Welcome. My name is Sharmila Dorbala. I'm the director of nuclear cardiology at Brigham and Women's Hospital in Boston and professor of radiology at Harvard Medical School, Boston. I'm pleased to present to you about how to perform cardiac PEC with a module focused on how to perform N13 ammonia perfusion studies. These are my disclosures. So what are our learning objectives today? Our learning objectives are to describe how to acquire and reconstruct N13 ammonia cardiac PEC perfusion studies, to describe how to prepare patients for this study, to learn procedures to optimize the radiotracer injection, and finally to share with you a few illustrative cases that highlight some of these points. Acquisition, reconstruction, display, quality control, interpretation, all of these are addressed in the ASNIC PEC guidelines. The camera setup and QC, dosing protocols, all of the details are provided in the guidelines that I suggest you refer to. The patient preparation for N13 ammonia PEC is similar to any other rest-stress myocardial perfusion imaging study. Patients are kept NPO for six hours and caffeine is withheld for 12 hours. Curefeline containing medications are withheld for at least 48 hours. The camera setup and quality control may vary from scanner to scanner. Therefore, we suggest that you follow manufacturer recommendations for these procedures. The camera setup and daily quality control of the PEC and CT equipment also per manufacturer recommendations. The setup of automatic injector for N13 ammonia is extremely important for quality control of the myocardial blood flow imaging studies, both within patient from rest to stress and between patients and for repeat studies as well. These are the suggested QC procedures as per the ASNIC SNMMI guidelines. Please refer to the document listed here for further details on this particular document. So what are the steps of a cardiac PEC acquisition? The three common steps are a scout or a localized scan, a transmission scan, which is used to measure attenuation and to correct for it, and an emission scan, which is our scan of interest. In hybrid PET-CT scanners, a scout scan is used to localize the heart. It is a CT-based image, looks like an x-ray. Basically, we identify the carinae, and if you start your PET acquisition at this point, for most standard PET scanners, the detector width is adequate to cover the cardiac sections on this slide. The transmission scan provides a specific density map of the thorax to measure and correct for photon attenuation. This is a measured attenuation, which can be obtained either before or after the emission scans. However, the key point is to remember to keep the table position constant for the transmission and emission scans. You can use either a radionuclide source, rotating rod source for attenuation correction, or a CT-based transmission map. It's critical to review both the registration of the emission and transmission images, as well as the quality of the transmission image. Most transmission scans take more than three to four minutes with radionuclide, and about 30 seconds with CT-based imaging. Free tidal breathing is recommended for these procedures. Here's an example of how a radionuclide source image would look like on your transmission scan. And on the right side, you'll see a CT-based image, which is much sharper compared to the emission image. This CT image is then transformed by application of special filters to develop this fuzzy new map and to match the resolution of the emission image. So these are the two types of the transmission scans. This table lists the various protocols and the steps for the transmission procedures. The slice collimation should be approximately similar to the slice thickness of the PET scan. The ECG gating is not recommended. Some investigators have tried breath-hold CT. They have tried gated CT, but all of these are not recommended because the image quality is no better. The tube potential can be adjusted. Some manufacturers provide for automatic adjustment of the tube potential and tube current, actually, more importantly, based on the patient BMI. So we would try to use the lowest tube current possible for any given patient. 10 to 20 ma is typically used. Breathing instructions. Shallow free breathing is the most recommended protocol. And the reconstruction slice thickness of the transmission CT should again match the reconstruction thickness of the emission image, 4 to 5 millimeters. When CT-based transmission maps are used, as shown here, metallic objects can cause beam-hardening artifacts on the CT. Because this CT is used to develop the transmission map, which is used to correct for the emission image, sometimes these artifacts may translate as hot spots on your emission image. Typically for cardiac perfusion imaging, this is not a problem because you see either ICD or pacemaker leads, and ICD leads cause more beam-hardening than pacemaker. And the typical location that they impact the image is this basal infraceptal area. Mechanical prosthetic valves in the mitral and aortic position typically are above the level of the LV myocardium and do not appear to interfere