It's great to be speaking about colon and colon EMR, and specifically about real-time endoscopic decision-making for colon EMR. We make a lot of decisions in colonoscopy every day during our detection, during our diagnosis, during our treatment. We're constantly making decisions, and how we make them, the confidence in which we use the evidence, their efficiency, that really makes the differences, I think, in our outcomes. In this particular case, knowing how to detect things, that thickening of the fold, a little bit of the erythema, drew my eye, but I can't see it. My next decision is, how can I see this completely? Would I use a biopsy probe? Would I have the cap? Do I retro? Because that's accessible to me in this spot. The retroflexion shows a reference standard with the scope, so it shows me that this is a small lesion. It looks having morphologic features that are concerning, that friability, that ulceration, the heaped-up edges. I'm using all of this constant information to input and make real-time decisions is what I'm going to do with this small lesion in the right colon at the hepatic flexure. Based on tools we have available, based on classification systems, all of these things give me confidence right here that this is a deeply semicostal invasive lesion, and that endoscopic therapy would not be beneficial, curative, and can be at high risk. What do I do in real-time? I tattoo this lesion for surgical referral, and I biopsy not at those heaped-up margins, but I biopsy that area that hosts the cancer for a confirmative for the surgeon. And I document in my report that this is not a lesion that's endoscopically amenable to endoscopic resection. So this shows a well-differentiated T1 adenocarcinoma with no lymphovascular invasion. In contrast, here's another example of decision-making we have to make. This is a little bit like what Dr. Cohen showed earlier. You see in front of you this large lesion. There's some vascular vessels coursing across there, some erythema. But if you actually take the longer view back, you can see up top that the vessels just stop. So this lesion I may not interpret well up front, but I'm using all of my tools, my detection tools, my diagnosis tools, to see the borders of this lesion from the distance. And then I'll get closer and look more at the vascular, the surface patterns to give me information. I'll expand it with air more to see if there's more prominent nodules to help determine optimal therapies. And in this particular case, even though this is much larger than the previous lesion, piecemeal EMR is an appropriate therapy for the tubular villis adenoma with focal high-grade dysplasia. So endoscopic mucosal resection is safe and effective for large nonmalignant colorectal lesions. We base our decisions on this assumption. We base this assumption on the evidence to date. And that we also consider the alternative, which is the surgical high morbidity and mortality after surgery for large nonmalignant colorectal lesions. We need to be mindful and informed, both as providers and as the patients, that referring patients for lesions that are nonmalignant have mortality rates approaching 1%, major event rates of 14%. So all of this information informs my decisions in real time. And so together, let us optimize the decision making. We'll have an endoscopic diagnosis and removal of colorectal lesions to prevent cancer incidents and need for surgery. I want us to interpret features to improve our confidence in diagnosis and resection by using image enhancement, applying classification systems, and distinguishing lesions that need surgery because that's important to distinguish. And then identify strategies for dynamic injection and select the best practices for resection. And in decision making, there are many components. There's the mindset, the technique, the tools. I'm focusing in the next six minutes or so on the technique more. And the reason that we need to think about classification systems and think about sort of algorithms in our decision making and help us put it into there is because a polyp is not a polyp. And even a nonpolypoid polyp is not a nonpolypoid. They come in these different morphologies and with potential different histologies and properties that can make endoscopic resection easy or challenging. And so using the PARIS classification system is key in informing us in that decision making path. Categorizing them into polypoid and nonpolypoid is important. We treat a penunculated polyp different than we would treat a sessile than we would treat a depressed. So these are important for us to consider. And in nonpolypoids, particularly, we need to classify that even further. So our mind sees the polyp, we see it's nonpolypoid, and then we want to say, okay, is this a polyp like Dr. Cohen showed earlier that had a nodule? So is this granular mixed type that has potentially a higher risk and we're going to approach this differently? Is this non-granular with the pseudodepression, which we could be most concerned about with the risk of submucosal invasion, as well as it's non-lifting properties and difficult to resect. So depending on these properties of a lesion up front, it's not size because a flat granular homogeneous that's 10 centimeters is easier to resect sometimes than a two centimeter non-granular. So it's not size that we