Chapters Transcript Video Beyond Echocardiography: Using Advanced Cardiac Imaging to Visualize the Heart Thank you, Stu and thank you to all of you for hanging in there through this till the end. Um And for being such a wonderfully engaged audience. So we've spent the last four hours inundating your brains with hopefully high yields clinical information. So we decided that we are going to end with some pretty pictures and just a little update on uh pediatric cardiac imaging and these are all every picture in this talk is a picture that we've taken on one of our patients. So, what we're gonna do is we're going to talk about some of recent remarkable advances in noninvasive diagnosis of cardiac disease in Children, particularly in cardiac MRI and CT, which is what I do. Um We're also going to discuss imaging safety considerations, particularly with regards to radiation in Children and Children with congenital heart disease. And we're also going to briefly review some advances in 3D printing and A I technology as it's been used with cardiac imaging, everything has beauty, but not everyone sees it as cardiologists. In order for us to be able to diagnose and treat cardiac abnormalities. First, we have to be able to visualize them even prehistoric peoples were aware of the vital role of the heart in sustaining life. This cave drawing is believed to be 30,000 years old. It is p purported to demonstrate where a hunter needed to aim. Oh, here, let me just get my laser pointer back where a hunter needed to aim in order to take down his prey. And it's thought to be one of the earliest known depictions of a heart by the 15 hundreds anatomists such as Da Vinci, gave us more detailed understanding of cardiac anatomy using cadaver specimens. It was really messy work because modern embalming techniques were actually not used during the renaissance and his specimens were prone to rapid decomposition. Still, he was able to describe the circulatory system, cardiac valves and struct uh structures and function and even coronary atherosclerosis, imaging of live hearts did not begin until the 19 hundreds initial assessment was invasive. So this X ray is kind of neat. It shows the first use of cardiac angiography. It was performed by a 25 year old German surgical intern named Werner Forsman. He inserted a urinary catheter into his own heart. As his horrified or nurse tried to stop him. You can see the catheter extending from his left cephalic vein into the SVC and down towards the right atrium. So eventually, uh Doctor Forsman actually won a Nobel Prize for his discovery. But in the short term, he was unfortunately fired from residency by his program director, echocardiography was for uh was first introduced several years later. So they weren't able to see the pretty echo images that we're used to seeing today, like what Doctor Dayton showed you. But instead they projected a single ultrasound beam through the chest back and forth and plotted its findings as a graph over the course of the cardiac cycle, we called it M mode. And so the lines in this mitral valve image actually demonstrate not the mitral valve itself, but actually the movement of mitral valve leaflets, which is the y axis over time, which is the X axis. With this incredibly basic technology, imagers were able to figure out a remarkable amount about the anatomy and the function of the heart. And we're even able to start doing fetal echo this way, echo and cardiac cath have advanced significantly since then. But inherent limitations to these conventional imaging modalities remain echo is our primary bedside imaging tool as pediatric cardiologists. But images are can be limited by poor ultrasound windows. As you can see in this patient with morbid obes obesity where it really just kind of looks like snow. Many cardiac structures, particularly the great vessels, the great veins of the heart are also very difficult to visualize through the chest wall. It's also difficult because we are dealing with a complex 3d shape that doesn't sit in sort of Xy and Z planes like everything else does in the body. Um And we're trying to quantify its size. We're trying to quantify its function by viewing it in two D slices. Let's talk about diagnostic cath, diagnostic cardiac catheterization is commonly, is commonly performed in pediatric cardiology. It gives you high resolution real time images but it's highly invasive and putting a catheter into the sem femoral vessels of a very small baby can cause severe damage. I actually had a young adult trunks, arteriosus patient a few years ago who was essentially disabled by her leg length discrepancy that resulted from occlusion of her femoral artery during a childhood cardiac catheterization. There's also significant ionizing radiation exposure. Um We're gonna discuss radiation more in a bit. The average cardiac catheterization can expose a patient to six millisieverts of radiation for reference A P A lateral chest X ray is 0.1 millisieverts. So 60 chest X rays worth of radiation. And then finally, most pediatric cardiac catheterization requires general anesthesia. And we know that not only are there potential neurodevelopmental effects of general anesthesia in young Children, but anesthesia induction actually uh carries a higher risk of cardiac arrest in Children with congenital heart disease compared to those without enter cardiac MRI. In the 19 nineties, cardiac MRI became a major breakthrough in being able to visualize the heart. So I'm not going to go into