Marc Dweck, MD, describes the advantages of using cardiac imaging to assist in the diagnosis and treatment of aortic stenosis.
Hi, I'm Professor Mark Quick and it's a great pleasure to be invited to speak to you here today. I'd like to thank Sanjay in particular for the opportunity to speak. It's a great privilege to meet to be here. Um So Sanjay's asked me to speak about ailes stenosis and about imaging left ventricular decompensation and how we can use that to perhaps optimize the timing of aortic valve replacement. So I'm just going to share my slides. Um So yeah, I'm the chair of Clinical Cardiology University of Edinburgh and my great interest is imaging the heart and in particular patients with aortic stenosis. Um as we all know, aortic stenosis is a common condition and it is characterized by severe narrowing of the aortic valve. You can see a beautiful example of this on the left hand side of the screen, but when we think about the need for aortic valve replacement, the development of symptoms, the development of adverse events, we also need to think about the my cardi um and the impact that that valve narrowing has on the heart muscle. But there's a lot of those factors that govern as much by the pathology of the muscle as the pathology in the valve. So what about aortic valve replacement? When should we do it? And um of course there's uh different thoughts on this. But generally the consensus is that we don't want to image. We don't want to replace the valve too early because then we are necessarily expose the patient to the complications of surgery or tabby and the the complications associated with prosthetic valves. The generation endocarditis and the calculation etcetera. But neither do we want to replace the valve too late because there's a real risk that we leave the patient with irreversible damage to their heart muscle that will govern their development of symptoms and adverse events for the rest of their life into the long term. So when we want to do it well we want to do it at just the right time. And that is when the eventual is just starting to decompensate, just starting to fail in the face of the increased after load. Um and that is the challenge that we face. How do we identify that exact moment? Well, the current guidelines recommend that we replace the valve when we have evidence of severe Abel's explosives, but also some evidence that the ventricle is starting to fail and most commonly that is on the basis of the development of symptoms. The classical symptoms of a latex gnosis, chest pain, sync api and breathlessness additionally recommends consideration when we have a reduction in the LV ejection fraction 50%. In the last guidelines. But actually the most recent up to date sc guidelines. 55%. So what about symptoms? I mean, that is the main indication for surgery in patients with symptomatic essence. Uh of course we mhm come back to this curve whenever we talk about symptoms and this is uh it's hard to do a talk and express without including the classic brown world curve and basically tells us that people's prognosis is pretty good up until the moment that symptoms develop and then um it deteriorates quite rapidly. So this is a good time at which to replace the valve And that's guided clinical practice for a long time. Um there are some things to talk about that when we think about this curve. And uh the first thing is the age on the X-axis and you can see that the patients in this cohort, you know, which is now over 50 years old. We're young. These are people largely with by customers of our disease without much in the way of co morbidity. A very different cohort of patients from those that we see in modern clinical practice who are frequent in their seventies, eighties and nineties with comorbidities that make symptom assessment quite challenging. But the other key point is that actually that curve is is made up. It's it's not actually based on clear data. It was based on dr Brown Fells. Brown was very insightful observations from a postmortem study. Um And that also means that all the patients on that curve uh died because these are people that were studied postmortem asked they died and he went back retrospectively to look at their case records. So of course this curve is important. It's been crucial in guiding practice, but it's not really a class one level a evidence. And there's certainly a need to try and uh do better to better optimize the timing of our surgery and to identify exactly when the venture has started to fail. And the other thing is that patients in contemporary practice ah don't necessarily have a good outcome following able to valve replacement. The instance of death before valve replacement is low, but following able to valve replacement actually the incidence of adverse events is quite high. So this is data from paris one of the leading half of our centers in europe, if not the world and from their data in patients who had high risk or evidence of left ventricular decompensation before they're able to valve replacement. In the years after the valve replacement, half of them were either admitted with heart failure or dead at two years. So this is really quite a bad prognosis despite people coming through a lot of our replacement, there's a large proportion of patients that don't do as well as perhaps we think they do. So can we do a better job. And I guess that's the whole purpose of this. Talk, how what how might we advance the field and get more objective evidence of what's going on in