with our emission image. Newer scanners now have a special protocol called metal artifact reduction. And metal artifact reduction can be applied to overcome these artifacts or these hot spots. In addition, if you use a dedicated PET scanner with radionuclide line source, this becomes less of an issue or actually a non-issue because we don't see beam-hardening from metallic devices. This slide just summarizes some of the points that we just discussed. The key point here is when you're looking at hot spot imaging, such as FPG PET for sarcoidosis, in those cases, these focal hot spots in certain areas can become problematic. So we recommend that you review both attenuation-corrected and non-attenuation-corrected FPG PET for hot spot imaging. For perfusion imaging, we recommend for you to use only attenuation-corrected images. Therefore, it is best to use either a metal artifact reduction or typically doesn't cause a problem as we discussed. What about the PET transmission scan? You can directly measure patient attenuation using a rotating line source. It takes almost a few minutes to obtain this. Free tidal breathing is recommended. Important to note that as opposed to a CT transmission map, which is always extremely high quality, a PET or radionuclide transmission scan is not recommended. A PET or radionuclide transmission map, the quality may depend on the age of the radionuclide line source. If the line source is old, then in those cases, what you might see is the counts in the transmission map may be low, and you may have to extend the duration of the transmission map, making the protocol a little longer. Otherwise, both radionuclide and CT-based transmission map appear to have equivalent diagnostic value. This table illustrates the acquisition parameters for N13 ammonia emission study. Remember that the recommended mode of stress test for cardiac PET imaging is pharmacologic stress. The typical dose for N13 ammonia is 10 to 20 millicuries, 10 millicuries for a 3D scanner, which is most modern PET scanners, and maybe up to 20 millicuries for some of the older PET scanners. It's injected over 20 to 30 seconds using an automatic injector. The images are reconstructed with a one and a half to three minute delay after the start of infusion, after the end of infusion, sorry. And remember that the camera must be started, the PET image acquisition must start first, and then the ammonia is injected. And we'll talk about why this is important for dynamic imaging. For patient positioning, we use a scout scan. As we discussed, this could be either a CT-based scout for hybrid scanners, or it could be one to two millicuries of a small bolus of ammonia injection to get a radionuclide south scan. TCG gating is required in all of these cases. And most modern scanners allow image acquisition in a list mode so that these list file can then be unlisted into a static image, a gated image, and a dynamic image. The typical scan duration for N13 ammonia is about 10 to 15 minutes. For 3D scanners, you can scan in as little as 10 minutes, and for 2D, you may go up to 15 minutes. Again, the attenuation correction or the transmission scan can be obtained either before or immediately after the emission scan. The reconstruction method typically is OSCM or iterative reconstruction. Reconstruction filters, again, have to be optimized to your scanner. You can start at a manufacturer-recommended level, but then adjust it based on your interpretation. The reconstructed pixel size is about two to three millimeters for cardiac PET. Here's a typical protocol. A scout obtained with CT, 10 seconds, a 30-second free breathing chest CT, non-contrast, non-gated low-dose for the transmission map. Then we start the PET scanner and inject ammonia and take a 10 to 20-minute emission image. A stress test is performed, typically vasodilator stress. We prefer to use RecurDynason because it's easy to use, and a single IV is sufficient. Dipyridamol can also be used with a single IV. But remember, if you're using adenosine as a vasodilator, you may need to place two IV lines, one for the adenosine infusion and the other for the radiotracer injection. The same steps are repeated, and a calcium score scan can be obtained in patients who have no previous documented coronary artery disease, stents, or bypass surgery. In our lab, we routinely perform a CT scan with end-diastolic CT imaging, which is gated, non-contrast, high-dose, following standard calcium score protocols in all patients undergoing cardiac PET, as long as they have no previous history of documented ischemic heart disease. One point to remember about N13 ammonia, because it has a 10-minute half-life, your rest and stress injections have to be separated by approximately 50 minutes if you choose to use equal dose. If you use 10 millicuries for rest and 10 for stress, then you wait 50 minutes to allow for background decay of the resting activity before you can get your stress injection. The second option, which is what we use in our lab right now, is to use a low-dose, high-dose protocol. This is very similar to what we do with system-A-B imaging single day. So we do a low-dose, 