Hi, everyone. We'll go ahead and get started. Thank you so much for being here. We have privilege to hear from Dr. Christopher Kramer today on imaging for nonischemic cardiomyopathies. 

Dr. Kramer is the chief of the cardiovascular division in the George A. Baylor distinguished professor of cardiovascular medicine. He got his MD from UC San Francisco and then went to the University of Pennsylvania, where he did internal medicine residency, a chief resident year, and then his fellowship in cardiovascular medicine. 

He came to UVA in 1999. And he's been the director of the cardiovascular imaging center at UVA. And under his leadership, UVA has become a world leader in cardiovascular imaging. He has six patents, 34 book chapters, over 230 peer-reviewed publications as an expert in this field. So join me in welcoming Dr. Kramer. 

[APPLAUSE] 

All right. Thank you for the kind introduction. Can you hear me in the back? All right. Good. So these are my disclosures. None of them relate to the topic of today's talk. 

So I'm talking about nonischemic cardiomyopathies today. But you have to remember, when a patient presents in heart failure, the first thing you want to do is exclude coronary artery disease or ischemia as the underlying etiology, because it's the most readily reversible cause of heart failure. 

And so the first thing you want to do is be sure that it's not ischemia. So how do you do that? In a low risk patient, you might consider a coronary CTA. In an intermediate to higher risk patient, you probably either coronary CTA or a stress test. In a higher risk patient, you would probably go directly to cath to exclude coronary artery disease. 

So once coronary disease is excluded, then you really want to know the underlying etiology. Today, we're going to talk about the use of non-invasive imaging, particular CMA, PET, and SPECT, in terms of making a diagnosis and also assessing prognosis and moving into choosing therapies based on imaging. 

So gone are the days where we have to use a biotome and take a piece of heart tissue to make a diagnosis in nonischemic cardiomyopathy. So advanced cardiovascular imaging has become the non-invasive cardiac biopsy. 

So we're going to start by talking about the techniques we use in cardiac magnetic resonance, talk about how we image fibrosis, talk about parametric mapping. And then I'll present to you five cases, different cases, of nonischemic cardiomyopathies. 

I will be talking about the use of gadolinium. And I'll just remind you that we can't use gadolinium in patients with GFRs under 30 because of the risk of nephrogenic systemic fibrosis. But I will say that now, with cyclic gadolinium contrast agents, the incidence of NSF has sunk to just about zero. But we still avoid it in patients with GFR less than 30. 

So the first technique that I need to tell you about is a technique called steady state free precession imaging. It's a way we do imaging of left ventricular size and function. Let's see if I can-- no. 

On the left is a two chamber long axis image in a patient with a dilated cardiomyopathy. On the right is a four chamber long axis image in a patient with dilated cardiomyopathy. And you see this technique was developed around the year 2000, when the scanners got fast enough to get this kind of image quality. 

So you can see very high contrast and noise between the myocardium and the blood pool. We see valves really well and trabeculation. So this has become the gold standard technique for measuring LVs, the volumes and functions, against which all other modalities are compared. 

The second technique I want to tell you about is the technique we use to assess scar or fibrosis in the heart. This is a technique that was worked on in the 1980s by Charlie Higgins, by our group in the 1990s. 

But it really wasn't perfected until Lan Simonetti and colleagues developed this pulse sequence around, again, at the turn of the millennium that ideally depicted the area of scar. And what they did is develop a technique called inversion recovery, which turns off the signal from normal myocardium. 

So let's see if I can get this-- I can't get the pointer to work. I can use the mouse, actually. So in this case-- can everyone see the-- is the mouse showing up? 

Take it all the way to the right. Keep going. There it is. 

Ah, OK. So you can see that the normal myocardium here is black so that we've turned off the signal from a normal myocardium. We inject standard clinical doses of gadolinium. And then we image 10 to 15 minutes later. 

So this is a patient with 50% transmural inferolateral myocardial infarction. And the myocardial infarction appears bright. Normal myocardium appears black. So why does the myocardial infarction appear bright? 

Well, it turns out that gadolinium accumulates in areas where there's increased volume of distribution, such as in collagenous scar. And there's also delayed washout in that area. So scar appears bright. 

