Welcome to the AASLV, ASGE, and DOSB course live Q&A session. I'm Jennifer Telford and I'm joined by my fellow program chair Mike Levy. We're also joined by our session presenters Kenneth Binmuller, Richard Wong, Julie Heimbach, Ian Gan, Praveen Chahal, Raj Shah, Jane Sababian, Suresh Chari, and Neil Mehra. Gideon Hirschfeld unfortunately cannot attend as he is speaking at another session for the meeting. We hope that you had a chance to view the fantastic course content prior to joining us. As you can see the presenters are here and they're ready to answer your content related questions. If you have not had a chance to view the presentation yet you can access them at the Liver Meeting Digital Experience at any time until the end of February. On the right side of your screen there's two tabs. There's a Q&A tab and a chat tab. To submit questions please use the Q&A tab and indicate which speaker you're directing your question to. We'll do our best to get as many questions answered as we can during the time we have available. Feel free to use the chat tab to engage with other attendees and to share resources. If we don't get a chance to answer every question today TLMDX offers a unique opportunity to connect with presenters through the platform. Navigate to your network and engagement center to search for a presenter by name. You may add them to your list of connections or request to schedule a meeting. We'll move on now to our questions. Thank you Jennifer. I'm going to get started with the first session management of gastric varices. I admit my bias because I use EUS so hopefully Jennifer will pick the same bias for Richard for non-EUS. Richard nicely pointed out that there are not many experts at EUS guided angiotherapy. What do you recommend in centers that have both EUS and non-EUS expertise? Do you think there are enough advantages that EUS is the procedure of choice? Thank you Mike for that question and thank you Jennifer for organizing this Q&A session. Look forward to it. So great questions and unfortunately you know we don't have proof that EUS guides us to conventional glue injection and that's how I had been doing it for decades before I started using EUS. What I want you to understand here it's not really about EUS being superior to conventional because there's no question we can see the varices endoscopically very very well and we had great results doing it using the conventional method. The issue for me was high migration or the embolization rates that unfortunately all of us using glue had experienced and my goal was can we reduce that embolization rate or ideally eliminate that because I personally experience severe complications from embolization because it can go anywhere in the body and if a patient has AV shunts or open foraminal valley it can go into the arterial system it can be devastating and fatal. So really for me the question is if we can do if we can reduce that complication then in my book it's superior but there but there is evidence that at least retrospective data this is from Indiana University so you may have seen the study last year comparing EUS guided glue injection with endoscopy guided glue injection and it's noteworthy that DeWitt's group found that there was a significant reduction in the re-bleeding rate not just for gastric varices was like a third or something but also for non-gastric variceal bleeding and they required much less glue to obliterate the varices so we need a randomized controlled trial like that to really validate the findings from that Indian study and we don't have that yet but there are conceptual advantages that I think deserve emphasis and the biggest conceptual advantage is you know yes we see varices and endoscopic is the tip of the iceberg and we don't know what's behind and for all of us that have been using EUS to guide pseudocyst drainage right we've kind of realized the advantage of actually seeing the pseudocyst versus just the bulge so we're seeing just the bulge with endoscopy and we can really visualize and document the obliteration of these varices and there are studies showing that the reading from varices correlates with our ability to completely obliterate the varices so I think we'll see the data that EUS is superior but I think the big limitation now is there are still a limited number of people that are skilled in EUS fortunately that's growing but you asked if you're in a facility with EUS and you know that has that expertise should you be doing this EUS guided I would say yes I mean if you do EUS why would you not want to have the extra set of eyes so to speak and actually see it with ultrasound it's endoscopic ultrasound it's a hybrid procedure absolutely and some use Doppler ultrasound probe to assess the response to therapy so with standard techniques don't use a probe one concern I have is you think that there's a limitation on the through transmission of the ultrasound beam it acts much like a stone and you'll get a false negative response how do we overcome that we use Doppler as well how can we overcome that limitation and better assess the response to therapy yeah I mean I think that probe is very it's very I very few centers I suspect have that app but that would certainly be better if you don't have EUS I think having that ultrasound probe that you can pass through your endoscope to at least sort of val confirm that you've obliterated that's truly a validation is yet to be proven you'd have to compare the Doppler ultrasound probe versus EUS Doppler flow and the Doppler flow you know gives you a view of the entire convolute and we see these convolutes that lie behind the surface surface meaning the mucosa right so yeah I mean we have to prove that that really is real but we usually do what's called instrumental palpation that it's hard to instrumental palpation and we don't really know it would have been interesting to do a correlation of instrumental palpation versus EUS obliteration based on no Doppler flow you know so having no Doppler flow you switch on the Doppler you see there's no flow you know that you've obliterated that varix and also the tributaries which is very important thank you Ken and Jennifer you want to I'm sure Richard has some ultimate views yeah Richard Dr. Wong do you have I think you're still on mute would you like to respond to Dr. Ben Muller's comments all right yeah we have for you as well yes well thank you very much Mike and Jen for inviting me to this symposium so you know I think what we need to do is so there is a shortage of manpower but I think we need to really be able to clone Dr. Ben Muller and Dr. Levy and if we're able to clone Dr. Ben Muller and Dr. Levy we'll have a wealth of really expert endoscenographers who are able to do this with EUS guidance but you know but on a more serious note you know it's it is you know it is hard you know we in our institution we have both EUS capabilities and we also do non you know EUS glue as well so we can actually we actually do both and I do EUS myself as well one of the things that we've recently had we've had some adverse events with the glue with the just a regular endoscope we've actually one of the endoscopes got glued so that the actual needle was actually stuck in the endoscope and this happened about a month ago and another endoscope more recently somebody had pressed the suction button and the glue coated the outside of it now that was with a regular diagnostic gastroscope and I would worry that if that would happen to a more expensive linear echo endoscope that it would cost a lot to repair that and some of these things just happen I mean we try our best you know glue is glue and we try our best not to have these things happen but sometimes these things happen so one of the things is that you know we obviously don't want to ruin any endoscope but if you are to ruin a endoscope it's probably better to ruin a diagnostic gastroscope than a expensive linear echo endoscope so that's one of the things to think about we also do a lot of these patients in the MICU and for us it's easier to go up to the intensive care unit with a Doppler probe which is what we have than to bring the whole EOS setup to the bedside in the MICU. Different institutions have different ways of doing it but those are some of the things that I think are important. Can I make a comment about the risk of damage to the scope? So when we inject endoscopically I would always see some glue exude out of the puncture site novarix or from the varix after injection and so I would have to wait a few minutes for the glue to solidify and keep my scope away from the puncture site. When you do it EOS guided I have never seen any glue exude out of the punctures and I think the reason is because we're going actually transesophageal no we're not going transgastric where the mucosa is very thin we're going transesophageal and we're going into the fundal varix typically when we're treating those which is a further distance away so none of the glue actually ends up exuding or exiting into the lumen of the esophagus that's number one number two we have a balloon over our transducer so even if some glue actually did out and had contact with the scope tip the balloon would protect it and we take one extra step of precaution and that is that we don't pull the FNA needle out immediately after we do our injection we actually extend it a little bit pull the whole scope out then wipe off any glue that may be on the tip of the needle and then we remove the