IBD Drive Time: Parambir Dulai, MD, on the Role of Biomarkers in IBD

In this episode of IBD Drive Time, Drs Millie Long and Raymond Cross talk to Dr Parambir Dulai about abstracts from the ECCO ’23 Congress on the role of biomarkers in the clinical care and research for inflammatory bowel disease.

Parambir Dulai, MD, is director of GI precision medicine and clinical trials at Northwestern University in Chicago, Illinois. Raymond Cross, MD, is professor of medicine and director of the IBD Program at the University of Maryland School of Medicine in Baltimore, Maryland. Millie Long, MD, is a professor of medicine, vice chief of education and director of the fellowship program in the Division of Gastroenterology and Hepatology at the University of North Carolina at Chapel Hill.

TRANSCRIPT:

Raymond Cross:
Welcome everyone to IBD Drive Time. I'm Raymond Cross from University of Maryland School of Medicine, and I'm here with my cohost Millie Long from University of North Carolina. And this is the third of our 3-part series of Best of ECCO in Diagnostics, Natural History and Epidemiology. And I'm delighted to have a return guest, my friend and colleague Parambir Dulai from Northwestern. So Parambir, welcome back to IBD Drive Time.

Parambir Dulai:
Thanks for having me. No, it's great to be back. I appreciate the opportunity. This was a lot of fun last time.

Raymond Cross:
Great. And it was nice spending time with you in Copenhagen. It was very fun. So I think you've identified 3 abstracts that you thought were particularly interesting and you're going to go over those. So why don't you jump right in and start with the first abstract.

Parambir Dulai:
Yeah. So I wanted to keep the theme of biomarkers as we begin to think about biomarkers and clinical care and research for IBD. So the first abstract I want to talk about is called PREDICTS. This says a study that looked at anti-integrin alpha V beta 6 antibodies to predict the diagnosis of ulcerative colitis and the identification of these autoantibodies up to 10 years prior to the disease and its post-association with disease outcomes. So this was a really well done study. They took an original cohort using a military database where they had banked serum samples prospectively, and they identified 82 patients who eventually developed ulcerative colitis and matched them to 82 patients who did not develop ulcerative colitis. And then went back and retrieved those samples and actually ran the ELIZAs to look for different antibodies that might be associated with the diagnosis or eventual development of ulcerative colitis.

So going back up to 10 years preceding the diagnosis, they were able to identify an antialpha V beta 6 autoantibody that was significantly higher in those patients who eventually went on to develop ulcerative colitis and the separation in the levels between those who developed ulcerative colitis versus those who didn't grew incrementally over the time period as they got closer to that eventual disease development. So it's a nice linear association for this. What they did from here is they did a really nice validation study where they actually had 3 separate validation cohorts. One was a Canadian cohort called the GEM Project where they're able to validate the difference in patients who developed ulcerative colitis and those who did not develop ulcerative colitis in an inception cohort and showed again a significantly higher value for the autoantibodies than those who did develop UC preceding the diagnosis versus those who never developed it.

And then in 2 inception cohorts at the time of diagnosis, COMPASS and OSCAR, which were North American, northeastern-based cohorts 55 and 104 respectively, they were able to actually look at the baseline value of these autoantibodies and its association with adverse UC outcomes defined as hospitalization, disease extension, colectomy or need for steroids, and escalation. And what they were able to see was that the patients who had the highest quartile of these autoantibodies at the time of diagnosis had the worst prognosis and those who had the lowest ones actually had the best prognosis with the lowest rates of complication. So this was a really nice study where they used multiple different cohorts and post-diagnosis to identify an autoantibody that could actually be used to predict the eventual development of the disease and that was associated with disease prognosis after the diagnosis.

Raymond Cross:
So Parambir, super interesting. I don't know much about this autoantibody, but is there some mechanistic reason why patients would develop autoantibodies to this that relates to pathogenesis or is this just something related to breaches in the epithelial barrier and permeability where we form antibodies or is it just unknown?

