, everything in medicine's a gamble. - But you're taking what's generally a range in your, for a population, and you are narrowing it to an individual. - Or. - The question is, I'm not saying you will not recur, and you will. Knowing that you have a, you know, five versus three percent chance of recurrence, does that change anything, or a five versus eight? - To me it does. Now this again, maybe is one of the bad games. - What was the last for you then? - Right, I lose money all the time. That's why I don't bet more than like five bucks, 'cause I know that I'm gonna lose 99.9% of them. See, if I was gonna lose 99.7% of them, I wouldn't bet as much. - Okay. - See, that's, all right, let's move on to our next, as Dr. Patton shakes his head at me. All right, let's move on to Dr. Patton's article. What do you got, Patton? - All right, so mine was the 31GEP parentheses decision diagnosis. A review of clinical utility and performance in a Mayo Clinic cohort, cohort by Pasava at all in the March 2025 issue of international journal of dormitology. It was a retrospective review of 65 patients diagnosed with melanoma from 2014 to 2022, a Mayo Clinic who also underwent 31GEP testing, specifically the decision diagnosis test marketed by Castle. So a small study, but I figured, you know, you get more granular data, that's C or data that Dr. Ferris mentioned. That was such a weird trial. They're like, we don't know how this test, order the tests in patients who better, we don't know how, but you probably just order it. I actually ordered it on myself, right? You can never be too sick. I feel better. So table one is a lot of the raw data, surprising finding of a higher percentage of class one patients, having a positive sentinel lymph node compared to class two patients. Now, when you look at that data, that is, they don't, they don't explicitly say this, but that is probably because they found class one tumors in legions that were T3BT4A, you know, like, why would you eat it? You know, those were probably sentinel lymph nodes positive. Now in practice, would you order Castle on those lesions? I probably wouldn't. So, but I think this kind of gets to like the problem I have with Castle in general. So that stood out to me. You go to table two, a higher percentage of class one patients underwent adjuvant therapy. It was a big difference, 11.1% of class one patients underwent adjuvant therapy compared to just 3.45 of class two. I think that's because of those, those T3, A, T4 lesions that happen to be class one, that you got the castle test and completely ignored it. So why did you get it on the first place? Figures 1A, 1B show how little to 31GEP test testing influenced management decisions, figure 1A, show that a decision on whether a perform set of lymph node was made before 31GEP results were even available for about 63% of patients. Like, why are you ordering the test? If you're making clinical decisions before the results are even back, why are you ordering that test? Figure 1B shows how close to 82% of patients with 31GEP testing no change in the standard of care was made. Figure 1C shows survival curves of relapse for your survival, melanoma specific survival overlap between class one and two patients. So I think this article sort of demonstrates how I feel about castle. It is way over utilized, way over utilized. What have you got to know that? How do you know how you do that? It is. It is. Listen, you could certainly say that the providers are misusing the test, but I think that the misuse of the test is strongly encouraged by castle. You guys have gone to the dinners and they're like, have every melanoma over 0.3 should be tested. Like, it's ridiculous. Now, NCCN guidelines specifically say similar to the surgical guidelines do not use this test in the everyday management of patients. This is experimental. It can be done in clinical trials and probably should be done in clinical trials. So we know more about it. But outside of that setting, do not order it. Though that's NCCN guidelines, the melanoma specialist, Dr. Ferris was part of it. You are the-- I was part of the NCCN. I was part of the melanoma prevention working group. Yeah, I'm saying like separate, separate thought. Right. Who are you going to believe? NCCN or me. Should I say it's you, Ty? No, I say it's me being you. I'm going to ask you a question. As of September 30th, 2024, how many times-- how many castle decision DX melanoma tests do you think have been ordered? 22,000. 183,000. What? There you go. There haven't been that many melanomas. They're bordering it on moles. I think WIT-High is probably unscrupulously using this. Well, it could be a melanoma. Better send it for castle. Which-- What's crazy is like the NCCN guidelines, Dr. Ferris was part of a group that they published, I think, in Jamaderm. They were as stringent against it. But they're like, yeah, we don't really know about this. Now this surgical group, people are ordering this test like crazy because they go to fancy state dinners and-- Willemar Nilly. They're ordering it willemily. I don't know that I think that it's because they go to fancy state dinners. I think the idea-- Oh, I totally disagree. I think that castle has done a very, very savagy marketing campaign where they published like that C or data, eight of the 12 authors were employed by castle. Like they're smart. They're really