It is my great pleasure to welcome back Dr. Amandeep Shergil, who is going to talk about endoscopic surveillance and management of colitis-associated neoplasia. I'd like to thank the course directors again for inviting me to speak today. This topic will focus on endoscopic surveillance and management of colitis-associated neoplasia. Here are my relevant disclosures. Over the next 20 minutes, we will review optimal approaches to surveillance in IBD, importantly reviewing terminology as it relates to documenting endoscopic remission, the technique for surveillance, as well as how we will describe detected lesions and the role of random biopsies. And then we'll touch on management of endoscopically detectable lesions, what resectable versus unresectable means, the role of parabolic biopsies, as well as surveillance. The first question is who we should be surveying, and almost all societies agree that all IBD patients after eight years, and at most after 10 years of disease onset, should have a repeat colonoscopy to reassess disease extent, and then eligible patients at that point would be enrolled into a surveillance program. Ideally, this is done in remission, because when patients have active disease, detection and interpretation of findings can be difficult. It's important, therefore, that we review what remission means endoscopically, as well as how that would apply to UC as well as Crohn's disease. So in UC, endoscopic remission is considered to be a male endoscopic sub-score of 0 to 1, and in Crohn's disease, it's an SESCD score of 0 to 2. Importantly, for both of these, it means the absence of ulceration. Standardized endoscopic reporting can help us understand who should be surveyed and whether or not there are diseases in remission. Ulcerative disease can be standardized utilizing the Montreal classification. In ulcerative colitis, patients who have more than just isolated proctitis, so those patients with left-sided or extensive colitis, Montreal E2 or E3 disease, are the UC patients who would be eligible for being enrolled in a surveillance program. In Crohn's disease, it's only the Crohn's patients that have the chronic involvement, so patients with Montreal classification location that includes colonic or ileoclonic disease, and importantly, it's those that have greater than a third of colonic involvement. So the Montreal classification can help us understand who should be enrolled in the surveillance program. We also want to use a lexicon or standardized descriptors of mucosal inflammation so we can understand if a patient is in remission or not, and I'd like to spend a little bit of time reviewing this lexicon. So the European Evidence-Based Consensus for Endoscopy and IBD put forth this description of inflammation lexicon, which includes vascularity, erythema, granularity, friability, bleeding, erosions, ulcerations, and stenosis, and these factors then feed into the endoscopic scoring systems that we would use for both UC and Crohn's disease. The mucosal vascular pattern is an important marker of disease activity, so when you have a normal mucosal pattern, more often than not, that's going to be a normal-appearing colon, both endoscopically as well as histologically. However, the first signs of underlying inflammation can be a loss of the normal mucosal appearance without the well-demarcated arborizing capillaries, and there can be a patchy or blurred loss of the capillaries to a complete loss of the vascular pattern. Erythema is defined as unnaturally reddened mucosa, and it can be from discrete or punctiform to more diffuse erythema, as noted here. Granularity is a mucosal pattern produced by a reticular network of radiolucent foci of 0.5 to 1 millimeter of diameter with a sharp light reflex, and it can be described as fine to coarse or nodular due to the abnormal light reflection. Friability is bleeding or intermucosal hemorrhage before or after the passage of the endoscope, and it can be from contact bleeding to spontaneous bleeding. An erosion is a definite discontinuation of mucosa less than 3 to 5 millimeters in size. It can be white or yellow in color, and it can have a flat edge, and it can be isolated or diffuse. An aphthis ulcer is described as a white, depressed center surrounded by a hollow of erythema, and some consider this synonymous with the erosion, and again, this can be isolated or multiple. And then frank ulceration is any lesion of the mucosa of unequivocal depth, width, or without a reddish hollow, and it can be isolated or multiple based on morphology. The ulcer size can be characterized as small, as less than 5 millimeters, medium, 5 to less than 20 millimeters, and large, greater than 20 millimeters, and then superficial or deep. So a superficial or shallow ulcer will have ulceration that's localized to the submucosa with no border, versus a deep ulcer, which goes beyond the muscularis propria, will have edges that are elevated greater than a millimeter. And stenosis is narrowing of the lumen. It can be single, multiple, passable by the standard adult scope or pediatric scope, unpassable, only passable after dilation, and we also want to include