Western AF Symposium 2024: Session 11 Roundtable

Time to Implement Advanced Imaging for Optimized Anticoagulation Strategies in Atrial Fibrillation

© 2024 HMP Global. All Rights Reserved.

Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of EP Lab Digest or HMP Global, their employees, and affiliates. 

Featured is the Session 11 Roundtable entitled "Time to Implement Advanced Imaging for Optimized Anticoagulation Strategies in Atrial Fibrillation" from WAFib 2024. 

Moderators: 

Fred Kusumoto, MD, and Jeff Healey, MD

Discussants: 

Edip Gurol, MD; Elaine Hylek, MD, MPH; Gaston Vergara, MD; Stavros Stavrakis, MD, PhD 

Fred Kusumoto, MD: All right, let’s start this next roundtable. My name is Fred Kusumoto, and I have the pleasure of moderating this with Jeff Healey. We're going to talk about whether or not to implement advanced imaging for optimized anticoagulation strategies in atrial fibrillation (AF). It's incredibly vague. So, when we think about advanced imaging, is it the brain, is it the heart? What is advanced imaging? We'll go from there. Jeff, do you want to say hello and then everyone can have a chance to talk? 

Jeff Healey, MD: Fred, thanks for the introduction. We have a nice group of cardiologists and a neurologist. Advanced imaging is going to start both in the head as well as the heart. So, back to you, Fred.

Fred Kusumoto, MD: Yes, I thought maybe we would start down the line and have everyone introduce themselves, and then talk about how they use advanced imaging in their practice now. I think that will start generating questions. So, Edip. 

Edip Gurol, MD: Hi, Edip Gurol here. I'm a neurologist at Massachusetts General Hospital, Harvard Medical School. Thanks for the invitation and for this very important discussion. I do see a lot of patients with AF who are at high brain bleeding risk or who failed oral anticoagulation. As stroke neurologists, we are mainly dealing with secondary prevention. So, these are, in general, higher risk populations for ischemic stroke and intracerebral hemorrhage. My biggest concern is the inadequacy of the CHA₂DS₂-VASc score and most other pure risk-based stratification systems to understand the embolic risk in people with AF. These scores work as well. They have nothing special for AF and they work as well for, let's say, large vessel disease or even brain small vessel disease as far as predicting incident strokes. Similarly, intracerebral hemorrhage risk is the most feared complication of anticoagulants and HAS-BLED and other scores have absolutely no specificity and not much sensitivity to predict that. I'm using brain MRI and the finding of even asymptomatic and embolic-looking infarcts or past embolic infarcts in a patient with AF is probably the strongest marker of recurrence risk. Similarly, the brain MRI also shows me hemorrhagic markers or high hemorrhagic markers, which are brain microbleeds, cortical superficial siderosis, and other markers of small vessel disease such as moderate to severe brain white matter disease, large perivascular spaces, and lacunar infarcts. When these markers are present, I discuss left atrial appendage closure as a potentially better approach to stroke prevention so that we can prevent strokes and ischemic strokes without disproportionately increasing the risk of potentially fatal anticoagulant-related brain bleeds. 

Elaine Hylek, MD, MPH: So, I would like to jump in and say that I think cerebral microbleeds have some significant data to support it as to whether or not the patient is at higher risk for intracerebral hemorrhage on anticoagulants. The problem is these patients are also at exquisitely high risk of stroke. So, if there is a group that I would think, left atrial appendage occlusion, that's probably them. I think all the sophisticated echocardiographic parameters that we measure haven't necessarily added much to the discrimination of CHA₂DS₂-VASc. We can argue that point, I guess. In addition to that, being here in Utah, I love the biology of the Utah score that's actually now getting to the tissue level of what's happening within patients' hearts, because it cannot possibly be difficult to understand that when you have healthy myofibrils that are being disorganized and separated by fat and fibrosis, that that's going to induce a lot of chaos, not only with conduction abnormalities, but also endothelial dysfunction. I think that a session on vascular biology would be so interesting, because we've learned a lot about endothelial dysfunction in the presence of neutrophils from COVID-19. We hardly study vitamin K, which is absolutely critical for vascular health. So, I think it’s a lot of different things. I would even throw out something really provocative, that I think that acetaminophen, when used consecutively in doses over and over again by our elderly patients, probably depletes glutathione and also leads to vitamin K dysfunctions. There's so many aspects of our clot triad that we don't know at the level of that heart surface, so it's really driving the initial clot formation. I think until we figure all those things out, we're going to be stuck with CHA₂DS₂-VASc C statistics of 0.7 and not 0.95. 