needle so those are all precautions and we haven't had this issue with damage to our echo endoscope there's no question about it being much more expensive just a comment about portability though our our eus equipment has now become so portable it's just on the cart on your standard endoscopy cart now we have eus integrated it's really become a hybrid procedure so that really shouldn't be a limitation bringing an eus scope to the icu i've done lots of eus guided glue injections in the icu with for bleeders that is never an issue for our staff to bring that equipment it's already on the cart right and some and some endoscenographers you know have argued it's actually you know easier than someone who's actively bleeding to do it via eus versus endoscopically from a from a visibility perspective that's a big practical advantage right that you're not dependent on an endoscopic view you don't have to clear off stomach you can just go right to eus guided you see the verses one other question that had just come up um that both of you could answer do either of you and patients that you're planning to do gluing on and sometimes we don't have that luxury of time because of the urgency of the procedure do either of you um routinely do cardiac echoes on patients to see whether they do have a patent for amenobally yeah no i you know that there's um there's lots of different practices in different institutions we actually don't um we we don't do that because we we you know we we don't do that and we think it it just delays things i know that some institutions do um get cardiac echoes and get bubble tests and get you know ct angiograms uh we have not done that and you know um we have not had any problems with that in the last you know since we've been doing it so um i'm hopeful that by adding the coil prior to the glue injection which acts as a scaffold to retain the glue in place where you inject it and eliminate the risk of embolization i'm hopeful that is truly um effective um and uh we haven't been doing it for the reason that at least we haven't experienced any further uh complications from embolization um but yeah i if you're using only glue that would be the ideal because i had a patient i literally stopped doing glue injection after i had one patient a younger patient who became hemiplegic because patent for amenovally and you know and and and stroked out that was devastating for me and i placed a moratorium on using glue until i came up with the idea of placing a coil first you know that was the impetus for me so i want you to understand even though i was doing us i really didn't see necessarily an advantage to be honest if glue would solidify adequately and avoid embolization then i probably would have never gone on to looking at us as the modality to guide this yeah i i think this is a really great issue and that's why why we're debating it um i think honestly i think there's a role for um there's definitely a role for us guided as dr binnemuller is saying um definitely a role for non-us guided and also for the radiologists the interventional radiologists which we haven't touched on in terms of um boto brto and tips so i think we do need some sort of guidance um some sort of randomized study of some sort to see which patients which patients actually do better with us guided you know glue injection as dr binnemuller is saying which patients don't need it and which patients may do better with interventional radiology approaches uh we don't have that and and a lot of it's just left up to you know who's available at the institutions correct yeah the local expertise that's right correct great well thank you maybe we'll move on to the next questions then we can circle back um as questions come up from the audience as well boy thank you both you know i'm going to move on now and ask a couple questions to julia if i can julie's one of our transplant surgeons in may and we appreciate you joining us today one question that always comes up is the risk and the implications of transperitoneal biopsy and i just want to be clear if a patient's had biopsy their primary tumor and it's positive does that now designate the patient as unreciprocable and non-transplantable yeah mike thanks for that question so what we're speaking about specifically is for perihyal or cholangiocarcinoma so i think it's important to to designate that up front the implications are different for other types of tumors but specifically what we unfortunately learned in our perihyal or cholangiocarcinoma experience this is a cohort of patients that come to us who are unreceptible so we offer them this neoadjuvant radiotherapy protocol and then there's a there's a pretty um long delay before they're able to access transplant and that may be how we were able to uncover this risk but you know because they were sitting there waiting before we could actually apply the therapy um and then we would find seeding along the biopsy tract actually saw this for i think six patients um in when we looked at the experience that we had so so based on that experience yes that would exclude a patient from the opportunity to have curative therapy um which would be transplant now would it actually preclude them from if they were a resectable patient who had a biopsy would it preclude them from having a resection you know i don't think there's a strong enough evidence to show that although i i don't think that the physiology is any different it's just that you maybe wouldn't detect the micrometastases and then maybe when you go in there to resect you sort of stir things up enough so that it's not an issue it's hard to say obviously there is a high recurrence rate after resection and whether that you know could be related to this transperitoneal biopsy in general i just would say it you know i would avoid it if you're thinking of curative therapy whether it's resection or transplant really try to not biopsy the primary tumor through the peritoneum either through the skin or through the duodenum or stomach across the peritoneum into the primary tumor us has just been so helpful to us in biopsying the nodes so that we can determine if the patient you know has metastatic disease prior to enrolling them and giving them this big bunch of radiation and chemotherapy that then ultimately they can't go ahead to transplant if they have lymph node metastases so this has been really fantastic to you know avoid putting them down this pathway which does have some you know it's good to cure the primary tumor with radiation but it's a big big amount of radiation that we wouldn't necessarily give to a patient with cirrhosis unless you have that chance to switch their liver out in the protocol so so us has a huge role to play for us in sampling the nodes but really don't sample the primary tumor i couldn't stress that enough and we still see it even today and so i hope i can keep beating this drum as much as possible you know that we do perform transperitoneal lymph node biopsy some of the lymph nodes are directly or in close proximity to the primary tumor is the risk of endotrachea not the same when sampling a lymph node versus the primary tumor well the thing is if their lymph nodes are positive they're not going to be offered the therapy so that's a rule out whereas if the primary tumor of course if you biopsy that you know that's fine as long as it's an endoscopic biopsy but if it's a transperitoneal biopsy then we would have seeded the peritoneum in a patient that would have otherwise had a good chance to go ahead so that's the stations matter so some designate well some use n1 and 2 based on the number of positive nodes so we'll say n1 or n2 relative to location and one being more local around the bowel ducts cystic duct etc and to perivascular duodenum pancreas does it matter which node is positive any node positive is positive yeah so for the transplant protocol that is true and in you know for a resection candidate um you know there is still some marginal benefit in resection even in patients with positive nodes so it's not a hundred percent recurrence it's you know sort of 70 percent recurrence so a patient with positive nodes who has a resection um you know especially those close by nodes they would still likely be offered resection outcome is definitely inferior to a patient that does not have nodes but it's still there's a chance to have some benefit but in the transplant protocol we we really can't offer transplant in the setting of metastatic disease no matter which nodal station it is in the radiation oncologists who are you know have been heavily involved in this protocol have argued that they actually can clear these nodes with their dose of radiation and they do give a massive dose of radiation so so theoretically that you know is a possibility so we're not actually looking at you checking the nodes that are just right next to them you know we have that sort of protocol of nodes that we ask you to look at and and then we actually stage them again just before we do the transplant with an open staging and or laparoscopic staging and sample the those closer in nodes just at the very end um just before transplant absolutely so if we're talking about high alert nodes cystic duct etc versus peripankreatic periduodenal any hope for the endosynographers to say how do I distinguish those nodes they're pretty tight in location on the u.s uh i know we've come up with a formula we need but what would you suggest uh designates the difference between the two in