Advanced Imaging of the Fontan, What is Driving Fontan Failure: New Horizons...

We should move on. Uh, it gives me great pleasure now to, uh, introduce Doctor Andrew Crean. Um, Andrew's, uh, professor of medicine and, uh, the SAB endowed chair at the, uh, University of Cincinnati. And Andrew's, uh, arrived, uh, uh, here from Cincinnati, um, and, uh, it's been a pleasure, uh, to already, sorry, from Toronto, I mean to say. Uh, and it's been a pleasure already, uh, Andrew, in many of our interactions, and I really look forward now to hearing, uh, uh, what, uh, insight you can have on the failing Fontaine from an imaging perspective. OK, well, thank you very much, Grucian, and good morning to everybody who's watching. Uh, so I approached this with a background as an adult congenital cardiologist, uh, but much of what I have to say, if not all of what I have to say, also equally applies to imaging pediatric patients as well. So as an imager, these are some of the things that uh go through my mind when I'm told that we have a Fontan coming in for an MRI or CT a Fontan who's not doing well, um, and we'll talk briefly about, uh, a number of these things. I'm gonna focus particularly on the plumbing because I think the plumbing is really what cross sectional imaging has the most to offer. I'm deliberately gonna stay away from talking about the liver because Doctor Veltman is going to talk about that, uh, in some detail at the end of the session. So we're all aware that when we talk about fontan failure we're not simply talking about a failing ventricle, although that's obviously part of the problem. Um, my adult cardiomyopathy colleagues are obsessed with uh systolic dysfunction in a way that we are not to the same degree. They want to know within 1% what the ejection fraction is and how it's changed. They're fascinated by the pattern of late adolinium enhancement and whether that portends a worse prognosis, and we can do all of these things in, in fontan imaging, but we rarely concentrate on. On the ventricle itself, um, interestingly, given a recent discussion earlier at the start about whether right or left, uh, morphological ventricles do better or worse, there's some recent data came out from, uh, sick kids in Toronto using a technique, an MRI technique called T1 mapping to look at diffuse fibrosis within within the ventricle in Fontan patients and in a population of adolescents, it was perhaps not surprising to find that the fibrosis content. Is higher in uh fontans than in normal controls and interestingly significantly higher in those with the systemic right ventricle uh so whether that whether the difference between the two groups expands as as we go on in life remains to be seen, but I, I'm glad I threw that slide in. What I really wanted to point out about the ventricle though is that for most adult patients with a fontan circuit, the ventricle is either moderately dysfunctional or severely dysfunctional, but it, it often doesn't change very much, surprisingly, it doesn't change very much over the years. And so one of the indicators for looking at the ventricle more carefully is if you hear that there has been or appears to have been a change in function, because that's not something that we see very frequently. And this is actually a a 22 year old man. I was reminded of this. This is a 22 year old man who came into Toronto, uh, last week with an atrial tachycardia, um, and with an echo which had suggested there's been a change in his function. Um, and when you look at this echo, you can see I get my mouse working, you can see that the lateral wall is really not contracting very well, and this represented a change from previously, uh, and as a prelude to his DC cardioversion for his atrial tachycardia, he went on and had some CT imaging. To make sure there was no thrombus visible within the atria. We didn't see any thrombus, but what we did see is an area of hypoperfusion in the lateral wall here corresponding to the area of dysfunction. So it looked as though he'd somewhat surprisingly had an infarction. Uh, and when he went on to the cath lab, uh, that's indeed what was seen. He had a truncation of the, of the mid circumflex which looked very much like an embolic, an embolic lesion. And so although embolic dysfunction is not a common feature of Fontan's, uh, we do see these kinds of embolic complications. I'd say roughly about once a year in our centers. This uh brings me on to what I really want to perseverate about this morning because, uh, thrombus imaging I think is generally not done very well, uh, particularly in adult Fontaine circulations, and this is a, this is a typical scenario. It's a, it's a Friday afternoon, uh, and you get a phone call from radiology and you know it's a Friday afternoon because the radiologist is smiling, which is the only time that you see that and. And, and he's, uh, he's doing the graveyard shift, the shift of death, the Friday afternoon acute shift that nobody wants to do, and the radiologists all have to rotate through it. And this, this is a musculoskeletal radiologist, and he's very happy because he's just seen a fascinating case, and he rings me and he says, uh, we've got what your residents admitted a a a