Our next session, which will be on Barrett's esophagus, 2024, it is a pleasure for me to introduce Dr. Vanni Kondak, who is the Director of Center of Esophageal Disease at Baylor and is an Adjunct Associate Professor at Texas A&M, who is going to be talking to us about best practices in surveillance of Barrett's. Thank you, Irving, and thank you to our course directors and to the ASGE for the opportunity to speak today. It is my pleasure to be speaking on surveillance of Barrett's esophagus. Here are my relevant disclosures. So as we know, the incidence and mortality of esophageal cancer has been steadily rising over the past several decades. Fortunately, we have a precursor lesion. Barrett's esophagus is a known risk factor for esophageal adenocarcinoma and is a complication of chronic reflux. It is the replacement of esophageal squamous lining by specialized intestinal metaplasia. And it requires two criteria for the diagnosis, both the endoscopic criteria as well as the histologic criteria. Endoscopically, we want to see salmon-colored mucosa in the tubular esophagus. And when we see this, we want to biopsy and perform at least eight biopsies to reduce any event of sampling error. On the biopsies, we want to see columnar-lined epithelium with the presence of goblet cells. Unfortunately, we do have a high inter-observer variability among pathologists for the diagnosis of dysplasia in the setting of Barrett's esophagus. And this is because dysplasia is often difficult to distinguish between inflammation, which is often present in the setting of gastroesophageal reflux disease in the distal esophagus. Dysplasia is often over-called. We do see that expert GI pathologists have less inter-observer variability. And therefore, we want to confirm all cases of dysplasia with an expert GI pathologist. Now, the exact stage of dysplasia is important because it informs our risk of progression to cancer, which informs our management decisions. When we look at high-grade dysplasia, we see that the annual cancer incidence is on the order of 6 to 12 percent per person per year. For low-grade dysplasia, the risk is 0.5 to 0.6 percent per person per year, albeit this represents a heterogeneous group of people. And for non-dysplastic Barrett's, the risk is 0.3 percent or less. High-grade dysplasia is the best marker that we have to identify who goes on to develop esophageal cancer. And therefore, this is the ideal opportunity to intervene to cure or prevent cancer. And it can be endoscopically treated. Our current standard of care is to perform endoscopic eradication therapy for those patients with Barrett's esophagus and high-grade dysplasia or intramucosal carcinoma. For those patients with confirmed low-grade dysplasia, they have the option for endoscopic eradication therapy or close surveillance. As well for indefinite dysplasia, they may have close surveillance. For those patients with non-dysplastic Barrett's, our recommendations are to perform surveillance endoscopy with a structured biopsy protocol every three to five years. Current society guidelines, such as the ACG, as well as the British Society guidelines also take into account the length of the segment to allow us to tailor the intervals, such that shorter segments may benefit from longer surveillance intervals. Now there are multiple guidelines. And I have taken the ASGE guidelines, the ACG guidelines, and the recent AGA clinical practice update on screening and surveillance guidelines, and have combined a lot of the best practice recommendations. But I've grouped them to make it easier to remember for myself and my fellows to the five L's for a high-quality endoscopic assessment, landmarks, length, look carefully, lesions, and levels. So our first L is landmark. We want to identify and document the diaphragmatic impression, the gastroesophageal junction, and the squamo-columnar junction. Any displacement of any of these would give us the opportunity to identify a hiatal hernia, Barrett's esophagus, or when we have both, Barrett's esophagus in the setting of a hiatal hernia. The second L is length. We want to look at the Barrett's segment length. Long segments are those that are 3 centimeters or greater. Short segment are those that are 1 centimeter or greater and less than 3 centimeters. We know that long segments have a greater risk of progression to cancer compared to short segment. We want to measure the length using a validated criteria such as the Prague classification system. This is where we take measurements from the top of the gastric folds to the circumferential extent of the Barrett's esophagus. And then we take another measurement from the top of the gastric folds to the maximal contiguous extent of the Barrett's esophagus, excluding any other islands approximately. And then we can take each of those distances and designate it with a C and M designation. So in this example, this would be a C2M5 Barrett's segment by Prague classification. Now we also want to pay attention to the esophagitis. And we can use the LA grading system for esophagitis to document esophagitis as well as where it is in terms of location and length. When we see grade C or D esophagitis, we want to avoid routine surveillance biopsies in the setting due to the inflammation. You can still take biopsies if you need to exclude malignancy. But otherwise, if you're taking your mapping biopsies, treat these patients with PPIs twice daily for eight weeks and then bring them back after the esophagus is healed. Now we're always fraught with the question in a regular Z-line. If I had a magic wand and I could do whatever I wanted to do, I would take the word irregular Z-line out of our vocabulary because I think it's confusing. But I would say the ACG guidelines are pretty clear on this. The diagnosis of Barrett's esophagus requires a segment be 1 centimeter or greater. In a patient with a normal appearing Z-line we shouldn't be taking routine endoscopic biopsies. And in the absence of visible lesions, patients with less than a centimeter of variability at the Z-line should not undergo routine biopsies. I like to rephrase this. Don't ignore the Z-line. Look carefully. And if you see a mucosal irregularity, target your biopsies there. Otherwise, in the absence of visible lesions, avoid biopsies. The third L is to look carefully. We heard this morning about high-resolution endoscopy and we want to use the best scope that we have for these patients. We want to be able to look at the detail of the mucosal lining. I like to use a soft distal attachment cap at the end of my endoscope to be able to look carefully between folds of the gastroesophageal junction and on fossa of the mucosa. We also have the option of using enhanced endoscopic imaging. Many of these modalities can be used at expert centers or for individuals who have experience. But the one I'll focus on is virtual chromoendoscopy because that's the one that is recommended for widespread clinical use for the surveillance and screening for Barrett's esophagus. Now we want to improve our technique in the esophagus. We want to perform suction, irrigation, use