Parambir Dulai:
Yeah, I think it's probably unknown to some extent, but I think the assumption here would be that we've always thought of trafficking as probably the first immune mechanism to proceed the eventual clinical manifestation of the disease because there has to be an insult. You have to have immune cells traffic to the local GI side of inflammation, and then that persistent source of trafficking is probably what sets up the cascade of continuous chronic inflammation. So it's actually not too surprising that it's an anti-integrin, which is related to traffic and immune cell binding and cell migration that they're identifying. But I don't think anybody's ever really worked out the mechanisms behind this. The reason why this is interesting is because it starts to get into conceptual frameworks of prevention and trying to actually identify high risk patients, maybe screen for those who might be at risk for developing the disease and actually starting to think about prevention trials.

Raymond Cross:
Yeah. So it's in line with what the foundation wants to do with interception, right? So we now have some data from GEM on permeability being abnormal before diagnosis, other serologic markers. So maybe there's a future world here in high risk individuals, maybe first-degree relatives where you can check these things and hopefully we'll have some intervention other than, I mean, I guess we could offer a Mediterranean diet and maybe someday we'll have better probiotics or something else, but maybe we'll be able to stop this in 10, 20, 30 years before it ever-

Parambir Dulai:
Yeah, I mean I think identifying patients who are high-risk for developing the disease offers not only an opportunity to intervene, but probably also it offers an opportunity to do targeted high-dimensional omic profiling to really profile those patients and sort of optimize the ability to identify true mechanisms that might be driving it. Because I anticipate there's likely heterogeneity, but this is really just offering a little bit of a flashlight in was otherwise a dark sea with a limited lighthouse for where we're supposed to be going.

Raymond Cross:
All right. Well said. And very interesting. Why don't we move on to the second abstract?

Parambir Dulai:
Yeah, absolutely. So going on with the theme of biomarkers, we talked about a biomarker that precedes the diagnosis and is associated with prognosis, but then I wanted to sort of highlight a study that looked at biomarkers that predict response to therapy. So this is called the EPIC CD study, epigenetic response biomarkers to adalimumab, vedolizumab, and ustekinumab in patients with Crohn's disease. So this was a well done study where they took 2 cohorts.

One was a discovery cohort of 88 patients, about almost 30 patients per treatment group, and the other was a validation cohort of 96 patients, again about 30 patients per treatment cohort. And they had really standardized assessments for endoscopic disease activity with the SES-CD, clinical disease activity with Harvey-Bradshaw Index, and then biomarkers of inflammation with fecal calprotectin NCRP pre- and post-therapy. What they then did is they looked at DNA methylation to see if they could find markers that might be associated with response to therapy, and also looked at the stability of those markers over time to understand if they were simply markers of pathways or truly markers of treatment specific models.

And when they looked at this and they broke it down into the performance characteristics, what was quite striking was that the accuracy of the drug-specific prediction model for ustekinumab had an accuracy of about 94%. The vedolizumab model had an accuracy of about 89%, and the adalimumab model had an accuracy of about 73%. So for ustekinumab and vedolizumab, the models were quite accurate. For adalimumab, there was modest accuracy. And what was most notable in this study was that there were differences in the models between the 3 cohorts and it wasn't just a general marker of response to anti-inflammation or just improvements in inflammation. So I think this brings us closer to the ability to personalize some of these decisions and personalized therapy if we can transform these markers into usable biomarkers in routine practice at some point.

Raymond Cross:
Wow. We've been talking about precision medicine for as long as Millie and I have been faculty and we haven't really been close, but this looks exactly like what we need right now. I guess the question is, it adalimumab specific or would it be anti-TNF in general? And likewise for ustekinumab, would that include P-19 inhibitors as well as ustekinumab? We can't answer that from what they presented.

Parambir Dulai:
No, so that's a good point and we can't answer that. The other thing that wasn't presented, which hopefully will be presented at a future time, is how these drug-specific models perform in predicting response to the other drugs. So they drive it based on the derivation core for each separate drug. And then it looked at whether the model predicted that same drug response in the validation cohort, so the ustekinumab model for ustekinumab, but they didn't fully look at whether the ustekinumab model also predicted outcomes for vedolizumab and whether it also predicted outcomes for adalimumab. So at some point, integrating the 3 models together to be able to say that for this given patient, this is the right drug is probably where it needs to go. And we haven't seen that quite yet.

