The next talk, so Dr. Iyer will be talking about screening and surveillance in Barrett's esophagus, sponge biopsy, or brush. Thank you. Okay. So I think before we get into this, I would like to go through a brief discussion on why should we be screening and surveying patients? So why should we be screening for Barrett's and why should we be surveying patients with Barrett's? Important for us to know. So if one were to summarize our current Barrett's esophageal adenocarcinoma paradigm is that we think patients who have reflux with some risk factors that we talked about before are at risk for developing Barrett's esophagus, which then develops through dysplasia, low and high grade dysplasia into adenocarcinoma. And we think we can intervene at multiple points along this paradigm. So the first is by screening or evaluating patients who have reflux and other risk factors to detect the presence of Barrett's and then put them under surveillance to detect the presence of dysplasia or early cancer. And then finally treat those with dysplasia or early cancer and prevent this progression of Barrett's to esophageal adenocarcinoma. That's at least a theoretical premise that we are operating on. The big reason why we are interested in esophageal adenocarcinoma is that its incidence has really increased over the last three to four decades, especially when we compare it to other solid cancers like breast, lung, prostate, et cetera. And not only is its incidence increasing, the mortality that is associated with esophageal adenocarcinoma has also increased. And when we look at overall, when we look at survival in patients with esophageal cancer, we still see that its survival is about 20% or less at five years. And really the only group wherein the survival has increased or improved is that of patients who have T1 or early stage esophageal cancer. We now have level one evidence that if we find dysplasia, whether it is high grade or low grade, and we treat it endoscopically, in this instance, high grade dysplasia and low grade dysplasia, these are two trials, landmark trials, one published in the New England Journal of Medicine, the other in JAMA, if we treat patients and we compare them to patients wherein only surveillance is being performed, we can reduce the risk of progression to esophageal adenocarcinoma. So now we have level one evidence to suggest, to prove, that if we endoscopically treat dysplasia, we can reduce esophageal cancer risk. Not only that, if we find early cancer and we intervene endoscopically, this is just one study that we published now almost 12, 13 years ago in gastroenterology, and this has been replicated by several other centers, that if we treat early cancer, we can now change the natural history of this disease. So T1 esophageal cancer, particularly T1A cancer, at five years, the survival is substantially better from the 20% we talked about before to 85% or 90%, and it is very comparable in the endoscopic group versus the surgical group. So we now have a rationale for the early detection of Barrett's and Barrett's related dysplasia and neoplasia. We can not only reduce incidence, but we can also improve cancer-free survival in these individuals. So this is the rationale, but we face several challenges. Our biggest challenge, I think, is not only challenge, but the biggest challenge is how can we screen for Barrett's? And the only technology we have available right now is really sedated endoscopy. And this was a cost-related study we did from a randomized study, a randomized trial that we performed in Olmstead County, Minnesota, where we looked at the 30-day direct and indirect costs of Barrett screening. And this was in a Medicare population, and this was over $2,000. And we know that this makes it expensive, not cost-effective. Endoscopy is invasive, requires sedation, requires a trained endoscopist, and of course is unsuitable for widespread application. And not surprisingly, studies have shown that only 10% of those who are eligible are actually getting screened. So of course, this leads to what we feel is a missed opportunity in early Barrett's detection. So if you look at studies which have looked at the proportion of esophageal adenoparsinoma diagnosed in surveillance, it's only about 10% to 12%. So the vast majority of patients being diagnosed with esophageal cancer are being diagnosed outside of surveillance programs, even though we know that 60% of all esophageal cancers have Barrett's present at diagnosis. And if you look at early-stage cancers, almost 90% of them have Barrett's alongside the cancer. So obviously, this is a huge missed opportunity, and this is also in parallel with studies that we and others have done showing that only 33% of all prevalent Barrett's is really under surveillance. So we are missing the vast majority of patients with Barrett's who should be under surveillance, but unfortunately are not because they don't know they have Barrett's. So what are some solutions that we can come up with? So the big drive in the scientific community has been towards coming