BSG IBD Surveillance Guidelines - part 1

[Music] Welcome to the B.S.G. from Topped Bottom Podcast series. I am your host, Dr. Shaheed Adin. A consultant gastroenterologist in Edinburgh with a special select expertise in inflammatory buildcies and the current chair of the IBD section of the British Society of gastroenterology. And I've led subgroups and surveillance and main guidelines. Today I'm really delighted that we'll be joined by Professor James East who led the IBD surveillance guidelines with Professor Morris Gordon. And in this podcast we will discuss encounter risk and stratification. And this will be followed by a second podcast focusing on the technical endoscopic options, MDT management, being led by Professor Pradip Pindari, who's the vice president and current chair of the B.S.E. IBD and DOSCII Committee. So I would like to extend a very warm welcome to Professor James East, a consultant gastroenterologist and endoscopist and director of the Bell Cancer Screening Program in Oxford with an academic interest in GI cancer prevention and co-author on a number of guidelines and consensus. So welcome, Professor East. Shia, thank you for that kind introduction. So I think I'd really like to ask you to start off and tell us the evolution of IBD surveillance and guidelines and how that started and where we have come to here today. So I think IBD guidelines actually go back, perhaps rather further than we think. And although the B.S.G. the current guidelines are described as the B.S.G. 2025 guidelines, strictly they are an update. And in fact, the first B.S.G. guidelines came out in 2022, a game of Matt Rattles guidance in 2010. Further guidance in 2019 that folded in scenic. So all of the major societies have, at some point, issued guidelines for inflammatory bowel disease surveillance, both from ECHO and the AGA and the ACG. And we've seen an evolution over time and that you can see that there is a direction of travel. So if we look at, say, you know, around 2000, we were doing standard definition white like coloscopies and rely heavily on random biopsies. But by the time we reached 2010, we've gone risk adapted. We were doing chroma and endoscopy and targeted biopsies. And then as we come through 2019 into the 2020s, much more personalized risk stratification and even the use of risk calculators. And similarly, on the timing of surveillance intervals, again, if you look sort of late 1990s, 2000s, if you had long-stained disease, you are on annual surveillance. As we got to 2010, we were sort of one to three or one to five years if we're outside the United States. And then in the most recent guidelines, we're certainly one to five years or possibly slightly controversially, even no surveillance for patients who are at very low risk. I think part of that we can see is driven by changes in IBD cancer risk and IBD cancer risk of death. And we can, you know, it looking back that there's the very well-known Olin Scandinavian population based data. If you were in the 1990s, your risk of dying of colorectal cancer was about four times out of the general population. But that has clearly reduced over time. And now it's maybe only 1.5, 1.6 times. And there are lots of good reasons for that. So patients are stopping smoking. Hopefully surveillance is a bit more effective. We're a bit better at timing collecting me. But I think we really know that the driver here has been better disease control. And particularly as we've entered the biologic era, we really, you know, patients have, you know, so often have normal looking colonoscopies. I think that's a really comprehensive analysis of where we have started and where we are now. And I think the reason that it's special to me is that when I joined the BSU, I had a committee in 2019. With that sole intention of asking so that if we can review the IBD surveillance guidelines, because in my own practice, I realized that I was doing a number of surveillance colonoscopies and actually wasn't finding inflammation or dysplasia or cancer. And that initial work stream, when we thought about it, was disrupted by the pandemic. And then Dr. Ian Pemman handed it over to Dr. John Morris and it landed beautifully with you as, you know, to lead that guidelines with Professor Morris Gordon. So I guess what we might do is just structure the podcast and think about sort of, why do we do surveillance? When do we do it? And how do we do it? And we might even touch on who should be doing it. So maybe just to ask you to help us understand, why do we actually do surveillance? And you've already explained a little bit about the population risk or the IBD specific risk here that helps influence that programme for us. So I think historically we've always felt that we were going to prevent IBD related colorectal cancer. And that's, you know, that's a good aim. But I think, you know, there are challenges to that. And certainly the data suggests, while we, while surveillance and that this is thinking about three-year lease events, there is some data that it reduces risk of developing colorectal cancer by about a third. It reduces death from colorectal cancer by about two thirds. And when you look at the data, the patients who are on surveillance are patients are developing early curable colorectal cancers. And the patients who are not on surveillance are developing advanced colorectal cancer dying. Therefore, if a cancer develops on surveillance that's still early in