mucolytics if we need to, to clean the surface of the lining of the Barrett's mucosa. We want to use insufflation and deflation to look for any abnormal contours of the esophagus that we can see in the wall. We want to use tip deflection to look carefully at the mucosa and look at it on fossa. So that way we can see if there's any mucosal irregularities. And most importantly we want to perform a careful retroflexed examination, looking from below to see if there's any abnormalities in the cardia or distal esophagus. In those patients with a hiatal hernia sac we should be able to get up into the hernia and look at the distal cardia from below within the sac. Now in this example we all just saw a cancer. The question is how many of us recognized a cancer? And this is where inspection is critical. We can use Barrett's inspection time as a way to improve our endoscopic examination. More time spent is associated with greater neoplastic lesions detected. It is recommended that we spend at least one centimeter per centimeter, sorry, one minute per centimeter for the segment of Barrett's esophagus. Then we want to train our eye to recognize subtle flat lesions. We are able to identify hopefully now that we can see this lesion from 12 o'clock to 4 o'clock that had an abnormal contour, maybe better appreciated with insufflation and deflation and also easily appreciated under narrowband imaging. We can see the mucosal irregularities under this narrowband imaging with near focus as well as the vascular irregularity. This lesion was resected with wide field EMR and was found to be an intramucosal carcinoma. So this case, it's important to see that we can take our time to interrogate the GE junction. Insufflation and deflation help define that abnormal contour and narrowband imaging with near focus was useful. Digital chromoendoscopy as mentioned earlier this morning is a filtered blue light. There's also post-processing platforms also available. NBI is probably the most commonly used. It does meet the PIVI criteria for imaging in Barrett's esophagus for the detection of neoplasia. We can see the example on the top right showing a regular mucosal pit pattern and the example on the bottom showing an irregular pattern concerning for neoplasia. The BING criteria has been described by international consensus to look at two features, the mucosal pattern and the vascular pattern and dichotomize them into two categories, regular and irregular. And any irregular appearance of either the mucosal pattern or vascular pattern could be a concern for neoplasia. Now unfortunately we're not doing that great of a job in terms of our surveillance endoscopy. We know this when we look at a meta-analysis of patients who were diagnosed with esophageal adenocarcinoma after they've had an endoscopy. And a quarter of patients were diagnosed with their cancer after an index endoscopy performed for nondisplastic Barrett's esophagus. Another VA study looked and saw that the rate of dysplasia detection within 18 months of an index endoscopy was over 4%. These concepts of post-endoscopy esophageal cancer and post-endoscopy esophageal neoplasia are things that we can look at and potentially will be quality metrics in Barrett's esophagus. Now it's important to take a look and not aim for that 60-second exam that was mentioned earlier this morning. In this particular case, this patient obviously has a long segment of Barrett's esophagus. The report that was initially made said no visible lesions, but the pathology report said low-grade dysplasia. So let's look again. The patient was referred with a diagnosis of low-grade dysplasia. And again, looking at this long segment of Barrett's esophagus, I like to perform a white light pull-through first. And then I go back and forth and look at each of the quadrants a little bit closer and focus on each quadrant a little closer, first with white light and then go back with narrowband imaging and near-focus. We can start to also appreciate again with the help of insufflation and deflation, there's an abnormal contour on the right of the screen. Going from about 3 o'clock to 6 o'clock we can see that that lesion actually is a raised lesion extending about 3 centimeters of that Barrett segment. And now we can see that same area with narrowband imaging with near-focus and we can appreciate the regular PIP pattern around that lesion and then the irregular mucosal pattern located the lesion. Now for those of you that were paying close attention, there is another lesion actually located at the proximal extent in this tongue right here at 11 o'clock. You can see a little bit of friability noted in that area, which is all I noticed on white light. But under narrowband imaging with near-focus we can appreciate that there's a distortion and loss of PIP pattern right in the center of that lesion. Both of these lesions were resected and both were intramucosal carcinoma. Lesion insufflation and deflation helped define abnormal contours as did virtual chromoendoscopy. And these were two cancers in different locations found in a case with no visible lesion originally. We want to find those lesions and we want to document those lesions and we want to, we can use the PARIS classification to document visible lesions. And it is important to do that and appreciate that visible lesions are associated with prevalent cancer. Secondary dysplasia by biopsy is associated with a 40% risk of prevalent cancer in that patient. Most of these cases are intramucosal carcinoma allowing us to manage both of these patients together. The risk of submucosal carcinoma is higher in those protruding or depressed lesions and it's rare in the absent of visible lesions. Importantly visible lesions require definitive diagnosis with endoscopic resection. We want to target mucosal irregularities in the setting of dysplasia with endoscopic mucosal resection. This patient was referred with high-grade dysplasia and has a short segment of Barrett's esophagus. Not really any obvious protruding or depressed lesions, but you can appreciate around 1 or 2 o'clock there is an area of mucosal irregularity. Looking a little bit closer again with narrowband imaging with near-focus we can sometimes appreciate a little bit more detail in terms of the mucosal pattern as we can see here. And we're going to see a focal loss and distortion of that pit pattern right in the center of that little tongue. This lesion was resected with endoscopic, and we can see here on a better defined under the retroflexed exam. And this lesion was resected and was again an intramucosal carcinoma. So looking at discoloration, loss of pit pattern and vascularity and recognizing this was a case where we can see that this lesion was upstaged from the biopsies but with a diagnosis of cancer on EMR. When Irving and I were both at University of Chicago we showed that over half the cases had a diagnosis change from the previous resection biopsies after a complete endoscopic mucosal resection. EMR is more accurate than biopsies and changes the diagnosis over half the time. There is less inter-observer variability among pathologists with