up with a minimally invasive screening tool. And why is this important? Because detection is not just a function of sensitivity. So you could have a very sensitive and specific tool, but if patients are not participating and there is no access to these tests, then detection is going to suffer. So the goal with coming up with non-endoscopic tests is to improve access, move the test from physicians to non-physicians who might be more available, improve access. Hopefully this will lead to better participation, and the lower cost might make this more cost-effective and promote reimbursement by payers. So we need all of these elements to happen in parallel before we can get this really widely disseminated into the community. So in order of decreasing invasiveness for screening, we went from sedated endoscopy to unsedated transnasal endoscopy, which unfortunately for a variety of reasons has not really caught on. And in the interest of time today, I'm going to focus on two technologies. The first is swallowed esophageal sampling devices combined with biomarkers. And the second is XA volatile organic compounds. So the first of these technologies combines non-endoscopic esophageal sampling, which provides a rich cytology sample of the entire esophagus using a variety of devices. So there's esophacab or the cytosponge, and essentially these are pieces of compressed polyurethane foam which are put inside a dissolvable capsule attached to either a cord or a suture. These are swallowed with a few sips of water, the shell dissolves, it releases the piece of polyurethane foam, which can be sized from either 25 to 30 millimeters. And then these are pulled out and they provide a rich cytology sample. Another device is called the isocheck balloon, which is a piece of, which is a silicone balloon, which is inflated, pulled across the G-junction, deflated, and then pulled out. Concept is the same. It provides us with millions of cells. And then these cells can be analyzed for biomarkers. And these biomarkers are a combination of either protein markers like trifoil factor three or methylated DNA markers. An advantage of the methylated DNA markers is that they can be quantitatively assessed. You don't need a pathologist to assess them subjectively, and they become a little bit more easily scalable. And we and others have done a series of studies looking at some of these markers and have shown promising accuracy. And this is just a summary of all the studies that have been done. These are case control studies so far. And again, these have used methylated DNA markers or trifoil factor three, which is a protein marker, both in the United States and in the United Kingdom. And several investigators have been looking at this. And again, these have given us a range of sensitivity and specificity, some of which are approaching or exceeding 90%, making them promising tools. All these case control studies have also shown that this is well-tolerated, they are safe, they can be done by nurses in the office in less than 10 minutes. And the biggest issue that we have to deal with is sometimes detachment of the sponge from the string. So beyond the case control studies, in the Lancet, Dr. Fitzgerald from the United Kingdom published the first community multicenter pragmatic randomized control trial. The setting of this was in the primary care offices in the United Kingdom, and 113 GP primary care clinics were targeted. Patients on PPIs for at least six months were randomized into two arms. One was a cytosponge arm, the other was the usual care arm. The usual care arm basically had patients who saw their GPs and got an endoscopy only if the GP felt it was clinically indicated, whereas on the cytosponge side, everybody was offered the cytosponge. So there were roughly about 6,500 patients in each group. On the cytosponge arm, about 40% were interested, 60% of them underwent the cytosponge test, of which about 230 were positive. If you got a positive test, you underwent an endoscopy. On the usual side arm, not surprisingly, very few people underwent an endoscopy. And what was striking, but probably not entirely surprising, was that you were able to pick up almost 10 times the number of patients with Barrett's on endoscopy. So that was the primary aim of the study. The primary aim was, how often do we pick up endoscopy, sorry, do we pick up Barrett's if you were to do a cytosponge on these patients versus the usual care arm? So it met its primary aim, not surprisingly. I was more intrigued by the fact that not only were you finding Barrett's overall, more Barrett's overall, but you actually found more dysplastic Barrett's and you found more stage one cancers, which could then be treated endoscopically or surgically. Not surprisingly, on the usual endoscopy arm, there were no dysplastic Barrett's found, and they ended up finding three advanced stage esophageal cancers who had to be treated with the usual chemotherapy radiation esophagectomy. The predictive value of a positive cytosponge was 60%, which was good. The sponge was safe in about 