5-millicury N13 ammonia PET-CT study, because we are using latest generation digital PET-CT scanner. We get excellent images, even with 5 millicuries at 10 minutes. We perform a stress test right after, then repeat the emission imaging immediately back-to-back. So a few words about the automatic injection of N13 ammonia. In order to get high-quality images, a good-sized, well-functioning IV line without kinks is important. The radiotracer needs to be injected as a tight bolus to obtain a good arterial input function. The rate of injection needs to be constant for rest and stress for all patients in the lab and for repeat patients within the lab. What we do for an automatic injection is listed here. We set up a 45-ml saline bag. We inject N13 ammonia into the line very close to the patient's arm and then run the saline at 2 ml per second, pushing the N13 ammonia at this rate into the patient. These are just the four commonly used pharmacologic stress protocols with N13 ammonia. Dipyridamol, as you all know, is a weight-based infusion. The infusion is administered over 4 minutes. We start the PET scanner at 7 minutes, and 5 seconds later, we start the N13 ammonia infusion. With Ragadenosone, it's infused over 10 seconds. It's a slow injection over 10 seconds, followed by a 5 cc saline. We start the PET scanner 60 seconds after the start of Ragadenosone infusion. At 60 seconds, we start the PET scanner. 5 seconds later, we infuse N13 ammonia. Adenosine is weight-based, 140 micrograms per kg per minute, infused over 4 to 6 minutes. At the midpoint, you start the PET scanner and image acquisition. 5 seconds later, you infuse the ammonia. What about exercise? Because of the 10-minute half-life of ammonia, you are able to exercise patients. So we perform a standard symptom-limited treadmill exercise inside the PET scanner room. The patient exercises maximally, at which point the radio tracer is injected. So the tracer is injected while the patient is on the scanner. We continue exercise for an additional minute, then we stop exercise. We wait for approximately 3 minutes from termination of the treadmill test to the start of the PET scanner acquisition. This is very important to minimize motion artifacts. Remember, some of these patients may exercise a lot, and their breathing patterns may be different, and they may be not very comfortable to lie down flat immediately after. So in order to minimize patient motion, we recommend that you wait about 2 to 3 minutes before starting the PET acquisition. Another point to remember here is that myocardial blood flow cannot be quantified with exercise ammonia PET perfusion imaging. So although a major advantage of the 10-minute half-life of ammonia is the ability to offer exercise PET, remember, the tracer is happening on the treadmill. The tracer injection is on the treadmill. So dynamic images cannot be acquired starting with the injection of the radio tracer. So you don't have an input function, an image-derived input function. Therefore, blood flow quantitation can be very challenging and is not clinically used with exercise stress. In patients where dynamic stress is required, the butamine infusion using standard protocols can be performed with N13 ammonia. And in that case, you are able to obtain myocardial blood flow with the butamine N13 ammonia imaging. So the next concept I want to introduce you was something called the pre-scan delay. So this is a time activity curve. The activity or counts are on the y-axis, and the time or the image frames are shown on the x-axis. The purple color represents the time activity curve in the right ventricle, the red in the left ventricle, the green in the left ventricular myocardium. So basically what you're seeing here is we place the region of interest in the right ventricular cavity, and you see that the PET scanner is started before ammonia is injected. So you have a blank period where there's no activity in the right ventricle. Then you see activity peaking in the right ventricle as the ammonia enters the right ventricle. And then it leaves the right ventricle and goes into the lungs. You'll see then the tracer activity peaking in the left ventricle, and then the counts going down. And finally, you'll see the activity entering the myocardium. Our signal of interest for perfusion imaging is the myocardial signal. So for static and gated images, we need images in this phase of the scan where myocardial counts are much higher than the blood pool counts. So your signal to contrast, signal to noise is much better. So the initial frames from the start of the scan until this point where myocardial counts go above the blood pool counts, this part is called the pre-scan delay. And this pre-scan delay is dependent on a number of different factors. So what are the determinants of pre-scan delay? The radio tracer half-life, arm to heart circulation time. So in patients with low left ventricular injection fraction, they have a long circulation time, and this can be prolonged. The pre-scan delay is about 90 to 180 seconds typically, and this is what works well for N13 ammonia imaging. The same time activity curves I'm