the u.s you know i i don't do eus mike so you've got to answer that question i don't know what it looks like to you i know what it looks like to me but well then you probably forget we had this discussion years ago but what we came up with was we would take the mid portion of the extra hepatic bile duct and so yeah we look at the papilla to our counterclockwise in the bulb to the hylum, take the mid-portion. Some centers will use an insertion of a cystic duct, which we don't like to do, because a cystic duct can insert almost anywhere, but we try to take that mid-portion if we can. Yeah. I think that's really important. You don't really need to be right up next to the tumor when you're sampling the nodes on this first pass. Well, thank you very much. Jennifer, you want to take us to our next speaker? Yeah. So thank you very much for your talk. It was great, Ian. So you do both ERCP and EUS, and your talk was directed more towards EUS, drainage of the biliary tree. But when would you choose, and someone who was not a good operative candidate, you wanted to drain their gallbladder, when would you choose to do ERCP and attempt to place a transpapillary stem cap through the cystic duct into the gallbladder versus using a luminoclosing metal stem and creating, via EUS, creating a passage between the gallbladder and the duodenum or the stomach? Yeah. It's a very good question. First of all, thank you for inviting me to talk, and the disclaimer to the talk was that it was originally designed as being one side of a debate as to whether or not to use EUS or ERCP for gallbladder drainage and biliary drainage. And then the other speaker didn't get their talk in. And so the talk that I gave was a bit one-sided. But I think that you really have to make a decision on a case-by-case basis. And I think that a lot of factors go into that decision. So for gallbladder drainage, I think that part of it is the environment that you work in. And you and I working in a Canadian medical system where our access to Axios is a bit reduced, at least particularly at my hospital. And that's just because of costs. And I do think that costs really do need to be considered. I think that transpapillary stenting is a very effective form of gallbladder drainage when it works. The literature suggests that it's upwards of 90% technically and clinically successful. But everybody's biliary anatomy is a bit different, particularly the cystic duct. And as Michael just referred to, the cystic duct anatomy can be very variable. And the takeoff can be angulated. It can be very difficult to get guidewire access into the cystic duct. Occasionally, you will actually have to use cholangioscopy to actually visualize the takeoff. But I think that a number of factors go into it. One is, I think, how long the patient needs to have gallbladder drainage. So if this is a bridging procedure, and it is merely going to be until they get their gallbladder out, and you're just really trying to avoid percutaneous cholecystostomy tubes. One of my slides in one of the talks that I give to the surgeons is a picture of my grandmother. Because my grandmother, when she was, this is 10 years ago now, when she was in her early 80s, had cholecystitis. And she came in really, really ill. And she ended up getting a PERC drain. And I mean, just suffered because of the PERC drain. And that thing was left in for a good four to six weeks prior to them actually doing the operation. And by all means, what I do want it to be very clear is that PERC drains are probably the single worst option. And I think educating the surgical staff in your institution about the downsides of percutaneous cholecystostomies is forefront. And then really to reach out to gastroenterology before consideration of asking IR to put that PERC drain in. I think that, again, if the idea is just a bridging stent, I think you probably should just go the trans-papillary route. I think if it's a matter of a patient with, and I think patient factors obviously are very important. So if the life expectancy of the patient is very short or is less than six months, then I think that if you have one chance of getting gallbladder drainage, it probably should be LAMS. What we really need a bit more data in is the long, long-term data with regards to LAMS, because we do know that these things can eventually get some granulation tissue, et cetera, whereas that's not much of an issue with trans-papillary stents. And I've had patients with trans-papillary stents for years and years and years and never had an issue with it. So I think the caveat to all of that is if the patient has a six-month lifespan, and if they're only going to have one procedure, maybe that should be LAMS, and you should go directly to lumen-opposing stents. So I think a lot of these factors have to go into it. It's how long you need that stent in there, patient factors, cost. Their life expectancy. And then we also have our personal biases as well, and I think that can't be overlooked. And me personally, I tend to, because I have a limited amount of capital means to have all the axioses in the world, as opposed to a few years ago when I was in Seattle and that wasn't much of an issue, I will actually do an ERCP first. If I don't succeed, then I'll consider doing the axios. But that's a very different situation from being in the States. But it's an international crowd, and I think that it's worth noting that that's an issue. With regards to... Yeah. Sorry, go ahead. I was just going to... I think with regards to biliary drainage, it just really depends on, again, the reason that you're doing the procedure. If the reason for failed biliary drainage with just regular ERCP is your just inability to cannulate, then I think you might consider doing a rendezvous procedure. If you were unable to even access the papilla because of duodenal stricturing or something like that, then I think you'd just go straight to LAMs, or I guess a hepatic adjustment. Yep. Sorry, Ian. I didn't mean to speak over you. I had to come up. With respect to being able to leave, being able to remove the LAMs, and whether a fistula is created that will now stay open, and what would be the time periods, what would be the timing of that if you decided you were going to remove a LAM after using it for gallbladder drainage? And could it be removed without necessarily having to go on to cholecystectomy? And then the flip side of that, if you do end up using a LAMs for gallbladder drainage as a bridging procedure for a cholecystectomy, do you need to remove it before the cholecystectomy? So generally speaking, we don't have any data. And so nobody's really reported on that. Nobody's really reported on removal of LAMs. It's generally been done as a one-way door, and we've left these in indefinitely. I'm sure if we asked all of the panels here, all of us would probably remove them at some point, but usually in the setting of something definitive being done. I have generally only placed LAMs in settings where it was thought to be either a palliative procedure or the most feasible way of providing permanent gallbladder drainage. And so there have been cases where the LAMs have gradually begun to granulate off, and I've had to place more double pigtail stents within the LAMs just to keep them open. But I haven't actually gone to the step of removing them unless there has been something more definitive planned. Great. Thank you very much. We will move on to the next speaker for questions. Thanks for joining us. You talked about in patients who have a Roo-Lim following transplant, you generally start with a colonoscope. If that fails, you'll try balloon-assisted. What do you do if it's post-bariatric where that Roo-Lim is very long? Do you even try a colonoscope? Do you go straight to balloon-assisted exam? Yeah, first of all, thank you, Mike and Jenny, for having me. So if you're talking about a bariatric patient where they had a Roo-NY bypass for weight loss surgery, you're talking a really, really long Roo-Lim in those patients. And at our institution, we don't have double balloon endoscopy, and the success rate of single balloon endoscopy with our current platform is fairly dismal. So my practices, I directly go with the EUS guided access of the remnant stomach, the aluminum opposing metal stent, and performing whatever is necessary EUS, ERCP, through that route. Do you think there's a role, actually for the exact same purpose you just mentioned, we sort of abandon straight to single balloon in those patients. We'll have radiology pass us a guide wire. We'll use the guide wire to help us get there and then cannulate. Is that worth doing, or no, just do an edge like you stated? So if I'm understanding you correctly, you are talking still about the bariatric patient with a long weight loss Roo-Lim. So the IR, they got you down. Again, the challenge is you have to still get to the anastomosis in those patients or a papilla if it's an intact duodenum. But oftentimes, the most challenging situation is reaching there with our current single balloon platform. I think if you can reach there, yes, another challenge is cannulating. So if you have an IR guided assistance, you can bypass the challenge that you encounter for cannulation, but reaching there still remains an issue. So yes, again, our practices, we generally go with the EUS guided access of the remnant through the pouch or Roo-Lim. And if you do that, will you then perform an ERCP in the same setting, or will you leave