Jeff Healey, MD: So, both for Elaine and Edip, before we move on to our other panelists, one thing I'd like you to comment on is this trouble we face, right? We have these great advanced imaging techniques, you get beautiful single-center reports or 2-center reports, they're technical, you require special equipment, special protocols. They're very hard to roll out into my world of large clinical trials and registries. You get core labs and you do your best, but you're dealing with 6 different MRI machines and different talent levels in terms of acquiring images, so they have a hard time bubbling up into guidelines because it's hard to get data on thousands of patients and really promote them. So, do you see any way around how we get these novel techniques into the mainstream and into studies? 

Edip Gurol, MD: If I could take the brain MRI part, that part is easy. Every single MRI vendor has the appropriate sequences to provide very good quality hemorrhagic markers when it comes to microbleeds and superficial siderosis or brain bleeds. They show chronic microbleeds, which means a microbleed or superficial siderosis that happens years ago, that's a very important marker, and you do see it on the brain MRI. So, for the brain MRI markers, for everything that I have discussed, they are very easily identifiable on any brain MRI in pretty much any machine that has even 1-Tesla or more. Right now in the United States, 1.5 Tesla is the standard of care. So, it’s no big problem for the things that I'm discussing. Otherwise, we have published plenty of papers showing that the higher your field strength, using 3.0 Tesla, the better sequences you use. Of course, you show more of these lesions, but even the basic sequences, they do help for me.

Jeff Healey, MD: So, I wanted to give Gaston and Stavros an opportunity to chat on this topic. They've done a lot of work in it. Gaston, you first, please. 

Gaston Vergara, MD: Yes, thank you, Nassir, for having us here. When I read about what we were going to talk about, I was a little puzzled about what I was going to say as a practicing electrophysiologist. I'm in private practice, so I don't have access to a lot of fancy modalities. Looking back, there's a lot of anatomic data with some degree of correlation between AF and anatomy and fibrosis and stroke. Nassir’s group published from the original DECAAF study a correlation between independent from CHA₂DS₂-VASc to fibrosis and stroke and Luigi Di Biase’s group published the anatomy of the left atrial appendage and stroke. But all those things, even though they're very nice, I cannot apply them in private practice. So, I have to resort to TEE. We like to talk about MRI and about all these interesting things, but as a practicing electrophysiologist, TEE is one of the tools that I use for those patients in which I don't know what to do. For patients with a CHA₂DS₂-VASc of 1, they're young, do I really want to put them on an anticoagulant? They're active. I still don't know if closing the left atrial appendage in everybody is a great idea. TEE allows me to assess the stroke risk. We all know about the size of the left atrial appendage, the shape of the left atrial appendage, the velocities within the left atrial appendage. I can look at other structures. I can look at the septum and see if there's a PFO or a hyperdynamic septum. I can look at the aortic valve, the ascending aorta, and I can get a pretty good idea of my patient's heart and decide if this patient has no conventional risk factors for anticoagulation, it might be worth considering anticoagulation with the caveat that it's less than perfect and the data has to be individualized for that particular patient. The other issue with TEE is it’s an invasive study, so I have to put the patient to sleep and do the TEE, but it gives me a lot of information. Usually, I get that when I'm doing a cardioversion, and obviously I believe in cardioversion. So, I believe in TEE. I like TEE. 

Stavros Stavrakis, MD, PhD: Yes, thank you for the opportunity to speak on this panel. What I would suggest is to leverage imaging modalities that are used for other reasons, like for pre cardioversion, TEE, or high risk of stroke. We can measure the left atrial appendage velocity, lower velocity, high risk of stroke. For patients who are undergoing ablation in our center, we do CT in all of them so we can look at the shape of the left atrial appendage. We can use MRI data and fibrosis. So, we can integrate all this in our decision-making. I would also like to bring up another modality, it's not an imaging modality, but we just published a study where we looked at biomarkers in the blood using metabolomics as a cross-sectional study. We found that there were biomarkers or some metabolites that were predictive of stroke. So, perhaps that's the way of the future. We measure metabolites in the blood in those patients with low to intermediate risk, and if they have it, perhaps we give anticoagulation. It’s in the future, but I think that's where we should be going. 