just sharing with you to show you how we can quantify myocardial blood flow with cardiac PET. So you acquire images in a dynamic mode, start the scanner before injection of the radio tracer, then you obtain a rapid sequence of multiple imaging frames. So this is multi-frame imaging over time, which is what is dynamic image acquisition. Using kinetic analysis and one to three compartmental models, adjusting for radioactive extraction and decay and correcting for partial volume-related underestimation of myocardial concentration of ammonia, we can quantify myocardial blood flow. You estimate it in ML per gram tissue per minute at rest. You can do that at peak stress, and the ratio of the stress to rest blood flow is used to calculate the myocardial flow reserve. This dynamic image acquisition can also be used to optimize the pre-scan delay. If you see a lot of blood pool activity on the ammonia images, you can go back to the time activity curves and extend the pre-scan delay and reconstruct the images to improve the signal-to-noise of the images. So how do you reconstruct the images? The transmission and the emission images are reconstructed. Again, use manufacturer-recommended protocols. This is what is listed in the ASNIC PET guidelines, just for your reference. How do you display ammonia images? So standard cardiac display using software. We perform quality control first, and then we review the images. So we look at the quality of the transmission and the emission images. We look at the overlay of the transmission and emission to understand appropriate registration of the images. Then we review motion or patient motion during the scan by looking at a sine loop of the dynamic image frames. So when you play the dynamic image frames in a sine loop format, the patient moved during the 10 or 20-minute scan acquisition, you can see the heart moving up and down. So these are the three quality control steps. So looking at the count density, looking at the registration of emission transmission, and looking for motion. Once this is all done, we review the stress and rest images, looking at the static images, gated images, pull-up plots. This is standard as with SPECT. Then we review the dynamic images to compute myocardial blood flow, and then the CT transmission images, both for coronary calcification and for ancillary findings. If a dedicated calcium score CT was obtained, that is also evaluated and calcium score is calculated. So these are the quality control steps that are used. And for technical considerations, remember to start the PET scanner before the injection of radiotracer. Remember to use an automatic injector. Make sure patients avoid caffeine for at least 12 hours before the scan. Vasodilator, as we discussed, is preferred, regadenosone. If using adenosine, use two IV lines. Remember that peak stress blood flow cannot be quantified with exercise stress and adequate quality control is critical. So with this summary of the technical aspects of acquisition and reconstruction, I wanted to then proceed to show some interpretation of cases, which will highlight some of these points. So we have a number of cases, and these are the key teaching points that we'll be addressing on the cases, showing you transmission emission registration, identification of patient motion, looking at some normal variants with ammonia, increased lung uptake, fixed lateral wall perfusion defect, then showing you cases to illustrate quality control of myocardial blood flow and flow reserve images. Then sharing with you a case of a normal perfusion with low myocardial flow reserve and how we interpret that and few challenging blood flow cases. So let's start with case one. So this is a 79-year-old woman with diabetes and hypertension, present with typical chest pain. She's on multiple medications, undergoes the rest, regadenosone, stress, ammonia. This was the dose of ammonia and regadenosone. Here are the perfusion images, stress images on the top, rest images on the bottom, polar plots here, stress, rest, reversibility, and this is the computer-derived score. What you'll notice immediately is that there is a perfusion defect in the anterolateral wall and the lateral wall of the ventricle, starting all the way from the base, the basal anterolateral anterior, mid-anterolateral anterior, and apical lateral, and that entire perfusion defect look reversible. Polar plots confirm the same and show reversibility in the anterior-anterolateral. So what do we do? We look at quality control and look for transmission-emission overlay. So this is the emission scan, this is the CT transmission scan and overlay, and clearly what you see is that the emission scan is not matched or aligned with the CT cardiaxial lobe. You can see that the lateral wall on the emission scan, the perfusion of the lateral wall, is overlying the lung field. Because of this, the lateral wall is under-corrected for attenuation because the attenuation coefficient of air is lower than the attenuation coefficient of air is lower than the attenuation coefficient of soft tissue. And this under correction potentially translates into a perfusion defect. So