So really, this is the first technique that we can really directly image the transmurality of myocardial infarction. We used to guess it by EKG. But now we can image it with this technique. 

So we have to understand the difference between replacement and interstitial fibrosis. This is from a paper Mike Salerno and I wrote a few years back showing the examples of the difference. So on the top left is a patient, again, with a myocardial infarction, subendocardial. You wouldn't biopsy him. But if you did-- him or her-- you would see this confluent fibrotic myocardium. 

The patient on the right-- so this is a late gadolinium enhanced image. The patient on the right is a patient with hypertensive heart disease. This is a post-gadolinium image here. But you don't see any scar. But if you biopsy them, they have this lacy, interstitial fibrosis. 

So we can't see it by late gadolinium enhancement, the technique I showed you earlier. But we can actually measure it quantitatively with a technique called T1 mapping. So how does T1 mapping work? 

So this is a cartoon from a review by Eric Schulbert a few years back showing how T1 mapping works. So the top left is a-- I lost my pointer there-- is a normal patient. So you have the blood pool on the left and then the myocardium in the top right. 

On the bottom left is a patient with interstitial fibrosis, a hypertensive heart disease. So you see this increase in this lacy blue collagen. Now, if you give gadolinium GD here, in a normal individual, there's a little bit of gadolinium in the interstitium. But in the patient with interstitial fibrosis, there's a lot more gadolinium that gets caught up in the interstitium for the reasons I mentioned earlier. 

Now, gadolinium in an MRI does what we call T1 lowering. So it lowers the T1 of the myocardium. So you can measure how much of the gadolinium is present by measuring the T1. The lower the T1, the more the interstitial fibrosis. 

And the last technique I'll tell you about is a technique called T2 mapping. So T2 in an MRI of the heart is sensitive to the myocardial edema. We have both qualitative and quantitative techniques to assess myocardial edema. 

On the left is the older, qualitative technique called T2 STIR. You can see this bright signal in the midwall of the inferoseptum in this patient with myocarditis. But we've progressed from imaging, which has its own artifacts, to mapping. We can actually measure the T2 of the myocardium on this image. 

So normal T2 is less than 60. If you put a cursor and measure the T2 in this bright signal here in the inferoseptum, it would probably be 70 or 80-- so high T2. So we can actually measure the amount of myocardial edema in the myocardium. 

So let's take all these techniques, put them together, and tell a story of how we use them in concert to make diagnoses in patients with heart failure. So the first case I want to share with you is a 64-year-old African-American female. She's had three months of dyspnea on exertion and leg swelling, and she presents to you in clinic. 

Her vital signs are as shown. She's got bibasilar rales and S3, no murmur, and 2 plus edema to the mid-shin. She had no risk factors for coronary artery disease and was considered low risk. And for that reason, we did a CTA, which showed normal coronaries, no obstruction. So she has a nonischemic cardiomyopathy. So the question is, what does she have? 

So here are her four chamber, three chamber, and two chamber long axis cine images. And what you see is severe global systolic dysfunction shown. Now, the only image that's moving here is the two chamber long axis, so inferior wall, apex, anterior wall. 

Let's see if I get the movies playing again. In the four chamber view-- it doesn't want to start. Go back and then forward. 

The four chamber septum apex lateral wall, three chamber [INAUDIBLE] septum, inferolateral, severe global LV dysfunction. EF probably 20% to 25%. So then we do our parametric mapping, which wasn't revealing in this case. And lastly, we image with late gadolinium enhancement. 

So we've infused gadolinium. This is 10 to 15 minutes later. And these are the images that pop up. And what you see here is that this patient has this midwall stripe of late gadolinium enhancement, shown here in the mid-inferoseptum, seen well on the four chamber long axis here and in the midseptum. 

You see it a little bit on the three chamber long axis. You see it here, on the more basal short axis, too. So this is macroscopic replacement fibrosis in the midwall of the heart. 

So the diagnosis here is nonischemic dilated cardiomyopathy. It turns out about a quarter of patients with nonischemic dilated cardiomyopathy have this midwall stripe of late enhancement. It probably represents old, healed myocarditis. 