EMR for the diagnosis of neoplasia. And while EUS is appropriate for nodal staging and deeper T-staging, it's inadequate for superficial T-staging. And the fifth L is to biopsy at multiple levels with a structural protocol. We often recommend the Seattle Protocol to target all visible lesions and then perform multiple biopsies in four quadrants at multiple levels, specifically every one to two centimeters and specifically every one centimeter if there's a history of dysplasia. Unfortunately there are limitations to this protocol. This includes sampling error, of course, but I think our real problem is adherence. We've seen in meta-analysis that there is poor adherence for this protocol and that is associated with increased risk of neoplasia. Also longer segments are associated with poor adherence and unfortunately we know that long segments have increased risk of neoplasia. We have some tools available to help us if we'd like. There's the Wide Area Transepithelial Sampling or WATS brush, which is a brush that we can place through the working channel of the endoscope and sample the lining of the esophagus of the Barrett segment. And then those cells are then analyzed by a computer to identify high-risk areas of dysplasia and then a pathologist confirms the diagnosis. In a meta-analysis we can see that 4-step biopsy detected dysplasia in 15.9% of the cases and then WATS provided an incremental yield of 7%. In those cases, specifically looking at high-grade dysplasia or cancer, we see that 4-step biopsies found about 2% and WATS had an additional 2%. It's also important to recognize it can go the other way around. So among those patients that were diagnosed with dysplasia by 4-step biopsy, 62.5% had WATS negative. So 4-step biopsy was required for the diagnosis of dysplasia, meaning you should be using this as an adjunct tool in addition to Seattle Protocol, not as a replacement. The other alternative situation of a WATS positive and 4-step negative is an interesting one. I think we need more understanding of what to do in those situations and it's probably an area where there's not complete consensus of what to do in terms of what's an actionable item. It is though a possibility for an adjunct tool, again, in addition to Seattle Protocol biopsies. So this is a summary of the 5Ls that I just described in terms of landmarks, length, look, and lesions and levels. Just want to mention a couple of other things in the spirit of the AI session that we started this morning. We thought that the AI might be a useful tool for our Barrett's examination. And we can see in this study by the Amsterdam Group which has one of the largest collections of Barrett's imaging that they did in case of 60 cases with nondysplastic Barrett's and 30 cases with neoplasia. The first section of this video looks at the overall inspection of white light and shows us this CAD-E detection tool with the green bounding box that allows us to see where the lesion is. And then the second part looks at lesion characterization with under NBI and gives a probability for neoplasia detection. And they found excellent sensitivity and specificity in both their image-based analysis and video-based analysis on this study. We can also appreciate clinical factors to help us with the risk of progression of disease to help us identify what that risk is for our specific patients. The progression in Barrett's or PIB risk score is based on male sex, cigarette smoking, Barrett's length, and confirmed low-grade dysplasia. We can see the annual risk of progression associated with these scores here. We can use biomarkers to help us with this risk of progression. Our group had performed a meta-analysis. And among those six case-controlled studies, we found an odds ratio of aberrant P53 associated with the progression of neoplasia of 3.4. And then among seven cohort studies, we found a relative risk of 17 for those cases with aberrant P53 for the progression of neoplasia. P53 is adopted by the British Society guidelines, but not in the United States as widely. And that's because of the variability of interpretation. It might be that multiple markers and a more quantitative analysis might serve us well. There is a commercially available platform called Tissue Cipher that looks at 16 features of the tissue with different markers and provides a risk score and a risk class of low, intermediate, or high. And we can see in a pooled meta-analysis that we can look at what standard of care, which is an expert diagnosis by a pathologist, and see an odds ratio of 3.4 for those cases of low-grade dysplasia compared to those cases of nondysplastic aberrants. And Tissue Cipher of high-risk versus low-risk designation actually provided an odds ratio that met and exceeded that of 7.8. So this might be a useful tool to help re-stratify some of our patients. Now I know this is endoscopy. I just want to say my last closing words. It's not just about what you do in the endoscopy room, but it's also what you tell the patient. Make sure that they're being prescribed the correct dose of PPI. Make sure they're taking their PPI correctly. Talk to them about reflux and counsel them appropriately with any adjunct measures as well as diet and lifestyle. I spend a lot of time on my reflux optimization with all my Barrett's patients. And we also want to counsel patients. Don't leave them with Dr. Google. Higher anxiety drives a desire for more endoscopic procedures, whether they need it or not. Higher risk perception drives a desire for endoscopic therapy, whether they need it or not. Discuss risk with them in layperson language. And this is a useful question prompt list for optimal communication. I also think it's a shame to bring that patient back earlier just because they want it to improve their anxiety. And if they're smoking you let them still smoke without telling them they need to stop smoking. So smoking cessation is also part of the conversation here. In conclusion, it's important to perform a high-quality endoscopic examination with the 5Ls. Detection of visible lesions is critical to diagnose and treat lesions. Don't ignore the variable Z line, but don't over biopsy it. Only target biopsies for mucosal irregularity. And we have computer-aided detection tools that are available and on the horizon for more. And there are additional risk stratification tools. And don't forget to not talk to your patient. Thank you. Thank you, Vani. I want to ask now Dr. Dennis Young, who's Director of Third Space Endoscopy at the Center for Interventional Endoscopy, Advent Health in Orlando, to come and give us some new insights in the management of post-resection strictures. So first off, I want to thank the course directors and certainly the ASG for having me today. It's truly an honor and pleasure. So these are my disclosures. Benjamin Franklin once said, in this world nothing is certain but death and taxes. But I think if he was with us today in this auditorium, he would agree that the third statement would be that esophageal restriction formation is pretty much certain after white field endoscopic resection. And it's incredible how