what do we do with this case? So when we see something like this, the scan should be not interpreted. You go back to the acquisition console, reconstruct the images again. So how do we do that? So most vendors and most scanners now have software where you can manually align the overlay of the transmission emission. And using that new registration of new alignment, you can develop a new MU-MAP, which is applied to the stress images in this case, and a new reconstruction is performed. So here's the new alignment, which is perfect. And then you can see the effect on the emission image. So using the new transmission map, that anterior interlateral defect is essentially completely normalized, no perfusion defect, normal stress, and normal rest perfusion. So why do you see misregistration of transmission and emission images? This we believe is typically caused by breathing changes from hyperemia during the stress emission scan. Remember, unlike SPECT images, where the vasodilator stress is performed, and then there's a 45 minute wait period before you acquire the images, with N13, the vasodilator is administered and the PET scan is started during maximal hyperemia. So while the patient's experiencing hyperemic changes, that can result in some breathing changes and changes in volume, and that can cause misregistrations. The most common location of this is interlateral and lateral walls. And since we often see this because of vasodilator, it's more common with stress imaging. Therefore, a lot of times these artifacts appear as reversible lateral wall perfusion defects. Very important to review the fused overlay of transmission and emission images to identify this artifact. The solution is to use software to correct alignment of the emission and transmission images and to use the new, new map to reconstruct the images. Next case, 59 year old woman with high blood pressure and a family history of hypertrophic cardiomyopathy. Presence with typical chest pain, dyspnea and palpitations. She's on the following medications. Again, rest, stress, N13 and ammonia PET. Here are the stress and rest images. What you'll notice at a first glance, when you look at the quality of the emission images is that the shape of the ventricle and the stress images is not round. It's an odd shape, irregular shape. And you'll also see that compared to the rest images, the walls are thicker, particularly in the lateral wall. So this should already be a hint to us that there's something wrong with this study. So what do we do? Look at the polar maps and it shows multiple regions of reversibility don't necessarily pertain to any given vascular territory. So we look at the overlay of the emission and transmission images and that appears to be excellent. One point here. So when you're checking for alignment, you need to look at multiple projections, short axis, four chamber, two chamber or axial sagittal coronal. But more importantly, something that I'm not showing you here is I go through the slices back and forth. So I look at the entire cardiac volume, making sure there's no misregistration, not just at this level of the short axis, but all the way from apex to base, all the way from inferior to anterior and septum to lateral. Then how do you check for motion? So we look at the dynamic imaging frames, which are played in a Cine loop format. So here's an example showing you the dynamic images playing in a Cine loop format, focus on the top, which is the stress. And you can see that the myocardial activity starts off at this level and gradually over time starts moving up. Okay. So this is significant patient motion during the stress images. Contrast that with the rest. You can see that the myocardial activity remains at this level throughout. Okay. So this case illustrates that there was significant patient motion and the cardiac position changed during the stress scan. We can also look at calcium score. In this case, there was no calcium score. So how do we deal with artifacts from patient motion during the image, emission image acquisition? So if you go back to the previous slide and we try to understand where the motion occurred, you'll notice that. So there's motion all the way here. You can see the myocardium is down here. And then in the later phases, there. So now it gets into the right position and then remains in that position. Okay. So what we do with motion correction is we go ahead and look at the imaging frames. And there's a way to delete all the images up to this point where there is no motion. And then we sum the last few frames to get an image without a motion artifact. So once you re-sum the images, you can then reconstruct the static and gated images using the re-summed images. So here's the reconstructed image, stress images with motion correction, stress on the top, rest on the bottom. And you can see now the ventricle is shaped similarly at stress and rest. The thickened lateral wall, we no longer see that artifact. Here's the stress images with motion on the left side. And this is without motion on the right side after we did motion correction. Look at the anterior wall here, this fuzziness, that's all