But what does the presence of fibrosis mean? Does it have any importance? Well, it turns out it does. This is a study from the Royal Brompton Hospital in the UK in London from Sanjay Prasad's group, about almost 500 patients who they followed from a chronic heart failure clinic for almost 10 years. 

And in each of these graphs, there are two curves-- one representing the 75% or so of patients who had no scar and the bottom one, in each of the curves, the 25% of patients who have scar in the heart. 

And no matter what the end point you look at, whether it's all cause mortality on the top left-- let's see if I can get the-- cardiovascular mortality or transplantation on the top right, sudden cardiac death or aborted cardiac death on the bottom left, heart failure death, hospitalization, or transplantation. 

No matter what the end point that you're looking at in nonischemic dilated cardiomyopathy, scar in the heart is bad for you. And that is true of every cardiomyopathy that I'm going to talk about today-- with imaging, with CMR. If we identify scar in the heart, that is a poor prognostic marker. 

And the reason for that is, number one, arrhythmic risk. It's quite likely that reentering arrhythmias occur from the edges of the scar. And secondly, increased fibrosis worsens both diastolic and systolic function and worsens progression of heart failure. So midwall LGE is an adverse marker. 

[INAUDIBLE] Kuravilla was an imaging fellow in our group. Along with Mike Salerno, we published this meta-analysis now about five years ago, in Circ Imaging, showing that with the presence of scar and in several studies, there was a three-fold increase in mortality, a three-fold increase in heart failure hospitalization, and then over a five-fold increase in composite endpoint. 

The composite endpoint included arhythmic endpoints as well as heart failure and heart failure hospitalization and death. So clearly, based on a number of studies, not just that one I showed you earlier, scar is an adverse prognostic marker in nonischemic dilated cardiomyopathy. 

We'll move now to case number two. Case number two is actually in our CCU just last month. She's a 19-year-old African-American female who works at a daycare center in Lynchburg. She presented to the hospital there with chest pain, fatigue, and dyspnea and, on presentation, was in cardiogenic shock. 

She was started on pressers. She had a moderate pericardial effusion that was tapped there and then transferred to our CCU. At the time of transfer, she was in shock. Her heart rate was in the 130s, her blood pressure in the 80 to 90's, 60 systolic. 

She had a low level troponin elevation, a very high BNP, a cardiac index that was markedly depressed, and an elevated SVR consistent with cardiogenic shock. She was stabilized over the next two days and weaned off of pressers and then sent for a CMR to evaluate for myocarditis. 

These are her cine images. They look somewhat like the last patient, but this is a 19-year-old. Severe global LV dysfunction on the two chamber long axis, four chamber long axis, and the short axis here, EF again 25 to 30. Why does she have this? 

So here's where parametric mapping comes in. So the top left image is a T1 map where we're quantifying the T1 of the myocardium. This is without contrast, so called native T1. 

So we know it on our scanner. And this was done on 3 Tesla. You have to know what scanner the imaging is done on to know what a normal T1 is. 

On our scanner is at 1.5 Tesla. And normal T1 of the myocardium is around 950. On a 3 Tesla scanner, the normal T1 is around 1150. 

This was done on a 3 Tesla scanner. We put a cursor around her short axis. And her T1 is 1352. That's markedly elevated. That's 200 milliseconds higher than normal. 

There are only three things that cause a T1 200 milliseconds higher than normal. One is an acute MI. She doesn't have that clinically or by imaging. 

Two is amyloidosis. She has acute heart failure, not chronic heart failure. She is 19. She doesn't have amyloid. And three would be acute myocarditis. 

We also did T2 mapping. Remember that I said that T2 is sensitive to myocardial edema. Her T2, normal T2, remember, is less than 60 milliseconds. Put a cursor in her midseptum, it was 72-- markedly elevated. 

So she has evidence of increased interstitial space by T1 imaging, evidence of myocardial edema on T2 weighted imaging, and her late gadolinium enhanced images were negative. She had no scar. This is acute. There is no scar. So she has myocarditis. 