the my card. Um um So I think the first thing to do is to go back to the pathology. So what is happening in the ventricle enables explosives. And um we're pretty familiar with this, but essentially the ventricle is exposed to an increased after load and that causes a hypertrophic response so that my sites increase in size. All thickness increases mass increases and that increase in wall thickness is a good thing. It maintains wall stress and it maintains contact performance for many years if not decades. However, with time, the hypertrophic response to compensate the muscle mass outgrows its blood supply, we get ischemia and we get cell death and scarring. And this. These two processes the cell death and scoring really drive the transition from hypertrophy to heart failure. Um and with it the development of symptoms, adverse events and the need for a lot of our replacement. So if we think about where on this scheme, we currently act where we we act on the basis of symptoms as people transition really towards the later stages of fibrosis. But as I said before, it can be quite hard to assess symptoms in patients who've got lots of comorbidities or indeed patients that uh immobile due to their capabilities. And so potentially here we have two other objective markers of left ventricle de conversation in the form of the my cardio scarring that develops and also the my site cell death. And we've been interested in trying to target those two processes in patients with aortic stenosis. So we'll take the scarring first. And essentially, if you want to detect my cardio scar, your imaging tool of choice is cardiac MRI. And like other conditions, we can see cardiac scarring in patients with aortic stenosis. We get a non infarct pattern of scarring in the ventricle. You see the scar is white areas compared to otherwise normal healthy black my card. Um and we can differentiate that panel scoring from my card infarction. So crucially we can tell if the sky is detail, takes notice all due to my cart and function. And there's been extensive validation work now showing that there's good correlation with histology. There are actually two forms of my cardio fibrosis in Abel's explosives. There's replacement fibrosis and that's basically irreversible areas of scarring that follow the my site death. And then there's also more diffuse forms of scarring. So called diffuse interstitial fibrosis which isn't associated with cell death and which crucially is potentially reversible. So we'll come to the interstitial diffuse fibrosis later on, that's a slightly harder thing to image. And first we're gonna talk about replacement fibrosis, they're irreversible scar that we see in able to explosives. And to detect this. We use the widely used now late gadolinium enhancement techniques. So this is a common MRI approach. We inject gadolinium contrast agent. It accumulates in areas of scar of increased extra cellular expansion and essentially increases the signal in those areas. So we see areas of white scar versus the otherwise normal looking my card. Um And as I said, we have this typical pattern. This non ischemic non in fact pattern that we can observe in patients with advanced they also explosives. And there's lots of papers covering this now and these papers have looked at what that type of scarring means, what it associates with. And essentially it associates with other markers of left ventricular decompensation. So we have an association with markers of heart muscle damage, things like troponin, uh wall stress, things like BMP. It's associated with less efficient pumping of the heart. So both echo and CMR measures of systolic and diastolic dysfunction. And crucially it's associated with patient symptoms and worse exercise capacity. So it does provide an objective marker of left ventricle de conversation that we know is due to the vow of rather than other processes that commonly occur such as uh my kind of functional indeed cardiac amyloidosis and is that important? Prognostic lee, well, this has been looked at again at lots of cohorts all around the world. Different centers and the answer is yes consistently. We see that patients who have this type of scarring have a worse long term outcome the patients that don't. And this is in term of in terms of hard endpoints. So all cause or cardiovascular mortality, both increased in these patients. This is data from a recent multi center study in the UK that over 700 patients, these people all image just before the able to valve replacement. And so these outcomes following surgical intervention. And again, we can see that the presence of this former late enhancement was associated with an adverse long term prognosis following a VR. Both all cause and cardiovascular mortality clearly increased And this was an independent predictor of events. Outperforming traditional echo markers, severity, injection fraction, symptom status etcetera. So one interesting thing about this is what happens to the scarring with time. And so this is a typical example from a patient. This is a guy that we saw who at baseline had severe aortic stenosis but was clearly asymptomatic. He did 12 minutes on the treadmill fit active person and had no symptoms. So we didn't send him for surgery. But he did take part in the research that he and he had an MRI scan and we could see clear evidence of late enhancement and scarring and eventual at baseline. So we followed the patient up. He was followed up as routine in the clinic. He was seen six months and then a