disappeared. And all in all, this is a nice normal study with motion correction with no perfusion defects. So this is possible with N13 ammonia if images are acquired in a dynamic multi-frame imaging mode. Works well with high count images. So motion typically occurs early on during the stress images because of the hyperemic effects of the vasodilator stress. And if that is seen, it can impact blood flow quantitation. Motion correction is possible with dynamic images and can improve the quality of the blood flow images. Motion after the first three minutes may affect the relative static and gated images. In this case, it's again, as I showed you with the case example, reframing is possible and that is likely to improve the static and gated image quality. However, if there's patient motion that is complex and throughout the image acquisition, sometimes we may not be able to correct for this motion and image acquisition may have to be, the stress testing and the scan may have to be repeated at a later point. Just a small word about motion during rubidium 82 PET. This is a short image acquisition. It's about seven minutes. Again, if it happens typically during the early phase, one to three minutes after the stress test, that is when you see peak myocardial rubidium counts. So sometimes if you have motion extending longer into the rubidium scan, your motion corrected rubidium may become somewhat uninterpretable. With rubidium, if the motion occurs at a later phase, not in the early, but at a later phase, it tends to have lesser effect on image quality because at minute four and five or six, the counts of rubidium contributed to the image are much lower because rubidium decays rapidly with a 75 second half-life. So slight differences in how motion impacts image quality in ammonia compared to rubidium. So to summarize, motion may occur during emission, which is what we discussed, and it can also occur during the transmission scan. It is typically very challenging to detect unlike SPECT images, typically seen in patients who are non-compliant or with inadequate instruction. Sometimes it is seen with the side effects of vasodilator. As seen in the previous case, when it occurs, it can degrade image quality causing apparent reversible defects, sometimes may render the images uninterpretable. Motion may alter the shape of the ventricle and the static perfusion images. How do you identify it? As seen in the previous case, you look at the review of the CineLoop format dynamic image frames. How do you deal with motion? The solution, careful instruction to the patient to stay still for the duration of scan. There's no simple software solutions. We can correct for motion by deleting the image frames, which are limited by motion and resumming the image frames without motion. If motion is noted during the transmission scan, transmission and emission images may have to be repeated based on the imaging system. Key point, do not review non-attenuation corrected images for perfusion PET at this point. Repeat study may be necessary in worst case scenario when the images are not salvageable. Just a normal variant with N13 ammonia, increased lung uptake. Here are the resting images, perfusion, and you can see clearly intense lung activity. Much more pronounced on the rest images compared to the stress images. This is commonly seen with N13 ammonia, particularly in patients who are smokers, patients with heart failure, low ejection fraction, and sometimes patients with inflammatory lung disease. When we see lung uptake, it's typically more pronounced in the rest than on the stress images. No simple solutions for this artifact. It may impact lateral wall perfusion, sometimes using the so-called non-shifted images, which is the images that you're reviewing are images that have already gone through quality control where the technologists have aligned the CT and the emission images appropriately. If there's lung uptake, which is significant as in this case, sometimes looking at the computer-based alignment of the CT and the emission images may help. Another normal variant that's seen with N13 ammonia is a fixed basal lateral wall defect as seen in this case. Typically it impacts perfusion of the lateral wall at the base and it is fixed as shown on the polar maps. In this particular case, you also see a little double shadow, which can sometimes be seen in patients who have hadopsical septal motion, either from paced rhythm or from prior bypass surgery. So this may be a combination of a fixed lateral wall perfusion defect. And a normal variant with ammonia that is exaggerated potentially by paradoxical septal motion. The key here is it affects the basal lateral wall and wall motion, wall thickening are not really normal. The solution is if paradoxical septal motion is present, exaggerating the fixed lateral wall defect, you can choose to look at the end diastolic images which sometimes help with this particular artifact. The next case, 82 year old woman with non-CAD prior PCI to the circumflex, multiple risk factors and N13 ammonia regadenosone stress. So here are the stress rest images, adequate overlay. And here are the