over years with the advancement of techniques and devices, we have really closed the gap with surgery in terms of what we can offer the patient. And endoscopic resection has pretty much become first line treatment of most superficial neoplasia in the esophagus. And the first time you perform some of these procedures, removing large areas of neoplasm, it's quite satisfactory thinking about the organ sparing procedure that you're providing the patient. But soon enough, you also start learning about the dark side of endoscopic resection in the esophagus. And you start learning more and more about this the more you do. And that dark side is the high stricture formation rate in patients who undergo resection, particularly more than 75% of the circumference. Importantly, this translates into significant morbidity. Many of these patients require repeated procedures, hospitalizations, health care utilization. So why does stricture develop? What's the timeline in the pathophysiology of stricture formation? So what we do know is that following resection, you have an ulcer defect. And within a couple of days, there's a strong inflammatory response with cell invasion. And then generally, approximately at the first week or two, now you start having epithelial regeneration, angiogenesis, and collagen fiber hyperplasia. And approximately by week three and four, you already have this remodeling of the esophagus from the fibrosis of the muscularis propria, and in many instances, like I said, leading to that refractory stricture. So what is the good news? The good news is that there has been several studies in the literature looking at post-esophageal stricture prophylaxis to date. What's the bad news? Nothing really seems to work, at least definitively and consistently. So let's start by looking at steroid therapy, because this is probably the most well-studied prophylactic management of esophageal stricture formation. So steroids, they inhibit that initial inflammatory response. They reduce proline hydroxylase activity and modify collagen enzyme activity, thereby reducing fiber-connective tissue formation. Now this is a table, obviously not all-inclusive, but summarizing, as you can see already, several studies looking at steroid administration for prevention after whitefield resection, not circumferential. So study designs, as you can see there, range from retrospective, most of these, to some prospective and randomized, and the protocol therapy is also quite variable, whether it's local injection or extended oral therapy. These studies look at extended mucosal defects anywhere between two-thirds to three-quarters. And when you look at the stricture formation rate, it ranges anywhere from 5 to 10 percent to as high as half of these patients. Now this table summarizes the data looking at steroid administration after circumferential endoscopic resection, and again, what you see is the protocol therapy is quite variable from local injections after resection to combination of local injections, oral therapies, and dilations. And what you can see here is a pretty grim picture. That stricture rate is extremely high, and in many studies, basically 100 percent. I didn't show in these two tables the limited data out there regarding other types of oral steroids such as topical budesonide or using fluticasone, but again, the data's extremely limited and also quite variable. And then on top of it, not only is the effect only modest at best, but we all know about the systemic adverse events, particularly accentuated in patients of advanced age and multiple comorbidities. So what about esophageal stents? The idea about esophageal stents is that if we place an esophageal stent, it can maintain patency through that mechanical radial force during the healing process, thereby allowing that esophagus to remodel while maintaining its patency. So again, data is quite limited when looking at placement of fully cover esophageal metal stents, and as you can see from these three studies that looked at stent placement after circumferential excision, the stricture rate was anywhere between 17 to 50 percent. So again, fairly high stricture rate. Now again, as I mentioned, the interpretation of this data is limited by the number of studies, study design, the very small number of cases. Importantly, we need to consider many potential issues associated with stent placement. Most of these patients, or many of these patients, are intolerant to the indwelling stent, whether that's due to chest pain or reflux. These patients often need repeated procedures, whether it's for stent removal or from stent complication itself, from migration, occlusion, perforation. And I put last one there as life-threatening pseudoaneurysms because perhaps this is one that we don't pay that much attention to, and I wanted to kind of illustrate that with this case. So this is a patient that, a 60-year-old man with a history of esophageal squamous cell cancer. He had a radiation-induced refractory stricture, as you can see there on the left. He apparently had on the gum multiple stricture dilations, as well as placement of fully covered esophageal stents. He did not tolerate the stents, he was refractory to that, and he underwent placement of a lumen opposing metal stent at an outside institution. The patient got lost to follow-up, and he returned approximately five months later to our hospital with severe hematemesis. The CT angio was negative, but the index suspicion for an aortic esophageal fistula was extremely high. We got the surgeons, vascular surgeons involved as well, but unfortunately, due to his extensive vascular disease, he was not deemed to be able to get an aortic stenting. And you can see it in the scopic exam there, the lumen is full of blood, and then we had to remove the stent in the off chance of identifying another source of bleeding that was not going to be a fistula, but unfortunately, it was massive bleeding as soon as we removed the lamps. A fully covered esophageal metal stent was placed, but the patient did end up passing away. Just to highlight the dark side of some of the things that we do, it's not harmless. So what about other therapies? Well, there's been some studies. Look at polyglycolic acid sheets. These sheets act as a mechanical barrier that you can put on the mucosal defect to protect it from the irritation from luminal contents. However, this is fairly difficult to apply. Most of the studies talk about application of these sheets along with fiber and glue and then trying to fixate them with clips or stents. And again, the data still remains quite limited. Now they've looked at autologous cell transplant, harvesting cells from the oral mucosa to develop and culture sheets and then replanting them in the esophagus. And again, this technology is still quite at its infancy, definitely not ready for prime time due to several technical limitations. So more recently, a self-assembling peptide gel was introduced into the market. This peptide gel, upon contact with fluid, forms a 3D scaffold matrix. And this has been approved for endoscopic hemostasis, although there was also