whether or not the stenosis is ulcerated or non-ulcerated. This lexicon then can feed into the endoscopic scoring systems for both UC and Crohn's disease to give an overall impression of the severity of inflammation. This overall impression, in addition to histologic staging biopsies, where we take at least two biopsies from each side of the colon, then gives an overall impression of both the extent as well as the severity of inflammation, which have implications for therapy as well as surveillance. So here would be a pictorial description of the myoendoscopic subscore. Again, mucosal healing will be zero or one. Zero is considered to be normal or inactive disease. One is mild disease, where you might see mild erythema, decreased vascularity, mild friability. And importantly for mucosal healing, there's the absence of erosions or ulcers. For a myoendoscopic subscore of two, or moderate disease, you're going to see marked erythema, a lack of a vascular pattern, friability, and erosions. And for a myoscore of three, more severe disease, that's going to be spontaneous bleeding as well as ulceration. For the simple endoscopic scoring system for Crohn's disease, or SCSCD, the ASGE SOP guideline on IBD has a chart that can be used to help us quickly calculate what the score is. When there is no active disease, that's an SCSCD score of less than two. Mild disease is considered to be an SCSCD score of three to six. And moderate to severe disease is considered to be an SCSCD score of greater than or equal to seven. In addition to performing surveillance colonoscopy when a patient is in remission, another important consideration is optimizing the bowel prep to optimize visualization. This will require use of split-dose bowel prep as per routine, and during insertion, aggressive water exchange to make sure that any residual stool or mucus is cleaned out. In terms of the optimal technique for detecting dysplasia during IBD surveillance, the standard was set by the CENIC consensus statement, which is a rigorous international consensus statement utilizing GRADE methodology to develop the following key recommendations. We should be using high-definition instead of standard-definition colonoscopes. We should be using chromoendoscopy with targeted biopsy sampling, which is superior to white-light colonoscopy with random biopsy sampling. Equipment-based chromoendoscopy, such as NBI, cannot replace dye-based chromoendoscopy. And using more accurate terminology for describing detected dysplasia as endoscopically resectable or non-endoscopically resectable is important. The technique for chromoendoscopy has been described, and there are many resources that are available for use and free through the ASGE Learning Library. The key are, again, having an excellent bowel prep and using aggressive water exchange during insertion to remove any residual residue, making sure the disease is in remission. The agent can be either methylene blue or indigo carmine, and the most efficient application is during withdrawal using the forward wash jet or pump. So approximately every 10 centimeters or so, what you'll do is you'll go ahead and pull back the scope using that water jet and spring to the antigravity side so the water can coat the colonic mucosa as it's cooling. Suction as you're finding your way back to that more proximal 10-centimeter area where you started the dye spring, and that suctioning banding out allows the dye to fully coat the colonic lining. And then insufflate and try and evaluate to detect any subtle lesions. The standard terminology to describe detected lesions will utilize the Paris classification and in addition include scenic descriptors. So we'll first classify lesions as polypoid or non-polypoid, and if it's polypoid, either pedunculated or sessile, and if it's non-polypoid, superficial elevated lesion, flat lesion, or depressed lesion. And then in addition, we want to add the scenic descriptors as to whether or not there's a presence or absence of an ulcer on the lesion. And then very importantly, description of the borders. Does it have discrete borders that can be seen, which will have implications for endoscopic receptability. Here's an example of how chromoendoscopy can help detect lesions. In panel A, we see an area of more focal erythema and friability, which after dye spring as seen in panel B, there is evidence of a very subtle flat dysplastic lesion where the dye is pulling at the edges of the lesion. And what's happening can be further seen here in panel F, where the dye is pulling in the inominate grooves of the normal mucosa, and dysplasia, which often has shallower and narrower glands, has less pulling of the dye, so you get this red and blue effect. Here's another example of that red and blue sign, and this is on the left panel, what was initially seen as a focal area of erythema, but after dye spring, was further delineated to be a depressed lesion. Serrated lesions, on the other hand, may stand out as being lighter, non-polypoid, where the dye pulls around the borders to highlight their edges, and often maintain that cloudy appearance with the mucous cap, which when combined