polar plots and you can see some apparent reversibility, mild reversibility in the inferior wall. So here are the blood flow images in this patient. You have stress on the top, rest on the bottom. You can see the region of interest that we place for the image derived input function. This is critical for blood flow quantitation because that is what tells the computer software program. What is the amount of radioactivity that's delivered through the myocardium out of which what is the amount of radioactivity that's retained by the myocardium? So it's a ratio of these two. Therefore, quantitation of myocardial blood flow is directly impacted by accuracy of this image derived input function. So here's playing the images in a Cine loop format. And you can see that this input function at stress has a peak activity of around 160 and at rest, it is about 80. So this is a low dose, high dose and 13 ammonia study. And here are the blood flow values, stress blood flow, LAD territory 2.95, CERV 2.98, RCA 2.41. And these are the corresponding rest values with flow reserves are super normal, 3.7, 3.7, 3.1, overall 3.5. These are very high flow reserves for an older adult patient as in this case. So basically the question here is, is this adequate? So how do you know if the input function is adequate? So here I'm showing you repeat reconstruction of this. And at this time I have moved the input function, which is more along the interventricular septum, sorry, more closer to the atrial wall. And I moved it to the more central portion where it captures the peak activity in the left atrium. And now you can see that the peak counts and the input function are nearly 220 here, significantly increased from the 160, all right? So once you increase your input function then your myocardial blood flow decreases and reflects more through myocardial blood flow in this particular patient, okay? So the flow reserve again is all about two, but look at the right coronary artery, the territory where we saw the mild ischemia, we could see that the flow reserve and the peak stress flow in the right coronary distribution was mildly decreased consistent with our visual interpretation of the images of mild inferior wall ischemia. So again, remember for the input function, once you place the region of interest, you move it around in the left atrium until you see this curve peak. So we need to pick the location of the input function where there is maximal activity in the left atrium that leads to a more realistic blood flow quantification. So I want to end with a couple, actually two or three challenging cases for blood flow quantification. So here's an excellent quality stress rest ammonia perfusion image, which looks completely normal. And here's the flow reserve values in this patient, very high peak stress flows, normal rest flows with high reserve flows, normal perfusion, normal flow. The interpretation is straightforward, completely normal, no epicardial, no microvascular dysfunction. What about this scenario where the perfusion is completely normal as in this case, but the blood flow values are diffusely reduced. They're reduced globally, they're also reduced regionally. And you can see that the flow reserve is all mildly reduced. So this is a challenging case. This is not uncommon, we see this in clinical practice and the differential diagnosis for this with a normal perfusion, low myocardial stress blood flow and myocardial flow reserve could indicate one of three things. The first thing could be, did this patient not respond to vasodilator? Was there caffeine on board? Is the patient non-responded to vasodilator? The second possibility is that this is a truly reduced peak stress flow indicating pathology in this patient. So if you exclude caffeine intake and you talk to the patient and they're saying that they didn't take caffeine, and in that case, then you say, this is not vasodilator non-response, the patient does not have caffeine on board. So this is true low myocardial blood flow values indicating microvascular dysfunction with normal perfusion and epicardial fornus. What about a case like this? So once you have done the quality control of your input function, you have checked the patient has not consumed coffee, then look at this case. There's a small lateral wall perfusion defect that looks completely irreversible. Transmission emission images were well-aligned, polar plots show a lateral wall ischemic defect. Here's the gated study showing nice normal ejection fraction, no regional wall emission abnormality. Look at the myocardial flow reserve here. The stress values, rest values are very high. Therefore, the reserve values are low. So here's an example of a case where peak stress blood flow is minimally reduced, more than 1.8 is normal, they're all nearly hovering around the normal range, but the reserve is significantly reduced because of high resting myocardial blood flow. So what is the cause for high resting myocardial blood flow? It could be either high blood pressure, tachycardia, or post-cardiac transplant state where these patients have high resting tachycardia. So the case that I showed you was a little