some data suggesting that this matrix would promote healing as well as tissue regeneration. So a group from Europe published this pilot study looking at the effects of this self-assembling peptide gel in terms of prevention of esophageal stricture after endoscopic resection, circumferential ESD, and a Porcine model. And what they identified was that there was a reduction in the overall proportion of patients who developed stricture formation, although not clinically significant, likely due to the small number of cases. So based on that study, we decided to conduct a prospective multicenter study here in the United States looking at the use of this agent for stricture prevention. We enrolled 54 patients, of which 43 ended up going wide-field endoscopic semicosal dissection. Five of these had circumferential excision, 21 between 75 and 100, and then 17 of these between 50 percent and 75 percent. And what our data demonstrated was that the overall rate of stricture formation at a medium follow-up at two months was relatively low at 21 percent, and not inferior if you look at that data that I presented from STERA administration. When we look at the rate of stricture specifically for high-risk lesions, those that underwent resection more than 75 percent of the circumference, it was 31 percent. So again, it's a reasonable stricture rate. When you break down the data even further, what you identify is that most of the stricture formation is driven by cases in which circumferential resection is performed, and the actual stricture rate for those that had at least some mucosal left was fairly good at 19 percent. So here's how we applied the gel after resection. That's an 80 percent circumference ESD, following the ESD immediately. Thereafter we applied the clear gel, and then this is the follow-up endoscopy in this particular patient approximately, I think, six weeks after the procedure. So what can we make of all this data? For one, we can say that the data is relatively inconsistent. We see different rates all over the place, and this is probably due to the heterogeneous definitions of outcomes and the protocols themselves, right? So some of these studies are looking at not only STERA administration, but also a number of endoscopic dilation sessions, and so forth. And then, again, we have to weigh in the efficacy of the treatment with the potential side effects and what that constitutes for the patient. So the bottom line was that we need large prospective comparative well-designed studies with a longer follow-up. We do know certain risk factors for esophageal cancer. As you can see from this slide, advanced age, location of the upper esophagus, the depth of invasion, so histology, circumference length of resection, as I already stressed, and also whether there's muscle injury or not. So when you look at the risk factors for esophageal structure, many of these are now within our control. But there are two that are modifiable, right, which is the amount of extent that we do and how carefully we do the resection. So the main learning point that I've come across over the years doing this is to avoid circumferential resection in the esophagus when avoidable. So given the lack of kind of guidelines or data on this, I decided to kind of just share my approach. Again, this is not necessarily evidence-based, but it's just what I do in my practice. And given our data and given some of the data on steroids, I do start the patients on local steroid injection at the time of the endoscopic resection. If it involves more than 75% of the circumference, I will apply the self-assembling peptide gel at that time. I think one of the most important aspects is I will bring the patient quite early. So I don't wait a month or two before I bring that patient in. If you looked at the timeline in the pathophysiology of stricture formation, by that time fibrosis has settled in. So you want to catch these patients at an earlier time when that lumen is starting to narrow. And I dilate them very aggressively in terms of how frequently I bring them back. So I may not necessarily go up significantly on the diameter on the dilation, but I will bring them back as frequently as every week to up to four weeks until I start noticing some significant improvement before I push them further out to two weeks and they are on a more PRM basis. Again this is referring to these refractory patients, particularly those with circumferential excision. So take-home message, again, I think the main one is that esophageal stricture formation at least till this day is nearly inevitable when you're dealing with wide-field resection, especially circumferential. So you should avoid it if possible. And one of the most important things I've learned is to discuss it with the patient up front. So when the patient comes to me to have a procedure done and I anticipate that this is going to be a problem, I discuss the benefits of the endoscopic approach as well as surgery. And I also tell them that we're going to become very good friends because we're going to be seeing each other quite frequently for these procedures. So again, schedule the procedures. Don't do it on a PRM basis. We still need more high-quality data to further guide us in stricture management. Thank you. »» Okay. So we're going to close this session with our last debate. By show of hands, for people that do treat Barrett's esophagus, if you have a patient with Barrett's esophagus and you see a visible lesion, is your first-line approach EMR, raise your hand rather than ESD, EMR? And how many do ESD as their first line for a visible lesion? Okay. So I think there's less doing ESD. So you have a longer rope to tow, Dr. Wang. So our first speaker will be Dr. Dragunov and he's going to start us off with a debate arguing for EMR. »» Thanks, Michelle. Actually I was very glad to accept this invitation having a debate with Andy. And he's a good friend and I hate to do this to him but I'll carry on anyway. And because I'm starting first, I'll set the ground rules of engagement. And it's actually only one rule. Anything that I have said in the past cannot be used against me in this debate. Okay, Andy, so this will be something that if you can avoid it will be great. We have come a long way. When I was a medical student, Barrett's with high-grade dysplasia equaled esophagectomy. And that was a fairly straightforward decision. And at that time we had a joke that endoscopy is talking about Barrett's esophagus related tissue resembled eunuchs. They knew what should be done but they could not do it. Well the things have changed. Now dysplasia and early cancer in Barrett's esophagus is exclusively handled by endoscopy. And one of the main reasons of that great success story is a wide-field EMR for the treatment of Barrett's. And that came into practical, became practical because of device evolution, the DUET device from Cook allowed us to do this fast, reliably, and with great success. So the current algorithm endorsed by the two guidelines that I'm quoting at the bottom from 22 and 24 is to remove all visible mucosal irregularities with EMR and ablate