year and a year he had developed symptoms. So he now had clear exertion of breathlessness and he was sent for a VR. Before he had his operation we repeated his MRI scan and we could see that actually the areas of scarring had clearly progressed. We have two new areas of replacement. My cardio fibrosis on the MRI here that weren't present just the year before. And that was a consistent finding across a large cohort of patients where there was rapid progression of this scarring in the patients where we saw that baseline. So what happens then after the able to valve replacement, what happens a year after he said the surgery? Well what's interesting is that whilst the fibrosis hasn't progressed is volume of fibrosis was much the same as it was before the surgery. Neither has it regressed. And again this is a constant finding in our study across lots of patients and in other cohorts. Um So this is important because it means that the scar and the fibrosis that you develop while you're waiting for your operation doesn't reverse it stays with you um For the long term essentially until you die. And that is important because we know this scar is associated with the adverse outcome and we know that there's a dose dependent relationship. So the more sky you have the worst year long term prognosis. So this is a quick summary of the fibrosis. So if you see a late enhancement uh evidence of non infarct fibrosis and MRI. It's telling you that there is evidence of left ventricular decompensation In patients with aortic stenosis, crucially that scar that you accumulate before surgery is non reversible and it is associated with the adverse long term prognosis that continues after valve replacement. The more fibrosis you have the worst your outcomes. And so this is not only a potential biomarkers left and triple D. Conversation but also potentially useful trigger too early and prompt bowel surgery with the rationale that you want to stop more fibrosis from developing and to favorably modulate patient's prognosis? There is a slight issue And whenever I could just talk I worry a bit about us doing MRI and all these patients with aortic stenosis. Is that a good long term strategy? And so along the way we have been thinking can we develop cheaper biomarkers either as an alternative to MRI or potentially as a screening tool so that we don't have to uh image everybody. And um this leads us back to our summary slide as the pathology and and this idea that we might be able to develop methods that target the maya sites cell death. And of course I'm thinking about that. The first thing we thought about was high sensitivity troponin essays. And we have a strong interest in that in Edinburgh with Professor Milt and Professor Newby. And so we did a study where we looked at high sensitivity troponin i in patients with able to explosives who had commented echo and MRI studies and and also had assessment of their coronary arteries. And what was interesting was that in this group of patients with aortic stenosis, your troponin levels. The high sensitive troponin levels are associated not with your burden of coronary artery disease but rather with the events that are going on in your muscle related to aortic stenosis and in particular rated two things. One was your left ventricular mass. So how much high perch for you has. And the second was whether you had my cardio fibrosis or not, I whether you had left trigger decompensation. And what was I think particularly helpful is that if you use the threshold of six nanograms per liter there are very few patients below that threshold had my codified process. So we can potentially use that as a screening tool. And we can say well look if you've got a value of less than six you don't you eventually find your heart muscle is pretty healthy. We don't need to further investigate. And when we looked at the outcomes associated with uh component then uh two predicted outcomes like most markers of left ventricular decompensation and the second cheap and cheerful biomarkers we wanted to look at was the BCG. The electrocardiogram which of course is quite all these patients. And we were particularly interested to understand the relationship between uh left ventricle hypertrophy and and the strain pattern with my cardiff fibrosis. So we've all grown up understanding how to interpret disease Gs. But what does the strain pattern on SDG actually mean? What does it refer to? And what we found was that uh basically it's a marker of my cardio fibrosis. So all the patients that we had in our cohort who had the left ventricular strain pattern on there. S. G. Had my quantify process. It was extremely specific tool for later handsome. It wasn't completely sensitive. There were patients who we missed who either had LDH with no strain or no L. V. H. But it was certainly specific if you want a sensitive tool, then perhaps you can look at the presence of left ventricular hypertrophy where nearly 90% of the patients had uh, fibrosis on their MRI scan so that the E. C. G. And the proponent of both potential insights into what's going on into the ventricle and to the presence of scoring, that we can potentially integrate into our clinical assessment. And again, a strange pattern is a very powerful predictor of outcomes in patients with aortic stenosis. So how do we put this all together? Well, this has been the challenge in developing the evolved randomized controlled trial. And this is essentially a randomized controlled trial that's ongoing, where we're seeing