Well, there's some cases of myocarditis that do-- acute myocarditis-- that do have late gadolinium enhancement. It really depends on the viral illness involved, because there are different patterns for different viruses. This is the first study demonstrating the fact that late gadolinium enhancement can show myocarditis. This is about 15 years ago, a German group. 

The late gadolinium enhanced pattern and acute myocarditis is very specific. Often, you see this basolateral midwall and subepicardial enhancement-- very different from myocardial infarction. Remember that in myocardial infarction, the wave front of necrosis starts in the subendocardium and spreads to the subepicardium. 

For reasons we don't understand, myocarditis affects the subepicardium of the myocardium. So you can differentiate the pattern of myocarditis from myocardial infarction. What this group did is they actually biopsied the L-- this study wouldn't be done in the US. 

They biopsied the LV in the areas that were enhanced and showed that indeed the enhancement was consistent with acute lymphocytic infiltration associated with acute myocarditis, so proving that the late gadolinium enhancement was inflammation, and acute inflammation at that. 

But our patient didn't have LGE. But that's OK, because the new criteria for diagnosing myocarditis with CMR, which I was privileged to help write and publish in JAC at the end of last year. The rewrite of the Lake Louise I criteria, which didn't work very well, because of newer techniques, we have what we call the Lake Louise II criteria for diagnosing acute myocarditis. 

So the main criteria are, number one, evidence of myocardial edema. And that's either by T2 weighted imaging or T2 mapping. As in our case, we had a T2 map that showed that her T2 was markedly elevated as well as evidence of increased interstitial space that can either be by the presence of late gadolinium enhancement, which our patient didn't have, or markedly elevated T1 on T1 mapping, which our patient did have. So the presence of elevated T2 and elevated T1 in our patient in the right clinical setting made the diagnosis of myocarditis. 

There can also be supportive criteria. She had global hypokinesis, which goes along with that diagnosis. And the presence of pericardial effusion can be helpful as well. But she did not have that. So putting these findings together, myocarditis was the diagnosis. Thankfully, her LV function recovered over the next several days, and she was discharged to home. 

Turn now to the third case-- third case was a 54-year-old gentleman who presented with dyspnea on exertion, had an echo that was markedly abnormal, and was sent for a CMR. This was his CMR, which is also markedly abnormal. So we have four chamber long axis on the left and a short axis image on the right. 

So what's the diagnosis here? Hypertrophic cardiomyopathy. So asymmetric septal hypertrophy as well as marked apical hypertrophy, normal wall thickness-- so here on the basolateral wall. In the short axis, you see the asymmetric septal hypertrophy. 

CMR is more sensitive than echo for more subtle forms of hypertrophic cardiomyopathy. Echo is certainly very useful in hypertrophic cardiomyopathy for measuring the outflow tract gradient assessing mitral regurgitation. We can do all those things by MR as well. What MR gets you that echo and no other imaging modality gets you, again, is the assessment of myocardial scar. 

So here is his late gadolinium enhanced image. He has confluent scar in the midwall and subepicardium of his anteroseptum in the area of most marked hypertrophy. So he had no other risk factors for sudden cardiac death. And I'll go over the risk factors for a moment. 

He was not offered a defibrillator. Unfortunately, two years after this imaging, he was on vacation in Tennessee, played golf, sat down to have lunch after his 18 holes, and died suddenly. So could we have predicted that based on the imaging? And we'll go over that in a moment. 

So what are the predictors we presently use for predicting sudden cardiac death? So in 2011, the ACC-AHA guidelines-- and these are being rewritten right now and likely will include imaging. The current risk predictors include the top three is the most important. And then the bottom two is the secondary markers. 

So family history of sudden cardiac death, syncope, markedly increased wall thickness of more than 30 millimeters-- remember, normal wall thickness is 11 millimeters. Over 15 and asymmetry gets you the diagnosis of HCM. And then the non-sustained VT or hypotensive response on treadmill testing are the minor criteria. 

The problem with these markers is they're really only the tip of the iceberg. They only predict sudden cardiac death. They don't predict heart failure. And the majority of sudden cardiac death events occur in patients like the one I presented to you who had none of these risk factors, or they might only have one of these risk factors. 