too much detail about how MRI machines work. Although I took a semester long course on MRI physics at the NIH. Um MRI machines utilize magnetic fields and electromagnetic waves to generate high resolution moving pictures of the heart. Because unlike doing an MRI of say your brain or your leg, the heart is in constant and sometimes irregular motion. So how do we, how do we compensate for that cardiac and respiratory motion through ecg Gating? So, Ecg Gating is the idea of controlling for motion by generating images at the same point in the cardiac cycle every time to create an image. And then you combine images from different time points to create a complete heartbeat. This takes a really long time, MRI is slow and the prettier, your pictures are the slower they are. That's always the trade off. So we control for respiratory motion because you're gonna have to breathe at some point by either having patients hold their breath for shorter acquisitions up to you know, 10 or 15 seconds or by also gating to a consistent part of the respiratory cycle. Cardiac MRI provides us with an amazing wealth of information about the heart. It gives us anatomy. More importantly, it gives us 3d data sets that you can manipulate, you can cut through in any plane. It gives you functional information about cardiac size and function about blood flow and about the presence of even scar or inflammation in the myocardium. Why do people refer patients for cardiac MRI S? So as we said, there's certain types of anatomy like great vessels that you can't see well. And so oftentimes, it tends to be uh congenital abnormalities that involve the great vessels TTR of flow. Uh Cardiac MRI is the gold standard for visualizing de transposition of the great arteries, aortic coition. As you can see here in this picture, vascular rings, single ventricles, et cetera. We can also refer patients without congenital heart disease in whom quantitative or any patient really in whom quantitative functional data is needed. So it can help you assess for, you know, myocarditis and cardiomyopathy. One of the neat tricks that we can do is that we can actually even noninvasively diagnose the histology of cardiac tumors just based on their different MRI signals, which has saved us from having to go in and literally open up a child's chest to biopsy a cardiac tumor. So here's a case example of a 14 year old boy with tetra of fallow just to remind you about tetra of fallow. So tetra of fall is really caused by this little guy right here. It is a piece of the ventricular septum that should have been right here, but it gets anteriorly deviated. It moves into the right ventricular outflow tract obstructs the outflow tract leaves A VSD and pulls the aorta, the overriding aorta over the septum. So structures downstream from this obstructed right ventricular outflow tract like the main and branch pulmonary arteries don't get to adequately adequately develop because as we cardiologists say, no flow, no grow. So surgical repair for this particular patient involved closing the VSD with a patch, opening up the right ventricular outflow tract with what's called a trans annular patch. So the way I like to describe a trans annular patch to patients, it's kind of like you take a skinny pant leg, you open it up at the seam and you make it a wide pant leg by adding additional material. Um but the problem with flaying open the pulmonary valve is that valves are there to open. So the blood can go through and they close. So the blood can't leak backwards. But if you've opened up your valve annulus, it's not going to coapt anymore. And so the patient ends up with what's called free pulmonary regurgitation. When you have all that back and forth regurgitation, it leads to right ventricular dilation and eventually dysfunction. So our 14 year old with this trans annular patch repair has recently been getting short of breath with exercise. So it leads to the question, is it time to replace his pulmonary valve to stop his regurgitation? So here are some MRI images of the heart. We call them sine images because they are beating heart images. And you can see how nice and clear they are compared to echocardiograms. Um You can see the dilation, this is the right ventricle here up against his chest wall. You can see the dilation of the right heart relative to the left from all of that regurgitation. You can also see it in the short axis here when we're kind of sweeping our way through the heart from the base to the apex. And in the third view, this is actually the pulmonary valve. This is where they've kind of cut open that pulmonary valve. And you can see the blood swishing to and fro across the regurgitant valve. So in order to get quantitative data, this is how we do it, we section the heart into individual slices of equal thickness. We measure each slice area, we tell the computer what's blood flow, what's blood, we tell the computer what's, you know, myocardium, which is the cardiac mass. And because we've already determined the thickness of these slices, we can figure out the volume of each slice and we can add them up in Sicily and we can add them up in diastole to get get volumes and therefore also their ejection fractions. So once we've done our volume and our ejection fraction measurements, we determined that our right ventricle when indexed to the patient's body surface area is severely enlarged. Look at the comparison there. Um and there is moderate to severe pulmonary regurgitation. 