the remaining flat Barrett's epithelium with radiofrequency ablation. There are issues with this approach. The recurrence rate is high. On the right-hand side you see the recurrence rate of Barrett's esophagus at the top line and high-grade dysplasia and cancer on the lower line. And the bottom line is that at five years we have roughly 50 percent recurrence of Barrett's esophagus. Complete resection is not possible in many cases when you are aiming for unblocked specimen, which is a great aim to be striving for. And you end up with fragmented specimen with lateral margin evaluation that can be compromised, which also though translates to deep margin evaluation, which can be compromised. And it may compromise cure if cancer is present. So you will say, hold on, Peter, you're getting confused here. You are arguing for EMR and I'm giving you a bunch of arguments against EMR. Those are the shortcomings of EMR. Well, actually, I'm sure not everybody is fluent in Greek, but synchoresis is a concession made for the purpose of retorting with a greater force. And that's exactly what I'm going to do. EMR has some definitive advantages. It provides post-procedure tissue diagnosis. Even if not curative, can provide staging. It's relatively easy, generally adequate to provide the depth of invasion. I'm not going to dwell about that because Andy, I'm sure, will give you an earful about that particular point. But most importantly, EMR is a time-proven strategy supported by extensive high level of evidence. And speaking of evidence, I'm going to concentrate on recent data. Hot off the press last month in American Journal of Gastroenterology, three referral centers look at their outcomes of recurrence after EMR or ESD. And that is after they have achieved complete remission of intestinal metaplasia. So the patient had EMR or ESD, and followed by RFA, had complete remission, and then they looked what was the recurrence. And the bottom line is that the recurrence was the same. In multivariate analysis, to no surprise, B recurrence was higher with a longer segment of Barrett's, the size of the lesion matter, and history of cigarette smoking. And Vanny already mentioned about cigarette smoking being important in Barrett's esophagus. So yeah, the mean visible lesion for the ESD group was larger, so there is some degree of comparing apples with oranges, but nevertheless, the bottom line of what we do is to avoid progression of Barrett's esophagus, and EMR does as good job as ESD. And also this retrospective data, how about prospective randomized control studies? This is kind of the gold standard. And we have one that compare EMR versus ESD. And if you look at unblock resection, which is the top line, of course, the ESD looked great. but if you look at the recurrence at three months, zero in the EMR group, and ESD had one recurrence. Perforation in the ESD group was seen in two patients. And of course, this study can be criticized that it was a small sample size, only 20 patients in each arm, short follow-up, higher perforation rate than other studies in the ESD group, and underpowered to evaluate recurrence rate. But it's a randomized control study, and as we say where I'm from, in a country where of the blind, the one-eyed man is the king. So we don't have a better study. This is the best available evidence at this point. So how do you put this all together? I'm going back to my earlier kind of how do we currently manage Barrett's esophagus, and I'm gonna expand on my earlier slide. We still want to remove all visible mucosal irregularities. And I wanna emphasize on the term irregularities, and Dr. Konda actually nicely showed that it is not nodules in many cases. It's just that the mucosa with dysplasia or early cancer is irregular and looks different than the rest of the Barrett's mucosa. So traditionally, we have done EMR, but now there is some role for ESD in highly selected patients. Then we ablate the remaining Barrett's with usually with radiofrequency ablation, but now we have expanding options of cryoablation and hybrid APC. When I was thinking of how to finish, it just came to me that this debate a little bit resembles that very old mirror light commercial. Is it great tasting, or is it less filling? I would say it's great tasting and less filling. Thank you. That's great, Peter, thank you. Sorry, it's still me. So now it's my pleasure to introduce Dr. Andrew Wang, who is professor at the University of Virginia Health System to argue the opposite ESD approach. Thank you, everyone. It's a great pleasure to be here with you. Thank you to the course directors, and I really appreciate the opportunity to talk about ESD for Barrett's esophagus. I have no relevant disclosures. So I actually have to take a step back. We described, Peter described rules of engagement, and I will have to modify them a little bit. I will promise to you that I will not describe or talk about eunuchs anymore in this talk, but I will talk about Dr. Dragunov's past. So basically, if I end with this single slide, I think I will have won the debate. Come on, what would you want for yourself? We're patients too. If you had any of these lesions, these two lesions describe large nodular areas in the esophagus arising from Barrett's with areas of depression and avascularity. These are cancers, esophageal adenocarcinomas. Is there any doubt you would not want Peter Dragunov to do ESD for you? Would you really be okay with piecemeal EMR if you had these cancers? So going back a little bit in time, 2019, Peter and I put together this AGA clinical practice update, and one of the key practice advice points for endoscopic submucosal dissection was that it should be considered in patients with Barrett's with the following features. Large or bulky area of nodularity, lesions with a high likelihood of superficial submucosal invasion, recurrent dysplasia, EMR specimens that showed invasive cancer with positive margins, equivocal pre-procedural histology, and cancers. Fast forward to this past year, and this is a really excellent guideline put out in Gastroenterology by Rubenstein, Sawas, and Wani, and they really looked in a very systematic way, did a systematic review of all the data for the PICOs, then meta-analyses to try to answer the questions. And for this debate, focusing on these data, these are meta-analyses that they conducted, especially for this guideline. They found that with the outcome of R0 resection, ESD is clearly superior, as we all would think, compared to EMR, and R0 resection is complete resection, margin negativity, microscopic margins, negative. Going further, ESD is also superior in their meta-analyses when looking at complete eradication of neoplasia for Barrett's, ESD superior to EMR, as you can see here. And then, if you kind of look at the data, and I'm trying to be fair, ESD and EMR, in this meta-analysis, appear to be similar for occurrence at one to two years, when you're talking about cancers alone, but we'll flip the page and discuss this a little bit further. So, in their recommendation, number five, with respect to Barrett's esophagus, patients with visible