the remnant in place into the remnant stomach and come back some period of time later? Excellent question. So it depends upon the underlying urgency of the indication. And my practice generally is I always, always try to do two-step. And the reason is, again, for the data that's coming out and what's already out, the single step EUS ERCP session carries a slightly higher risk of complications like stent migration perforation. So if you're dealing with a really emergent situation, the options are, if you're not comfortable, go with the PTC route. You have time in your hand. You can then go on to do it in a two-step fashion. But if you have tools and techniques like suturing, et cetera, that you'll need if things go awry like stent migration or perforation, then you can think about doing a one-step. But the most important thing is educating the patient about what are the risks, what are the alternatives, how things can go wrong, and if you have tools at your disposal and you are technically adept in managing those complications should they arise. You talked about nicely some different techniques to help you get a scope where you want it. You did not mention you can use the internal accessory channel scope stiffener. Do you think that can help you get there? And what are some tricks to using it? Yes. So I think, again, if you're using, you can try, you're using colonoscope. If the issue is more of a looping and redundancy, then you can try to stiffen the endoscope. If you're using a single balloon, you can use the internal stiffening wire. I mean, in my experience, how helpful are these? I would say flip of a coin. Again, if you are, you know, if you're using a colonoscope, you can advance the balloon, which I didn't mention in there. You can advance the occlusion balloon, inflate the occlusion balloon, hook it, try to pull yourself through. You can advance a biopsy cable through the accessory channel. If you're using single balloon, you can advance the stiffening wire. So, yes, all these techniques are at your disposal. But, again, chances of these working if you're dealing with really redundant or highly angulated root limb, which is not uncommon. Patients, especially with a lot of adhesions and scar tissue, I would say about flip of a coin. Again, there's no data to support the numbers I'm giving you. This is purely personal experience. This may be more controversial, and it's nice, Julie, that you're on the line. You may have some strong opinions about this. In a patient post-transplant who has a root limb, it's not always easy to get there. It can be quite long and complicated exams. Some patients require multiple, numerous exams. Is there a role for EOS-guided LAMs, so Illumina Posing Metal Stent, to the root limb to facilitate ERCP? Yeah. Hey, thanks. So, you know, we've actually just identified a couple of patients that have had that fun experience that you're describing with long exams who are going to try this. I think there is a role. It certainly wouldn't be like we just heard in the patient with the RUE where, you know, with the previous gastric bypass, where that would be the first thing you would do. But if it is a patient that has had multiple stents and, you know, it's been quite arduous, you know, this could be something to try. Most of the patients that we have this, unfortunately, are the radiated cholangiocarcinoma patients. So whether that's the ideal patient population to try it in, you know, we'll have to let you know how that works. But I think for certain patients, it's a really good idea. The other thing to know, though, is the patient with the RUE limb is typically the patient with PSC who has an abnormal underlying bowel to begin with, to be honest. It's, you know, UC plus, you know, some other inflammatory situation. So that is probably also something to consider, that that RUE is not perfect intestine to begin with. So something to think about. I do think from a patient perspective, any opportunity that they can have to have an internal stent as opposed to an external stent is certainly what they would like. Probably, is that something that you're offering patients now or they'll wait and see what your approach? Yes, I think especially if you're dealing with a patient who may require multiple endoscopic procedures. And if you see that the first time you ended up in and it required all that you could offer, you know, it was a three hour long procedure, multiple tools, accessories, and now you have to leave a stent in, come back, multiple dilations. Yes, those are the patients that you would want to consider creating a fistula for a more straighter route and shorter and possibly safer procedure. Thank you. Very interesting. Thank you. So Raj, your talk was very interesting. I think that the, you know, the use of a colitoclostomy to evaluate these structures is, you know, has been made great advances for us in looking at the bile duct, certainly. But, you know, it's not always completely straightforward. And I was wondering if you had any technical tips that you could share with our audience regarding, you know, biopsying in structures in the bile duct and particularly distal bile duct. So being able to get nice visualization if it's a tight structure, and then also being able to advance the forceps out, back in again, you can really just get very small pieces with the biopsy forceps with the colitoclostomy. So you usually want to do multiple passes down to get enough tissue. I was wondering if you could, any tips on your technique that may be helpful for others. Well, thank you for that question, Jennifer. I think these are all really nice technical points that can be helpful to address. And I think that, you know, the biopsy forceps has a span of about four and a half millimeters or so. There is not a spike on it, but I found that if you can stay fairly close to the lesion, that when you open the biopsy forceps for a second piece of tissue, usually you can get two pieces with one kind of pass. If you will. And as far as the advancement of the forceps, sometimes when the assistant, if it feels like it gets stuck, it may get stuck in the bending portion of the duodenum within the duodenoscope. And the other time is if you have the clangoscope dials locked, then sometimes there'll be difficulty advancing it across the tip of the clangoscope. But I would say that if you feel a hang up on the duodenal end where the bend is on the duodenoscope, then a rapid open enclosure of the forceps can be helpful to advance it. The other thing that I may do is if it's trouble negotiating a tight stricture, is I would pull back on the clangoscope a few centimeters or so, to see if you can straighten out and advance the forceps even further, meaning straighten out the clangoscope and pass the forceps further in, and then advance both the forceps and the clangoscope together. And sometimes that can help you get across the duodenal bend. In some cases where you bridge the stricture into distal stricture, then I'll advance the clangoscope up toward the bifurcation, against trying to straighten out or reduce the angulation in the duodenum to then advance the forceps during that time. So I think those are techniques that can help and advance the forceps, can also help and improve the yield of the tissue that's sampled. We usually try to get a minimum of six pieces. Some data has suggested that even one or two biopsies can be sufficient if you have immediate on-site histology or sort of a frozen section, if you will. But we haven't incorporated that in our practice, but we generally try to aim for about six visible pieces of tissue in the formalin jar to be sufficient. If the tumor is softer, I've also done multiple biopsies with one pass. And what you do is you'll do multiple biopsies, avoid suction while you're doing that, or irrigation. And then with the multiple biopsies, then we'll use a polyp retrieval, sort of a trap, and put that on the suction end of the clangoscope. And then after we have multiple pieces, we'll then suction and retrieve some of the tissue that way as well. So that can be another technique to try to improve the amount of yield. Sometimes the question is related to dilating a structure prior to doing clangioscopy to try to traverse it. And we try to avoid dilation, balloon dilation prior, just to reduce the risk of any traumatic changes or oozing or bleeding. But in some cases, it's still necessary. And if so, we'll try to use the smallest diameter possible, so a four millimeter balloon dilation. It may or may not get you all the way across it, but it might get you a little closer to the middle of the structure. And you can see a distinct difference in the tortuous vessels, neoplastic type of tumor vessels, as you would, or then you would from a traumatic dilation and some oozing related to that. But in some cases, what we'll do and what I would encourage is if there is a plan to do clangioscopy with the initial structure evaluation, to try to do it without any dilation, to look for neovascularization, to get some characteristics of the structure prior to dilation and the alteration that can occur from the balloon. Can I ask a question? Thank you. Yeah, of course, Ian. I was interested to see the contact friability as a potential characteristic on cholangioscopy. How do you assess that contact friability? Is it contact with the cholangioscope itself or when you're taking the biopsy or how did they actually define that? Yeah, and that's been a