Elaine Hylek, MD, MPH: Just a comment on the biomarkers from the RELY trial, I believe, as well as ARISTOTLE, there were some studies looking at the CHA₂DS₂-VASc 1 group to try and better refine the risk within that lower risk group, and BNP and troponin were found to be somewhat helpful. I don't think that's used much in clinical practice. But we learned yesterday, and I absolutely loved the lecture from Uli Schotten where he talked about endomysial fibrosis, which I think that has to be the key, getting to that level. He was talking about measuring BNP 10, if I'm not mistaken. So, I do think that there hopefully will be biomarkers that may reflect that state of architectural remodeling. 

Jeff Healey, MD: It's a great idea to leverage. I think that's a great point, Stavros, to leverage what you already have with echo. We have known for decades that left atrial size predicts stroke risk. We know people forgot that in SPAF III, about 600 or 700 transesophageal echos are done and the thickness of your aorta was a pretty good independent marker of stroke risk; it's marking atherosclerosis, obviously. But we all appreciate that CHA₂DS₂-VASc has modest abilities—the C statistics are not wonderful. But we are tethered to it. So, in a person with a CHA₂DS₂-VASc of 3, paroxysmal clinical AF, but their atrial appendage looks good, they don't have much fibrosis in their atrium, their atrium is relatively small, and the biomarkers are all low, who has the confidence to move off the guidelines there? Again, I guess I don't for the most part, and it's for lack of data, but maybe we should be more bold. Elaine? 

Elaine Hylek, MD, MPH: When you think of the research, having been in this field for way too long, all of these were discovered by dichotomous variables, yes/no, it's there or it's not. In addition to that, this was not exactly elegant research, just looking at the big database from some Medicare or whatever. I think that's part of the issue. In the earlier AF trials, most of those patients did not have these brief bursts of AF. For ARISTOTLE, we required an ECG 2 weeks apart that showed AF. We couldn't enroll enough people, so then I think we extended it to 4 or 6 weeks. So, these were individuals enriched for AF. So, it's an interesting question, but even still, it looks as if you have already demonstrated the risks are still similar, so I think we're kind of stuck with CHA₂DS₂-VASc. 

If I could just say one more thing that I want to correct with the guidelines, the guidelines actually suggest to maybe look at some other factors like renal disease. One of the back stories on CHA₂DS₂-VASc was supposed to be serum creatinine, and probably because of misclassification or how people were documenting CKD, they couldn't get it to be independent, so it became a sex category. But we've known from Dan Singer's work and others that renal disease is an important component of stroke. Obesity is not. Obesity has been shown in ENGAGE that as far as stroke risk, there's a paradoxical effect with weight. So, one wouldn't be saying, your BMI is 37, I’d better give you an anticoagulant, because it's interesting, it falls apart for stroke risk but not AF risk. 

Fred Kusumoto, MD: Mina Chung is going to talk a little more about the guidelines, but there is no question that in that back room, there is a lot of discussion about all these issues and the fact that the discriminatory characteristics of the CHA₂DS₂-VASc and anything else, even if GARFIELD-AF, you like it a little better, depending on the score that you like, it certainly is more worldwide, etc. But still, it's moderate at best and really just trying to provide the clinician nuance to be able to make decisions based on these sorts of things. 

Edip Gurol, MD: I totally agree with Jeff's point. Essentially, you have a patient whose AF is diagnosed through symptoms or by the PCP using an ECG, so whatever methods diagnosed people to enter the anticoagulant studies or left atrial appendage closure studies, CHA₂DS₂-VASc of 3, we do either start anticoagulation, long-term anticoagulation, or we consider left atrial appendage closure. The problem arises, I think, for the primary prevention. The problem mostly arises for a lower end of CHA₂DS₂-VASc or CHADS₂. Also, in people who have device-detected AF. I think this is where the problem arises. I had a recent generation Apple Watch for a while. So, if that told me that I might have AF, let's say I get 1 or 2 external cardiac monitors and nothing shows, or I get an insertable cardiac monitor and 3 months down the line, it shows an episode of AF, I don't have many risk factors outside of obesity, so what do I do now? That's the problem, and I think these are the situations. In those cases, I would really favor the use of AF burden, which is a very big plus with insertable cardiac monitors, which I believe we should use a lot more. We should collect more data from these devices. But all these new cardiac imaging methods and brain imaging could show asymptomatic findings that might tell us about an embolic stroke even if it was asymptomatic before. I think these things come into play. 