smart doctors. They know how to do this data where it probably favors the test. And they have bombarded the literature with paper after-- The surgical oncology group said that. They actually have a graph. When you look at it, you're presented of the authors. That either work directly for castle or who have clear conflict of interest. It is really, really high. This is one of the most egregious-- now, the thing is, is I do think castle-- I think castle is appropriate. I think those T1B that are castle 1A, you could say, let's not do a set on node. There was that decide study. I think that stuff's good. I think there's a utility for castle. But it just gets so-- it's so egregious to be overutilized because of savvy marketing. Let's say that we're discussing, because I want to have a higher rate. All right, so first, let me just give you-- and I'm so, man, the universe just helps you out sometimes. So today, like two hours before the podcast, I'm looking through all the articles that published today. And there were two articles in the journal of the National Cancer Center Network specifically about GEP. So the first one-- and this was just an ad-- they were abstracts. There's no full text yet. But it was-- the abstracts had the-- the job of just having the information I wanted, right? Metastasis-free survival-- first one, metastasis-free survival prediction with the 40G expression profile test in patients with container squamous cell carcinoma risk stratified according to the National Comprehensive Cancer Center guidelines. So basically, they say that it's unknown if the 40GEP, which I think is the castle squamous cell carcinoma one, if it had value beyond the Brigham Women's Hospital staging. And the Brigham Women's Hospital staging is a very simple staging, like how big is the tumor, how deep is the tumor, like stuff like that? It's like four or five things. Then if you have any of them, you're high risk. So they-- primary container squamous cell carcinoma with people who had at least one high risk factor based on a previous study looking at the-- the GEP combined with a retrospective cohort from two academic centers. And patients who received adjuvant radiation therapy were excluded. So then 1,412 patients in the analysis by NCCN risk stratification. So NCCN has their own risk stratification. And just by the way, as for everybody out there listening, if you're not familiar with NCCN, you absolutely want to go to their website, register, and it gives you access to all of these really well-done sort of protocols for how to manage pretty much any cancer you can possibly imagine. And so the NCCN, Scrimmousel Carcinoma thing, is pretty similar to the Brigham women's. It's things like location diameter, how well defined are the borders? Is it primary recurrent? Are they immunosuppressed? It's been radiation therapy in the past. Is it rapidly growing? Were there neurologic symptoms? Was it poorly differentiated? What's, is it one of those acantholytic deals? Is it go to the fat? Does it go deeper than the fat? Is it parry, neuro? So things like that. So they looked at the NCCN guideline, if they were high risk or not. And among what they would call very high risk patients, the metastasis-free survival was 87.3%. Now, if you were a class one, it was 91.9%. If you were a class 2A, it was 74.9%. And if you were a class 2B, it was 58%. So big differences in the metastatic rates. And then among the high risk patients, or not the very high risk, but the high risk, the ones were 97.4%. The next class, the 2A were 90.5% and 71.4% for class 2B. So it definitely added prognostic information based on the way that they looked at it. Now, there is a very similar additional article. And that SEC one, whenever, by the way, whenever I look at the list of authors, is a powerhouse list of authors, none of them. They might all work for Castle in some way, but none of them are actually employees of Castle. The second one was a melanoma article, not nearly as high-powered a group of authors here of academic radiation oncologists and whatever. But this one, the 31GEP identifies patients with early stage. One in 2A, Ct. N. Smellinoma, and Oman high risk of melanoma-specific overall mortality. And basically, they took stage one or 2A, Ct. Smellinoma, that was tested with the GEP. This was from the CER program. And the people who had a class 1A had a higher five-year melanoma-specific survival than those with 1B/2A or 2B, 98.8% versus 94.7 versus 91.6. And significantly, that higher overall survival with both whatever you looked at, multivariable as well. And so they showed that in these people have a very, very good prognosis to begin with. This gives you a little bit better prognostic information. And I'm going to kind of stop there, because to me, this summarizes everything I've ever heard that makes me think, "That seems like a pretty good test." It's pretty much summarized here, that when you retroactively look at it, the way that I think of this test being used. So first, I'm going to stop for a second and bring on our guest, Dr. Whitney High. So for anybody who doesn't know him or hasn't had the pleasure of getting to listen to Dr. High lecture, so Dr. High is a very frequent speaker on the lecture circuit at Falclinical Maui Derb. If there's a good meeting, he is the Dermpath lecturing there. A brilliant, brilliant guy. And his faculty at