description by mostly our East Asian colleagues that with the original cholangioscopy mother-daughter type of initially fiber optic and then video scope as well. And my interpretation of that literature is that it would be contact with the cholangioscope. In my personal experience, you know that with guidewire access, there can be some minimal trauma related to that. But I do think that whether it's with the cholangioscope or with the touching of the forcep, it's a pretty apparent sign that we see. And usually it's associated with some neovascularization or dilated tortuous tumor vessels. So I think it's a little subjective and it's hard to call that a definite malignant criteria. But several papers on cholangioscopy will utilize the easy contact or contact friability as one of the malignant features, which is why I included it in the talk. I would say and I thought about that as well watching the talk and with the definition of friability. But also we're often doing this after the patient has had a stent in situ as well. It may not be the first time we're evaluating the structure and that can make it challenging too. That can make it more difficult. I think the contact friability is probably more pertinent for the index case or at least the prior to stenting because the stents themselves also can lead to some stent associated changes that can and the way to distinguish stent associated changes from say an IPMN, which you can't see biliary IPMN as I know the panel is aware of that, but for the audience. And the difference that we have found in our data is if you have fairly uniform homogeneous projections that are even velocipering, but if they're uniform without vascularity on the fronds, more than likely they're stent associated. I would still consider biopsy of it, but I then reference the scout film that showed where the stent was and where the internal flanges were on the stent and how it relates to the appearance of the lesion on the clangioscopy to see if there's any correlation. Certainly you can get some stent associated changes at the tip of the stent as well, but often it's at the flange, so you can do that correlation as well. But really look at the vascularity and the uniformity or lack of vascularity and the uniformity of the lesion to help and distinguish stent associated changes from neoplastic. That's great, thank you. With respect to these indeterminate layer structures, where do you incorporate EUS into sort of your algorithm of evaluating these structures? If it's a first go around for a malignant obstruction, we uniformly begin with EUS to look for surrounding adenopathy, liver metastases to sample for staging purposes, and also if it's an extra hepatic structure, and it was nicely delineated earlier in the session about avoiding transduodenal FNA of hyler or perihyler type of lesions, but if they're distal and you think that it would be appropriate for sort of a look over section type of management, extra hepatic, then FNA is probably fine to do at the primary mass. So we do utilize EUS fairly regularly for the index, and I would say that if they've had ERCP on the outside, or if we've done it here, we weren't necessarily suspecting malignancy, but it turns out we want to exclude it, then we'll do EUS at a second session, again for the same rationale. But then if EUS is not available at your institution, or it's a, you know, a 4 or 5 pm kind of hat on ERCP, and you're wanting to achieve drainage with ERCP as a primary modality, and the only modality, then we generally would do multiple tissue sampling methods, which would include brush cytology, fish analysis, as well as fluoroscopic guided biopsies. And when we do the fluoroscopic guided biopsies, I have turned to using primarily pediatric forceps that seem to do pretty well. They're a little more compliant and malleable to get across the elevator and across the sphincterotomy, and I'll leave a guide wire in place to allow for a side-by-side passage of the forceps. So I think with those three modalities, we can get a pretty good yield, diagnostic yield, and we have data to suggest that once you add clangioscopy visual impression, then our yield for detecting malignancy can be over 85 percent. But with those three modalities of cytology, fish, and histologic sampling with fluoroscopic guided biopsies, then those can all be a very nice first line if EUS is not available or utilized at that time. Thank you very much. We move on to the next speaker for questions. Thank you, Ross. Well, James, your palm trees look a little bit better than my buckthorn in the back, but it's the way it is. When it comes to fish interpretation, what is the threshold above which you feel comfortable saying this is malignancy? Is it serial pounds only, isolated, tetras only, tris only? What's that threshold you want to see in the right clinical context? Well, thank you, first, Dr. Lefer for chairing this course. An excellent question. That's a tough question to answer. And because, like so many of these other questions that have arisen, you can answer it in a vacuum with some level of precision, perhaps, or certainty. But so many of these cases, or all of these cases, really have so much more context to consider. But to try to give a concrete answer, I generally am not very impressed by a single polysomy result, unless, again, the context is so supportive that there is malignancy that to doubt that result, you run into a high likelihood of really making the wrong decision or recommendation or interpretation for that case or that patient. So that being, so aside from that scenario, which that itself can be defined differently by different individuals and different centers, I do like to see serial polysomy, because we know it's been published that nearly 50% of patients will not have serial polysomy. They revert to not having polysomy, or for some other reason, polysomy is not demonstrable on on subsequent ERCPs, and maybe it's a sampling issue to some degree. So I think either seeing serial polysomy or multifocal polysomy, which that too has also been published, is really, that's where the level of confidence begins to become quite solid with regard to there is an advanced neoplasia or advanced dysplasia here, if not frank carcinoma. And generally, so I think, but we don't really want time to tell, we want to intervene earlier. So I think having either multifocal or serial polysomy in the right clinical context, these are the right clinical context, is sufficient for me personally to say this is compatible with advanced dysplasia. It may not be sufficient, though, and I go into this in some of the slides to say this is carcinoma, go ahead and start chemotherapy, especially if it's palliative. Now that becomes a different territory where a higher level, a higher bar met, at least according to most of our medical oncology colleagues. So great question. The answers depend on who you ask, and that's mine, but the caveat is that there's a lot of variables to entertain or to consider. And in general, the less diagnostic fish interpretation, so the prizomas, tetrazomas, et cetera, will that force you or impact your timing of repeating the RCP? Will you bring them back earlier for resampling or not? How does that influence your practice? So the subset of patients who don't have polysomy but have some lesser or less worrisome result. So again, it depends on the context. I'm not terribly concerned usually about trisomy. I think many of us have seen that in very benign situations which have proven themselves over the test of time, five years out, four years out, ten years out, doing well, no masses, no evidence of malignancy. Tetrasomy, depending on the context, I may be a little bit more suspicious or wary of. And so I would make a case that there is a RCP. I would definitely capitalize on that opportunity or seize that opportunity and really want to reassure the clinical team, including the patient, that this is moving in the right direction or not in the right direction, at least. And with those centers that are performing routine psychology alone and now adding fish, can all that still be done on one brush and you split the specimen? Does it require an extra brush? Do we even have data that well-addressed this issue? Sorry, I missed the first portion of the question. What was that? So when you think about just routine psychology alone versus routine psychology and fish analysis, can we perform all that on one brush or do you take an extra brush dedicated for fish and one dedicated for routine psychology? And do we have any data supporting either approach? There may be data that supports one or the other. I'm not familiar with those data if they do exist, but I think it's reasonable to believe specimen you're able to harvest. The more specimen you have to evaluate and therefore you increase your likelihood or the likelihood of making the right call. So if the specimen is quite scant, especially because if it's only based on one brushing or one brush, they miss the opportunity to diagnose what would have been in the second brush. I think I'm not sure solid data that say you do need the second, it does make a difference. And the non-tissue means of diagnosis. Some might say that you guide where your biopsies are, but do your hepatologists and transplant surgeons, will they rely solely on non-tissue means of diagnosis when the tissue diagnosis is not there routine psychology or fish? So our surgeons and hepatologists are much more, I