fontan and for query PE and we've just done the study and there's massive thrombosis of the fontan circuit. And I'm like oh my God, I'm picturing surgeons and ECMO and all sorts of things and so I said, well, what, you know, what sort of state is the patient in? Oh, she's she's sitting up and chatting, she's fine. So, OK, well I'll I'll just come down to the scanner if that's OK and and take a look with you. And this is the sort of, this is the sort of thing that you may see, and we've got some axial CT images showing what looks like a large thrombus sitting in the right atrium in coronal view. It looks like we've got total obstruction of the fontan pathway, and then it also looks as though we've got multiple pulmonary emboli. So this looks like a complete disaster. The only thing that's odd is that the patient seems pretty well, really. And so I politely say to the radiologist, do you mind if we, if we, you know, repeat the scan and perhaps just change the timing a bit? And when you do that, what, what you notice is that what you thought was massive thrombus perhaps isn't so massive. Suddenly, on this arterial phase, the, the right atrial portion doesn't look as huge as it did before. And when you wait a little longer, The right atrium looks entirely normal, equally what you thought was a pulmonary embolus in the left lower lobe. Now showing normal enhancement. So to answer my question, what does the Fontan circuit have in common with a good joke, the answer is timing. Timing is everything in the imaging of the fontan circulation, and usually we're too quick, uh, and this is a problem that, that I, I, I bring it up and I'm gonna perseverate about it because we see it time and time and time again, particularly with on-call PE studies done in the middle of the night. Now it's not just a problem in atriopulmonary fontans, it's also a problem potentially in bidirectional glen shunts, and this is another case. This is a case that was done overnight where this was read out as showing a subtotal occlusion of the left pulmonary artery. You see it looks very dramatic on the MIP image, but in fact, if the resident had looked more carefully at the timing bolus, he or she would have appreciated that what's going on here is simply streaming between a pacified and unopacified portions of the pulmonary artery. Uh, this is also an issue for looking at the, the inferior limb of the circuit. Uh, Radiologists, particularly when they're imaging the chest, very want to image in the pulmonary arterial phase, particularly if the question we've asked them is, does this Fontan patient have a pulmonary emboli? And so that we we'll usually acquire an image that's designed to look at the pulmonary vascular bed rather than the IVC. And when you do that, of course, you will get images that look like this where. The inferior portion of the conduit appears to have a central thrombus. Again, all we have to do here is wait maybe 1020 seconds, and we can see that the, uh, the apparent thrombus fills in is simply on a pacified blood that hadn't yet been opacified. And if, if there's any doubt at all, we can make measurements of Hounsfield density to reassure ourselves that this is changing in uh in intensity and therefore clearly cannot represent thrombus. So how long do we have to wait? Well, the answer is sometimes we have to wait a really long time. Uh, this is a patient of mine, uh, who had one of the larger atrial pulmonary, uh, fontans that I've seen, and we, we've, if we go here from left, sorry, from around in this direction like this, you can see even at 4 minutes, we've still got pooling of contrast. We're not really sure what's going on up here. Is the thrombus here? Is the thrombus here? It took a full 7 minutes. Before we were able to say confidently there was no thrombus whatsoever in this ventricle, so sometimes uh patience is, is important. She Jiggle them, yeah, you could, you could actually put them prone sometimes, and sometimes I've done that. I have put them, I have put them prone on occasion to try and jiggle them up a little bit. Here, here's another one, which, uh, looks slightly odd. We were, we were called and said we think there's a thrombus in the atrium, and that there is certainly something that looks like a low density and it looks unusual, looks rather symmetric, though, which is perhaps a clue that there's something else, uh, going on. And the beauty of CTF, of course, is that it's an isotropic data set, which means that we can reconstruct it after the fact, after the patient has gone, and we can look at this, this abnormality in multiple different angles. And on this occasion as well, the, the data was also gated, which was very helpful. So when we look at this rather beautiful sepia image here, which is not how it came off the scanner, you can see that what we're looking at actually is a conduit, a rather unusual intraatrial conduit. With swirling turbulent flow coming back uh from the venous circulation, giving us this impression, false impression of a thrombus on the axial images. Sometimes though it can be very difficult and even with the best will in the world using MR, it can be difficult on the initial images to know whether we have thrombus present