43% of the blood flow that for flows forward is actually coming back. So beyond a certain size threshold, the right ventricle is actually unable to remodel to a normal size even after you put in a new pulmonary valve and eliminate the regurgitation. So when you get this far, this meets established criteria to go to the operating room. Another key part of our um cardiac MRI is that we get Mr angiograms, we use gadolinium contrast, which is actually different than the contrast we use in cat scans. And this Mr angiogram, sorry, it's I don't have control over how fast it's spinning, try not to get dizzy. Um So this Mr angiogram is actually just showing you the right heart. I have cut away all of the aorta and the rest of it. And you can see this is the right ventricle leading to the right ventricular outflow tract and the right ventricular outflow tract. Sorry if you can follow it is there's narrowing down proximately, but you can see that the pas are actually really big and kind of bubbly looking and that's from all of that to and fro flow. And so all of these things when this patient goes to the operating room for his pulmonary valve replacement could potentially require revision. So, MRI is not without its drawbacks, implanted devices like stents and artificial valves can cause artifact, obscuring. Pictures of the heart implanted devices like pacemakers are becoming increasingly MRI compatible but are not always. Also gadolinium contrast has been kind of considered problematic in recent years. It was the subject of a 2015 FDA black box warning that said that after multiple exposures, there has been, there have been studies showing that there is a risk of gadolinium metal deposition in your tissues in your brain, in your skin, in your kidneys, even with, you know, a completely intact uh blood brain barrier. And so as a result, you know, the jury is still a little out on exactly what this means for people. But as a result, we have cut back significantly on our use of MRI contrast, particularly with repeated administrations of contrast. Thanks to new technology as well, we've been able to do a lot of non contrast MRG grams and find other ways to get around using gadolinium. Finally, MRI S are long tests to get all of this data. We take about 3000 pictures. It can take up to 90 minutes. That is going to be really hard for kids under 10 years of age who are being asked to lie in a narrow tube and hold their breath and kids who are developmentally delayed, that's that's gonna be pretty difficult. And so these patients generally require sedation and it's not just sedation. We actually have to sedate and paralyze them with general uh an anesthesia and intubation so that we can control their breathing. So a lot of these uh a lot of these limitations have actually been addressed by cardiac CT. Cardiac CT is a much more recently adopted tool in pediatric cardiology. It's seen dramatic growth in the last decade. So, um the fizz database is uh that you guys may be familiar with. This is a hospital pediatric in patient utilization database and fizz data from the 2000 and tens found that pediatric cardiac CT actually only started growing as of about 2014. But by 2017, it had already completely surpassed MRI. And that's because it's able to get around some of these limitations. So first of all, cardiac CT is very high resolution, MRI is usually about three millimeter slices up to eight millimeter slices. Whereas cardiac CT is less than one millimeter, it also acquires these an atomic images in one heartbeat, ecg gating controls for cardiac motion. Um You can avoid gadolinium contrast because we're usually using um iodine based contrast, which is the same stuff we use in the cath lab and you can typically avoid sedation unless you have a really wild patient. The big drawback of course ionizing radiation. So a lot of attention has been paid recently to the rapid rise in the use of medical imaging involving ionizing radiation. On one hand, best radiation is no radiation as a cardiologist and radiologist dramatically pointed out in the New York Times a few years ago, medical testing gives us cancer and this particularly affects congenital heart disease patients. So, studies have shown that in congenital heart disease patients by the age of 64 their cumulative cancer risk has actually been shown to be about 50% higher than in the baseline population. And part of the reason is because of we think is because of all the radiation that they receive. So this was a study from a few years ago that showed that in patients who had surgical congenital heart disease, um less than the age of six. So these are patients who had at least one open heart surgery at some point in their first six years of life. Their median imaging radiation dose just in those first six years was 2.7 millisieverts, which doesn't sound crazy, but look at the range, it goes up to 77 millisieverts to place this into context. The average Hiroshima survivor was exposed to about 210 millisieverts of radiation. So we're really getting up there. On the other hand, though cardiac CT can provide you with high quality imaging while avoiding all those higher risk alternatives. So in recent years, the American College of Radiology has established a campaign to limit pediatric radiation exposure across the board. It's called image gently. Um It emphasized being judicious and optimizing our protocols. So we want to make sure we are doing the right test that will affect our management at the right time in the