21 on FDG PET imaging and infective endocarditis. I'm Fabien Yachil and I'm working at the European Hospital Georges-Pompidou in Paris in France. Here is my disclosure slide. So the overall objective of this lecture is to provide you the ability to read and interpret cardiac FDG PET acquisition in patients referred for suspicion of endocarditis. And the learning objective will be to describe the characteristics of normal and abnormal FDG PET uptake on PET acquisition in patients with a suspicion of endocarditis, to identify the characteristic of false positive study, and to explain the components of an appropriate report for FDG PET acquisition in patients with a suspicion of endocarditis. I will first start with discussing prosthetic valve endocarditis, native valve endocarditis, and then we'll discuss infection of CID. So just a little bit of background. Why is this imaging important? It's because an increasing number of patients are being implanted with prosthetic material or cardiac electronic devices. And these patients are often fragile and they tend to have a higher prevalence of infectious syndromes or fever of non-origin because of their age or because of their diseases. And an important clinical question is when you have a sepsis in this patient is whether the device itself is infected, or if not, if it has been infected secondary to the sepsis, which has another origin. This is a quite difficult clinical question in these patients. Echocardiography and blood culture play a key role in the diagnosis of infection of implanted cardiac material, but the infection of this material can now be specific on the morphological image. So in patients with a suspicion of endocarditis, we use a probability score, the modified criteria that will help you to classify patients into a definite possible or rejected endocarditis. This is a probability score, but to move from a probability score to clinical decision can be also sometimes difficult. For example, in patients with a suspicion of infection of cardiac prosthetic material, you will have to discuss if you initiate long-term antibiotic therapy. If you have infected material, it can be up to six weeks or even three months, or even to discuss if you should refer the patient for redo cardiac surgery. Or in patients with a suspicion of infection of CID, same, you will have to discuss extraction of the device if you have a high probability of infection. In some cases, at least give a long-term antibiotic therapy. And then if you extract the device, there's questions when you can reimplant the device if there's an infection. So cardiac and whole-body CT angiography can help find these endocarditis around the valves. But FDG-BET-CT imaging allows also for detection of inflammatory cells with a high sensitivity. We will discuss what could be the role of FDG-BET-CT imaging in this context of diagnosis of infected endocarditis. So just a little bit of background on FDG-BET. I think it's in the rules, but to a reminder is that FDG is a radio-labeled sugar, and it's taken up by the good receptors. In fact, activated inflammatory cells, they have a high consumption of sugar, so they will have a high uptake of FDG. When you look at the heart, it's important to limit, prevent the uptake of FDG and the normal myocardium. And this can be done now. It has been shown that you can prepare patients at least 12 up to 24 hours before the imaging, if possible, by a so-called low-carb, high-fat diet, where the patient shouldn't have any sugar and you can eat only proteins and fat before the FDG injection, and at least six hours, and if possible, 12 hours. And these are the best conditions to limit the physiological uptake of FDG in the myocardium. And this will provide you better images with higher image contrast. Usually when you look at infection, you perform your imaging 60 minutes after the FDG injection, and you will do a whole body acquisition. And now with the new devices, you can do this in 10, 15 minutes, up to 30 minutes. So what's the literature regarding the role of FDG PET and the infection of prosthetic valves? So there was a first paper by a group of Marseilles in France with 72 patients that were referred for suspicion of prosthetic valve endocarditis at least one month after the surgery, and they found FDG PET to have a high sensitivity and relatively good specificity for the detection of infection. More important, the use of FDG PET as a major criteria changed the classification of patients with a possible prosthetic valve endocarditis from in about 24% of the patients into confirmed endocarditis. And also, if you add FDG PET, the presence of FDG uptake in the valves as a major criteria in the modified criteria, they improve the sensitivity of the criteria from 70% to 97%. You have on this image an example of an infection of an arctic tube and going down to the arctic valve where you can see the high FDG signal surrounding the prosthesis, which is due to supportive tissue. And here you have a shim showing you how FDG PET is changing the classification of patients. In most cases, this is a typical classification that you have in patients with endocarditis, usually 50% to 60% of the patients with a possible prosthetic valve endocarditis. And if you add FDG PET to this patient as a major criteria, you will change the classification of 20% to 30% from possible to definite because of the high sensitivity of this imaging. Another important feature that you have with FDG PET imaging, it's a whole body imaging with a high sensitivity for the detection of both infection and also sometimes tumor. In the case of infective endocarditis, you have