neoplastic lesions, they do say that EMR or ESD could be pursued really based on lesion characteristics, but I'd like to highlight, again, patients with bulky neoplastic lesions, or those highly suspicious for at least T1B, that's submucosal invasion, depressed lesions, those with amorphous avascularity, are deemed candidates for endoscopic resection, and may benefit more from ESD over EMR. Patients, of course, who failed EMR might benefit from ESD, but really, why do EMR first, if you have the ability, do it right the first time? So, these are some data that have come out since the guideline was published, and I think some of these are actually important. This is from a Canadian collaborative, I believe, and they looked at a multi-institutional cohort of patients undergoing EMR and ESD for high-grade dysplasia and cancer, and they found that compared to EMR, the ESD group had larger lesions, as you might expect, more cancers, deeper submucosal invasion, and with respect to neoplastic recurrence, this is recurrence at a pretty good follow-up up to three years, the following complete resection, recurrence was significantly higher in the EMR group, 13% versus 2%. This is a much larger study than the one that Peter quoted. The cumulative probability of recurrence in three years was 18% versus 4%, and more EMR patients actually underwent RFA afterwards, so even with RFA, 18% versus 4% when you're talking about highly dysplastic lesions. These are data that I believe Dennis, Peter, I, and a bunch of others in the audience contributed to that we presented recently at DDW, but just as we were looking only at cancers, invasive cancers that underwent ESD at three-year follow-up, and really, the point I'll leave here is that really it was comparable, similar outcomes to surgery, esophagectomy, and patients who had curative resection by ESD did just as well. So I believe this is a paper Dr. Waksman participated in, led by our Australian colleagues, and it took 19 experts in Barrett's esophagus from eight countries. Experts were asked to look at images and try to prognosticate invasion, and when you actually have these great experts, they can very reliably differentiate invasive cancer, intramucosal or submucosal cancer from high-grade dysplasia. However, when you ask them to differentiate T1A versus T1B, mucosal versus submucosal, they cannot reliably distinguish T1A from T1B, so this truly has implications about what we're talking about when you select endoscopic resection. Oops. So this was the ASGE guideline that came out about a year ago, and Dr. Dragunov also helped to put these together, and the ASGE guideline, very appropriately, I think, suggests that for early-stage Barrett's nodular dysplasia, well-differentiated, non-ulcerative, with lesions more than 20 millimeters in size, suggest ESD over EMR. Per the AGA guideline that I referenced earlier, this is an important point. EMR should be sufficient for T1A intramucosal lesions. ESD may be more effective for T1B lesions, but one does not know the depth of penetration until after the resection. So despite 40 years going by, we still don't have a time machine to go back and take back your EMR when ESD could get you the answer. This is just my final case. This is an elderly gentleman who's not a candidate for esophagectomy. He had a lot of comorbidities. Here you see a large lesion nodular with avascularity. This is a cancer. Here we're just performing ESD, and it's a pretty straightforward one. We use a tunnel technique, and as you can see in the coming few seconds here, very well we're able to look at the submucosal planes. We can dissect very close, right above the muscular dyspropria. Typically we use the bottom 1 3rd of the submucosa, leaving as much as you can to help stage, and in the case of a T1B cancer, really get a R0 deep margin. And so here we're able to dissect close, right above the muscular dyspropria, and this lesion came off pretty easily in a near circumferential manner. Final cuts. This ended up being an R0 resection of a T1B invasive submucosal cancer. The margins are negative, as you'll see. This patient actually had lymphovascular invasion, and it was a poorly differentiated cancer. Lateral margin's clear, deep margin's clear. G3, which is poorly differentiated, plus LVI. However, he was not a candidate. It's actually now been five years, and he is totally fine, cured of his disease. So again, I'll come back to where I began. Come on, if it was you, if I, actually for me, if I had invasive Barrett's adenocarcinoma, I would buy my ticket to Gainesville right now. I would 100% want Peter Dragunov to do ESD for me, and I would not be okay with EMR. Thank you. Oh yeah, we wanted to do a post poll. So how many now, given nodular dysplasia and Barrett's, would send the patient to either of the two doctors that just presented? Raise your hand. Yeah, I think you guys swayed them. Okay, so we, the course directors have been having a discussion also with our ASG staff. We are behind. There is coffee in the back of the room. I want to give people a chance to ask questions if they want to, and hear some of this. So if you could quietly get up, get your coffee, and come back, we will do a few questions, and then move on with the next section. That's going to be interesting. It'll be fine. Can I get all the speakers from the last session back on the stage, please? Go ahead. While they're sitting here drinking their coffee. Right. It'll be fine. Just a couple. It's just a couple. There's so many. No, you don't. There's like 7,000 people. Yeah, Ollie, you're hilarious. Oh my God. Peter, I'm sorry I didn't give your title and your institution. My brain was still in the survey. You can sit wherever you want. Yeah, just sit. Okay. Are we missing somebody? That's everybody? That's everyone. I think this is it. That's it. We're good, I think. I was really happy. Like I told you. Okay, so actually, the first question was mine. This is for Vonda, and in disclosure, I don't do Barrett's therapy. You mentioned poor adherence in terms of Barrett's surveillance. Is that on the part of the patient or the endoscopist when doing surveillance biopsies? Oh, I'm sorry. Could you repeat the question? So in part of your talk, you talked about that there's poor compliance with surveillance. Is it that the doctors don't do a good job, or is it that the patients don't come back for surveillance? Thank you for that question. So it is the doctors that do not do a good job with adherence to the Seattle Protocol biopsies. So they usually take less biopsies and they need to put them in less jars than they need to, and adherence to this is probably mostly due to time and effort. It has been shown that a dedicated intervention of teaching physicians about the Seattle Protocol and why the Seattle Protocol is important does improve adherence among the doctors, and that was shown both with that meta-analysis that looked at multiple factors for adherence retrospectively, and then also it was shown in a randomized control trial in the ACQUIRE trial. It was published in GIE last year. Great, thank you. This one's for