the University of Colorado, both does Dermpath and sees patients. So like the kind of ideal patient to eat to the person, because he's got to have the conversations with the patient, just like the rest of us. So like you're normal Dermpath who's out there telling you, "Oh, you bet, could be a drug rash. You better take them off all their drugs." And they have no idea what a big pain in the ass that is. So, Dr. High, welcome on the show. Thank you, thank you for having me, Matt. So you've heard of a robust debate. So, and I'm a little concerned because I know certainly no less about skin cancer and have managed less skin cancer than Pat and Ferris. But I, so I think of it as being, if I had a really low risk melanoma, I would want to have castle. And if it came back that I had a 0.4 millimeter melanoma, and was like, okay, who am I very worried? But I had a bad castle result. That I might, I probably would want to send a live note biopsy. That's the one setting where I think of it as, it would never make me not do something, but it might make me do something if I had a low risk tumor otherwise, and then it had a bad castle. Squam is still, I think of it the other way, that it's probably not going to drive any manate, and what I would do. But if I had a high risk squam, it might make me feel a little bit better if it had good castle results. Right? So I think we've heard that Dr. Pat and Ferris think the tests highly overused. What, Pat and Ferris, anything, so I guess I'm just going to stop there. What do you deal with this year, the pathologist, you're the clinician, you kind of, sure see a lot of these results coming in with lesions. And what do you think? >> Yeah, so a lot of the data that premises that underlies the castle test was taken from the University of Colorado archives here. So before I even came, and after I came, they were taking a lot of our melanomas through Coutanian Sanctology and using those to learn about the issue and develop the test. So some of the tests is probably locked down on Colorado melanomas, which is interesting. That said, it's always struck me a strange that our Coutanian Sanctologists who deal with melanoma all day long, they don't want us to use the test. They say when we need the test, we'll go ahead and order the test. So that's kind of an interesting thing to me. They don't utilize a test that they contributed highly to. So I think that's interesting. I have personally followed a course that's a little bit more like Dr. Patten was describing where, from time to time, I'll come to a melanoma and twine with an evis, and it's difficult to decide if it's all melanoma, top to bottom side to side, or as part of it an evis, and part of its melanoma, would that impact the depth and would that impact what I would do for the patient? And then I've used the test even ordered it myself. But I don't routinely do it on every single melanoma that's 0.3 millimeters or deeper. And while I heard what you were saying about, if you had a melanoma of those 0.4, wouldn't you just want to know? Well, interestingly, like some of the more recent things, and I'm not an oncologist or anything like that, but the recent papers about priming the immunologic pump and actually letting melanoma sit in the lymph nodes to, and giving the PD-1 drug while there's melanoma still there, rather than waiting and removing all the melanoma and then giving the PD-1 drug, that's all kind of interesting. And I think it impacts this discussion as well, just a little bit, and that maybe if you missed melanoma in the sit-alive node, would you be really missing a lot or would you be actually waiting for it to become clinically detectable such as you could give the PD-1 drug with the pump prime? So-- I got to jump in there for one second. And Pat, I don't know if you guys got the same spiel from Zateli as I did whenever we were residents. When we were residents literally 20 years ago, we wrote that we wrote that, but John Zateli, and he was one of the big anti-central lymph node people. And that was literally what he preached all the time was that by taking out that sentinel lymph node, we are reducing this person's probability of having a robust immune response. Now, that probably got to be a better argument once we got adjuvant therapies. But I mean, Petta Ferris, did you guys get the same spiel from him whenever you were rotating with him? Oh, yeah. So I think we have data from MSLT1. And patients who got a sentinel node didn't do worse. So I have a hard time believing that the sentinel node makes the patient do worse. I do think though that in the era of immunotherapy to Whitney's point, you would want the opportunity for neo-adjuvant therapy, right? So the idea that having some-- They're the first to find neo-adjuvant because they do not have everyone in the world knows. Yeah, it means immunotherapy, but in this case, immunotherapy given prior to definitive surgery. So neo-adjuvant therapy from melanoma is PD1 inhibitor, usually three doses of pemberlizamab, then surgery. But like surgery for the local lesion or sentinel lymph gobyopsy. This is when you have measurable disease, so usually in the lymph node basin. So it's really not sentinel node positive data, are Sentinel node disease. it's not, you know, what I'm thinking about. I'm thinking about it is clinically apparent. Well, what I'm thinking about, so for our