to use this imaging to identify sometimes the origin of infection. You have an example of a colic cancer, but you can also find some infection of the teeth. You can also identify septic embols with a high sensitivity. Here you have an example of a spondylolisthesis. And you can also identify a mycotic aneurysm in the legs. What you cannot really analyze is a brain because you have a high FDG uptake in the brain. And so you cannot detect very well the septic embols or mycotic aneurysms. So this is a limitation of FDG PET that you have to know. And usually you have to perform a brain MRI to look at septic embols or mycotic aneurysms in these patients. An important also value of FDG PET is that in patients with a sepsis, it can help you also to identify other causes than endocarditis of sepsis with abscesses or other causes of this of this positive blood culture. But with this, it is really high in prosthetic valve endocarditis of FDG PET, but in patients with native valve endocarditis, the sensitivity is a bit lower. This can be explained by the fact that the signal that you detect on PET with the FDG is mostly due to separative tissues. And so you have abscesses and you have quite often abscesses around prosthetic valves. You will have a high signal, but the signal is much low and you have only vegetation because they are less activated inflammatory cells than vegetation. They are smaller and mobile and so often the signal is low in this kind of lesions. So this explains why the sensitivity is often low in active valve endocarditis with FDG PET. Things to know is the causes of false positive and false negatives. What you have to do when you look at FDG PET imaging is that you look first at the intensity of the uptake. Usually you quantify this with SUVs. Second, you have to look at the attenuation correction because with prosthetic material, it can have some artifacts overcorrection due to the prosthetic material. And so you have to look at the images without attenuation correction to make sure that it's not an artifact. And the fact that you have a persistence of the signal on the non-attenuation corrected images is in favor of a true signal and possibly of infection. Another feature that you have to look on the PET is the aspect of the signal, whether it's heterogeneous or homogeneous. The presence of heterogeneous signal around the valves is in favor of infection. In particular, if you have on CTA or an echo, some kind of thickening in the same areas is very suspicious of infection. There are some causes of mistakes when you acquisition that you have to know. One of the most difficult ones is probably the use of biological glue because you have often some really chronic, relatively high inflammation due to this biological glue. It can give you really an aspect of infection, even though it's mostly inflammation. And then you have some also can have some uptake in the myocardium next to the prosthesis or in the vessel wall due to vasculitis that can give you also some false positive aspects on the PET imaging. And there are also some cause of false negative that can be due to certain germs, where the intensity of the signal is sometimes lower. And the other point is that if you wait, the longer you wait after the initiation of the antibiotic therapy, the lower the signal. So if it's possible, it's better to start to try to image this patient as soon as possible after the initial initiation of the therapy to limit this decrease in signal in patients. And of course, like I told you, if you tell me a small patient and the signal is low, it can be also a cause of false negative. So I would just illustrate what I just told you with a clinical case. It's about a 60-year-old patient that presented persistent fever. She had underwent an artic valve replacement for symptomatic arctic valve stenosis for four years earlier. And two weeks earlier before admission, she had a flu that was complicated by bronchopneumonia and was treated in the antibiotics. But the fever persisted at the end of the antibiotic therapy, and she started complaining of back pain in the night. At clinical examination, she showed moderate fever, had a normal cardiac excretation, had no signs, peripheral signs of endocarditis. White blood cell count was 10,000, and the CFP was at 45, and the repeated blood culture were all negative. Here are the images of the transversal visual echocardiography in this patient. And you can see her prosthesis, biological prosthesis. And here you can see next to the prosthesis some kind of thickening next to the prosthesis. And in this case, it's difficult to interpret this aspect. It can be hematoma, some fibrotic tissue, or some superlative tissue. There were no other abnormality or abnormal findings on this transversal echocardiography. So if we look now based on the clinical, bacteriological, and echocardiography features in this patient, how would you classify, how would you reclassify this patient? So you have here the list of the modified eukaryoteria. And so she had some mass next to the valve supporting structure. So it could be considered as one major criteria. And the bacteriology was negative. And she just had the, was implanted with a biological cardiac valve. And this can be considered predisposing heart condition. So this is a minor criteria. And so you collect all these different criteria. And then based on this criteria, the patient, as you can see, had one major, one minor, minor criteria, and then was classified as having a possible infective endocarditis. So the question in this patient is, should you treat this patient for six months, six weeks, sorry, because you