or not. This is a good example. This is a patient where the sine imaging is really very difficult to make out what's going on here. There's a lot of swirling, turbulent flow. The angiogram. Well, this is a MIP image, and a MIP image can sometimes hide things within the image, and so it's not a good way to look at the data here you'd be fairly confident there's nothing to see. I should mention that this, this patient, by the way, has had a transesophageal echo which says that there's swirling flow but no evidence of thrombus. So we've already had a TE which says that there is no thrombus. I routinely ignore those reports. They're as likely to be right as they are wrong in my experience, particularly with the AP Fontans. This is the same patient, and we've, we've just imaged in a slightly different way and now we're starting to wonder whether there is in fact something sitting here in the atrium, and then you give gadolinium and you wait 10 minutes and do late enhancement imaging, and we can see there's clearly a huge thrombus, uh, sitting right here. Now, over time, we follow this and we can, I've got here a mixture of CT and MR pictures just to make the point, it really doesn't matter whether you choose to use CT or MR for this purpose, they're both exquisitely good at looking at the presence of thrombus, and in fact with CT it's very easy to measure a volume of thrombus and work out how you're doing as time goes on on anticoagulation. I have to say though, sometimes the gating is a disadvantage. This is the same patient and when you gate these images and you look at how these lumps are moving, uh, and held on in some cases by a fairly slender threads, it makes you rather nervous. Um, fortunately, we have a tried and tested approach to dealing with this kind of problem, uh, which largely involves doing this. And putting them on heparin and hoping for the best and that's what we did with this patient, and actually she, she got away with it, not, not totally without sequelae, and we have to be very alert to new symptoms because there's obviously a risk of embolization, uh, whilst we anticoagulate and hope for the best. By the way, if there's anybody out there from a German speaking country, I learned during my research last night that this doesn't mean anything in German speaking countries, and this is the symbol for fingers crossed, so fingers crossed. This is a um an interesting man who came in a couple of years ago with facial swelling and when we did his imaging, his, his heart, his heart showed no evidence of thrombus, but in fact he had a left sided glen shunt, and he had quite a large thrombus sitting here, uh, causing subtotal occlusion of the shunt and he was responsible for his facial swelling. And so, uh, we took a decision to give him TPA and over the next uh week or so, uh, his thrombus basically melted away. And I show this case because what I found interesting about it was the effect of the resolution of the thrombus on his albumin levels. So here we can see, we can see that he's, this is before we thrombolyze and we can see that he's offloading down his hemiozygous system and this will return to the IVC and presumably increase hepatic venous pressures. And when he first comes in, his albumin is down. I can't read these numbers, but it's down in the sort of the mid-twenties somewhere. We put him on heparin for a couple of days and his albumin comes up a little bit. We thrombolyse him and his shunt opens and his, his albumin totally normalizes. I think that's just an interesting link between the anatomy and the physiology that I haven't seen before. The other area where I think imaging comes into play or cross sectional imaging comes into play is in dealing with the fontan who seems unexpectedly desaturated. Now we all expect our fontans to be a bit desaturated. Many times the coronary sinus is emptying into the systemic chambers, and so we'd expect people to be a little desaturated, but people in their mid-80s without open fenestrations are unusual and therefore we're looking for reasons for that. We've heard a little bit about the interventional. From, from Brian about some of the interventional options and and what we're really looking for in particular are things such as vena venous collaterals where these are collateral vessels from the systemic veins to the pulmonary veins we've got an example here here we've got uh a communication in this case between, uh, this is actually an unusual communication between a a hemiozygous and a pulmonary vein here. Um, but we're also looking particularly in some of the older patients with classic glen shunts for, uh, arteriovenous malformations because of the absence of hepatic factor, and we can see examples of these things here. These are very easily seen by the CT or MRI angiography, but very difficult to detect in any other way. Um, and then there are things, things that we shouldn't miss that we do miss from time to time, and this is a patient who was bluer than expected, and we didn't know why, and what was missed was this left SVC which was draining to the coronary sinus, and Doctor Goldstein has already shown you an example of that, um. This is an, this is a similar case where the echocardiographers