think, understanding or less incredulous than our own. And maybe it makes sense why that may be. So I think if they feel that we have exhausted our options, multiple rounds of ERCP, brushings, biopsies, cholangioscopy, and other ancillary non-tissue means, they take the compendium of data as the best we have and what we have to act upon. Our oncologists, on the other hand, will really, even if it's highly suspicious, highly atypical glandular epithelial cells, even in the absence of PST, so with really no good reason to have such cells, that's not acceptable or adequate to make a diagnosis of carcinoma and initiate chemotherapy. So it depends really on who the audience is and to a large degree, too, what's the patient context. I could just add here for the OPTN, there's clear criteria and they do not require a tissue diagnosis in order to proceed with a transplant. So that is sort of in policy. Malignant appearing stricture with one of the following, which would be elevated CA-19-9 in the absence of cholangitis or a mass or fish polysomy. So those would be the OPTN criteria. And Julie, have we ever looked at our resected specimens in those two different groups, one where we have firm diagnosis and one a bit softer in terms of percentage that have residual tumor? Yes, we definitely have. And we've also looked at recurrence rate in those cohorts. And we've excluded those that did not have a tissue diagnosis before, but then had one after. And the bottom line is that the recurrence rate actually is quite similar. So that's interesting. The dropout rate is higher for those that have an established diagnosis prior to, which probably just means they maybe had a little bit worse tumor starting out. But at the end of the protocol, the outcomes are the same. Thank you. And Jennifer, you want to go? Suresh might have some thoughts on some of this, but please. Thank you, Mike. Suresh, IgG4, cholangiopathy. I was wondering whether with patients who've diagnosed with IgG4 disease, whether in the bile duct or the pancreas, do you screen them for other organ involvement? Or do you have someone who has a colleague who has an expertise in that area, and you have a rheumatologist in our city who is very interested in this, that you send them to screen for other organ involvement? Good question. So for IgG4 cholangiopathy, other organ involvement is crucial for the diagnosis itself because it's very difficult to make it simply on the bile duct alone. And the most common organ involved is pancreas. And so looking at the pancreas closely and making sure you have a reasonable suspicion that there is also pancreatic involvement is great for making the diagnosis of cholangiopathy. Beyond that, whatever imaging techniques we use often pick up, for example, you do a CT of the abdomen or MRI of the abdomen, and you pick up a whole bunch of these intra-abdominal other organ involvement. The head and neck ones are visually seen, like a swollen eye or lacrimal gland or parotid gland. So is the chest that, do I do a chest CT looking for something? No, I don't. But if it's a chest CT is already done, or the abdominal CT picks up something in the lower chest, that's useful. I don't do a PET scan looking for something unless I don't have any organ to biopsy and I'm looking for an organ that lights up. But the yield is low and these existing data don't provide a clue that there is another organ involved. So there's no benefit to finding asymptomatically involved organs for the sake of treatment because they all respond to the same treatment. And if there's a minor involvement, it'll disappear with steroids. So from a therapeutic standpoint, you don't need to look. From a diagnostic standpoint, if you have one that is biopsiable or has easy access, that's great. But I don't go looking for it. Very good. Thank you. And when do you decide to start a patient on maintenance therapy? If they respond to their initial therapy, steroids, how do you decide whether to give them a chance off maintenance therapy, whether to start more maintenance therapy? What's your management strategy with these patients? Yeah, that has evolved over time for us quite a bit. Initially, our thought was to just give a course of sorts and stop and see what happens. And in the very first paper we published, we showed that biliary involvement was the highest risk factor for relapse. So nowadays, if there is... So the biliary involvement comes in three flavors. One is purely a distal myeloduct stricture, which looks like pancreatic cancer induced stricture. Other one looks like cholangiocarcinoma. It could be extrapartic, intrapartic or both. And the last one looks like PSC. Now the pancreatic cancer one is usually responsive to steroids and I don't necessarily use that as a indicator for maintenance therapy. If that's the only stricture there is, this is in the distal myeloduct. But anything proximal to that, anything intrapartic, I strongly consider giving maintenance therapy even upfront, starting with having discussions with the patient that you are somebody who will need maintenance therapy because the risk is about 70% of relapse. And so start discussing that right away. And if the patient is amenable, then strategizing that upfront is useful. Because if you're going to use either thioprine or mycophenolate as your maintenance therapy, then if you wait for relapse before you start that, then you're going to go through this whole thing all over again. So to avoid doing this over and over again, you discuss that in patients with proximal myeloduct involvement or intrapartic involvement. Which medication do you typically- Yeah. So there are three different options. The Japanese have published that data on just using a load of steroids for maintenance, but what they show is that there's a 25%, 30% relapse rate. What they don't tell you is what they do with those patients. I've asked them directly, what do you do with those patients who relapse on steroids? And they don't give you a clear answer. So what we have found is that, I mean, those are the kinds of patients who will need maintenance therapy. So the question is, is it one of the traditional ones, like either thioprine or mycophenolate, or should it be a Tuxman? All our data shows that if you go down the thioprine, mycophenolate route, there's a 35% to 40% chance you'll find out after probably two more relapses that it was not the right drug. Either because there's intolerance to the drug or because they relapse on the drug. And most importantly, neither of those two drugs as an IBD induce remission. So they're only maintenance drugs. They're not for induction of remission. So if you're in the middle of a flare and you just put them on azathioprine and don't give them the sores necessary to bring them into remission, they won't go into remission. So they need this prolonged course of sores to make sure that disease goes into remission, and then you hand over to azathioprine or mycophenolate. I've done that, but I've found that relapses are pretty extensively investigated. If they are not being seen by you, by me, and they're going somewhere else and getting relapses, they got a huge workup again because relapses look like cancer. And most patients get an EOS and ERCP and MRCP and stenting and stent comes out like you spent $20,000 investigating a relapse. So if you look, I was against rituximab upfront because it was expensive. But when I look at the cost effectiveness, it probably is equally effective to give rituximab and make sure a patient doesn't relapse. So nowadays, the ones that I use rituximab upfront are those who have multiple organs involved, those who have already gone through a hell of a time getting the disease into remission, or those who have very high IgE4 levels. And even the biliary involvements, I propose using that. But if the patient says, no, I don't want that, I want to try azathioprine, do that. So that is a subset that will benefit largely, greatly by top-down approach. And others, you can work your way up, so. Great, thank you very much. Can I ask a question of Dr. Tariq? Sure. Dr. Tariq, I think maybe you touched on this, but now there is a biosimilar that the FDA has approved for rituximab. Do you expect that that might change your practice to perhaps favor a top-down approach since cost is now to a large degree taken out of the formula? Yeah, actually, I don't need much convincing that top-down approach is better. It's just that the ability to get insurance approval is not uniformly there. I've never had a refusal, but I see a lot of people complain to me about putting rituximab even on the algorithm because they say they can't get it. And in other countries like Italy, it's extremely difficult to get rituximab for anything. So if you look at that study, it's a very, very small proportion of patients that are on rituximab. So I wish Genentech had done the right study upfront with the randomized control trial and got it approved, but it was a small fry for them, and so they didn't do the right trial, so it's not FDA approved. But there's another company that's coming up with a CD19 antibody, and they're actually going straight for FDA approval and doing a randomized control trial. So in future, we will probably have an approved drug that can be used by a biologic agent and much easier to then use that as a first-line approach. So. Thank you. Cheers. Well, Neil, outstanding lecture on AI. I'm looking forward to seeing what you produce in the future. And one question