2000 sponges, one detachment, 4% had a sore throat, acceptability score was high at nine on a scale of one to 10. So the next group of the next technology that is being evaluated and developed for the detection of Barrett's is exhaled volatile organic compounds. So this basis, it's identification of Barrett's on metal oxide sensors, which will interact with exhaled volatile organic compounds. So again, you can see this patient or volunteer breathing in and out of this handheld device for about five minutes. The interesting thing is that you don't want to lose any of the exhaled volatile organic compounds through the nose. So you actually have to wear a nose clip and you have to breathe in and out of this. And in this handheld device, you have these metal oxide sensors, which then develop a digital breath print using artificial neural network analysis. And then I was able to differentiate between cases and controls. And in a paper from the Netherlands, sensitivity of 91%, unfortunately with a fairly modest specificity, but when you look, increase the specificity, you can gain some sensitivity. And again, as you can see in all comers, the area under the curve was 0.91 in patients with GERD, it was relatively low, but again, proof of concepts study more than 99% of patients who were offered this technology actually underwent the procedure. So that's a big plus. So again, these are technologies that are promising, are at various stages of development and could be coming down the pike in terms of finding patients with BARRIT. So let's switch gears and move to dysplasia detection. So dysplasia detection right now is the sole determinant of our management recommendations. So if patients have no dysplasia, we are recommending surveillance. If they have low grade dysplasia, we are recommending either endoscopic therapy or surveillance. And if they have high grade dysplasia, we are recommending endoscopic eradication therapy. This is really based on the risk of progression that we understand is based on dysplasia at baseline. Now, unfortunately, there are several limitations to the detection of dysplasia and surveillance. We know that most dysplasia or cancer in BARRITs is subtle. It is difficult to recognize. We know that unfortunately, despite guidelines, despite guidelines to follow the Seattle Protocol, we are not compliant with recommendations. Even if we are compliant, pathologists have a very high inter-observer variability, which is why we recommend that at least two pathologists, at least one GI pathologist should be reading and confirming diagnoses of dysplasia. The pathologists have a very low inter-observer agreement, particularly for low-grade dysplasia. And what does this lead to? This leads to missed dysplasia or cancer, almost a third to a quarter of our dysplasias and cancers that are prevalent are unfortunately being missed. So this is just a reflection of how we are doing our four quadrant biopsies. And as I like to show in this cartoon, even if we were to take all the biopsies that we are recommended to take, we are sampling maybe only 5% or less of the at-risk Barrett's mucosa. So most of the Barrett's mucosa we don't sample. So is there a way of improving the sampling of the Barrett's mucosa better than just using Seattle protocol biopsies? So this is where a brush-based technology comes in. The technology is called WATS, Wide Area Transepithelial Sampling. What this uses is a endoscopic brush, but it is a bit more stiffer and it samples a bit more deeper than the typical cytology brush that we use. So in addition to our Seattle protocol biopsies, we can now sample the mucosa going up and down the entire length of the Barrett's mucosa. And once these samples are obtained, they are put on slides. The slides are sent to the center in New York where an artificial neural network then creates the, it assesses the entire slide and it then focuses the pathologist's attention on the most abnormal 200 cells. And then there are cytology criteria for dysplasia and non-dysplastic Barrett's which the pathologist can use and then report this out as either non-dysplastic Barrett's, low-grade dysplasia, high-grade dysplasia for cancer. So there are several studies which have looked at the increase in the yield of dysplasia with Barrett's compared to that of biopsies alone. And I'm just going to highlight one study. This was systematic review and meta-analysis where we pulled results from seven studies in over 3,000 patients. This just came out. And again, it is not a surprise to me that if you look at this summary slide here, so this is the blue is any dysplasia, the orange is high-grade or cancer. And you can see forceps biopsies are detecting dysplasia in about 15% of patients. But when you add works, you increase the yield of dysplasia by about 7%. And I like to keep this as absolute increase because relative increases are somewhat difficult to interpret. Even when you look at high-grade and cancer, which is a group that we are really focused on and we know we all agree on in terms of needing treatment, you're