The other thing to note-- that the event rate, the sudden cardiac death event rate in HCM, is low. It's certainly a lot higher than the normal population. But it's less than 1% per year. So you're trying to predict relatively rare events. 

The European Society of Cardiology has a different panel of risk predictors. So they have a risk calculator which includes those three markers that I mentioned earlier-- family history of sudden cardiac death, non-sustained VT, unexplained syncope. I think someone may have leaned against the lights there. And then, as well, you put in age, maximal LV wall thickness, left atrial size, and LVOT gradient generally measured by echo. 

So these criteria were developed in 2014 and then tested in studies over the next few years in about 3,000 patients. And what they found was that the observed and predicted were pretty close in the lower risk patients. But the ESC criteria overestimated the risk in intermediate and high risk patients that the observed sudden cardiac death rate was lower than that was predicted. So clearly we have room for improvement. 

So how does advanced cardiac imaging fit in here? Well, there are all kinds of patterns of late gadolinium enhancement in HCM. These are six different patients. Patient A has no late gadolinium enhancement. 

Patient B has confluent late gadolinium enhancement in the anterior wall and anteroseptum. You might take it for a myocardial infarction. But it's not, because there's subendocardial sparing, which never happens in MI. And there's also this patchy midwall area of scar here, in the RB insertion site in the inferoseptum. 

Patient C has the classic form of scar in HCM, which is RV insertion site midwall and subepicardial LGE often shown in pathologic studies. Patient E has the same. Patient F has massive hypertrophy of the inner septum and very fluffy, diffuse myocardial scar. 

Well, it turns out somewhere around 50% to 60% of all patients with HCM have any degree of LGE. So we can't use just the presence of scar as a marker for implanting a defibrillator. Why is that? 

Well, I just told you that the sudden cardiac death rate is less than 1% per year. So you can't put a device in half of the patients. You'd be putting in a device in way more people than really need it. So we have to do better than just the presence of scar. 

Mike Salerno and I published a meta-analysis on this topic back in 2013. But it's been superseded by this later meta-analysis published in 2016, which included significantly more patients, additional studies, but which showed that the presence of LGE was associated with 1.8-fold increase in all cause mortality, a nearly three-fold increase in cardiac death, a more than three-fold increase in sudden cardiac death, and a two-fold increase in heart failure death. So clearly, again, scar in the heart in HCM is a bad marker. But how can we use it in terms of risk prediction? 

So the first shot across the bow here was this study published by Marty Marin's group in Circulation in 2014-- about 1,100 or 1,200 patients followed for up to five or six years. And what they showed for the first time was that there was a cutoff of amount of scar that was predictive, that if you have scar that subtends more than 15% of the left ventricular mass, that was associated with reduced freedom from sudden cardiac death event. 

Now you have to be careful here. The y-axis here doesn't go to 0. This is 90%. So the 10-year event rate is, excuse me, the five-year event rate is about 10%, even in those with the most scar. That's 2% per year. That's double or triple what would be expected in the average HCM patient, which is about 0.7% per year. 

So late gadolinium enhancement, more than 15% of the myocardium by LV mass is a marker of adverse outcome. You can see that this progressive stepwise increase in sudden cardiac death events as you go from no late gadolinium enhancement to less than 10% to 10% to 20% to more than 20%. So clearly the extent of LGE based on this one study is predictive. 

So that brings us to a study that I'm co-principal investigator of along with Stefan Neubauer at Oxford. It's an NIH supported registry in HCM called HCMR, which aims to improve prediction of outcome in HCM including standard clinical predictors-- the ones I mention-- along with CMR in every patient with late gadolinium enhancement as well as T1 mapping, extensive biomarker panels, as well as comprehensive genetics. And we have Core Labs for each of these. That's a website if you're interested in finding more about it. 

Recruitment started in 2014. We completed recruitment of 2,755 patients from 44 sites in North America and Europe in April of 2017. And we're now at about a mean of 2 and 1/2 years of follow-up, closing on three years of follow-up. We recruited over 80 patients from UVA. And Mike Salerno was the site principal investigator here at UVA. 

