Thank you so much. I would like to thank ASG and the organizers for inviting me to come and present today. I think that we've learned over the last years that ADR, or adenoma detection rate, is really an important quality measure. It's associated with interval cancer and interval cancer death. So we made several, let's say, moves to improve the ADR. In our practices, we've learned from Dr. Rex how you can make it happen in your own practice. Despite the fact that our ADR are meeting our current benchmarks, we are still facing some post-colonoscopy colorectal cancers. So this session is called a breakthrough to excellence. So if you think about the breakthrough, AI is certainly a breakthrough, but the question that I would like to discuss with you throughout this case is whether it will lead us to an excellence in preventing colorectal cancer. So I will start. These are my disclosures. So I will start with this case. So you can see that AI is already on. You can easily spot the ileocecal valve. And immediately, when we pull back, both our eyes and AI can really easily recognize a lesion, a sessile lesion on a fold in an ascending colon. But then we pull back. The quality of bowel prep maybe is not optimal. We have Audrey Calderwood here. So it's probably not Boston 3. It's probably Boston 2. But then I have a question to the panel. Can you spot another lesion? Is it right in between those folds there, Michael, in which there is a subtle redness and an erosion right on top of it? Yeah. So we are probably talking about this lesion. So then I will move on. And the AI is still on. So it is on. It showed some folds. We are looking at the lesion, which was really nicely spotted by Pratik. Then the AI helps us to recognize some stool in the colon. But it doesn't show at all a lesion that we were talking about. And just have a closer look how it looks like. It even spots the tip of our snare. Then we are looking really close on this lesion. And this is like a 2C lesion, a small 2C lesion, which was not spotted at all by AI. We lift this lesion. We can appreciate how it looks like. Even right now, after injection, AI fails to recognize the lesion in the middle. So we will proceed to remove all of this lesion with a snare. After generous injection, we use a snare, as recommended by Doug Rex, 15-millimeter snare, stiff snare. We engrasp the entire lesion. Is that what you recommend all the time, Doug, 15-millimeter standard? For everything. I mean, everything, right? Brainwash. We chop it off. And here it comes. Let's wait a little bit longer. So here it is, how it looks like. So the point is that I would like also to emphasize or to give you some take-home message is that it is really important that we have to know what we are looking for. I mean, the new technology might also not be knowing what we are looking for. So we have to train our eyes and really nicely spot lesions which are difficult to spot and which may further on translate into post-colonoscopy colorectal cancer. Question. What made you decide to inject that first, as I guess, and lift it, as opposed to just taking that pilot, that particular one? Could you repeat your question? Yeah. I guess I don't know that I would have necessarily injected that one. What was the decision? Yeah. I think that with this kind of depressed lesion, colorectal 2C lesion using a Paris classification, I think that we want to inject to ensure that we can ensnare as much as possible and have a generous margin. I mean, it doesn't look like a cancerous lesion, but it was a high-grade dysplastic adenoma. So you want to remove it on block and with reasonable removal of normal mucosa. So again, I think, and Doug, perhaps you can comment on your paper in Gastro about the recognition of these lesions. But just to Michael's point, and even though this is entitled Breakthrough to Excellence, and I do think AI is a breakthrough. But also remember that this is not magic, right? I mean, AI is not God. It is being trained in a specific way. And so you want to make sure that's part of the thing that I think all the societies want to know is how are these models trained? Because right now it's just a black box. Was the training done on different types of polyps for that machine? Was there a good representation of these 2B type or 2C types of lesions as opposed to only 2A types of lesions in there? So I think that's an important point to recognize. Doug, you want to make a comment on that? Yeah, we looked at about, I don't know, a thousand consecutive polyps and found that about 2% of them were just not detected. And they tend to be larger and flatter like this lesion, particularly large, flat serrated lesions. We published this in Gastro just a few weeks ago. But at the same time, at the other end, there clearly is a group that they're smaller on average that are detected. So it does increase ADR, but this is something that has to be fixed. And as Prateek was saying, it should be fixable by getting the right training set, a big enough training set. It's just a computer program and it'll only recognize what it's been trained to recognize. So I think we'll see generations of CADI. I was impressed that it improved your ADR, Doug. It did. Yeah, by my calculation, it did still improve my ADR, but at the same time, there's a set of lesions that it misses. And hopefully in the next couple of years, we'll see these problems fixed. Thank you so much. Okay, thanks very much. Okay, I'm going to skip the first one because I think we've beaten up the serrated thing pretty hard. But I want to ask about this case. This is a patient who is referred, and this is in the cecum right next to the ileocecal valve. There's a couple of flat lesions here. You can see some bumps