curable, this is not a surveillance failure. But the corollary of that is that potentially we can't prevent all of the cancers that are developing. The biology of IBD associated colorectal cancer is clearly different in molecular genetic terms to sporadic cancer. And it is probably substantially faster equally. We're trying to detect subtle dysplastic lesions, which are the ones that are, that are IBD-driven colorectal cancers. There is a risk of missing and therefore cancer developing in between. And we can see that a little bit in the post-colorectal cancer data where patients with inflammatory bowel disease have a post-colorectal cancer that is six times that of the sporadic population. And while there are methodological reasons, because we escape them quite frequently, why that might be the case, nevertheless, that probably does reflect that we can't beat the biology. I think that's an interesting point to think about as opposed to cancer prevention and detection and how that actually aligns. And I think one of the things that maybe we need to focus on as our community is actually we're trying to detect cancer early. Because the biology is different and for us to actually change that biology, we'd have to work a number of years backwards to work at what time point we can actually intervene with the way we actually treat disease optimally to actually stop cancer from developing in the first place. So I think that's nicely summarised why we actually do surveillance and the purpose of surveillance. I can maybe just push it a little bit and say when should we start surveillance? So this is difficult. I mean, historically we had the Eden et al metronalysis where we saw sort of inflection pointed about 10 years. But when you look at the data in that metronalysis, the sort of midpoint of it is sort of the 1950s and 60s. So that's very, very different to our current practice. So I think the key thing is we should choose that the start point is the start of symptoms, not from where you get diagnosis. So if you've had symptoms for eight years, you're already ready for surveillance now. Why we chose eight years, other guidelines have chosen somewhere between six and 10. And we think that you need sufficient inflammation to drive your cancer risk. And therefore to some extent it's a slightly arbitrary number. Again, if you look in the, for instance, teen jess's population base work, the cancer risk only starts to become significantly above population risk out to maybe 12 or 15 years. But equally, you know, that there are some patients perhaps who have had kind of slow burn disease where we might not really have known. And they, you know, they seem to develop cancer early, equally patients with PSC clearly that the risk and that kind of curve divergence of cancer risk starts quite early. So that they're on surveillance from their disease inception. And I think that's the really a neat way to think about it. It's been very clear in the guidelines this time that we start to eight years from the onset of symptoms, taking to account in that burden of disease people have before they get diagnosed. But for PSC, we do it at diagnosis. And that's because we believe that PSC in itself as an independent risk factor for colorectal cancer. What about stopping surveillance or suspending surveillance? Because that is something I think we have to consider. And firstly, because as you've rightly demonstrated that cancer risk is not the same in everybody, it's dynamic. We've also got a huge number of patients now who I will have had disease for a number of years. And if you estimate that from a population perspective, I suspect about half of the population in the UK, which was about half a million that have IBDS, suspect you know 250,000 patients might be in need of surveillance. And you've also then got to think about someone's frail to index and whether they actually receive any benefit from going through regular colonoscopies. And the burden that gives to patients and services. So I guess the new thing in this guidelines, we've started to think a little bit about how we risk stratify better. And part of that, we've kind of started to discuss when to stop surveillance. So I wonder if you can maybe elaborate that on a little bit. So I think you're absolutely so fundamentally surveillance, it is about trying to improve patients quality and length of life as our presence. and practical medical interventions. And so if patients are frail, and they have well-controlled disease, they've got a low risk of developing cancer, they've got an increased, or low risk of developing cancer in their lifetime. But I think we substantially underestimate the burden of surveillance canaloscopy. It's not that patients don't like surveillance canaloscopy. Patients hate surveillance canaloscopy. So, putting elderly patients through relatively high risk procedures with bowel prep and sedation, if it's not absolutely clear that it's going to offer them benefit in their lifetime, is not good medicine. And so we need to think about, or we need to move away from kind of sort of wrote a following of surveillance guidelines to see the patient in front of you. There's strong emphasis in the guidelines that these are decisions to be made with patients, and that we take into account, what is their realistic life expectancy? What are the risks for that particular patient if they have renal failure or significant cardiovascular risk disease? What