I have is, how do you prevent your model from being overfitted and focusing on features that are not relevant? Yeah, well, thank you first for inviting me to be a part of this great panel. It's really been awesome. Yeah, overfitting with these models is sort of the monster that keeps us all up at night, for those of us who do these machine learning algorithms. And there's a few things that we can do. One is that when we're actually doing our training and validation is we actually, during our hyperparameter searches, we actually purposely choose specific factors like learning gradients and things like that to make sure that the model isn't too smart. So a lot of these models, like I was showing, especially after 2015, they're so deep. They're knowledge-based. The technology is so impressive that they can memorize these very, very small parts of an image, and they can basically be too smart. And they actually memorize things that are not clinically relevant. So if you can actually control the rate by which they're learning, you can actually stop your model from overfitting. And the other thing that we did in our study is, and I didn't really talk about this in our paper or in the talk is, we actually use the occlusion heat maps during the validation set where we can actually in real time say, gosh, this model is looking at a different part of the screen. It's looking at text. It's actually not looking at the structure of interest. Now, the issue with that is that, potentially we're reintroducing our own biases as clinicians. And that's part of what's great about AI is we're trying to avoid that. We want the AI to find things that otherwise we wouldn't consider. So there's a delicate balance there. The other thing you can do is you can try and be a bit more conservative with your models. And I think as any of us who are reading these journals, there's so many articles coming out these days about AI models that are being developed and everyone's doing it a little bit differently. You might hear terms like cross validation or static validation. When people create these datasets that are not very conservative, where they allow patients to be in both the training and in the validation, and sometimes even in test datasets, you're allowing there to be this possibility that you're overfitting, that the computer is actually memorizing a particular patient and not a disease process. So that is where you can get yourself into trouble. But I think there's a lot to learn. I think, especially with some of these deep learning models, they've only really been around for about four or five years. And I think we're all still learning what's the best way to sort of keep it under control and prevent it from overfitting. And what's the near-term goal for this technology? Do you see this as being applied following an examination? Is real-time examination going to be permissible in the near future? Yeah, I think the goal for a lot of the scope makers is to try and create something that they can put on a device for real-time acquisition and analysis. I think when you talk to these folks, they say image processing and image quality can only reach a certain limit. When we talk about our TVs for existence, we've gone from 1080p to 4K to 8K. And at some point, we're not gonna be able to tell the difference. It's just gonna look great. And that's kind of what the endoscope processors and these other ones are going through, is at some point there's going to be a limit by which it won't matter how good the image is from their end, because to the human eye, there's a limit. So it then goes on to beyond image enhancement to image analysis. And that's where I think these scope makers are gonna be very interested, is creating on-site devices. Olympus just put out a product that can do on-site computer-aided diagnosis for colonoscopy. Medtronic just came out with PillCam Genius not that long ago, which provides real-time AI. So I think creating a real-time device is very important, but I think as clinicians and researchers, we should be trying to create the best models possible. And I think what we're doing in our lab is we're in parallel to creating real-time devices. We're also trying to create the best models that are the most accurate. Because once you actually create the scaffolding and the architecture to perform real-time analysis, putting in a newer model that's able to actually perform real-time diagnostics isn't necessarily as challenging. But I think the challenge is curating an accurate data set, getting correct data, data that's solid, that has gold standards behind it. That's the real challenge, and that should be the focus for all of us. Well, you mentioned two applications already underway, colon polyps, PillCam. Is that currently or likely seen to be a billable activity or service? Who pays patient insurance? Is it bundled care? How's that addressed? That's a great question. And I think, you know, from going to a few talks, the AGA Tech Summit had a recent talk, you know, before COVID where the FDA was actually brought to the representative to discuss AI applications. And they've made it quite clear that the FDA is interested in AI development as long as a physician remains involved. I think that's the key, that it should be computer-aided diagnosis and not replacing a physician. We've seen some recent examples where CMS is providing compensation for physicians, specifically radiologists who use AI platforms. So I think once AI is paired with a physician, and, you know, there's been some amazing trials. Tyler Berzin's group just did a randomized control trial looking at colonoscopy for polyp detection. Once we get more and more of this data, I think billing will soon follow. We're almost struck by how rapidly the technology is changing. I suspect the models will change as quickly. So what happens then? Are institutions now, as we say, CPs or psychology, et cetera, are we going to have to save the image assets to allow re-application of the new models? Yeah, that's a good question. I think, you know, first, I think the data, and I think there's so much that needs to go into this. I think there's a lot of ethical questions about who owns the data. Does the institution own the data? Does the patient own the data? You know, does the person who just decides to make a model out of it, do they own the data? Does the scope maker who buys the product? I think there's a lot of ethical questions that need to be answered. I think when we're actually curating these data sets going forward, I think institutions that are interested in model development should have plans in place for how to not just record images, but to record videos as we kind of talked about briefly in our talk, but also in our paper. And actually Bill Carnes' group at UC Irvine just demonstrated this, that video data seems to have superior performance than image data. And a lot of us don't routinely record US videos. We take physician-captured images, which, and I talked about this briefly with some other folks, you know, when we take an image ourselves, we're biased, right? Because we're taking an image that we think is relevant. The AI might not agree. And that's where video data is actually, I think, superior because you're giving the computer raw data to look at. And that's where I think videos come in. In terms of how we actually implement this, I think, you know, it's gonna take a little while. I think for some institutions, perhaps those that don't have expertise for some disease processes, perhaps uploading videos online to a secure server is the way to do this. And we've seen that with radiology, you know, these folks who actually contract out their CT scans to an outside third party to actually read the CT scans and return a result. Who knows? This is really a new frontier. It'll be really interesting to see where we go. I found it very interesting with the AI and colonoscopy, not even what you, diagnosing what you do see, but rather saying, you know, I'm sorry, doctor, but you didn't look behind that folder. You missed seeing that segment there. You're withdrawing, you know, too quickly through. And I found from a quality assurance perspective, that quite an interesting application that I wouldn't necessarily have thought of. Yeah. And I think there's some really interesting data out there about, oh, you actually need to clean this part of the colon more. You haven't achieved this level of Boston bowel prep yet. And again, I think this is why the FDA has actually been very interested in all of this, because like you were mentioning, from a quality assurance standpoint, it potentially could help clinicians not replace them. Yeah. That's great. We had a couple of questions from the audience and a bit of time left. There was one that had come up. So this was related to treatment of gastric varices. And this is not something that I've done before, but the audience member was asking about detachable snares for the treatment of gastric varices. And also, you know, people's experience with that, but also whether there'd be any comparison data with gluing or with coils. And so maybe Richard, I'll ask you first. Yeah, so it's interesting. So there were some reports of using detachable snares, mainly from our Japanese colleagues. This was a while back. I mean, there've been reports of it. I don't think they generally work that well, that, you know, it's like doing bad ligation of gastric varices. It works as a temporizing measure, but the