were having trouble telling us where the coronary sinus was emptying. And again, one of the advantages of cross sectional imaging, particularly CT, is that the spatial resolution is so exquisite, we're able to manipulate these data sets and really see very subtle features here we can see almost a sort of highly stenotic, if you like, coronary sinus emptying into the left atrium, and that was confirmed subsequently on catheter angiography. And I think Dr. Goldstein makes a good point. Of course, the other thing that we can do at the same time when we do this kind of imaging is look at the, uh, the upper circulation, make sure there's a bridging vein between the left SVC and the right SVC if we're considering intervening on this. So other issues. I think most of the other issues we see, but I don't think imaging is the main focus. So, uh, you know, we can, we quite often see fibrosis within the atria in Fontan patients, particularly the adults. We may see pulmonary venous compression and of course we're going to see what echo probably shows us better, which is systemic AV valve regurgitation, but by the time we're seeing these kinds of things, many of which present with arrhythmia, we're really looking at adult Fontan patients who may be close to needing either a revision or consideration of transplant. Uh, I've put this case in because I wanted to emphasize that the patient who presents with a heart failure phenotype, particularly an edematous patient, does not necessarily have pump failure. Uh, this is a patient from the Royal Brompton Hospital in London who presented, uh, with anasara, and this is a, a Bjork type fontan where there's some, some kind of conduit has been placed between the right atrium and the right ventricular outflow tract, uh, and there is narrowing of the con of the conduit. You can see that. Marked by yellow asterisks you can see there's extensive pleural fluid and there's also acidic fluid, so this patient came in pretty waterlogged. Uh, you can also see that flow within the atrial conduit was slow and turbulent, so obviously high risk for forming thrombus. Uh, the interventional cardiologist put in a stent which is difficult to see because of artifact from the metal and MRI, but what you appreciate immediately is that the pleural fluid is gone, the ascites is gone. Uh, and in fact even the appearances within the atrium look, uh, more encouraging, suggesting that flow has been, uh, optimized. So we need to remember that when a patient comes in waterlogged, it doesn't mean like an adult patient that their pump is necessarily failing. There may be other reasons, and I think cross sectional imaging is, is useful for establishing what those can be. Now this is something that somebody has already mentioned, plastic bronchitis. I throw this in. I've never seen it. I haven't seen it in 10 years, uh, in an adult patient. I'd be curious to know whether any of my colleagues have seen it in adult patients. I don't know whether we're missing it, um, but I have never had a patient tell me that they're coughing up funny things. Uh, having said that, it's easy to look for by CT. CT is obviously fantastic for the airways, um, and we can produce very clear pictures. These are, these are from a pediatric publication, but it's very easy to see evidence of, of abnormal material within the airways. So that's just a whistle stop tour. I don't have a lot of conclusions. I think my main conclusion really should be that the imaging should not be done by the musculoskeletal radiologist, but should be done by somebody with really, really an interesting an understanding of the anatomy and physiology. Uh, and I'd really like to emphasize that, and this is a personal opinion and the panel may disagree, but I do not believe that in adult patients echocardiography is adequate for thrombus in a fontan circuit. I've seen too many cases where echo says there's thrombus and there isn't, and an equal number of cases where where echo says it's all clear and that's not the case. So I'll leave it at that and I'm happy to take any questions. Thanks, Andrew. That was, uh, superb. I, I have a question about TEE. Do you think the problem with TEE is a problem with the technique? In other words, the, uh, either the resolution or the quality of the images or, uh, or is this a problem with how the TEE is done? In other words, the technical skill. Uh, of the, the operators, in other words, getting to different parts of the circuit that you wouldn't necessarily on the standard TEE approach, you may get. Um, I, I just wonder, I mean, should we throw TEEs in the bucket because we do a lot of it obviously pre, um, pre-cardioversion. Uh, so what's your, what's your great questions and I'm gonna give you a very biased answer, so. I, I think it is a bit of both. I think it's partly the technique and partly the operator. Clearly I think an experienced operator who understands the anatomy is more likely to detect an abnormality if it's there. Uh, it also depends a little bit on the circuit that you're imaging. If you're imaging a lateral tunnel or an extra cardiac conduit, I think you can say it's normal without even putting the probe down 99% of the time, and you'll probably be