So our methods paper was published in the American Heart Journal in 2015. But we've now analyzed all the baseline data from this study. And we're one minor revision away from final acceptance with JAC. I kind of jumped the gun here and put it in press. I hope I didn't jinx ourselves. 

[LAUGHTER] 

This is the crux of our findings, the central illustration from this paper. So what we found is that 80% of the patients had one of these two morphologic patterns of HCM-- either what we call reverse septal curvature, meaning the thickest portion is actually in the midseptum. It's about 40% of patients. Another 40% had isolated basal septal hypertrophy, which is the one you think about most commonly. But it's about 40% of the patients. 

Now, again, we did genetics in all of these. And about 36% of the whole population had a positive sarcomeric mutation that we could identify that's been known and associated with HCM. So it turned out that the patients with the reverse septal curvature morphology were at a much higher rate of being sarcomere mutation positive. 

They, like this patient, tended to have more late gadolinium enhancements. So this is an LGE image. You can see the patchy LGE in this very thickened septum. And they had more interstitial fibrosis on T1 mapping. 

Interestingly, they didn't generally have significant LVOT obstruction, because again, most hypertrophy is in the midseptum, not in the basal septum. Contrast that to the other larger group or for a group of 40% or so of patients who tended to be sarcomere mutation negative. They had isolated basal septal hypertrophy, less fibrosis, both replacement and interstitial, but more LVOT obstruction, because it's the basal septum that's hypertrophied. 

So it will be very interesting. We need to go out probably five to seven years, in terms of follow-up, to have enough events to develop a predictive model and find out what is most predictive of sudden death. Is it the extent of scar? Is it a combination of the morphology, the scar, the sarcomeric mutation? 

We know sarcomeric mutations are associated with, in general, a worse outcome. But this is really the first study to put all of this information together. And we hope that, at the end of it, we'll have a smartphone app that you put all this data in. In the end, you come up with the patient's five or 10-year risk of sudden cardiac death and heart failure. 

We are, as of the last four or five years now, a HCMA Center of Excellence at UVA. What does that mean? So HCMA is the Hypertrophic Cardiomyopathy Association. It's the patient advocacy arm of HCM. 

It's run by a survivor of HCM. And she's lost many family members. She's actually required a heart transplant herself. She's a force of nature. She basically runs this association on her own. And she accredits institutions around the country in terms of HCM care. So we were approved as a HCMA Center of Excellence, again, four or five years ago. 

Myself, along with Bob Battle and Matt Thomas, a genetic counselor, put the group together to take care of these patients. It's a multidisciplinary team involving individuals from the cath lab, EP services, TCV surgery, pediatric cardiology, genetic counseling, and then imaging, of course, to help make the diagnosis and prognosticate. 

So in 2010, if a patient in Charlottesville was diagnosed with HCM and they call the HCMA, and they said, where should we get our care? They would say, well, you go to Duke or Hopkins, because those are your closest Centers of Excellence. But now they come to Charlottesville. 

We're getting patients from all over Virginia, West Virginia. The Hopkins Centers had some issues lately. So we're getting some Hopkins patients, patients from Pennsylvania. So we have patients coming from all over the mid-Atlantic to our Center of Excellence now. 

So turn to case number four. It's a 56-year-old African-American male, presented with incessant ventricular tachycardia, admitted to the CCU. The VT was quieted down on Amiodarone. His vital signs are as shown. He had bibasilar rales and an S3 and 1 plus pedal and ankle edema. And he was, once the VT was quieted down, he was referred for CMR. 

And this is his CMR. Four chamber long axis, two chamber long axis, three chamber long axis. We see this very small, focal aneurysmal segment at the apex, discoordinate LV function, severe LV dysfunction, EF 20% to 25%. You see that the basal anterior wall is unusually thin, so unusual wall motion pattern. 

So I stop the movies. And then we look at the LGE pattern. It's a very distinctive LGE pattern. These are the late gadolinium enhanced images. You have this inverse, wedge-shaped scar in the lateral wall, in the inferior wall, and this very distinctive, very bright basal anterior and basal septal LGE that you might confuse for an infarct. 

But this wedge shape can't be an infarct. Again, broader in the subepicardium, narrower in the subepicardium. MIs don't present like that. 