have a, you think there's a high probability of infection of the valve or just careful monitoring of the evolution of the lesions in the valves. In this case, we perform FGT PET imaging to try to evaluate whether there's a truly an infection in this patient. And you can see we do, we perform whole body PET imaging. And here you have the PET with attenuation correction, the CT and the PET CT, the fusion of both. And you can see there's a rather diffuse high signal, the SUV, max SUV was at 6.6, so this is relatively high, higher than five. And if you look at the images without attenuation correction, you can see the signal persisted on the non-corrected and attenuation corrected images. And on whole body PET imaging, there were no signs of symptoms. So what you have to do when you look at these images and then write also in your report is you have to define where the FDG signal is. So here we have a signal that's in the region of the optic prosthetic valve or magical prosthetic valve. You have to give the intensity of that FDG signal, so it was a 6.4 high, 5.0. We found that we've seen on the non-attenuation corrected PET images that the signal persists on these images. And I've described the aspect of the signal as diffuse rather than heterogeneous. And I've done also this analysis of the whole body PET CT acquisition to look for septic embols, portal of entry or alternate diagnosis for the sepsis. And so the report, what I will write is due to the high signal in the valve, I would just describe that the aspect and intensity of the FDG signal in the optic valve region with no septic embols. But you have also to be careful and just say also that the results of that FDG PET imaging should be integrated to the clinical history of the patient and the results of the co-cardiography and bacteriology. In this patient, what we've done is we perform the white blood cells spect imaging. This is your regulable white blood cells and this is highly specific of infection and can be quite useful if there's a question whether it's inflammation or infection like in this case where there is a diffuse signal, really high. And this patient quite interestingly, there was in fact a high accumulation of radiolabel white blood cells in the same area of the FDG signal. And this is really highly in favor of an infection. And this patient was referred for cardiac surgery and the presence of active infection of the optic prosthesis was confirmed intraoperatively. So now let's move to the diagnostic of CID infection. So the CID, you can have an infection of the pocket or the lid. For the pocket, what has been shown by the group of Bordeaux in France is that you can make the difference between infection and post-implantation inflammation. You do this measurements of the maximal counts in the pocket and you do a ratio with the signal in the lungs on the non-attenuation corrected images. And as you can see on this graph, the infected patient, most of them had a relatively high ratio, higher than 2.5, whereas the patients with no infection didn't have this high signal. There's still some signal due to the acute inflammation after the implantation of the pocket, but it's most cases much lower than when you have a true infection. They found there was a meta-analysis performed based on the different articles that had been published with FDG PET imaging for the detection of pocket infection. And they found overall a high sensitivity of 93% and full specificity of 98%, which are really good diagnostic performance for the detection of infection. What about the lid? I think the lid is a bit more challenging because, of course, the structure is smaller in the background, but it's most cases around the lid. In some cases, it can be very easy, like in this example that I show you where you can see a high FDG signal along the lid. This is clearly in favor of infection. And in addition, what you have to look when you look for an infection of the lid is also at septic embols. So here, the example of a septic embol in this patient, which appears as a really focal FDG signal in the lungs or multiple focal FDG signals in the lungs. And this is really highly in favor of an infection of your lid. In some cases, it can be more difficult. The signal can be weaker. And what I advise you is some places, the signal of the blood is lower, as you have here an example. And sometimes you can also then see better the low contrast of the FDG uptake along the lid. And you have also to analyze then the non-attenuation corrected images. And if there's a persistence of the signal on the non-attenuation corrected images, this is also in favor of an infection. But in some cases, it can be quite difficult, I have to tell you, and you have to be very careful. And that the sensitivity, as has been shown in this meta-analysis of FDG PET for only lid infections, a bit lower than for pocket infection. As you can see, 65%, even though the specificity is quite high, 88%. So if you have a negative FDG PET study, and partly if the patient has already been treated for a few days or weeks by antibiotics, you have to be careful and you cannot exclude informally that there's not an infection of the lid. So I will illustrate these two points by this clinical case. It's a 55-year-old man that was admitted for suspicion of CID infection. He had a history of ischemic dilated cardiomyopathy with a severe left ventricular dysfunction. He had been implanted with a CRT-ICD device four years earlier in primary prevention. And there was a discussion of his corruption on the cardiac transplantation list because of the poor evolution of his left ventricular ejection