with the dye, can help to draw our attention to these often very subtle lesions. Since the publication of the Scenic Consensus Guideline, there have been multiple studies comparing high-definition white light endoscopy to chromoendoscopy, and really inciting a debate over the optimal method for IBD surveillance. In the most recent systematic review and meta-analysis looking at this question, dye-based chromoendoscopy detected more patients in dysplastic lesions than standard-definition white light endoscopy, but no significant difference was observed between dye-based chromoendoscopy and high-definition white light endoscopy, or NBI. While the main outcome did seem to numerically favor dye-spraying chromoendoscopy, although procedure time is longer with dye-spraying chromoendoscopy. Regarding many of the other virtual chromoendoscopic methods, there's still not enough evidence to support or not their recommendation. And if we look at the meta-analysis, so certainly, again, for standard-definition white light endoscopy, we're seeing clear superiority of chromoendoscopy, but for high-definition white light endoscopy, again, the data are a bit more mixed, and it's really driven by one study, again, of a really highly expert endoscopist that seems to be able to detect lesions in the absence of chromoendoscopy. And again, if we were to look at these studies in more detail, and specifically at this Anderson 2018 study, which has been included in the standard-definition white light endoscopy section, but actually was a study comparing chromoendoscopy and high-definition endoscopy, the Alexanderson study was a single-center prospective randomized controlled trial which compared high-definition chromoendoscopy to high-definition white light endoscopy out of Sweden. So whereas in the prior systematic green-bed analysis, it was thought to be a standard-definition comparison based on the abstract, its final publication, it was a high-definition white light comparison. And in this study, the yield of high-definition chromoendoscopy was two times that of high-definition white light endoscopy, 11% versus 5%. And while, again, it did demonstrate that chromoendoscopy procedures were an average of seven minutes longer, the high-definition chromoendoscopy was more efficient than high-definition white light endoscopy detection in macroscopic lesions per 10 minutes of withdrawal time. And one of the most interesting findings of this study was that 20% of patients with dysplasia were identified only by analysis of random biopsies. So until we have a CNIC-2 to really review these most recent literature very rigorously using brain methodology and with a multi-international consensus, the data would, again, continue to suggest that the way to optimize dysplasia detection is with chromoendoscopy. But the question of how we maximize dysplasia detection really brings in that question of what is the role of random biopsies. And in talking about random biopsies and the CNIC consensus, the panelists could not reach a consensus on this issue. 60% voted to abandon random biopsy when using chromoendoscopy. 25% voted to perform random biopsies given the concern for missing dysplasia in a small proportion of patients. And in the CNIC analysis, dysplasia was detected with random biopsies in approximately 10% of patients undergoing chromoendoscopy or high-definition white light endoscopy and on targeted biopsies in the other 90%. The opinion piece by Hamad et al. helps to put these findings into perspective. For instance, if we think about, in rational decision-making, which technique you would prefer to choose, something that's going to yield an expected value of 91 gold coins or something that's going to yield an expected value of nine gold coins, I think most of us would agree that we would pursue what would yield the 91 gold coins. We can argue that even more valuable than finding gold coins is finding dysplasia in an IBD surveillance endoscopy. And so which technique then would we utilize, a technique that's going to yield dysplasia 90 times out of 100 or one that's going to yield dysplasia 10 times out of 100? That's one way to think about chromoendoscopy versus random biopsies. The yield of chromoendoscopy is going to detect dysplasia significantly with more gain than random biopsies alone. However, if you think about techniques to maximize dysplasia detection, that would be the combination of both chromoendoscopy with random biopsies. So again, the role for random or these non-targeted biopsies would allow us, in addition to chromoendoscopy, to maximize dysplasia detection, and they may be considered for a select higher-risk group. In one study of 1,000 patients undergoing chromoendoscopy with random biopsies, the yield was 0.2% per biopsy, but on multivariate analysis, random biopsy only detected dysplasia was associated with a personal history of dysplasia, concomitant PSC, or a tubular appearing colon. So in this high-risk group, we can consider the addition of these non-targeted biopsies to maximize dysplasia detection. Given how important this terminology is, I just want to take a minute to review this expert