you, Dennis. Do you utilize steroid injection for near-circumferential resection in other areas like the duodenum or even after cold EMR? I don't. You also have to be careful with the steroid injection because just like it affects the healing process, I'm quite careful about being in the duodenum injecting steroids. I have used some of these topical agents for whatever it's worth. Obviously, there's no data on it, but just like in the esophagus, the duodenum's another area that can be high-risk of stricture formation. Yeah, I think especially if it's circumferential. I think this one could be for anybody. When you guys are doing EMR versus ESD, is there any role for EUS prior to resection? Peter and Ann, take it. Sure, we'll take it together. I think it's very clear that EUS is not perfect. It doesn't mean you can't do it, but for mucosal versus submucosal, it's not super accurate. If you're worried about something that has lymph node metastasis, I think that's a different question. That can be an accurate exam. However, usually if you see something with a stricture, I think this made it into one of the guidelines, it's already T3, and so you probably don't even need it. So it's optional, but I would not make your decisions based on it. Yeah, I can expand on that. I definitely agree that, and I believe Dr. Konda made that point, that EUS is not particularly good in distinguishing T1A versus T1B disease, which is where the money is. That said, I pretty much routinely do EUS, and it's mostly to exclude advanced disease or a lymph node, a lymphadenopathy. But also, probably even more importantly, is not the EUS part. It is the EGD part that I'm repeating at the time of EUS, which can give me great information following the rules that Dr. Konda set for us, careful observation. People sometimes overstage or over-exaggerate lesions. Sometimes lesions are completely missed, and I believe before embarking on a procedure that has still significant risk of complications, particularly with wildfield EMR or ESD, second-look EGD by expert is well worth it, and I use that opportunity to combine with EUS, the EUS literally taking five additional minutes, if that. There's a question for Bonnie for surveillance. Biopsy first, then what's first, then biopsies, what is your, how do you do it? I think there is some variation, but I do targeted biopsies first, then I do the random biopsies, and then I do the WATS. And if in the middle of doing a random biopsy, I wanna take another targeted biopsy, it's quite easy to do so. Ask her that one. So does the positive P53 change your surveillance interval, because it's positive P53 will treat? I will take those opportunities for those high-risk stratification tools, whether it be P53 or tissue cipher. If they are positive, I will bring that patient back, and I will offer them a rigorous endoscopic evaluation. I'll use enhanced endoscopic imaging. I'll consent them for endoscopic mucosal resection, and I will do my best to see if there is any evidence of dysplasia, and I might shorten the interval. However, I don't think that we are at a stage where, I think there is a balance of what these procedures do, and there are risks of what these procedures do, and I take this as a sign that I have to do my best to determine if that patient is harboring any dysplasia, but I do not take it as an open door to going forward with treatment if there is no evidence of dysplasia. There's a question for Peter and Andy. Why do studies comparing EMR to ESD use R0 resection as the endpoint? This seems like an inherently biased outcome as it's an inherent limitation. Well, I argue for EMR, but let me now try to explain the benefits of R0 resection. And the colon is a good area where we have extensive data of high levels of recurrence after EMR, and the reasons are simple, because we leave stuff behind. And we have gotten smarter over the years, and we have improved our EMR technique in the colons. We ablate margins and so forth, which has been a game changer, but even with that, recurrence is around 5%, which is not bad. On the other hand, if you look at any ESD study with R0 resection, the recurrence is consistently significantly lower than 1%. So the benefit of R0 resection translates into far lower rate of recurrence, and secondly, adequate histopathologic evaluation. And to add to that is a basic oncology principle. If you can remove cancer in one piece and you know that you have removed it unblocked, and you have a histologic negative margins, you know that you have provided cure, and that affects your follow-up intervals, which also is not to be ignored, how frequently you scope the patient. Andy? Yeah, I totally agree. If you think about piecemeal EMR, for margin negativity, it's really our endoscopic visualization, and we take on faith that our overlapping pieces have got it all. When you think about ESD, it's actually a bit different. We're looking endoscopically, but you have pathologic margin negativity, and you're pretty much certain that everything in between there is completely gone. And so it's a little bit different, and when you look at the studies that we quoted, the Canadian study, they actually did very good EMR, but they posit that there was some microscopic stuff left in between. When we do RFA, I don't know that everyone RFA's over your clean scar, and so perhaps you leave that behind. The study that Peter quoted in the Red Journal was actually very good when I saw it, but they didn't have a lot of cancers in it. So when you're throwing in high-grade, maybe early T1A with not nodules, I mean, RFA should do the trick anyways, but when we're talking about real nodular cancers, I think it's different. Anything, anything that's the, so I think when you talk about field defect in Barrett's esophagus, it's difficult to tell, right? Because you're resecting a lesion within the Barrett's esophagus, but when we're specifically talking about cancers, R0 resection matters. If you're talking low-grade dysplasia, high-grade dysplasia, it might not be that significant differentiating EMR versus ESD. But having said that, even certain studies looking at EMR versus ESD has shown more inconsistent margins with EMR, even for smaller lesions. I mean, ultimately, I want to give credit to whoever asked this question, because it's an example throughout medicine where we go for, you kind of, yeah, R0 versus R1 matters, but what ultimately matters is whether the patient dies from cancer or not. And that outcome, it's very hard to study because it requires a very long follow-up. Same like hypertension. I mean, yeah, we control blood pressure, but ultimately, do they die from heart attack or strokes? It's a much better outcome. It's just those surrogate markers do translate into final benefit, as I alluded earlier, on multiple levels. Thank you. I think we're gonna end the questions here. We're actually gonna take a five-minute, I mean five-minute bio break.

Barrett's esophagus

esophageal cancer

endoscopy

Seattle Protocol

EMR vs ESD

AI in surveillance

P53 risk markers

patient communication

biopsy protocol

treatment optimization