right way to maximize benefit to our patient. We're also guided by a principal called a Lara. We are optimizing our settings to attain as low as reasonably achievable radiation dose that will still answer the clinical question. Um Actually, as part of this campaign in 2018, I was involved in putting out the first set of guidelines for radiation dose management and pediatric cardiac CT. And we've been increasingly working to try to benchmark and try to kind of consistently bring that dose down across the field. We utilize several methods to minimize radiation in our scans. So typically the prettier the picture, the more radiation. Um but you know what, we don't really need pretty pictures. This is an adult cardiac CT in adult cardiac CT. You are looking for fine details of coronary atherosclerosis. Most of often in pediatric cardiac CT, we are just looking for anatomy. We don't need beautiful images, just diagnostic ones honestly. If you be if your picture is too pretty, it means you're using too much radiation. So we use much lower X ray settings. We also actually what's kind of neat in the last few years is that we've started using artificial intelligence algorithms in post processing to improve image resolution. So you can see this here. And unfortunately, I'm not sure the screen makes it clear how different pictures are compared to this. Um But the picture on the left is basically the image. Yeah, and it really doesn't show up all in here. The picture on the left is the image that we took with lower radiation in a patient with um a history of repaired aortic COMT. And we used artificial uh intelligence algorithms to sharpen it up on the back end kind of like a very high end Instagram filter. Um We also minimize scan coverage. There's no reason for me to scan the kid's whole chest. We will scan as little as 468 centimeters just enough of the heart to answer the clinical question. Another final way that we uh that we decrease our radiation exposure is that we use medications like beta blockers to lower patients heart rates that we're decreasing cardiac motion artifact making our images sharper. So we don't have to use as much radiation. So we've successfully employed these techniques in our cardiac CT program here at Atlantic Health. I've listed here the average radiation doses for common exposures again for reference P A and lateral chest X rays, 0.1 millisieverts. I don't think a lot of people realize that we actually are all exposed every year just by virtue of living to natural background radiation that's from naturally occurring elements in the soil and the upper atmosphere. And it actually differs by what country you live in in the United States. It totals about 31 chest X rays per year. Interestingly, even flying round trip from New York to L A gives you about a half a chest X ray's worth of radiation just from what's naturally occurring in the upper atmosphere. So it kind of makes you scared to be a frequent flyer. Um In that context though, when you look at it, this is our average for an infant cardiac CT, we've been able to scan our patients with remarkably low radiation doses, especially compared to the old standard of diagnostic cath. And so this actually is in continuity with Lauren Rosenthal's case that she presented because it's the same patient, but we're going to talk about her from an imaging standpoint. So again, this was a real case that happened in 2022. It was actually all over the news. Um A 15 year old girl collapsed during track practice. As Lauren had mentioned, CPR was initiated, she had an appropriate shock with an ad and she had returned to spontaneous circulation and we met her when she was transferred by helicopter to us. And so my colleagues found a coronary anomaly on her echocardiogram and we did ac T to confirm the diagnosis. And so this is a picture of what Lauren had been describing on our CT. So just to orient you, this is kind of zoomed up. But when you look at a cardiac CT or any sort of CT scan, you have to picture the ta the patient lying at a table with their feet sticking out towards you. So this is the patient's left, this is the patient's right, this is their chest and this is their spine. And here we are seeing their heart and we focused the contrast to light up the left side of the patient's heart. This is the patient's aortic root, which is where and you're seeing it in cross section. And this is where the coronaries come out normally, your right coronary comes out at about 11 or 12 o'clock and wraps around the right side of the heart. Your left coronary typically comes out at around 54 or five o'clock and wraps around the left side. But you can see here that your left coronary, you can see here that your left coronary is actually coming out from next to the right coronary from the right aortic sinus. And unlike the right coronary, which comes out straight, it comes off at this like really acute angle and it's narrowed and it follows this intra arterial course between the aorta and just below the main pulmonary artery. And you can also see that it's closely up against the aortic wall and it's actually embedded in the aortic wall. So as I said, this was, as we mentioned, this is an alo anomalous left coronary from the right aortic science. The left coronary follows that intramural course, intraarterial course, sorry, and an intramural course within the aortic wall. This is actually a 3D is the same thing, it gets compressed during exercise. And these patients have a 6% per year