out here and some deformity of this fold. And so these are lying right on a scar where previous EMR with electrocautery has been performed. So basically, there are two areas here of recurrent polyp. Michael, you described using a cold snare for this. Can we agree that we don't have a randomized trial or comparative data of how to deal with these lesions? You'll have to specify which Michael. Professor Burke. No, we'd do this hot. So I'd inject that just at six o'clock there. That's not scarred. So it is a scarred recurrent lesion, but I'd inject that and chop that out. So you'd inject right into the scar? I mean, you know what? It oftentimes won't lift when you inject it. Right at six o'clock there. I don't think that's scarred just there. Scarring's up the top there. I think the scar extends all the way underneath this to the ileocecal valve. Just pretend that I'm right for a minute. That's hard for Dr. Burke to imagine. I can't agree with other people. So you would use electrocautery, but there's no comparative data, right? No. I would also say that when we're removing large lesions by cold EMR, which is associated at least for adenomas with a higher recurrence rate, and we still have a scar, right? I mean, just because you remove something cold doesn't mean there's not going to be a scar when you come back. Okay, so I just asked the panelists, how many would remove this cold right off the bat? Cold? Okay. And how many would go hot? Okay. All right. I went hot too. But I will tell you, I always go hot on recurrences. I'm trying to get the hot versus ... Okay, so here we are. So one thing is that we're using a bigger snare because hot snares are, as you already heard, they're a little bit ... The braid is a little bit thicker. And so there's our first piece coming off. And I think there's possibly a little nubbin that I didn't get. So the snare, in my view, is tending to bounce. Here we're trying to take the second piece off. And I just want to get at, what's the best way to approach these? I find that the snare often sort of bounces. It tends to slide, even when I push pretty hard when we're dealing with a scar. And so we've left some residual disease. Here there's probably a little bit right here, there may be a little bit down here, maybe a little bit right there. So what's the next step? Well, with the snare initially, with your assistant, you don't need to look at what you're doing. You just put the snare around the polyp and then we go for a full redout. So we don't even watch. And I just watch the assistant's hand. And then when she says, or he or she ... So a full redout means you're, by suction, you're getting a redout? Yeah, it'll collapse everything. Because now, because clip closure is so effective, we don't need to worry about perforation. Never worry about perforation. Just go for gold. How many in the audience are still going to worry about perforation? There's a few hands out there, Michael. Take all the gas out. So you put the snare down and you close a little so you see that it cinches in. And then you take all the gas out and you just watch your assistant's hand. And they tell you when they have resistance. And it's usually the same sort of distance. And then you re-inflate and you see that you may have actually captured what you're after. And then you can cut through hot or cold, whichever you prefer. That's a weird normal move. You know that I sometimes will, because I have a cap on the end of it. Yeah, you can do that. I'll put the cap up and suck up into the cap. Yeah, that works great as well. Okay. Well, I didn't do that. I don't know why. I think I was having a hard time approaching it closely. So I decided to switch to Evulsion. And I'm going to do this hot Evulsion. So Evulsion, you use it, I know, less than I. How many are fans of Evulsion? How many, when you do Evulsion, do you do it cold, by show of hands? Most of the people. Anybody do hot Evulsion at all? Okay. Thank you. What about the audience? Thank you, Amrita. Okay, so. Do you want to pull the audience? So also, we don't have a comparison of hot versus cold Evulsion, right? I mean, I'm going to make the argument for hot Evulsion that I think it's also very safe. And that, for one thing, it leaves the field dry while you're working. You know, you're not bleeding. There's no bleeding. I actually just rotated the forceps. If you take the electrocautery out and you spin the forceps counterclockwise, you can actually get them to rotate, which sometimes is an advantage for grabbing the tissue. So for hot Evulsion, we have this on endocut I, on the 141 setting. So it's almost pure cutting current. But on endocut I rather than endocut Q. And then we grab it and then tent it up and then just tap the yellow pedal and the tissue will peel off. That's taking a little bit longer than it usually takes to peel off. So, Doug, what do you think is the advantage of hot here versus cold Evulsion? You know, the main thing is, I guess, you know, I mean, we're dealing, we don't have a randomized trial. So we're dealing, I think, I think to the audience, it ought to be some of some interest as to how to approach this. But one advantage is, look at the field. It's totally dry. There's not a drop of blood on it. And we're getting down into, this is submucosa in here. We can see, you know, the residual disease well. So it's probably partly what we're comfortable with. But I just want to bring out, I think there's more than one way to skin a cat here. But in any case, so it looks like all the, at the time, I thought probably all the disease was off. So are we done now? Michael, you would be, would you get the snare tip out and burn up the margin at this point? Yeah, yeah, go for it. Big lot of snare tip