review, AJA Clinical Practice Updates Recommendations, the terminology to use to describe the types of biopsies that we obtain during surveillance. So again, the targeted biopsies are going to be biopsies of suspicious or subtle mucosal abnormalities to rule out dysplasia. The non-targeted biopsies are going to be those biopsies of non-suspicious areas to rule out invisible dysplasia that we might take in that higher-risk group, so to use the term non-targeted instead of random. And I think in all of these colonoscopies, we should be considering taking histologic staging biopsies. These are biopsies of macroscopically inflamed and uninflamed areas to assess histologic disease activity and extent, because this may inform our future surveillance recommendations. Once we've utilized chromoendoscopy to optimize or even maximize dysplasia detection, the key is how to manage endoscopically detectable lesions. And so what we want to know is, are these lesions resectable or are these lesions unresectable? The easy ones are the polypoid lesions that have very discrete borders and are clearly endoscopically resectable, likely with a simple cold-snare polypectomy. And I would urge us all to consider, even for the diminutive polyps, cold-snare polypectomy to ensure clear margins. Here's an example of a non-polypoid lesion that was detected during chromoendoscopy. So it's slightly elevated and it has discrete borders indicating that it can be resected, although likely may need some advanced endoscopic techniques and should be considered for referral to an advanced endoscopist. And here is what a lesion like this might look like after resection. And with complete resection, these lesions can have good long-term outcomes. This lesion, however, again, may have some slight elevation to it, so a non-polypoid, slightly elevated lesion. However, the borders are very indistinct and it's hard to see where this polyp starts and where it ends. Therefore, this lesion is likely not amenable to endoscopic resection. Ultimately, this decision can be made by an advanced endoscopist. The older practice also used to recommend peripolyp biopsies, such that any polyp resected during IBD surveillance would require biopsies of the resection site to rule out a potential field effect. However, we now know that visual inspection by trained endoscopists is likely sufficient with a yield of 0 to 5% for unsuspected dysplasia. All of these lesions, polypoid and non-polypoid, will require close surveillance. And while the data on optimal surveillance recommendations is continuing to evolve, this is the most recent expert review AGA clinical practice update, which recommends that for lesions that are less than two centimeters and deemed to be resectable, endoscopic resection with continued surveillance can continue. In large lesions, complex lesions, or with local recurrence, endoscopic resection that is more intensive every three to six months for the first year can be continued versus consideration of surgery. The decision to resect will be based on lesion details, local expertise, and disease activity. If the lesion is considered to be unresectable due to size, location, or features of invasive cancer, and or there's invasive cancer on histology, surgery is recommended. And in the case of invisible dysplasia or a subtle poorly delineated lesion, the histology should be confirmed with a psychopathologist, inflammation should be aggressively treated, and diasporae chromoendoscopy should be performed by an expert. The U.S. Society of Guidance recommends surveillance colonoscopies every one to three years in our patients with IVT. The patients that we consider for the longest intervals are those patients in endoscopic remission with histologically normal mucosa and no dysplasia on two consecutive high-quality surveillance colonoscopies with chromoendoscopy. In summary, to optimize detection of colitis-associated neoplasia, ideally surveillance should be performed when the patient is in remission, as defined by the endoscopic score, and with chromoendoscopy and targeted biopsies of any detected lesions. Histologic staging biopsies should be obtained throughout the colon to determine optimal surveillance intervals. If a lesion is detected, resectability should be determined. If there's a circumscribed border with no evidence of invasion, a targeted biopsy of that lesion can be performed, or you can consider tattoo and referral to an advanced endoscopist. Consider maximizing detection in high-risk groups such as those with prior dysplasia, PSC, or tubular colon with chromoendoscopy and non-targeted biopsies, and determine the optimal surveillance intervals when lesions have been resected, whether or not that's going to be continued annual surveillance or more intensive surveillance, as well as a multidisciplinary discussion for when surgery should be considered, for instance, in the case of multifocal invisible low-grade dysplasia or high-grade dysplasia. Thank you for your time and attention.

endoscopic surveillance

colitis-associated neoplasia

Montreal classification

chromoendoscopy

dysplastic lesions

follow-up colonoscopies