Fred Kusumoto, MD: So, we have a question from the audience about specific LAA morphologic features that are high risk for stroke. We're getting close to the end, so I want to make sure that these questions are answered. So, Gaston? 

Gaston Vergara, MD: The data show that non-chicken wing morphology is associated with an increased risk of stroke. Again, it's one of those things where, in clinical practice, I don't look at that specifically, but we all know certain left atrial appendages look ugly for lack of a better term. I was going to comment about CHA₂DS₂-VASc as a clinician. I don't think all CHA₂DS₂-VASc is the same, and we've discussed this before. CHA₂DS₂-VASc of 2, because you have a previous embolic stroke, is not the same as a CHA₂DS₂-VASc for hypertension and diabetes. I think there's somehow a disconnect to how in CHA₂DS₂-VASc, everything is weighed the same. I get the sense that it's not quite the same.

Fred Kusumoto, MD: Well, that's for sure. I mean, there's no question that there's a ton of data, both from meta-analysis and standard data to suggest that in fact they all have different strengths, with age being really the strongest more than anything else, and certainly vascular disease. Perhaps sort of more than, let's say, coronary disease, which is probably at least in most analyses sort of the least. But having said that, though, the data is really mixed and frankly hard to interpret.

Nassir Marrouche, MD: So, I put together this panel for a reason. We have a neurologist, an internist, a trialist, and people who see patients every day. Every day I go to work, and every day everybody in this room has to decide what to do next for their patients, with a yes or no on blood thinners. That decision is made legally with the guidelines. That's what we're living in today. So, the question to you is, what are the next steps that need to be done? Should we go to Nazem Akoum and tell him, let's do MRIs in everybody and take this fibrosis image? Or should we do echoes in everybody? Talk to Wael Jaber about the new AI echo machine to get a 15-minute echo image and predict the future? What ARTESIA and NOAH showed us, as you said today, is the next 3 months will define the future. You're going to give us subgroups who will profit from anticoagulation or not. There's no one size fits all. So, what's the next tool we're using? Is it imaging? Is it an AI model? Is it me and the patient discussing if they like apixaban or rivaroxaban? What's the next tool we need to use? 

Jeff Healey, MD: I can jump into that. I think it's going to be all the tools, probably starting with the tools that most people use, but I think it's absolutely the time to rewrite this, because today's AF is not what it was 20 years ago, right? We have AF picked up on a watch 5 minutes out of a month. 

Nassir Marrouche, MD: Who is rewriting it? 

Jeff Healey, MD: Who is rewriting it? I mean, there was a good AFNET/EHRA group that got together in the fall, or late August or early September, to put at least a framework together what that might look like. It will take a lot of data because, again, there's all this different AF that we never saw before, but I agree with you, Nassir, it's a lot of work that needs to be done to reinvent it, but I think it is time that we reinvent it.

Elaine Hylek, MD, MPH: If I could jump in, I would say we overemphasize bleeding as a field. We forget to mention that the mortality for strokes in AF is 24% at 30 days, and there may be safer anticoagulants coming down the pike, the factor XI inhibitors, but we'll have to wait and see. 

Fred Kusumoto, MD: To answer your question, Nassir, it was an intentional piece. You can argue about some of the numbers, but 1 and 2 percent is sort of the cutoff with whatever seems to come up. The CHA₂DS₂-VASc got stuck in there because people felt strongly that clinicians needed something for guidance, but it really that wasn't my area where I put a lot of work in the guidance, it’s not my area of expertise. Nonetheless, it was specific to try to provide the clinician with this 1%-2% number to then use new technologies. So, thank you very much. 

The transcripts have been edited for clarity and length.