And now, it is my pleasure to introduce our next speaker, Dr. Peter Dragunov from University of Florida, and he has been a pioneer in third space endoscopy and has helped introduce endoscopic submucosal dissection to the West, and he will be speaking about that now. One of the greatest achievements of flexible endoscopy and its impact has been the way we currently manage Barrett's esophagus. When I was a fellow, patients with Barrett's and high-grade dysplasia had to go to surgery. Now, the vast majority of patients can be successfully managed by endoscopy. The current algorithm for therapy of Barrett's esophagus consists of removal of all visible mucosal irregularities with endoscopic mucosal resection and ablation of the remaining flat Barrett's epithelium with radiofrequency ablation. The situation gets a little bit more complex if we're dealing with intramucosal or superficial submucosal invasive cancer, then surgery remains the prime modality, but EMR can be utilized as therapy or as a diagnostic staging procedure. We do consider systemic therapy in these patients, but that has not been a well-established strategy and definitely not supported by a vast number of data. We have a problem, though. If you look at our pre-procedure staging tools, we can use EGD with high definition, MBI, EUS, CT, PET, CTLE, OCT, Watts, brushing, and of course, biopsy. The very fact that we have that many tests tells us that none of them is perfect. And indeed, that is the case because we frequently are not absolutely sure what exactly we're dealing with as far as the degree of dysplasia and as far as the depth of invasion. In addition to that, a high recurrence rate has been well documented by multiple studies with our current algorithm, and this is just one of the latest studies which shows that in about five years, roughly 50% of the patients will have a recurrence of the intestinal metaplasia and a good but still significant minority will have the bad guys of high-grade dysplasia and intramucosal or submucosally invasive cancer. I should not have to spend too much time to try to convince you that a fine steak is much better than hamburger meat. By the same talking, I don't have to spend too much time to convince you that the specimen obtained by ESD is better than the specimen obtained by piecemeal EMR, yet in our literature still there is the claim that EMR histology is adequate for evaluation of the depth of invasion. I want to point out to you, though, that this frequently is not the case. The question is why? And the first reason is that ESD gets deeper into the submucosa compared to EMR, and that is why you frequently get these readings from your pathologist that there is at least intramucosal cancer. The reason they cannot commit that there is deep submucosal invasion is that there is no submucosa to be seen. Even if there is submucosa on the EMR specimen, because of the curling of the specimen and specifically of muscularis mucosa, it may be hard for the pathologist to evaluate whether there is cancer cells invading in the submucosa. The problem is further aggravated by the fact that there could be a duplication of muscularis mucosa in Barrett's esophagus. Compare that to ESD specimen where clearly you can see the muscularis mucosa and the pathologist will have much easier time evaluating that. Our early observation was reconfirmed by a study that was conducted in the Netherlands. 25 EMR specimens were reviewed by two pathologists that are experts in Barrett's esophagus. The two pathologists were blinded to the original findings and the original reading. Importantly, discordance was considered only if both of the study pathologists' interpretation differed from the original report. And that is a very high bar. Despite that, discordance was present in roughly 50% of the cases. And importantly, the vast majority of discordance was observed for T1B lesions, which is that very important category that we really want to separate are they T1A or are they T1B and also how deep they invade into the submucosa. Because that means from practical perspective, the difference in between observation versus sending the patient for surgery. Even more recent data support the notion that EMR specimen evaluation is suboptimal. In this particular case, the author compared EMR with ESD specimens and equivocal lateral margin was reported in 42% of the EMR specimen. But even more importantly, equivocal vertical margin was reported also in 42%. Of the patients with equivocal histology, 11 were reported to have at least intramucosal cancer and of those four went to surgery. Two of the four, which is 50%, had T1A disease and that could have been easily handled by endoscopy alone. So those two patients got unnecessary surgery. In our own experience at the University of Florida, and this is unpublished data, we have a discrepancy between pre-ESD biopsy and ESD histology in 62% of the cases. So equivocal biopsy and inconclusive EMR specimen evaluation is frequently seen and it is an unfortunate routine part of our practice. We have accepted EMR to be utilized as staging procedure and we have incorporated it in our current management algorithms. The question is, can ESD be utilized the same way? And the answer is yes. In this study from Japan, patients with suspected invasive cancer, which all should have been treated with surgery, underwent either surgery or ESD based on physician and patient preference. So this is retrospective non-randomized study. To no surprise, ESD showed very low curative resection rate of 20% because all of the patients were considered to have deeply invasive cancer. Importantly though, diagnostic ESD did not have negative impact on surgical outcomes. If you look at the surgical group, in 20% pathology was T1A with no lymphovascular invasion and they had no lymph nodes on their surgical specimens. So those 20% could have been successfully treated by ESD. So you have 20% that had curative resection with upfront ESD and another 20% from the surgical group that could have been