risk of sudden death. It's part of the reason why coronary anomalies are the second leading cause of athletes sudden death as Lindsay had showed. Um so this patient went to the operating room, I'm not going to go into the details of how the surgery was done, but it's essentially something called in called a coronary unroofing procedure, which actually created a new ostium from the correct sinus and kind of pulled her uh coronary out from inside the aortic wall. And on follow up imaging, she had good coronary flow, good cardiac function. I actually just did a cardiac MRI on her last week and she has no scarring in her heart. Um One last example of AC T this is a 25 week infant from our NICU who developed the dreaded high mortality prematurity, complication of pulmonary vein stenosis. Um Cardiac CT is a mainstay of evaluating and surveilling pulmonary vein stenosis. We were able to use low radiation CT to diagnose him and to follow him serially. So you can see here this is his left pulmonary veins, they're kind of hanging by a little string. Um after surgery, they looked much better. So this is after the first surgery, you can see that they're much more open on his left. Um And then so unfortunately, he had re stenosis. So he had um stenting. You can see the little stents right in there. And then subsequently, he had instant re stenosis requiring balloon angioplasties. So you can actually see within these stents, little black lines and that's neo internal hyperplasia. And we've been able to safely image this patient in a serial manner. Finally, let's very briefly talk about some emerging technologies and managing congenital heart disease. So 3D printing and modeling hold great promise in pediatric cardiology. So, repairing complex cardiac defects requires extensive presurgical planning. You need to go know all of the details before you go into the operating room, because you're not going to waste time on your bypass pump, figuring out the anatomy. And so conventional imaging has limited ability to depict intracardiac anatomy and especially 3d spatial relationships. And so, MRI and CT based patients specific 3d scale printed models will allow you to literally hold the heart in your hand to plan repair. Um This is actually the first picture that I'd shown of the patient with the uh with the giant coronary fistula at the beginning of my talk. Um And so this is a 3D model of it and this actually is a hollowed out version that we gave to the cardiac catheterization docs to practice putting in a little plug to plug up this coronary fistula. And this is actually a new story from 10 years ago on the first 3d printing case I ever did in New York, which actually went viral. And it was what was kind of amazing was that the information it provided with the surgeons significantly improved the outcome for the baby and allowed him to do the repair in one stage instead of multiple stages. Oops. Oh Is that not playing? Oh, it was working before, I'm sorry. So since then, we've actually moved beyond physical models to augmented or virtual reality guided procedure planning. So this would have been really cool to watch this video, but surgeons and interventional calf docs, they now actually combine 3d digital models, echo images and cath angio s. And they can literally use these little joysticks to cut through the heart and practice their procedure even before they set foot in the or or the Cath lab. And finally, as with many other fields of medicine A I and machine learning are playing an increasing role in cardiac imaging. So machine learning mimics humans most valuable skill which is the ability to learn and improve from data. So I already showed how A I based algorithms can reduce radiation exposure in CT by sharpening up imaging quality. A I can also be used to identify coronary artery plaques and calcium with greater accuracy than us cardiologists. And it can even be used to noninvasively model coronary blood flow. So I promise this is the last patient I'm going to talk about. This is an 18 year old woman with Williams syndrome. So you may remember that Williams syndrome is associated with supval VR aortic stenosis. Um And this was actually a patient of Doctor Donnelly's. In her case, the supval VR narrowing resulted from in this membrane. You see this little line right here in her aorta, this is her aortic valve, this is that narrowed super valva uh area. And you can see that that membrane is actually cinching down her super valva area and it's obstructing flow to her left coronary artery which is kinked around it. Um So what we did was we used uh based on her cardiac ct, we used machine learning based software that combined computational flow dynamics, pattern recognition and machine learning to noninvasively predict perfusion to hurricane coronary artery. Thank you so much for hanging until the end. We really appreciate it. And I'm so sorry, I think we've probably run out of time for questions. Um And so if anybody has questions, I'm sorry, I ran, I'm the one who ran over. Ok, perfect. And so um thank you all for spending your Saturday with us. We we really hope that you uh gained a lot from it. Um I just want to thank again the planning committee, Jeff, our course director and our entire team. I am so proud of our program. You know, this is a program that Doctor Donnelly had built and made so amazing and we continue to build on it and we really appreciate the opportunity to take care of your patients. Thank you. Published Created by