dirt, for our dirt and listenership. So if I get a melanoma that's 1.4, send it in 1.4 millimeters deep, am I okay to go ahead and resize that as quick as I can? Or do I want them to send them to Onk first to see if they should get neoagevent before even that local re-excition? No, you should not do that. So one point for-- I'd like to get for anybody who's listening and you're not watching. As I'm asking that, Dr. High is over there, like vigorously shaking his head. No, no, no, no, no, no, no, wait, don't wait. Get the melanoma out. No, I think if somebody has a 1.4 millimeter melanoma, I would argue that is a patient who is, you know, by NCCN guidelines, a candidate for Sentinel node biopsy, and you probably want somebody to have a very informed discussion about the pros, cons, risks, benefits. You certainly do not want to excise it. You want them to have the opportunity to have Sentinel node biopsy offered. Okay. What about like a 0.8 melanoma would you-- So 0.8 to 1 is where I think this becomes really tricky, right? So if you look at who went into MSLT1, you had to be at least 1 millimeter. So we don't have a prospective study on 0.8 millimeter, you know, randomized Sentinel node biopsy. So we have sort of moved the needle or the threshold down to 0.8. You know, that's probably where the benefit or the prognostic value of Sentinel node maybe isn't quite as prospectively validated. And there are, you know, like Mike Marquetti, who was at Memorial previously, did this kind of really good reanalysis of the data, the castle data, and said, you know, if your default is Sentinel node for everybody for T1B lesions, like 0.8 to 1 millimeter, this is really a place we could probably reduce the Sentinel node utilization by about 40%. So, you know, these cases where it's borderline, you know, should we where it's sort of discuss and consider, maybe that's where and to me where this could make a big difference. So there, and I think that the benefit is probably primarily there. I mean, remember Matt, for your T1A, your 0.4 millimeter melanoma, where we want the information, the positive predictive value was not even 5%. So if you had a high risk and you say, "Oh, I go get a Sentinel node," the positive, the chance that you would actually end up having a positive Sentinel node was less than 5%, even with that high risk result. Is there risk to a Sentinel node? So they got, you know, what bad things happen to people who get a Sentinel node? I mean, in particular, if you're looking at lower extremity melanoma lymphedema, and that could, you know, there's any time there's, you know, there's surgical risk of infection, there's anesthesia, there's cost. I mean, there's a lot of potential risk to a Sentinel node, particularly lower extremity once. The lymphedema rates not in signal. So, I would have jumped back over to you. So the whole reason we've got this damn castle test is because you derm path suck at pigmented lesions, right? There's study after study showing if you send it to 10 experts, you get, you know, three, you're going to die, three, I don't know, and three, you're going to be fine. Like, are you guys getting any, like, why aren't you guys getting better at diagnosing melanoma and giving us? Why do we need this damn castle test? You know, that's a super interesting question. And it's actually something I've thought about a lot recently. So, so let's just accept that the castle test can decide that some melanomas are really, really bad and some melanomas are really good. It doesn't even have to be the castle test. Let's just say that genetic test, there was one developed at Wuhan labs. Yes, that Wuhan, and released last week in science scientific reports, and they have their own prognostic test there. So, so everyone has their own prognostic test, but let's assume you can slice and dice these melanomas into really bad actors and not so bad actors. What does that say if you believe that and I see everybody nodding on the call, if you believe that's possible, then, then what do we really know about what melanoma being an on or off zero or one binary system? How can melanoma just be either binary, you have melanoma, or you don't? If there's all these different spectrums of melanoma ranging from from benevolent or, or, you know, barely milling that melanoma to horrible alkylia from a lesion the size of six millimeters, it's crazy to think that then if we think melanomas, this broad spectrum of disease, how are we empowering people by looking through this tube right here? And to decide what is melanoma and what isn't, it's really, really bothersome if you really think of the logic of the argument, the more you think melanoma can be broken down into bad actors and good actors, the less likely it is that I'm going to be able to peer into these tubes and decide if something is melanoma or not. They're actually work against each other. So I think that's a huge one. Why do we need like why do we buy up C9 benign things for everyone malignant, right? Why is it not just obvious? Like it's hard for us to decide clinically what's a melanoma and what's not. I mean, there's very obvious melanomas clinically and then there's things that we all get surprised by, right? And I've had like, I can think of one, you know, Neville, melanoma that I buy upseed on a patient and I almost didn't do it. I just thought it was a trauma, I think it