bit of infralateral ischemia in a patient who had prior cardiac transplant, and this study was performed nearly 10 years after cardiac transplant. So again, important to not only look at the reserve values, but to understand what is causing this lower reserve values. Is it a reduction in stress flow? Is it a reduction or is it an increase in rest flow? In this case, it was a high resting flow. Next case is a 55-year-old male with premature coronary artery disease, had a prior infarct and is sent to the LAD and PDA. His angiogram showed patent stents and circumflex stenosis of 90% without a graft. So he underwent rest stress and 30-minute morning effect. And here are the images of stress and rest. So this patient shows a large and severe perfusion defect involving almost the entire anterior wall, the entire septum, the entire inferior wall, as well as the apex that was fixed. All the plots confirm the same. So what's the challenge in this case? The relative perfusion images are very challenging to interpret because you're only normally perfused myocardium with the lateral wall, highly unlikely that we can provoke ischemia on the relative perfusion images. So this is a real challenging case. So could blood flow help in this case? The answer is yes, but note here that the patient also has high lung uptake. And as we discussed earlier, lung uptake can be more pronounced at rest compared to stress. And that's what you're seeing here. And that is interfering with his blood flow quantification. So look at his LAD, CERC, and RCA. His LAD stress flow is significantly reduced. RCA is significantly reduced. But look at the circumflex flow. Both rest and stress flows are increased. And this is clearly artifactual because of added counts from the lung uptake. So this is a true challenging case where the relative effect was not helpful and absolute flow quantification was also very challenging because of the lung uptake. So with this, I believe maybe one of the last cases, a 65-year-old known coronary disease prior strain to the LAD. High blood pressure, diabetes, referred for dyspnea and hypertension, multiple medications, stress, stress, and ammonia. So here are the stress and rest images. Large reversible perfusion defect involving the entire anterior wall, apex, the entire septum. Okay? So basically the entire anterior wall, the entire septum, the apical four myocardial segments, two apex, completely reversible, confirmed on the polar plots. The polar plots also suggest that there may be some lateral wall islands of ischemia and there's transient ischemic dilation of the ventricle when you look at the images here. What about the blood flow? So here's the resting flow, here's the stress flow. And you can see the LAD shows almost absent vasodilator reserve, so 0.89 to 0.96. But the circumflex territory has substantial vasodilation and the RCA territory has intermediate vasodilation. So this is an example of a case where the patient likely has multi-vessel disease. We would predict that the LAD disease was very severe. The RCA disease was somewhat intermediate and that the circumflex is likely abnormal because the peak stress flow is less than 1.8, which is abnormal, but this is less severe than the LAD and RCA. So patient went on to have an angiogram. His RCA was occluded. LAD was showing 80% disease with a 1%, sorry, with a 99% disease in the diagonal. And this was fixed by revascularization. And his circumflex showed a 60% intermediate grade stenosis. So I want to now conclude by giving some summary of all these cases and of the image acquisition and reconstruction protocols. So image acquisition and processing for N13 ammonia cardiac PET perfusion study are well-described in the ASNIC-PET guidelines. Follow carefully the patient preparation instructions, which are very similar for both PET and for vasodilators, PET myocardial perfusion imaging. It's important to optimize the injection of N13 ammonia to get high quality quantitative myocardial blood flow values, which are reproducible within the patient from stress to rest and between patients and at repeat studies. Careful interpretation of the N13 ammonia perfusion imaging and flow quantitation, taking into account the various normal variants and artifacts is really important to improve the accuracy of the test. I now have a few references listed here for those of you who are getting into practice. The ASNIC imaging guidelines, SNMMI procedure standard for PET on your cardiology procedures, a lot of the material that I shared in this talk comes from this publication, which is available online on the ASNIC website. The other reference is the normal variants and pitfalls in cardiac PET CT. This was published recently in seminars in nuclear medicine and I've taken a number of key points from that publication as well. With that, I want to thank you for your attention, for attending this talk and for listening to me. Thank you.

N13 ammonia perfusion studies

cardiac PET imaging

patient preparation

camera setup

cardiac PET acquisition

artifacts in interpreting images

myocardial blood flow quantification

challenging cases

ASNIC-PET guidelines