there, baby. Yeah, burn away. Okay. You can't do enough snare tip. It's so safe. I wouldn't do this. I absolutely would not do this. So Michael, Michael hates APC. So I show this video just because I... You're just trying to bait me. Michael, why don't you like APC? Because we showed in a randomized trial in a porcine model that it's, the depth of destruction is unpredictable. And so it's not as uniform as a snare tip. And the electricity arcs to unpredictable locations. And it's expensive as well. It's expensive, yes. Compared to the snare tip. Okay. And I'm going to give you that. I think I should have probably gotten the snare tip. So three nil. Three nil. But it's been shown to reduce recurrences. No, it's, APC is an independent risk factor for recurrence. But that's only if you use it to ablate visible polyp. I know, but that's the main way people use it. Not the margins. If you use it to ablate the margins, it's been shown to reduce recurrences. But it's probably not as effective as the other technique. Okay, so we have to clarify, right? If you use anything to ablate visible residual polyp, whether it's APC or something else, you're going to get a higher rate of recurrence. But now we're talking about margin treatment. It looks like we've removed the whole thing. And our choices to burn it up would be snare tip soft coag, which is clearly less expensive. And in Michael's big trial, it worked better. And we don't know for a recurrence. And I, you know, sometimes I guess I like that brown char. That's not a good excuse, is it? But I still use it sometimes for treating the margin of a recurrence. But I'm going to stop now, Michael. But when, you know, you probably do a lot of barbecuing, right? How do you know the meat is cooked? You know, not by the surface. You can't tell whether your piece of beef is cooked or not. You can only tell by the time it's on the rotisserie. Every time I want the APC catheter, I'm going to remember the Barbie analogy. But there is, I think, one reason why people are using APC. It's relatively easier than soft tip coagulation. I mean, soft tip requires more tip control. Therefore, especially in this situation, when you have a little bit unstable position and difficult access, then APC may be a way to go as well. I hear what you're saying. I'm not sure I fully agree. I kind of like the fact that you're in contact, you know? But you're saying you have a very light touch on your STSC, right, Michael? Yeah. I heard you say that. Yeah, just a light touch. But the movements, it's a good way to start to learn the movements for ESD. Yeah. Because the movements are very similar. Okay. So we've treated our recurrence. And, you know, in our experience, this, I mean, even when it looks perfect, I would not, I know you say 100%, but we have, it's about 90% for us. We follow this patient up in another year or so, and in a small percentage of cases, we'll see something. But let's talk about it right now. Are we done? Or should we clip this? I don't know the answer, but I did, I did clip it. No, you have to clip it. You have to, you know? So when you're describing it, you feel like you have a type two muscle injury when you do, because Michael would do this with cold evulsion if he decided on evulsion. Then he's going to burn it with the snare tip on soft COAG. And then you've sort of charred it white. You can't see the muscle. So now you're going to clip it basically to reinforce it. Right. So again, I'm missing the point. Why are you clipping it? Because we don't know, because, so when we do EMR, you know, like the usual situation, you see a blue homogeneous mat. And that mat tells you there's no injury to the muscle. And if we don't see the mat, then we use topical submucosal chromoendoscopy. We put the blue dye on to make sure everything stains blue. And then if you have fibrosis, a scarred area, you don't know whether the muscle's been injured or not. And we showed in that big publication in, it was based on 1,000 cases published in Gut in 2017, that there's a risk of delayed perforation. It's very low. So your own individual experience won't tell you. You know, the risk is very low, but it exists. So I think whenever you see, we call that deep mural injury type 2. Whenever you see that, type 3 is a target sign. Type 2, you can't be sure the muscle's not injured. So you should close that. Just that area. Not the whole defect, just that area. Yeah. Because that looked very superficial, what he removed. I mean, there was nothing deep in there. Can you say 100% the muscle's not injured? 100%, can you say? I thought there were no perforations in Australia. This turned out to be a good question. We just published a paper of 100 perforations, colonic perforations, after EMR. But we fixed them all. Yeah, so. So Pratik, I would do it for the same reason. It's on a scar. The layers are fused together. And I, you know, I did put some thermal injury in there between the snare and the hot avulsion. So it just, you know, it makes me nervous enough that I can't tell for sure. So I closed it. So there you go. Okay. You know, we've only got a few minutes left. Michael, want to show some cases? Yeah, well, just on perforation. Okay, good. Something you're not worried about. You shouldn't worry about it, so long as you look carefully. So this is the reality of perforation. If you don't recognize it, then it's really a disaster. But if you recognize it, it's nothing to be worried about. So we have, you know, instrument-related perforation. This is a disaster, but it can be very difficult to fix. But now, with the clips that we have at hand, you know, you can so easily fix this. And this is a 12-month follow-up of this particular patient that had a sigmoid colon