was in a kilter, right? Like that, these things happen. So you have to wonder and just answer this, you know, to yourself, but how many of those PT1A's aren't melanomas at all? So it's kind of mind boggling to think that you have this spectrum of disease, 6000 different mutations and not only, you know, you got to even think of these mutations in kind of a factorial way. Like what if you get this mutation before that mutation? Is that as bad as you get this mutation before the other mutation? So, so if you're looking at this spectrum of melanoma, probably it's impossible to really look at them under the microscope and just decide on their morphologic features, their appearance, what's malignant, what's not? That's crazy. So I mean, I think we've used morphology because it's all we've had, right? We've only had morphology clinically. We've only had it histologically. I mean, it seems like we should be in the reason why I'm interested in these tests and, you know, do I think we should use them on every point three millimeter melanoma? No, but I think we have to move toward genetically defining what is melanoma. Yeah, they will happen. That will be the way of the future. But here's a glass slide stained with H.E. This is the same technology we've been using for 138 years. So if you went into a patient room and said, I'm going to analyze this lesion and I'm going to do it with 138 year old technology, well, wouldn't the patient be like, "Did I do anything better?" Anything else happened in the last 140 years you could do to the tissue, Dr. High? And then they get angry if you do any stains. So it's a no-win situation for the germ path. We either use the 140 year old technology or we do some stains and get everybody angry at us from. So it's a real messed up world here in germ path. Have you worked with any AI on this? I mean, it seems if humans have a difficult time where the one's training AI, AI is never going to figure this out. Have you seen anything with AI? I haven't personally done it, but I've been interested in any AI people out there. Go ahead and give me a call. But I'd be personally interested in doing those kinds of things. You know, I've seen people even when I was a young resident 25 years ago, I saw people that wanted to look at nuclear size and they wanted to bend things on nuclear size, number of interest spying as cells, all those kinds of things and go after it mathematically. You know, there's been attempts to do it molecularly, sort of like the work I did on gadolinium that we were talking about before the show. There's been all kinds of different approaches to this problem, time of flight, mass spectrometry, all those kinds of things. And none of them have worked perfectly. Is there anything getting close? Is there anything even in terms of not prognostic information, but yes, no. Is there anything? Well, you know, I think, you know, the neatest thing I've had in the last 10 years I've been doing germ path is praying, immunostaining, praying, immunostaining. It's just a simple single. I mean, it has to chemical study, you know, that's a hundred to two hundred fifty dollars depending upon where you work. But if something's praying expressive, it doesn't make it melanoma, but I sure want to take a longer look at what, why would that thing express praying? That's kind of strange. Considering that 90 plus percent of melanoma expresses praying and 90 percent of benign lesions don't, I think it's a pretty neat thing. It's more powerful than P16 or H&B 45 or any of those other things we've used through the years. But the only thing I would say is I think it's powerful in your hands where you are an expert dermatopathologist and you're not like, oh, here's a pigmented lesion, I'm going to get praying, right? You're like very closely targeting it. And you have usually pathology probably sent to you by melanoma experts. I mean, you probably have a variety of people who are doing biopsies, but you have, you know, a dermatologist who's chosen. a clinically suspicious lesion, an expert dermatopathologist who said this one is borderline. And so you're using it on a selected case where, you know, there's already like some pre-test probability that that could be melanoma. Whereas if you just started being like every pigmented thing I'm going to get preyed on, it would probably lose its value, right? Right. And certainly we don't want to say everything that expresses prey in this melanoma, it's just not true. But when I do see some prey in something, if I'm going to override the concept that it's melanoma, I want to be really, really aware. Boy, that thing was preying expressive. Let me see it one more time. Oh yeah, yeah, it's okay. But I do think it's an interesting thing. And more powerful than P16, more powerful than H&B 45. You know, with the, what used to be the myriad test, we would get an indeterminate result and you still have to decide, is it melanoma or is it not? And you got some weird signatures from Blue Nevi and so Blue Nevi got some really, really strange genetic signatures from that test and yet it wasn't melanoma. So you don't want to take anything as thumbs up, thumbs down. If that were the case, we'd have dolphins looking through these tubes, you know, there are things where melanoma now. You got to have some clinical judge. Two chirps, acumen, two chirps for melanoma, one chirp for Nevis. But