right. Most of the problems come with the atriop pulmonary fontans in my experience. And for those, if yes, I think, I think transesophageal echo to answer the question of is there a thrombus in an atrial pulmonary fontan, I think TE is a waste of time. I wouldn't do it. I wouldn't do it prior to cardioversion. I'm, I may be an outlier in that view. It just, it took me roughly 10 years to convert my colleagues in Toronto and many cases where. The CT or the MR were right and the echo were wrong. Uh, now I would say we've, we, we've largely moved away from transesophageal echo pre-cardioversion, um, and the other, it's simply a practical thing as well is I mean we can get a cardiac CT within about 1 hour. It's a lot more difficult to organize space on the CCU to to do a sick Fontan for a, for a DE. So personally I, I would be quite happy to cardiovert a Fontan patient on the basis of a CT scan and wait 10 minutes. Sorry, and wait 10 minutes or turn them upside down. Oh, you rarely have to wait 10 minutes. Yes, do. What happens if, I mean, do you think you have the resolution to see a 3 millimeter thrombus in the left atrial appendage? Yes. I do, so the, the spatial resolution of cardiac CT is at worst 0.5 millimeters isotropic. Yeah, but I'm just talking about, you know, you've got the anatomy, you've got the pectinate muscles, you've got this, I mean, um. It sometimes seeing something flibby flobbling around in the left atrial appendage, it clarifies that it's not supposed to be there, whereas if you've got a relatively static image. I have never yet seen a case where we've had to go from. CT to echo based on doubt. Now you could, I suppose you could accuse me. Well, that's right, I, I realized you were gonna say that I shot myself in the foot, but I've seen plenty of, I have, what I can say is I've seen plenty of cases in the other direction where somebody has said your side, I'm not sure about this. I'm on your side. The problem is when you have something, this is always always the problem with imaging, when you have a test that is regarded as the standard of reference. How do you prove that something new is better except by experience or or outcome someone has a stroke. So, so in, so in 10 years we have not seen that happen. I accept that that's anecdotal, um. And I can't prove it to you in any sense because there isn't any other standard of reference against which I can judge. I mean, not in our field because we don't have enough patients, but you would think in, uh, you know, adult AF you might be able to do that study. Yeah, cos it's usually the left atrial appendage that's the, the culprit. It, you know, having done a lot of this, it's sometimes difficult enough when you've got a really good TE and somebody with a structurally normal heart to know whether it's a pectinate muscle or, so, uh, anyway, let me just throw it into the mix. The the other thing I would say the pectinates are what caused the problem, but of course on CT because you can. You can after the fact, look at that left atrial appendage from any angle you want. You can follow the pectinates. You can see that these are muscular structures and, and, you know, on a, on one view on a TE it may look like a a freestanding round thing, but you turn it around on the seat, you can see it's connected to a wall. And so actually that question, although reasonably that rarely arises as a problem. We've got one question from the audience, I believe. Yeah, but Mike Takahashi from Seattle, I'm assuming they've answered your question, which is how do you compare TE to MRI for sensitivity. So, I think you've just answered that question. It is also very operator dependent, or at least its interpretation is that so I mean it's, it's less operator dependent because the reason for that is that when you do a transesophageal echo, you've only got one operator. And the image, as you know, the images you require are what you're stuck with. With, with a CT, as long as you have the timing OK, the data sessions throwing up and vomiting and coughing and spluttering. I mean, is that CT, is that CT or is that no, that's, that's the tea. No, there's no doubt that there's there's the downside to doing it, but I just wonder whether it will ultimately have the sensitivity, you know, and, and where am I more worried about the clot is that exactly, the negative predictive value is everything. And you know, you only have to, I mean, until you've done 500 of them. And not had a stroke, you probably don't know yet. So there that that may be true in the congenital population. There are, there are large series now in the adult AF population. Oh what comparing? Yes, OK. But, but as we pointed out, the comparison is irrelevant because, you know, our hypothesis is that CT is better. So there are outcomes up to I think about a year now, and that may not be long enough, but there are outcomes up to about a year suggesting that that you don't see stroke when CT is called the study, and I guess I'm less concerned about clot in the conduit. Uh, as far as stroke is concerned that I am in the, uh, systemic or pulmonary venous atrium rather, yeah, yeah, so depends on the patient and the circumstances very much. OK, we're getting very late.