whether using MRI assessments of left ventricular scarring and the conversation can optimize the timing of aortic valve surgery. And it's a study that's being conducted across 25 centers in the UK. We are over halfway in recruitment and this is the study design. So essentially the idea is that we want to take people with asymptomatic, severe adult explosives. People that don't currently have an indication for a clear indication for a little valve surgery. And then we screen them. So we are site, we used to opponent, uh, some sites don't have component and we use the CG there. But in our site we used proponent. And essentially, if your opponent is less than six, we say, well, your heart muscle is healthy, we don't need to do anything else. We'll see you back in six months and we'll assess your symptoms and proponent again. However, if your troponin is elevated, then we say, well, okay, there is some evidence of microbial injury. What's the mechanism of this? Is this due to your aortic stenosis and left ventricular decompensation? And these patients get an MRI scan and then we see if they have evidence of the non infarct, mid wall scarring and in those patients that do, we randomized them to either early surgery or to be, or the standard of care, which is essentially waiting for the development of of symptoms or reductions in ejection fraction. And we're following the patients up for three years and the primary outcome is all cause mortality and unplanned aortic stenosis related hospital admission. So as I said, we're well over halfway and the study is still recruiting. Um, there's been some delays with covid, but hopefully we're over the worst of those uh, at the moment. And um, and yeah, one interesting thing is that actually the screening step looks particularly effective. So we we had kind of estimated that around 50% who had the MRI scan would have fibrosis after the screening step, But actually it's more like 70 or 80%. So this appears to be a very nice way of merging the sensitivity of biomarkers with the specificity of advanced imaging and I think this is a paradigm were very interested in looking at in the future. Of course, lots of different disease states. So that's the cultural that's ongoing. But what about the future? Is there are there future directions that we can take that might even improve the assessment of left and trigger D. Conversation further. And there's two areas that I want to speak to you about that. There are lots of different techniques but I'm going to focus on on to. And the first is another MRI technique which this time looks at not the replacement fibrosis, the replacement scoring but instead diffuse interstitial fibrosis. And as I said, this is potentially an even more interesting stage of disease because it's it's reversible. So if we can pick up this stage before the development of irreversible scar then potentially we can leave the patient with a completely healthy eventual. And there are different techniques for doing that. They all rely on this T. one mapping approach. But from the T. one mapping you can get different you can different different measurements out. So you can get a native T. One measurement. You can get a measurement following the administration of galleon contrast and you can get an extra cellular volume which is a ratio really of the native and post contrast T. One values in the my card in versus the blood pool. And this is just a quick summary study that shows that there is some maybe some slight differences in native T. One values. Other studies have shown that there is perhaps a small difference, There's very little difference in the post contrast imaging. They on their own aren't particularly useful but the CV with the added uh I guess sensitivity the administration of contrast provides allows you to differentiate signal in patients with aortic stenosis and healthy volunteers. And this gives a marco essentially of the percentage of your mark hardin that's made up by the extra cellular space. So you can start to divide the my card um into the matrix volume and the cellular volume and express that as a percentage. And the follow on from that is this market called E. C. B. Or the index extra cellular volume. And this isn't a fraction. This is just saying, well what's the volume of fibrosis, Any eventual index to your size and index of how big you are And you can derive this very quickly by just multiplying the fraction by your my cardio volume, which we measure all the time. And this is potentially it's related to easy fraction but it's potentially more sensitive in detecting changes because monitoring fractions is somewhat complicated, particularly when the my sites are changing as well. But if we just look at the absolute volume of of fibrosis, then that's potentially more useful. So we've used these approaches and in particular I CV to track changes with time. So this is a study that we did before you can see in the black bars with different parameters. What happens to each of these measures in patients with mild, moderate and severe disease. So peak velocity is obviously higher in these patients. The LV mass is higher in these patients. Uh the index and the change. So these are changes. So the change in uh in these volumes increases with time and that's an interesting feature available extensive statuary disease progression appears to accelerate with more uh severe disease. And then when you replace the valve you see a reduction. Obviously your people philosophy reduces. Obviously