fraction. And now presented fever, and there was a suspicion of the CID infection. Clinical examination confirmed the present 38.5 degrees Celsius fever. Cardiac excretation was normal. There were no congenital signs of infection in the pocket, no peripheral signs of endocarditis. There was a slight increase of CRP at 34, but there were three blood culture positive for staphylococcal arrhythmia, and it may have been medicine-sensitive. Here is the results. There are some few images of echocardiography. You can see the severe left ventricular dysfunction. And on the leads, there were no mobile mass detectable on transesophageal and transthoracic echocardiography. In this patient, we performed FDG PET imaging to look for an infection of the CRP. And what you can see is quite a slight, small FDG uptake here in the right atrium next to the lead, which was in favor of infection. And another point was that he had a septic. You can see a septic embole with this focal FDG uptake in the lungs. And we performed this late acquisition PET images, and there was a signal along the ventricle lead here that persisted on the non-attenuation-corrected images. So this is highly in favor of infection. The patient was then discussed whether we should extract the device. But it was very, very fragile, and there was a risk of, there was an old device, there was a risk of tamponade, and then there were two anestheses, anti-diarrhea therapy with sasiciline and angiotensin. But this was not successful, because one would like, on the transfer visual co-choreography, you can see there's a development of a mobile mast was not present a week earlier, so it means that we were not controlled fortunately. And the patient was registered in a super emergency list for cardiac transplantation and transplanted two weeks later with removal of the CID device. One year later, he's doing well. He didn't present any recurrent infection. So just to summarize what's important to include in the report when you do look at for an inpatient with suspicion of CID, you have to really analyze carefully the pocket and also all the path along the leads. You have to just quantify the signal similar to a prosthetic valve, if it's a high signal, if the max SUV is higher than 5.0, you can use a patient corrector around the material, that there's a persistent of the signals. It's also in favor of an infection, not an artifact. And importantly, if you have a CID, you have to really analyze carefully the lungs to look if there's any pre-receptive embols. And similar for prosthetic valve, you have to also look for alternate diagnostics for the sepsis, the positive blood culture or the inflammatory syndrome on the whole body acquisition. So the conclusion of this, the report of the imaging was that the presence of signal along the leads of the CID in association with primary septic embols is highly suspicious of CID infection. So just to summarize, take on messages about FDG PET imaging in patient with suspicion of endocarditis, in patient with prosthetic valve, FDG PET imaging provides a high sensitivity for the detection of perivalvular infection. And it's very useful, particularly if you have doubtful or negative transdermal echocardiography with a high probability of infective endocarditis. When you look at the imaging, you have images, you have to really describe well the location of your FDG signal in the heart and next to the prosthesis. Measure the intensity of the signal, typically with maximal SUV. Describe the pattern of the FDG signal, whether it's homogeneous or focal heterogeneous. And look if you have them to other morphological imaging, the results of other morphological imaging like echocardiography or CTA to look if there's any lesions that could match the signal that you see on FDG PET images. You have to perform whole body FDG PET imaging and analyze them carefully to search for septic embols, mycotic aneurysms, and the possible portal of entry, as well as alternate infective or inflammatory causes of the infectious inflammatory syndrome. You have to be aware of the causes of false positive and FDG PET imaging in patients with cardiac match. And therefore, you have to really integrate the analysis of FDG PET imaging with the clinical, biological, and other imaging results in the endocarditis team. For CID infection, it's quite similar. You have to look carefully at the signal in the pocket and along the leads. And you have to describe the location of the signal, and if there is any signal along the leads, and the intensity of the FDG signal. You have to look carefully at the whole body PET imaging, but for CID, more carefully to the embols. And you have to also search for portal of entry and look at alternate infective and inflammatory cause that could explain the inflammatory infective syndrome. You have to be aware that the sensitivity of FDG PET imaging for only lead infection is lower. And so, if you perform the FDG PET imaging after a few days or weeks after the initiation of antibiotics, you have to try to do as early as possible the FDG PET imaging for these cases. And be aware that you might miss some case of lead infection with FDG PET imaging. And similar to prosthetic valves, you have to integrate the analysis of FDG PET imaging with clinical, biological, and the results of other imaging, ideally in a multidisciplinary endocarditis team. And with this, I would like to thank you for your attention, and feel free to contact me if you have any questions regarding this topic.

FDG PET imaging

infective endocarditis

prosthetic valve endocarditis

native valve endocarditis

infection

cardiac implantable devices

CID