perforation. And then we get other types, you know, with advanced lesions that may invade the muscle. This is a clear risk factor for deep mural injury. So we have to recognize the injury here. But generally, with EMR, the perforation's tiny. You just have to see it. And so you have to look at the defect very carefully. And it's often related to transverse colon location. We published this in that gut paper and advanced histology. Because of submucosal fibrosis, the lesion becomes adherent to the muscle layer. And particularly if it's in the transverse colon, it's not relatively fixed. So, you know, the whole muscle layer can be invaginated in the snare relatively easily. And the other thing to be aware of, this is an ESD-related perforation, is you must close the full injury. So here, we've closed this up and we pat ourselves on the back. But we didn't recognize that actually the muscle here is a bit damaged. And here, it's just sublethally injured. And then the next day, the patient got terrible pain and had a perforation at 30 hours. Luckily, we're straight on to it and, you know, laparoscopically repaired. So this is the problem there. But we can also embrace deliberate perforation with EMR technique. So with small early cancers, we publish this as well. We deliberately excise the full lesion with a snare. We make an injection and we take a large snare. And we deliberately take the whole lesion and the bulge around it. And this is the margin. And we can cure a lot of early cancers that way. And we just plan to close it up. So with the advent of CO2, it's become very safe. And this is this type 2 injury. The muscle may be stuck onto the submucosa and there's fibrosis. So... Michael, can I ask you, with this type, these kind of injuries that you're talking about here, because some of them are clean through holes. They're type 4 injuries. Yeah. And so you close them up and then do you admit them all or do you let them go home? No, you can let them go home. Yeah. Yeah. So in that study, the 100 cases, 90% of them went home. So this is... So just see, we'll go back to the beginning of this video. This is a good video. So I just go previous and go back. In those cases, Michael, do you give antibiotics or any other treatment when they go home? We give them antibiotics. But I wouldn't say, look, you know, I don't see you're okay. Away you go. You've got a bit of tummy pain, but you'll be fine. No, we keep them there most of the day. But if they're fine, I think they can go. And you don't need to do a CT if you've closed it up and you're very comfortable with your technique. But let's watch this case. This is just while we've got a bit of time. So this is the start here. This is one of our fellows from a couple of years ago in action. Transverse column, like I said. And then... So if we just... Michael, you're blaming the fellow. You're blaming the fellow for this? You can admit it was you. Look, I've made many, many mistakes, so for sure. But we'll pause it there. What did people probably saw just for a moment there? Did you see what we should be worrying about? Jennifer, maybe you want to say? Prateek distracted me and it seemed... Audrey? I think we saw like a little blackness, but it seems to have closed up now. No, yeah, just we go from here. So play, sorry. So that's the piece. And yeah, so you see there's no hole, but it's not blue anymore. So he should have used the water jet here, but... It's a him. He doesn't. He re-excises the whole thing. What do you think is going to happen here? Look at that. And at this point, you know it's not going to be good because the snare is taking a long time to cut through. So now we do have a hole. So the first thing is, so what, Doug, what do you do when you see? I mean, you probably see one of these every day in your practice once a week. Guys from Australia. So, yeah, but I mean, I would always go for hemostatic clip closure because even this is a small hole and, you know, it's just it's just faster because you stay right there. You make sure nothing goes out the hole. And in my experience, it's it's really reliable and usually start at one end or even off the end. So you're you're not even over the hole and draw the tissue together. Place the clips really close together. Make sure they're buried all the way up to the stem of the clip. And it seems very reliable. Marco, any thoughts? Yeah, I think that I mean, crucial thing is to have an overlap of normal mucosa that you can drag over the hole. Then it makes it really effective. And the other point is that, I mean, you shouldn't play just one clip or two clips. Usually you require several clips to close it really fully and efficiently. Yeah, you want to have you want the whole thing reinforced and you want to be absolutely sure you've got it just solid. And then you can be you can be confident if the patient's fine, they can probably go home. I would admit this patient, though. I would not send him home. Yeah, this is America. Yeah. Yeah. We got lawyers here, man. I'm not saying you should send them home, but you could. But I also think the key here is recognizing, right? Which you. Yeah, that's right. So the whole thing, if I could leave the audience with one message, it's all about careful interrogation of the defect and recognizing that there's a problem and then that then then you can handle it. You deal with it and you're good. OK, on behalf of Audrey and myself, I'd like to thank, you know, all the panelists, speakers, moderators. Excellent. I mean, this was one of the best sessions we've ever had on colonoscopy, guys. So thank you all very much.

adenoma detection rate

colorectal cancer

artificial intelligence

lesion detection

perforation management

complications prevention