Millie Long:
That's interesting, Parambir, because that's exactly what I was thinking. I was like, "Gosh, this needs to be done, flips to the other drugs so that you can truly understand is it characteristics of the patient that will respond to any drug or is it drug-specific?" That'll be hopefully a great follow-up analysis for them.

Parambir Dulai:
Yeah.

Raymond Cross:
We're always thinking about sequencing, what to start first, second, and third. And I guess I wonder with this biologically, I don't know if it makes sense, but I wonder if this signal is stable from therapy to therapy. In other words, if a therapy fails a patient and then you go to a second line, is it slightly different. We can't again answer that with this.

Parambir Dulai:
Yeah. No, that's a great question. They did look at the stability of the markers over the initial exposure period and during that initial phase and they did show that the markers were quite stable in the early phases. So it's not clear yet if exposure to a drug will change these markers in a manner that would allow you to understand what the right next drug is. But it is definitely something that over time I think as they begin to grow this cohort, they hopefully will be able to integrate a lot of those questions in.

Raymond Cross:
Great. I'm going to turn it over to Millie for your third abstract, but before that she has some announcements.

Millie Long:
Yes. I want to make sure our listeners know that IBD Drive Time is now available on both Spotify and on Apple for podcasts. And so we hope that you'll find us there. I also want to make an announcement just to remind our listeners that the AIBD regional series not only has in-person sites, but it also has a series of virtual conferences. The next one is June 15th to 16th and registration is free. So I would encourage you to join and log on for some IBD education. We really find these regional series to be really clinically oriented and really great clinical discussion of cases and really up-to-date information on some of our newest drugs available in clinical practice.

And as always, I want to make sure that everyone knows that our podcasts are sponsored by Advances in IBD and also the Gastroenterology Learning Network, where you can find our podcast as well. So with that, Parambir, I'm going to turn to your third abstract to present from ECCO. I'm finding this really interesting. I feel like you're holding up a crystal ball and telling us what some of these things that we may have in our arsenal in the future to help to take care of our patients. But this one's a little bit different because this has to do with stopping therapy. Is that right?

Parambir Dulai:
Yeah, so sort of just keeping with the theme, you accelerate through with the biologic to get the inflammation under control and the big question becomes, when can you decelerate and how safe is it to decelerate or are you sort of just hitting a curb in a corner and you're going to go off the side of the cliff here if you stop the therapy too soon? So this is the SPARE study looking at predicting risk of short-term and mid- to long-term relapsing Crohn's disease patients who are stopping infliximab. So there was a previous study called the STORI Trial, which some of the listeners might remember, which showed that there was a short-term and long-term risk of relapse among Crohn's disease patients, which could be predicted in some capacity by distinct biological profiles and short-term meaning within 6 months of stopping and mid -to long-term meaning greater than 6 months after stopping.

The recently completed SPARE trial, which also included 211 patients across 60 sites in Europe and Australia, looked at the impact of stopping infliximab or corticosteroid for remission out to 6 months and less than 6 months. And in that study there were 3 arms in the SPARE trial, those who continued combination therapy with infliximab and an immunosuppressant, those who stopped the infliximab or those who stopped the immunosuppressant. So the goal of this study was to take both cohorts together and actually try to identify markers or biomarkers that might predict either short-term relapse or mid to long-term relapse with stopping infliximab with the goal of being able to identify those patients who may be safe enough to stop versus those who may not be and might be at risk for later term relapse or to define when you might be able to sort of reintroduce it or consider monitoring them.

In short, the take-home point from this was that the genes that were associated with risk of short-term relapse not surprisingly were really genes associated with inflammation IL-6 CRP. And then they also identified a couple of them that were associated with complement activation and markers of dendritic cells, which is new and unique. And specifically they found a marker that was reduced, which was the complement component MASP-1.