are the risks of them having repeated colonoscopies? And what are the benefits? And we know that for patients with well-controlled disease who haven't got dysplasia or other significant risk factors, actually their risk is very low. If they have population-based risk, if you are in the population, you get 75, actually we don't really recommend further surveillance for you. And so, equally in our IBD cohort, we should apply the same logic. And I think that's a really good way to think about applying that same logic. As well, we don't have direct evidence from randomized controlled trials. We do have evidence from the work that's come from St. Marks and from other, I think the Danish team have produced some information that allows us to think about risk. Can you be elaborate on that a little bit? So I think you're leading to Matt Rutt's recent work on the thinking about the risk of polypectomy in patients with limited life expectancy. And I think this has been a super helpful piece of work that essentially matches up your risk of dysplasia becoming cancer versus your risk of having a severe adverse event. And that means you know, a kind of re-admission to hospital type event for patients who come for colonoscopy and therapy and it's with and without anti-carragulation. And you know, what that shows is for certainly for many older patients who are carrying comorbidities measured by the Charleston comorbidity index that actually risk of intervention rapidly starts to exceed the benefits in terms of their risk of colonorectal cancer. And while that's not directly applicable to IBD dysplasia, I think conceptually it's a very helpful way to see it because I think doctors underestimate the risk of intervention and they overestimate how long people are going to live with good quality of life. Yeah, and I think that's a lot's beautifully coined about how we maybe need to start thinking about the surveillance programs and inflammatory bowel disease because I think there's going to be a mismatch between demand and our capacity and the perceived benefit of this. So thank you for summarising that. And I think aspirationally I would love us to have that kind of framework in that paper thinking about someone's frail to index and the benefits or the risk of cancer over time. So I guess what I might ask you to then maybe talk about a little bit is about how our cancer risk estimation is different this time compared to what we did in the previous guidelines. So historically we mainly had data on single risk factors and that relatively kind of corraled those into a series of low moderate and high risk groups and then from that it floed how frequently you get surveillance with lower risk groups having five years then moderate at three years and high risk at one year. And at that time there was almost no multivariate analysis available although it's clear just from first principles that those risk factors must interact inflammation, previous dysplasia, age, sex, other things, post-inflammatory problems and strictures. And so in the current guidance we have, you know, we've looked again at those single risk factors, tried to use more modern data because again some of that data goes back well into before the 1990s which doesn't reflect modern practice and where possible use multivariate models but that still is problematic that you know you're broadly getting a single risk factor. It's the top Trump's thing. What's your top what's your best card and that's what defines your surveillance interval. And so it was very exciting just as the guideline development group was gathering that a Dutch group led by Bazarmburg produced a multivariate risk score that included eight different risk factors and used them all together and rather than that feeding into a surveillance interval it just gives you a percentage risk. So in the paper they they've they've been able to calculate your five year or your ten year risk and you can you know you can adjust the variables so as patients pres you know diseases progresses over time maybe they get worse inflammation or they develop dysplasia you can include those things so that the risk that their risk level changes and potentially their surveillance interval. What the paper didn't do was make that clinically you know enable us to operationalize it clinically and therefore part of the task of the guideline committee was to say out at what levels do we consider risk to be high enough that we should trigger three yearly surveillance one yearly surveillance or even high enough that actually you know surveillance is inappropriate and we should consider collectime and so there was a risk thresholding exercise that's actually quite challenging because you have to do it before you see the data to try and reach risk thresholds from our code of the new place that trigger certain surveillance intervals so that was one challenge the other part is that essentially the model that the Dutch group produces is on an Excel spreadsheet and that's not usable in clinic and therefore we built an online risk calculator that essentially gives you an access to a cut down version of the model it only allows you to look at the five year advanced code of the new place at risk and in areas where the data set was quite thin so extremes of age we didn't have many 90 year olds with with inflammatory bowel disease equally beyond 30 years we don't have much data for more than 30 years surveillance so in