treated with ESD, which adds together that if you use upfront ESD as a staging procedure, 40% of the patients could have avoided surgery. We clearly need to learn more about ESD in Barrett esophagus and we looked at the technical success of the procedure in a meta-analysis published in 2018 and basically shows that the unblocked R0 resection and curative resection rate is very much in line with accepted standards from the Japanese literature. In addition, complications were fairly unfrequent with the main one being stricture, which is an area that clearly will benefit from improved management strategies. The question though, is EMR or ESD the better clinical strategy because we can measure unblocked resection rate and R0 resection rate and of course with ESD they are better, but does that make any clinical difference? And actually there is a randomized control study in 20 patients that were randomized to ESD and 20 to EMR and it showed the expected better performance of ESD as far as unblocked R0 and curative resection, but the recurrence in three months was not significantly different. One case in the ESD group and no cases at three months in the EMR group. I want to point out to you is that this study has multiple limitations. For one, the sample size is very small, only 20 patients in each group. It has a very short follow-up, only three months, clearly inadequate to evaluate recurrence. A higher perforation rate was reported than multiple other studies of ESD in Barrett and the study was underpowered to evaluate recurrence rate. So in a way, this study, although it was a randomized control study, provided us with more questions rather than answers. And we have continued to try to answer some of these questions. EMR versus ESD for superficial esophageal carcinoma was evaluated in this meta-analysis and it showed the expected high unblocked R0 and curative resection rate, but also showed a very low recurrence with ESD. Importantly, the superior effect of ESD in unblocked resection, curative resection, and local recurrence rate was only manifested when lesion size was greater than 20 millimeters, which makes sense because smaller lesions can be successfully removed unblocked with EMR in most cases. We wanted to evaluate the clinical outcomes of EMR versus ESD and we conducted a multicenter retrospective study, which is currently under consideration for publication, 140 EMRs, 85 ESDs for advanced Barrett's neoplasia. All patients underwent radiofrequency ablation for residual Barrett's after endoscopic resection. Yet again, ESD comes far ahead as far as unblocked R0 and curative resection rate and significantly less recurrence was observed in the ESD group. There was no difference in adverse event rate. More recently, we published our experience with endoscopic submucosal dissection in North America and that was a prospective multicenter study that included a total of 692 patients. I'm presenting here only the data pertaining to esophageal ESD. The average specimen site procedure time are outlined here with unblocked resection rate of 97%, R0 resection of 86%, and curative resection in 71%. Complications again were uncommon, specifically serious complications, and the stricture formation was seen considering the large size of the resected specimens. All strictures were managed endoscopically. ESD allows resection with EMR may not be feasible. So I want to give you a couple of cases, and this is one example where ESD was performed for this nodule, but histology showed at least intramucosal carcinoma with positive lateral and deep margins. This is the post-EMR imaging. You can see the clip here that the endoscopist placed for bleeding, and the area was removed with ESD and showed intramucosal cancer with negative margins, so we had curative resection. In this particular case, there was extensive nodularity, and you can probably appreciate that approximately 50% of the circumference is occupied by this large nodular area. Obviously, the larger the area of nodularity, the more likely is to harbor an occult cancer. Therefore, unblocked resection is recommended, and in this case, actually, there was no cancer, but there was only high-grade dysplasia. One can argue that if this was removed by EMR, we may have had inconclusive histology, and we may have not had definitive answers as we do after the ESD. In this case, EMR clearly is not an option, and the patient probably would have ended up with surgery if it wasn't for ESD, but this is the original nodule that was removed by EMR. In this case, the patient had two endoscopic ablations, and when he came for the third ablation, this bulky nodule was seen. The patient was referred to our institution for ESD, and actually, the patient had two nodules, one here at 11 o'clock and another one at 5 o'clock. I felt that it would be very difficult to remove the two nodules as independent specimens unblocked. Therefore, we did a full circumferential resection, and this is the final specimen, and pathology showed well-differentiated intramucosal adenocarcinoma stage T1A with negative lateral and deep margins, so this is R0 curative resection. To summarize, ESD in Barrett's esophagus should be considered in patients with high-grade dysplasia with irregularity more than 15 to 20 millimeters in diameter, with depressed areas such as Paris IIc, or very bulky protruding lesions. ESD can also be considered in intramucosal or superficial submucosal cancer, equivocal histology on biopsy, and prior EMR with positive margin. Finally, recurrent lesions after RFA or EMR should be given consideration for ESD. These indications are summarized in a document put together by the American Gastroenterologic Association, and this was published in Clinical Gastroenterology and Hepatology. It discusses the indications for ESD with particular attention to the focus on the U.S. endoscopy's practice patterns. Thank you very much.

Barrett's esophagus

endoscopic procedures

endoscopic mucosal resection

radiofrequency ablation

endoscopic submucosal dissection