When they looked at markers that were associated more with the risk of mid- to long-term relapse, it was the protease inhibitors that were actually decreased in that group and the growth factors in anti-inflammatory cytokine IL-20. So this study, although it didn't provide a clear-cut guidance on who would be at risk for relapse and when they would be at risk for relapse, I think it very clearly identified a few distinct pathways that were associated with relapse and that were associated with variability and timing of relapse. And the goal here I think that the authors were trying to gain was an understanding of the biology that drives Crohn's disease relapse with stopping of the infliximab and maybe considering ways to either personalize the decision or to transition off of infliximab to something that might be able to maintain that healing in the long run.

Millie Long:
No, it's great. And I think that of course we all have in our practice some patients who are incredibly stable once coming off therapy, but gosh, I've seen some terrible relapses in some individuals. So I think data like these, the individualization, and again these themes of personalization, whether it's selecting the right therapy or determining who is appropriate for deescalation, I think are really important. But we're kind of not there yet, is my take home to this. Is that right?

Parambir Dulai:
Yeah, we're not there yet, but I think the one thing that the audience should take home from this is that the overwhelming pathway that was associated with short-term relapse in less than 6 months was inflammation. And I think oftentimes we forget to do our due diligence in looking for multiple different sources of inflammation, both with endoscopy, histology, and biomarkers with CRP and calprotectin. And so I routinely, when a patient asks to consider stopping infliximab, ask that we first confirm they're truly in deep remission before making that decision because it gives us an opportunity to provide them the best chance for success if they do go forward with that. And I think at minimum, this sort of reiterates that point for the audience to think about is don't just stop the medicine without making sure that the disease is truly, truly fully under control.

Raymond Cross:
I completely agree, but let me take the flip side of that. So you have this patient in this precious remission, the deepest of remissions and we're all trying to get them there. It's so hard to get them there and that's the patient we're going to stop. So I mean they're the ones you would consider stopping, but they're also the ones that have the very best results. So it always gives me pause when I do that because I just don't know that I can ever get them back to that again. But at the very least, telling patients that there is emerging research that's going to help guide us on potentially who we can stop therapy in the future is helpful because this concept that they're on the drug forever and can never come off is very unsettling to patients. So even having a little bit of data like this I think is helpful.

Parambir Dulai:
Yeah. And I completely agree and it is a difficult sort of conversation to have. It's not something we necessarily want to do. I do think it's something we're going to need to address in the near future because as we saw in the recent risankizumab trial, there was a very high rate of maintenance of response for those randomized to placebo after getting induction to risankizumab and we don't really quite understand what that carryover effect and maintenance of response was out to 1 year. So I think as the pathways become more selective and our treatments become much better, thinking about whether we're going to cycle back to episodic treatment which we moved away from because of immunogenicity, it is going to be something that we may need to reengage with at some point.

Millie Long:
Very fair. Lots to be learned in the future. Well with that, I do think we need to unfortunately wrap up this IBD Drive Time, but so Parambir, you are our second time guest so we can't ask you kind of our typical closing questions. So I want to ask you one about ECCO because I'm going to venture to guess that many of our listeners in the US at least don't regularly attend ECCO, myself among those individuals. I have actually not attended ECCO in the past. So what do you think is it about ECCO that our listeners should really consider attending this meeting? What did you enjoy most about the experience of going to ECCO?

Parambir Dulai:
Yeah, so first of all it was international travel, so I got to see somewhere new and fun every time. I think the other thing is that the conference is very well put together and the attendance is quite substantial with, I think they had 7,000 people this year all around IBD. And then they have a killer party at the end of it that is to die for and you have to buy your tickets early as many people are scrambling towards the end to get into. But it's just overall just a very well put together collegial, interactive experience and I've enjoyed it and I've been going annually now for a few years.

Millie Long:
Well that's great. Maybe some of our listeners will add it to their list of IBD conferences that are a must attend. I think you and Ray have convinced me so I will likely be attending in the future. So thanks very much Parambir for joining us and we look forward to seeing you and hopefully having you a third time eventually back here to IBD Drive Time.

Parambir Dulai:
Yeah, no, thank you so much for having, this is a lot of fun.