durability isn't there. But there's also reports of using a detachable snare or bad ligation and also injecting glue as well. So in other words, not just the detachable snare, but do dual combination therapy. That data is really rudimentary. It really isn't there. Part of the issue with glue injection though, I was gonna say is that there's so much heterogeneity in glue injection and what technique you use. You have to really look at the methods and materials of exactly what technique you use, what types of glue do you use, and, you know, do you dilute it, do you not dilute it? Because that affects the polymerization rate of your glue, which can also impact systemic, you know, migration or systemic embolization in terms of the risk of that. Because you really, when you're doing it, you really want a glue that solidifies in the varices, a glue that solidifies right in the varices. Because if you have a glue that takes 16 seconds to polymerize, 16 seconds is a very long time in endoscopy. And that glue that you inject in the gaseous varices could well be somewhere else. Can I make a quick comment to the question here? So firstly, I was very intrigued by the use of these loops a few decades ago when the first reports from Japan came out, because I've been enamored with loops all along. Still am, but they're very, very floppy and difficult to deploy still. But having said that, I think it's really important to distinguish esophageal varices from gastric varices. Band ligation works very well for esophageal varices and the reason why is because the varices are located in the lamina propria. The gastric, and they're small, right? So they're a diameter that can be sucked in entirely into the cap attachment. By contrast, gastric fundal varices certainly, the typical ones, are large, but more importantly, they are located in the submucosa. So there you're really seeing the tip of the iceberg. This is why sometimes these gastric varices are misinterpreted as being gists, because they don't have that violaceous hue to it like the esophageal varices do, and the red whale signs, they're much, much deeper. What that means is that it doesn't matter whether you use band ligation or you use loops. You're not gonna be able to really get the entire varix inside of your loop or your band. And if you're not getting the entire varix, then when that thing necroses as it strangulates off the varix at its base attachment, it's going to bleed. You're gonna have exsanguination. And so that's the caveat about using bands and loops for gastric varices. So that would be my recommendation. I have to say something about the glue because the difference in polymerization time. Richard's correct. These differ in their polymerization time. And so what I use is the 2-octylsinoacrylate, which has a three to four-fold longer polymerization time compared to the N-butyl 2-cyanoacrylate, the histochrome, the original. But the reason why I chose that is because I use the coil and glue technique. And here the idea is that I'm literally dripping in, drop by drop, the glue and giving it time to attach to the wooly fibers, to the synthetic fibers on the coil. So there's a reason for choosing a glue with the long polymerization time. Now there's the added benefit that our nurses aren't all nervous about having to inject very quickly and the glue solidifying inside of the catheter or the FNA needle in this case. So that's another advantage. But the real reason is that I want to drop it in, drop by drop, so that it attaches to the fibers. If you inject vigorously, it'll just flow away. The fibers won't catch the glue. So that's the reason for that. It's a drop by drop. And typically I'll inject it over 30 to 45 seconds. So it actually is very leisurely, no stress to inject the glue. Yeah, no, I think Ken brings up a great point. I think, you know, as endoscopists, we have to understand in what platform we are doing glue injection. I think it makes sense for, if you're going to use coils with the fibers and stuff, what Ken says about using a glue that polymerizes a little bit less quickly. But whereas if you're doing, you know, direct endoscopic injection without coils, I think you want a glue that polymerizes right there and then, and doesn't take too long, you know, to decrease the risk of systemic migration. When I came to the U.S. from Germany, where I was using the histoacryl, histoacryl wasn't available in the U.S. And the only one that was FDA cleared, of course, not for varices, but for topical, for skin closure was Dermabond. That's the 2-octylsinoylate. So that's all I had that I could use that was at least FDA cleared. So that's why I started using that. But the question really is, it's never been answered. Would a drop-by-drop injection, a very slow injection done conventionally, endoscopically, will that produce equivalent or even better results compared to the very rapid injection that you do with N-butyl 2-cyanoacrylate? Because we don't know. I mean, it could be that you actually reduce the embolization or the migration risk just by going drop-by-drop, because there is vigorous stream. You see that on the U.S., how the flow is. So if you're like injecting really vigorously, there's a fast injectate there, maybe it'll just flow away and won't have a chance to solidify on site. So we don't know the answer. Maybe someone would do the study and tell us. Right. Good point. Excellent, thank you. And then quickly, just cleaning the scopes. That is an issue, and we're facing a question now. We use acetone to clean the scopes, and we use nail polish removal. It's cheap and it works, but some centers, and we're questioning whether it has to be medical-grade acetone, which is extremely expensive. What do both of you do? You mean to clean the scopes if you have glue on them? Afterwards, we always clean the scopes, the channels, the elevator with acetone. We use nail polish remover, which is about 70% acetone, it costs nothing. But medical-grade is extremely expensive, and we're getting a push to start using that. I would say, at my institution, we haven't routinely done that after glue. But what we do, and I don't know, this is just theory, but what we do is we actually flush olive oil through the accessory channel, the biopsy channel first, with a thought that that may, again, it may decrease the risk of glue sticking onto the scope. It hasn't been proven, but that's something that we do. I learned this from a colleague who actually used silicon oil way back when to do the same purpose. But as we know, silicon oil is hard to get off the scope and clean, so we don't use that anymore. But we've been using olive oil. We don't routinely use the acetone. I don't know what Ken does. So when I was in Hamburg with Nibsahandra, we always flush the channel with silicon oil, and then we wiped some of the silicon onto the end of the scope. We didn't necessarily wash it off afterwards unless if there was something seen, then of course we would. But we would inspect and see if there was any glue on the end. But we intentionally kept the tip far away from the injection side to avoid contamination. That was the ideal. But I have to say, since doing EUS-guided, which is a long time now, maybe 10 years, at least 10 years, I have not done conventional anymore. And I'd probably have to relearn it. Come to you, Richard, and learn it again. But basically, there, I just wanted to emphasize, if you just push the F&A needle, the sheath, I should say, the F&A needle's retracted within the sheath. But if you just advance the sheath forward, then pull out the whole scope, and then wipe off the tip of your sheath, then, and you can use anything for that, really, just because I don't even see glue on that end. Then pull it out of the working channel after everything's had a chance to fully solidify. What's really key is that it's solidified, because if it's solidified, it'll just come right off. So it'll come off very easy. So that's key, that you don't want it. Yeah, anyway, so that would be the protocol right now. Yeah, just one last point, the key thing, I mean, we teach a lot of trainees at our institution as well as many other institutions, but we've actually, more recently, we've taken the suction off the scope as we're injecting. We're taking the actual suction device off so that even if you press the suction button, nothing sucks up. So that's what we've done, because we have a lot of trainees. Yeah, that makes sense. Good tip. Thank you. Well, just a couple of housekeeping items, first of all. If you've not had a chance, and this is for the attendees to view the presentations, or if you missed some of the live Q&A, all of the content will be available on the living meeting on demand until February 16th, 2021. And again, at the conclusion of this session, you'll receive information on how to obtain a certificate of attendance, CME certificate, and or ADM mock points. I want to thank Jennifer for all the work. You had an outstanding job. I want to thank everybody behind the scenes at AASLV, ASVE, and certainly the speakers. Fantastic job. I think we've all learned something today, and to the attendees as well. Thank you. Yeah, thank you. Thank you. Thank you, Michael. Thank you, Jennifer. Fantastic job.

endoscopic procedures

stiffening wires

multiple biopsies

tissue samples

bile duct assessment

gastric varices treatment

appropriate techniques

treatment materials

gastrointestinal endoscopy

research

training

proficiency enhancement