I would argue if I showed you a separate caratosis and you're like, that's a sub K and then you stained it and it was praying positive. You'd be like, it's still a sub K. Absolutely. So I would also argue if I give you a 1.8 millimeter melanoma and say that's a class two way, you go, all right, that sounds like that could have some risk for metastasis. I would say that a 0.3 millimeter, a 0.4 millimeter melanoma, that's a point, that's a class two A, to me is more in the world of a sub K with some praying versus the really atypical looking lesion that has praying. Well, yeah, unless you're barring that it's you caught it really, really early and it's really, really horrible looking from a cytology standpoint or something like that. You know, I would tend to agree with you. But part of the reason I've wondered if the oncologists don't like us to use this test very often is because most of the tests come back with a benign or a less aggressive subtype and they don't have to have a long comprehensive discussion. But when they have one of those thin melanomas that has a 2A or 2B and the patient says, okay, well, you know, we got this result. What are you going to do for me? And the oncologist says we're going to follow you very, very closely. And I think the patient's like, no, seriously, what are you going to do? You're going to squish me with some juice. You're going to radiate me. What are you going to do? And you just keep seeing we're going to follow you really, really closely. I think it's very, very disconcerting to the patient and the encounters become very, very odd and long and lengthy. And I think that's why the oncologists don't necessarily want us to do it to every single thin melanoma because it creates really, really strange, long, painful encounters where they just keep saying we're going to follow you very, very, very closely. And where we don't have anything to offer anyway. That's what I mean. Yeah. All right. So we're going to cut that part of the discussion off now. So now that we've finished our discussion on gene expression profiling and skin cancer, prognosis in general, we've got a really special treat for you over on scholars in medicine. So for those of you who are younger, you may never even heard of a condition called nephrogenic systemic fibrosis or NSF. But for those of you who were practicing 20 years ago, especially during residency in an academic center, you saw these people come in who skin was just turning into wood. And Dr. High is sort of the person who put together the pieces of what was causing that. And that's why we he's a big part of why we don't see it anymore. If you want to see that here, that full discussion from us, head over to scholars in medicine. Again, that scholars in medicine.com. And you can get the full story of how they figured out that gadolinium from MRI contrast was causing that horrible new skin disease. All right. So let's jump on to our last segment of the show, Dr. Patins trivia and see if Ferris can continue to keep me in the zero winds column. All right. So histopathology focus is our stains named after people. I called it call me by my stain. I appreciate that. All right. This day's pathologist noticed that after staying some bacteria resisted decolorization leading to a classification of bacteria still used to this day. Graham, Graham, Christian Graham. I was trying to. We're going to like, I know, I'm going to predict that Dr. High has. I had a strong feeling that that's going to be. You know what? I'll give bonus points. If you get through it, you know, for for for a bonus, you know, what we do is the Graham stain into the microscope now is really a brown brand stain, a little bit variation in the grand and stame. Oh, oh, named after Dr. Brown and and Brent. Very good. All right. Number two, an American pathologist. He served as the chief of the laboratory branch at the Carville leprousarium during the late 1950s. And in 1969, published the role of histology and the study of leprousy Hansen. He, Hansen's older than that. Pass it fast. Not a fast. Fast. You got me. Hansen is older than, than, than, than the United States. Yeah. You had zeal meal sin. And then they modified it and called it the what's thing? No idea. We George fight fight. Sure. Sure. Fight stain. Fight same with peanut oil. All right. Last one, which could end up in a tie if Ferris and Zayers can can bounce back here. One half of this silver nitrate based stain was named after this pathologist who also had a tumor of the parodid gland named after him. The stain is used to detect Bartonella, H. Pylori and syphilis. Stary. It's worth it. Damn it. If the shore. Orthon. Yeah. Stary. Obviously they collaborated. Stary. His middle name was chronister, which that was the most fascinating thing I found about Stary. But yeah. All right. That's a top. Nice. Let's come back. I continue my no wins streak. All right. Dr. High, thank you for joining us on this. Yeah. Thank you. And for our listeners, if you're, I want to thank you for joining us as well. If you've got questions, comments, ideas for toppings on the show, shoot us an email at questions in terms of drug.com. And I hope you learned a few things. I hope you laughed once or twice. But mostly I'm hoping you're planning to join us again next week. Until then, I'm Matt Zyrus. I'm Tim Patton. And I'm Laura Ferris and we are Derms on Drugs.