your LV mass reduces, but so too does your burden of interstitial fibrosis. So your I C V goes down now. What's interesting though is that you're the volume of your mass goes down more than the volume of your fibrosis. So actually your ec fraction goes up, which is a bit confusing but if you think about it in terms of just the burden of the fibrosis that it behaves exactly how you would expect and confirms that this type of diffuse fibrosis is indeed reversible. So that's interesting. But you know, is this a prognostic market? Does this tell us about life contribute to the conversation? Is it a useful marker? And so to address this important question, we performed a big multi center study. Lots of centers from all around europe north America and asia and 440 patients. And we look basically at CCV fraction and I CV values. So the first thing is that this is the first real multi center T one mapping study and there were lots of challenges if you're thinking about native T one. But E C V fraction is pretty robust across centers. The measurements are pretty similar. And so we were able to do this multi center study using ECW type measurements and what we found is that light late enhancement PCB fraction and I E C B both track with other markers of left ventricular decompensation. So track changes in differences and now the mass we volumes your replacement fibrosis and ejection fraction. So again this is a mark of 11 triple decompensation. And similarly again we found that light late enhancement, it is a marker of an adverse prognosis. In fact it was the strongest marker of an adverse prognosis in this cohort. Outperforming ejection fraction, even outperforming late enhancement um uh in the prediction of all cause mortality. Of course we now have machine learning and so you can do lots of clever stuff and start to really tease out the different effects of different variables and start to see some nonlinear associations and trends look at which of different co linear variables might have stronger effects. And so we did that in a study uh that use the similar populations that previous study as well as a validation cohort. And essentially what we found was that again fraction was a very powerful predictive outcomes as was late enhancement. Uh and that there was a cut off in the c. V. that was associated with an adverse prognosis, not cut off was 27%. So patients with any TV fraction over 27% were associated with poorer outcomes and left ventricular decompensation. So that's useful because if we want to time a lot of our surgery, you know, ultimately we do need to make a dichotomous decision. We always want to keep things as continuous variables, but ultimately we need to make a black white decisions. So we have a cut off that might allow us to do that now with the Cv fraction. So I think the T. one mapping area is exciting. It may hold advantages even above late enhancement. The data supporting that isn't quite strong at the moment is late enhancement. But it's getting there and I think in the future we may use it more. And then the final thing that I think it's really exciting is that with pet imaging now we can start to measure activity. So we can do this with lots of different traces, targeting lots of different pathological process processes. And we now have a tracer for fibrosis activity. So it's called it's called happy. We can do it with flooring flooring dated radio Liggins. We can deal with galleon radio label labels but essentially it tells us about activated fibroblasts. And there's some very exciting data that's coming out showing that you get very high signal in areas of keep my card and fortune. This is a patient in fear of my car and far to extremely high signal in the infarct zone, but also in a more diffuse forms of michael injury, including hypertensive heart disease. So this is a recent case report getting very high signal. So I think this can be extremely interesting and able to explosives and are about to start study looking at this tracer in patients of able to explore this with the potential to tell us not only about the fibrosis that is already developed, but developing fibrosis, which is really the stage of the disease that we want to target because it's the stage that we might most effectively um, altar with intervention. So, um, in conclusion, thank you for your attention. Um, we've talked about able to diagnosis and the importance of the my card and particularly in timing, able to valve replacement. And I think it is worth remembering that the optimal timing of a photograph replacement has never been tested in a randomized controlled trial. And so we don't know when the best time to operate and to replace the valve truly is. And there are challenges with the current paradigm of waiting for symptoms to develop. And so I think the time is now here for randomized controlled trials to investigate new ways of optimizing the timing of our surgery and those include trials of uh, MRI based decision making algorithms where we're using the presence of scarring to time our surgery. Um, it's an exciting field, there's lots of new developments uh in uh in progress. And I think I think the logic is sound. And so I think this is an area that there will be more and more research conducted over the years. So, um, I want to acknowledge all the people that we've worked with over the years in developing our thoughts on this and in developing the trial and helping with the trial. And I want to thank you for your attention. Thank you very much.