areas where the data set was thin and confident intervals would get wide we've restricted it but it does it otherwise it will allow you essentially to use that model by entering the risk factors with little radio buttons and then it will output a five year advanced code of the new place your risk. What it doesn't do is output a surveillance interval and deliberately so because there are other factors to consider you'll see on the flow chart there are special circumstances and equally the risk model doesn't include the risk of first-degree relative with code of cancer so you need with the patient to sit down and say this is you know this is your five year risk these are the other factors we would consider and we think this would then lead to we could consider a three year or one year surveillance interval critically it is quite likely that this will not be the same as the top trumps interval so it's it's a different methodology and by combining risk factors actually you get a much more precise estimate that is personalized to the patient and we think probably the guideline committee supported this as as our preferred strategy but it's a big change and so that you know both options are available and that we hope with time as people get used to multivariate risk calculation and there is further validation that we move strongly towards that as our future strategy. And I it's up for somebody who likes data I actually like the fact that we can actually translate our clinical experience and skills into that risk thresholding exercise where we came out with percentages about you know either odds ratio has a ratio or the actual absolute risk that somebody has for cancer and then I think it's very unique that the work that you actually and Gora for the lead from Oxford to be fair has transformed that information into an online calculator because it becomes immediately clinically useful for everybody. I guess what you're saying is that we prefer that model because it integrates a number of different risk factors. It is you know a difference of what we were doing before and so it was always important for us to try and provide the two different approaches. So our highest top Trump's model goes for the odds ratio. It has a ratio over population-based risk that we understand from patients in IBD whereas the multivariate risk category is much more personalized to the patient and gives you I guess an absolute type risk for that patient with that be a fear. That's a kind of dichotomy that you get and certainly quite a lot of experience with patients suggests that this idea that it is precise and personalized is quite appealing and particularly that it also it uses present day data. So the idea that for the multivariate risk model, the data is actually relatively modern, whereas for a lot of the single risk factors, it goes back quite a long way. And even though we've tried to update it, that some of the data is still perhaps reflects practice from a different era. Yeah, and I think the feedback that we've had in the community so far, within the UK and far, I feel, is that they do actually like using CalCla later. We've planned a webinar to go through some of the more detailed information about how it's built the limitations that you've discussed, some of them, and where we might go in the future with this. But I think it's certainly the direction of travel, certainly for a complicated disease like IBD that is dynamic and changes over time. And so this is something that actually I think is very welcome to buy many in the community. If I maybe go on and then just ask you a little bit more technical aspects of some of the different things that we've talked about. So we've talked about, I guess, who should get surveillance while we actually do it? The purpose of counter-risk ratification, how should we actually do it? But this won't necessarily include any of the endoscopic procedures, but how should we actually do surveillance? What would be the two or three key messages that you would want the audience to be aware of and used to actually implement in their clinical practice? So I think that there have been a few changes kind of around endoscopy. Certainly thinking about bowel preparation. It turns out, patients don't like having surveillance can't be, they really, really don't like it. And one of the things that they find most distressing is about preparation. And so the guideline contains a new metronalysis that looks at bowel preparation. And I think there are a couple of key messages that come away from that that low volume pegs, so two litres or less, equivalent cleansing to four litres, but much better acceptability. And then there are some new bowel preps out, so peak or sulfate or all sulfate-based preps, which interestingly, so this is data that post-date the ESG-200019 bowel prep guidance. And in those studies, so comparing two litres of peg versus oral sulfate versus or peak or sulfate, again, equivalent cleansing, but maybe a bit more tolerable, so easier to finish the prep. And so this is not, you know, we're not trying to choose one prep or the other, but to give patients a range of choices, volumes, flavours for something where they can find an acceptable prep for their procedure. And, you know, fundamentally, without good prep, surveillance is impossible in futile. So we want them to have good prep. First time gets available, it's done. Hopefully forget about it for a few years. It's