almost doubling the rate of dysplasia that is being detected. So you have about 2.3% here on forceps biopsy going up by about another 2% with Watts. However, some caveats, Watts was false negative in almost 60% of cases where forceps biopsies identified dysplasia. So you have to understand that these are complimentary technologies. I don't think we are at a point where we can say we will only do Watts. I think Seattle protocol biopsies still have a role. Most of the incremental yield was not confirmed by forceps biopsy histology. All the Watts 3D pathology is still read by the CDX, which is a company that has developed and provides this service. So it is unclear as to if the same results can be applicable if non-CDX employed pathologists can read these slides. And then of course, does Watts 3D diagnose low-grade or high-grade dysplasia have the same risk of progression as forceps biopsy, low-grade or high-grade dysplasia? And then I should also note that a reason for this big gap between what Watts picks up as any dysplasia versus high-grade or cancer is the addition of a criteria called crypt dysplasia, which the significance of which is not yet fully understood. So what are some considerations I would have with the use of Watts 3D as an adjunct to Barrett surveillance? I think it should remain an adjunct with Seattle protocol biopsies at this point. I think we should be reconfirming Watts only low-grade, high-grade or cancer with forceps biopsy with a follow-up endoscopy where we do a good examination, we use narrowband imaging or any other technique of virtual chromoendoscopy, use a cap and maybe refer the patient on to an expert center for treatment. As I said, the significance of Watts 3D crypt dysplasia is unclear. And then a recent retrospective study which actually came out in GI endoscopy did look at the natural history of patients with Watts 3D low-grade dysplasia. And indeed it does look like they are at a higher risk of progressing to cancer or high-grade dysplasia. So I think we are beginning to generate some data that is solely needed in this field. Last part of this talk is in terms of prediction of progression. As I alluded to before, our recommendations of progressing to high-grade dysplasia or cancer in those with Barrett's are not really personalized. We are only using dysplasia as the sole risk factor for predicting progression, even though we know that there are several risk factors such as age, male gender, smoking, body mass index, length of the Barrett's, which are also important in predicting progression. We also know that patients with non-dysplastic Barrett's progress. What are the predictors in them? And unfortunately what is happening is that surveillance as we currently practice is not cost-effective given the low rates of progression. So if we are able to personalize prediction of progression, we can then personalize management. So in those who are at low risk, we could say you either decrease the frequency of surveillance or stop surveillance. In those who are at high risk, we can bring them back for more intensive surveillance or maybe even proactively ablate them. And this has a potential of making Barrett's management more cost-effective and decreasing costs of ineffective or unnecessary surveillance. So there have been attempts at making and creating and validating Barrett's progression scores. So this was one such attempt by Dr. Sharma's group using the BEST cohort, which is a multicenter cohort of almost 4,500 patients. And using these cohort, they came up with a progression in Barrett's or PIB, risk progression score. And this combines male sex, cigarette smoking ever, which is a combination of current or past, Barrett's length and confirmed low-grade dysplasia. And using this, they have come up with a risk pyramid approach for defining the risk of progression. So if you have low risk, you are between zero to 10 points. Your annual risk of progression was only 0.13%. Intermediate, 0.73%, 11 to 20. And high, more than 20, higher risk of about 2%. And what they showed in the study was that in their test cohort, the C-statistic or the model accuracy was about 0.7. So remember, C-statistic goes from 0.5 to one. The lesser it is, the worse it is. The closer you are to one, the better the prediction of the model. And again, as they were able to show, compared to a low-risk group, which is in blue here, the high-risk group does have a much higher risk of progression. And their suggestion was that if you fall in this low-risk group with zero to 10 points, you could maybe even discontinue surveillance. If you are in the intermediate or high-risk group, you should continue surveillance. Now, as I alluded to before, this is in a referral center cohort, is using some of the more common clinically available data, but it is unclear if you remove the low-grade dysplasia in patients with no dysplasia, can you still use this score? So this is where there are some newer technologies. I would like you to be aware of. So this is called the tissue cipher assay. This is a lab-developed