The brightness and the location of the scar are generally indicative of sarcoidosis. Myocarditis could be in the differential as well. But it turns out this patient had a history of pulmonary sarcoid. So that was certainly consistent as well. 

You can actually see here some lung infiltration from the sarcoidosis here. So this is a cardiac sarcoid. It's very distinctively very bright late gadolinium enhancement. Also, very commonly, basal anterior, basal septal involves a conduction system. And that's why many of these patients get AV block, et cetera. 

Well, again, like the other cardiomyopathies I've talked about, scar in the heart is bad for you. And that was first shown by the Duke group in this paper in 2009. DE means Delayed Enhancements-- the way they call late gadolinium enhancement. As you can see, the patients with late gadolinium enhancement had lower event-free survival over a three year period and a lower cardiac survival over the same period-- so clearly a marker of adversity. 

And again, a meta-analysis from Mike Salerno and our group, published a couple of years ago-- 11 studies, 805 patients with three year follow-up combined outcome of all cause mortality and arrhythmogenic events, six-fold increase in the combined outcome and a strong trend to an increase in all cause mortality if you have scar in your heart with sarcoidosis. So someone with a sarcoid LV dysfunction and scar, based on the latest Heart Rhythm Society recommendations, indication for an ICD. 

PET is also useful in sarcoid. And it's complementary to MR. And the late gadolinium enhancement in MR may represent burned out granuloma, old granuloma, or active inflammation. You can't tell. The information the PET gives you is the information about presence of active inflammation. 

This is a paper from the Brigham Group. 118 patients with suspected cardiac sarcoid-- they did FDG PET for inflammation and rabidium PET for perfusion. 60% of the patients had abnormal PET. And they followed the patients for a year and a half. There were 27 VT events and eight deaths. And a positive PET was predictive with a hazard ratio of nearly four independent of LVEF and extracardiac sarcoid. 

So if you look at this graph, the green curve represents those patients with sarcoid in a normal PET. The blue curve are patients with one abnormality on PET, either abnormal perfusion or presence of inflammation. And the red curve is abnormal both parts of the PET, abnormal both inflammation and perfusion. So the worse the PET, the worse the outcome. 

107 patients from the same group had undergone both. 85% of the patients had late gadolinium enhancement. 66% of the whole population had evidence of active inflammation for deoxyglucose uptake. 48 were reclassified based on the PET data. 

So they proposed an algorithm that you start with cardiac MR. And if you see positive late gadolinium enhancement, you go on to FTG PET, because you don't know whether the LGE represent active inflammation or not. If it is active, that patient may be a candidate for immunosuppressive therapy. 

Lastly, we'll turn to case number five. It's a 78-year-old white gentleman with a history of carpal tunnel release a couple of months ago, six months of progressive shortness of breath, vital signs as shown. The chest was clear. There was no murmur or rub or gallop. He had 1 plus edema. And his echo showed moderate concentric LVH with an EF of 30%. 

Here is cine SSFP MR images showing severe primarily concentric LVH, although there's a little bit of assymetry, which can be seen in this disease that's not HCM. But there's severe concentric LVH and severe global LV dysfunction with an EF around 30%. 

And these are his late gadolinium enhanced images, which are quite different from any of the other LGE images I've shown you to this point. The LGE is diffuse. It's not particularly bright. In fact, it's very hard, in these patients, to null the normal myocardium. 

About twice per year, I'll get a call from the imaging fellow at the scanner saying the scanner is broken. It's not working. I can't null the normal myocardium. I said, we have the diagnosis. 

What's the diagnosis here? Amyloid. Yeah. This is amyloidosis. So there's this classic pattern of diffuse, transmural late gadolinium enhancement, inability to null the normal myocardium like you do in all other patients. 

There's also this pattern in some patients of diffuse subendocardial rather than transmural LGE that's indicative of amyloid. No need to biopsy these patients in general, because the accuracy of a CMR is quite high against endomyocardial biopsy. 

Now remember there are two general types of amyloid-- AL amyloid, which is light chain amyloid, and ATTR, or transthyretin amyloid, which is generally a disease of elderly patients. It can be hereditary. There's a hereditary form and a wild type form. 