great, so that's a very key message for us to focus on and try and improve overall. What are there any other key aspects that you'd like people to take away? So I think perhaps one of the controversial bits of the guidelines is about taking caudrantic biopsies. And I know many people have, I think, felt beyond quite sad that this seems to have reappeared. And, you know, part of the reason it's reappeared is because it is clear that even with chroma andoscopy and targeted biopsies, there are still some patients, they tend to be high risk patients, who benefit in terms of displays detection from additional caudrantic targeted biopsies. And that the two key groups here are patients of PSC and patients with previous dysplasia where yield is increased. Part of that may reflect the pathologist slightly changing the goal posts on what counts as dysplasia. So this idea of non-conventional dysplasia that can be very subtle and is often endoscopically invisible. If we're going to detect that, particularly for PSC patients, taking biopsies is critical. And so many people have felt that's a retrograde step, but I think the data, you know, there are a number of studies that support that fairly consistently. And in fact, it's not a huge proportion of your IBD patients. So the combination of patients of PSC and patients with previous dysplasia is probably only 10% of the surveillance population. So we're not saying you have to take caudrantic biopsies every time, still probably chroma and oscopy with high definition, light is our optimal detection technique with targeted biopsies. But in this small group, it looks like there is additional value in caudrantic biopsies. Yeah, I think that's a very good point. I guess the other sound bite is just to say that we still want people to take biopsies for disease activity assessment. Because I think that's something that's often just forgotten when you're doing a surveillance procedure. There's no targeted lesion that you need to biopsy, but you still need to remember to take some biopsies from the right left and rectum for disease activity, because that will help us inform the future of what the next surveillance would be, especially if we're going to use histological inflammation and grading to actually inform that. You're going to have another podcast with Professor Bindari talking about the end-scoping modalities and sort of how you manage dyslysia. If I can maybe just ask you, who do you think should be doing surveillance? And I think the other new thing about these guidelines is that we did actually consider service provision and we did this time round think about or it just introduced that topic that we need to have some training or teaching in this aspect. So who do you think from your experience and who's someone who's done lots of endoscopy both in IBD patients and non-IBD patients and that experience should bring, who do you actually think should be doing these procedures? So I think I think as you lead to, we should be thinking about an IBD surveillance service. It's not about individual endoscopists although they're important. So it's how you run your service and whether everybody should be doing delivering canoscopy as part of an IBD surveillance service or whether it should be restricted to a smaller group of individuals. And I think certainly the guideline development group felt that possibly a degree of restriction probably is appropriate. You do need to have appropriate training even if you can technically do diespray in the sense that you can apply the dye to the colon in a consistent manner. You then have to have the experience to interpret those images and turn that into either biopsies or a section or a strategic management plan. Historically, I think it's fair to say training in IBD surveillance to say there's not much of it. It's been very ad hoc in your lucky if your mentor happened to do it for you but in terms of structural training, there is none. The only real piece of training, a mariaturic teachers, has produced an optical diagnosis web tool that certainly is evidence-based in terms of improving performance and can be recommended. But I think it might be something for Jag and the BSU to look at whether we actually need training and not necessarily accreditation but appropriate experience. And again, one of the things that patients want, they want pain control, they want good communication and they want to be confident in the technical skills of the endoscopies doing their procedure. So to deliver what patients want, we need to be appropriately trained and deliver that service as probably a smaller group of endoscopies. And just on that, and I know this is a point map right, it feels very strongly. Not only is this a, it's actually quite one of the most challenging diagnostic exercises, a badly scarred post-inflammatory colon that you've got to examine and find one bit of dysplasia and feel the post-inflammatory polyps. It's very hard work, it's tiring. Whether we should have a full list of four IBD surveillance canoscopies are on a list because you do need a bit of extra time. But equally, the people who are doing IBD surveillance canoscopy are a bit like doing family cancer surveillance. You need to understand it's not about this canoscopy. It's about their lifetime of disease and therefore it is critical that you don't rough the patient up and give them a bad experience. And so that means doing it, you know, a technically good