test that is done in a CLIA-certified clinical laboratory. It uses a series of biomarkers. There are actually nine of them. In combination with nuclear morphology, this is done on a multiplexed immunofluorescence scanning platform. And this uses a combination of 15 markers, biomarkers and morphology, and comes up, the platform uses, so it's not subjective, it is an objective quantitative platform, which integrates these 15 features to provide a risk score going from zero to 10, and then divides patients into a low-risk class, an intermediate risk class, and a high-risk class. And these reports can then be provided to the provider and are actually clinically available. And we recently performed a pooled analysis of four international multi-center studies. These are studies done at the University of Pittsburgh, Geisinger Medical Center, Cleveland Clinic. Dr. Thotta is a co-author on several of these papers, and these have been either case control or nested case control studies involving patients from the Netherlands, as well as from the SERP study that I alluded to earlier on, which was a randomized study, randomizing patients with low-grade dysplasia to either ablation or surveillance. So we looked at a total of 475 patients who had either indefinite dysplasia, low-grade dysplasia, or mostly non-dysplastic barrets. And we performed a pooled analysis of these progressors, which we defined as patients who went on to high-grade dysplasia or cancer, more than a year after initial diagnosis, compared them to non-progressors. So about a third of our patients progressed and the remaining did not progress. And they all had their baseline tissue cipher risk class assessed, high, intermediate, or low. And about 16% of patients had a high risk. Of these, 12% had non-dysplastic barrets, 22 had low-grade dysplasia, and only five had indefinite dysplasia. And we looked to see how well did the baseline tissue cipher score, whether it be low, intermediate, or high risk do in predicting progression when adjusting for some of the clinical variables like gender, length of barrets, presence or absence of hiatal hernia, and age. So when we looked at the performance characteristic of this in the entire cohort, and we divided, we dichotomized the cohort into either high risk versus intermediate or low risk. And again, what we found was the specificity was pretty high. So very few false positives. If your test is positive, you are at high risk of progressing. The sensitivity remained unfortunately a little low. So the sensitivity was only about 40%. So what that means is that the test was only able to pick up only 40% of the progressives. Now, if you widen the definition of what a test is positive, so if you combine the high risk and the intermediate risk class, and you compare that with only the low risk class, then you gain sensitivity, 55%, but you lose a little bit of specificity. Remember, we talked about what happens in patients with nondisplastic barrets. Here again, the same pattern applies. Very specific, not so sensitive. If you widen the class of patients, you're gonna call it positive, you gain some sensitivity, but you lose a little bit of specificity. So the advantages of a test like the tissue cipher is that it could be done on, it can be done on paraffin tissue. It is automated. It is quantitative. So it is not subjective. The cutoffs for calling the test positive or negative are validated in multiple independent sample sets. As I showed you, it is very specific and can be used in patients who have no dysplasia. Its utility in those with low grade dysplasia is somewhat lower because we know we are going to recommend ablation to some of them, but if you are on the gray zone on whether you want to recommend ablation or not, it could be used as an adjunct to just histology. Limitations of the test is its modest sensitivity, and you would really need to think about how to improve it. Could you use multiple levels of biopsy? Could you use biopsies from more than one endoscopy? Or is it better to more comprehensively sample the Barrett's mucosa and go to brushings or maybe even some of these capsule spun samples? And of course, the cost of the test has to be balanced against any change in clinical management. So in summary, minimally invasive, non-endoscopic tools for Barrett's detection are making progress. As I have shown you, these are accurate, they are safe, they are well-tolerated, and I think may enter the clinical realm in the near future. Dysplasia detection remains the cornerstone of Barrett's management. I cannot stress enough the importance of careful inspection, sampling any visible abnormalities, performing adequate surveillance. WATS, as I said, is a promising tool, but I think needs a few additional data points to validate its role. And lastly, risk stratification tools, both clinical and biomarker tools are making progress, and I think there is still need to improve sensitivity. Thank you.

Barrett's esophagus

esophageal adenocarcinoma

screening

surveillance

early detection

dysplasia

WATS