In terms of the pattern of LGE, there is some overlap. But in general, ATTR patients tend to have more transmural LGE compared to subendocardial and absence of LGE as compared to AL patients who are pretty evenly spread among the three. But based on the MR pattern alone, one can't differentiate the amyloid type. And we'll talk more about how we do that going forward. 

So LGE and amyloid, I should make the distinction. Compared to the other diagnoses that we've talked about, it's not scar in amyloid. It's increased interstitial space. It's from the beta pleated sheets that are deposited in the myocardium increasing the interstitial space. Gadolinium gets caught in there, hence the diffuse LGE. 

So in this, the one meta-analysis I think I've shown you that's not ours, this is from seven studies, 425 patients, two-year follow-up. 73% of the patients had LGE, which was associated with a fivefold increase in mortality. So the presence of LGE is a marker of death in [INAUDIBLE]. 

And the pattern of LGE is also associated with prognosis. This is data from Mariana Fontana at the Heart Hospital in London. They have the largest population of amyloid in the world. 

Absence of LGE, a better prognosis. Subendocardial LGE, an intermediate prognosis. Transmural LGE, the worst prognosis. Now remember, transmural LGE is generally associated with ATTR subtype, which also has a worse prognosis than AL. 

The other concept that you should be aware of that's very useful in many of these patients-- so amyloid is associated with chronic kidney disease. If the GFR is less than 30, we can't give gadolinium. We can't look for LGE. But turns out that native T1 is often, as I mentioned briefly earlier, is often abnormal in amyloid. 

In fact, in definitive cardiac amyloid, the T1 in this study from the Oxford Group, normal T1 was 950 at 1.5 Tesla. The patients with definitive biopsy-proven amyloid has a T1 of 1,150. So it's one of the, again, the three diseases that cause a T1, native T1, 200 milliseconds greater than normal. So we can diagnose amyloid in the right clinical setting without gadolinium using native T1 mapping. 

Now to differentiate AL and ATTR amyloid, remember the pattern of LGE doesn't necessarily work as well as we would like. So there's a SPECT technique called pyrophosphate scanning. It was used in the 1970s and 1980s to identify myocardial infarction. It turns out it's useful for identifying ATTR amyloid. 

This is an example of a pyrophosphate scan from our institution, lent to me by Jamie Bourque. You can see very high uptake over the heart of pyrophosphate and no uptake in the contralateral side. And you measure the ratio of left-sided chest uptake to right-sided chest uptake, more than 1.5 being abnormal. And this patient was 1.65, diagnostic of ATTR amyloid. 

In ATTR, in this study of over 1,200 patients with almost 900 confirmed cases of ATTR, myocardial pyrophosphate uptake was 99% sensitive, 86% specific for ATTR. The few false positives were mostly due to AL amyloid. Now, for AL amyloid, you're also going to look for light chains, do an [INAUDIBLE]. 

So that's the other test you're going to do, blood test. But you're also going to do a pyrophosphate scan to look for ATTR. And it turns out that the heart to contralateral ratio is prognostic, that if the ratio is greater than 1.6, the outcome is worse than if it's less than 1.6. 

Why is it important to identify ATTR amyloid? Well, it turns out, for the first time in May of this year, the FDA approved the first medical therapy for ATTR amyloid called tafamidis. It's a TTR stabilizer. It actually pulls the amyloid out of the myocardium and has been associated with improvement in survival. 

So for the first time, we have a drug that is effective in ATTR amyloid. And this is why it's going to be so important to not only make the diagnosis of amyloid, but to make the right diagnosis of the type of amyloid, ATTR versus AL. 

So in summary, I hope I've shared with you during this grand rounds that advanced cardiac imaging has become the non-invasive myocardial biopsy. We don't need biotomes anymore, except for perhaps post-transplant patients. We can use cardiac MR, PET, and SPECT to make the diagnosis and assess prognosis and impact therapeutic decision-making in nonischemic dilated cardiomyopathy, myocarditis, HCM, sarcoid, and amyloid. And with that, I'll end. And thank you for--