endoscopy with perhaps underwater intubation and position changes. But critically, you just need to give a lot of sedation. So I start in titrated doses now for pretty much everybody with ulcerative colitis or cronocies. When I'm doing IBD surveillance, I start with 500. And if there's any indication that the patient's not enjoying it, I give more sedation. And I think I'm reasonably technically good at a loss of this. But, you know, the patients have had significant bowel inflammation, if they have this right sensitivity, being technically good is not enough that they actually need proper amounts of sedation. And, you know, and we should consider proprfol, if patients can't have a comfortable examination because they're going to have to have this again and again again. If you have PSC or annual surveillance, so you've got 40 colonoscopies to have. But I think there is very clear data that patients find the whole process, distressing, they blame themselves, they feel lost of control. But critically, they're very worried about pain. And I think we seriously underestimate this and I think the data supports that. And so you just, you know, if you think you're giving enough sedation, just give a bit more. And to have a very lay threshold to go further, even beyond 500, assuming, you know, the other physiologic parameters are okay. and equally think about full patients who are still finding it difficult. Think about pre-prefall and critically ask your patients what their experience is. So this idea which is quite strong in the guidelines, in fact very strong and in fact is the idea that we should collect patient reported experience measures. So that essentially means, you know, asking them a questionnaire to do once they've once they're back at home. You can use a validated questionnaire like the Newcastle IBD endoprem which is quite long but covers a lot of ground. It doesn't have to be every patient so it's not going to destroy your service by having to do this. But once in a while for all of your IBD endoscopists and therefore individually and as a service to have feedback on what the patient experience is and then to take action on what the feedback is. So if patients are still saying it's still hurting me and pain is a big problem even if that's in a reason, you know, if it was 20%, but if you have to have a colonoscopy every year and 20% of the time it's going to really hurt. I don't think, you know, that's not really acceptable. It's going to be very hard to get people to come back and if people don't come back that's how they get cancer. And I think that's a very well made point is that we need to make sure that the patient experience actually from their perspective is a good one that we can offer. And just to be clear, I guess, when you say five and a hundred, you mean five milligrams of metaslam and a hundred of fentanyl. Yeah, absolutely. Or equivalent. Or equivalent. And I think just for readers or sorry listeners who may not be aware, when we talk about Professor Matt Rutte, he is the clinical chair of JAG. JAG is a joint advisory group in the UK who are the organisation that sets standards and quality for endoscopic procedures. And as you shared with us, we don't have one specifically for inflammatory bowel disease at the minute. If any listeners are looking for a little project, this would be a great one to get started on. Absolutely. So I guess that's been a really comprehensive discussion about our guidelines. I'm sure there's a lot of cells that would not be able to cover just some final thoughts from you. What would be your final take-home message from all of this that we've discussed? I think one of the things that came out amazingly strongly in the guidelines is this idea of concordance. So it's not about telling patients to attend. It's helping patients be concordant with their surveillance interval. And if you look across international data, patients only turn up the appropriate surveillance time about a half the time. So we need the entire IBD team. So that means that means you use the clinician, your IBD nurses, your medical secretary and the endoscopy booking team all to help patients attend on time. Because we know that there is no cancer prevention if the patient doesn't attend. Yeah, and I think that's a fundamental point isn't it, is that we need to restructure our services, our expectations, patients expectations, to actually deliver an IBD surveillance program that's actually fit for purpose. So we also asked chat GPT what they thought about our new guidelines. And James, do you want to tell them? I thought they're very complimentary. They said overall improvement, more personalized evidence-rich endoscopy-focused and patient-centered. And certainly that last bit I think is critical. And actually that's not a bad summary. Okay, yeah, I agree with that too. So thank you for joining us and hope that you've enjoyed that podcast. As I said, there will be a second episode that will be more focused on the endoscopic and tech aspects of managing dysplasia and shared decision-making. Just to get a plug, we do have a cancer estimation webinar that's planned and I hope you will be able to join us. If you're not able to join us live and please do watch it on demand. So, for us to reach, thank you for coming to speak with me today. I really enjoyed that conversation and hope that the listeners enjoyed too. To you to thank you, it's been a great pleasure. [BLANK_AUDIO]