consider, we're mindful of it, but it's these other properties that inform our decisions of our optimal treatment approach. So this is just an example of that. Here's a lesion here and you can see it injects so nicely. This is a flat non-granular homogeneous type lesion, and you can see it nicely lifts and then it can be nicely resected in a piecemeal faction for a curative resection. As opposed to this non-granular lesion, you can see non-lifting properties. We're more concerned about advanced histology. We want to get this in an unblocked fashion. Can I get this in a snare? Can I capture it in a snare and resect it unblocked? Because that's what I want to do so I have accurate pathology. So you may try this and if it slips, then you would transfer over to potentially like a hybrid ESD or such in this case. But you can see here with the snare and with some intention and some patience, this can come off in one piece. And so now we're able to get an unblocked resection here and have an accurate pathologic staging of the lesion. So there's another class of lesions that we really need to be aware of that the features of these are very important in our practice. And if you just take a moment and imprint as many images as you can to help you with your diagnosis interpretation skills, these are all deep submucosal invasion. These all have features of this. Features of redness, firm consistency, expansion, fold convergence, deep depressed areas. These are all features of deep submucosal invasion and these are not amenable to endoscopic resection. There's a risk of lymph no metastases. There's a risk that there will be significant bleeding without cure or perforation. So be aware of and imprint these images because these are ones that are better served with surgical resection. So let's move on to the next moments for the EMR technique. And there are really some key things along your decision path that help you with this technique. So optimizing the position to six o'clock, that's going to be beneficial in every scenario. Getting an injection up and molding it so that a flat lesion can become more polyploid, allowing you to get it on block when possible, but not being greedy. So another decision tree that you'll see is when should you take something on block and when should something be piecemeal resection. Using a stiff snare is very important and placing that snare so that you get a normal margin. Again, taking the time here to capture normal around. And then once you have it, being decisive in how tight your snare is and loosening it to release any entrapped muscle, but then tightening it again and a quick step on the pedal allows you not to have much burn. So you can see minimal burn there, clean edges, and you can assess those efficiently for any residual. So moving to the dynamic injection, why is this helpful? It helps, as I said, to mold that semiucosal into that polyploid shape. It distances those deep semiucosal arteries and musculopropria, and it can facilitate efficient inspection of the defects. So here is a flat lesion here. How do we make that into this? Well, you want to keep the needle tip right under the, at the top of the semiucosa. And so when you start to inject, you want to pull your needle catheter back closer to you so you always stay right at that. And as you're doing it, you're directing the path away from you. So you're almost pulling it and looking up and suctioning the lumen so it's more pliable. And then you can see here, if it will move forward, it creates this nice bleb. Here's another example. So here's a lesion that's over a fold. How are you going to feel confident in where the borders are? Or how will you feel confident in, in removing this in one piece? So you can see here again, my decision is I'm going to inject right there so that, and then I'm not going, I'm going to pull back the needle as I'm injecting. So look, it expands. And then I'm going to flip to the left a little because I want more fluid going in that direction. And now you've made this nice big bleb and can get your snare around it. So I think here's a schematic of that. I have just a few more to show the last moments of, again, you want to optimize the position to six o'clock. That's a key first step. Get your diagnosis of the predicted histology. Be intentional of where you inject for dynamic injection, keeping that fluid. And then as well as when you are going to do the snare, keeping it in a plane, parallel plane to the defect and, and having that stiff snare, again, taking the time so that you can get enough tissue, but not too much tissue. And in this particular case, you'll see in a moment, one of my decisions is you want the, always to have it close to your scope and you want to be twerking to the right just a little bit. So that snare goes against the somnucosa when you're closing. So I think we'll stop there with the take-home points being characterized colorectal lesion size, morphology, and predicted histology in real time. You want to choose appropriate strategies to optimize the safe, efficient, and complete resection. And you really want to be intentional in your decision-making. Use existing classification systems, use existing evidence for your decisions. Thank you.

real-time endoscopic decision-making

colon EMR

visual cues

treatment approaches

PARIS classification system