Good morning. My name is Andrew Crean. I'm a cardiologist at the University of Ottawa Heart Institute. Now, my job today is to lead you through Module 12, which is looking at how we interpret and report PET myocardial perfusion studies. I'd also like to acknowledge Dr. Ali Peddarazadeh, who's one of our imaging fellows at the Institute, who helped me significantly in putting this presentation together for you. Myocardial perfusion imaging is one of the most commonly performed tests in the world for the diagnosis of ischemia. SPECT imaging has historically been the workhorse, and as you likely know, is a method of assessing relative myocardial perfusion. That is to say, we try to compare one part of the image to another part in order to work out where delivery of tracer is reduced, and then to infer from that which artery might be stenosed. There are many problems with this approach, not least that all conventional SPECT tracers have been shown to significantly underestimate both the severity and the extent of disease. Technetium, which is now the most commonly used tracer for SPECT MPI, is inherently limited by low first pass extraction fraction, particularly at high heart rates. So this can limit the precision and the accuracy of the tracer during stress. Even small differences in extraction fraction can adversely affect the recognition of the presence and extent of ischemia by SPECT. So this is where PET enters the picture. Now superficially it may seem similar at first glance. After all, with PET we still look at tomographic slices of relative perfusion, and we still assess ventricular size and function from gated images, just as we do with SPECT. However, PET is a very different technology using very different tracers. As you will see, the most exciting thing about PET is the ability to quantify absolute myocardial blood flow. This then allows us to quantify global and regional flow in mils per minute per gram of tissue. The good news is you already know at least 50% of how to report a PET study if you've been regularly reading SPECT cases. So first of all, I'm going to briefly review the overall reporting structure and what information it's recommended that be included in your report by the American Society of Nuclear Cardiology. After that, I will move on to discuss the slightly trickier part, which is how to interpret and report the quantitative PET findings. So let's start by talking about something which on the face of it doesn't seem that interesting, but is actually very important, and that is the imaging report itself. Most of the buzz surrounding nuclear cardiology is based on better cameras and newer tracers and that's understandable, but what matters most to the referrer is receiving a clear, well-structured report that is comprehensible by a non-expert. In many cases, this will be the only form of communication between the reader and the referrer, and as such is a critical component of any imaging procedure. Importantly, it also becomes part of the patient's legal medical record, yet another reason to provide concise but clear and appropriate information. The imaging world as a whole has tended to move away from so-called free-text reporting. Instead, structured reporting allows for the transmission of large amounts of information in such a way that the referrer knows where to look to find what he or she wants to know. This approach tends to limit verbiage and is generally appreciated by referrers. This slide summarises the information that is required and we will briefly explore each one of these headings in the slides ahead. This is quite straightforward. It is required for the name and the location of the hospital to be at the head or foot of the report. It's also recommended that a contact phone number be included, as well as any details regarding accreditation status and the accreditation body. Demographic data that must be included is shown here and is mostly common sense. Although patient sex and ethnicity are recommended rather than required, it also makes sense to include these, particularly if you work in a unit that is in any way research active. Some of the items listed on this slide may not apply outside of the United States, but clearly the study date and time, the report date and time, and the electronic signature of the reporting physician are universally required. No imaging study should take place without clear knowledge of the primary indication, and there may be relevant secondary details to include as well. I think it's helpful in a perfusion report to include a little bit about the type of pain the patient is having, whether that's typical or atypical, for example, if that information is available. With an ever increasing focus on using the right test for the right patient at the right time, it has been suggested that appropriate use criterion should also be included in the report as a potential form of feedback to the referrer. It can be helpful to know about patient medications, especially those taken on the day of the test. Cardiac factors are very helpful and knowledge of conditions such as diabetes, dyslipidemia and hypertension can be highly relevant when assessing both macro and microvascular disease by quantitative PET. Relevant prior cardiac testing may also be appropriate to mention. For example, a recent cardiac CT study with a questionable stenosis in one area. If the reporting system is fully electronic, most selections can be made from drop down lists, which makes it very quick and leads to a uniform reporting style. Details should be provided about both the stress agent and the radio tracer employed. Most PET perfusion studies are done using rubidium, in which case there is no choice but to use pharmacologic stress. However, qualitative ammonia PET perfusion studies may be done with exercise, and in such cases the exercise protocol should be described. I am sure that this is very familiar to you. We have to collect and report all hemodynamic parameters in order to assess the effects and the effectiveness of stress. We also need to report the onset, type and duration of any symptoms that occur, as well as provide the reason for which the study was terminated. All perfusion studies are performed with ECG monitoring, so we need to describe any resting abnormalities which may affect test interpretation, such as left bundle branch block and LVH and strain pattern. Arrhythmias such as AF should be described, and the presence of high-grade AV block excluded before the test commences. There is a little bit more to report with a stress ECG. We need to describe the time of first onset of changes, where these changes are, and how long they persist for. We also need to comment upon arrhythmias that may be seen during the test, such as atrial or ventricular tachycardia, and even things like ventricular bigeminy or extrasystoles. Finally, we should state at what point in the recovery period the ECG has returned to baseline. It is clearly vital that we describe clearly what stress agent is being used alongside which perfusion tracer. Both rest and stress doses of tracer should be given in the report. Finally, it should be stated whether CT or line source attenuation correction is being used for the study. Qualitative assessment of relative perfusion images starts with a general summary statement about whether perfusion is normal or abnormal. If abnormal, we then need to comment on both the extent and the severity of the defect. Both of these may be semi-quantitatively described, as shown here on the slide. Of course, much more should also be added to the report. The location of the perfusion defect should be described based on the AHA 17-segment model, as well as whether it is fully reversible, partially reversible, or fixed on rest imaging. Both visual and quantitative transient ischemic dilatation should be looked for, and its presence or absence noted. Abnormalities of myocardial wall thickness and resting or stress-induced segmental dysfunction should be commented upon. Some centres choose to give additional information using parameters such as the summed stress and rest scores, as well as the summed difference score, since together these provide a description of the extent of ischemia and infarction. However, in order to better improve communication with a referrer who likely has little or no understanding of what such parameters may mean, many sites will choose to convert summed scores into more readily understood metrics, such as the percentage of the LV that is ischemic and the percentage of the LV that is infarcted. When commenting upon LV function, it is helpful to do so with a visual description that matches the quantitative value calculated by the software. On this slide, we see the typically used set of cut-offs together with the appropriate text description for the level of function measured. It is important that everybody within a lab is using the same thresholds to define categories as well as the same wording to describe different levels of LV function. Stress-induced wall motion abnormalities are important to look for and recognise. Since imaging occurs at the time of stress, abnormalities represent true ischemic dysfunction if they are not also present at rest. The location and extent of such abnormalities should once again be described clearly in segmental language. Not all centres choose to report LV volumes, but if you do, these should ideally be indexed to body surface area for interpretability. All academic centres and any centre that reports data to a central registry should ideally collect and report these items. Rest function should of course be reported using the same thresholds and language as we have just described for stress. Similarly, resting wall motion, thickening and volumes should all ideally be included in the final report. Now in this section, we are going to move on to one of my favourite things about PET and indeed one of its greatest strengths, that is to say the ability to tell us in fully quantitative fashion about global and regional myocardial perfusion. So before we go any further, I would like to make you aware of this publication which came out from the American Society of Nuclear Cardiology at the start of 2021. This is really a tour de force in terms of being a practical guide for how to start approaching the understanding and reporting of myocardial blood flow measurements from PET. I strongly recommend that you spend some time looking through this if you are going to be doing this as part of your clinical practice. There really is many, many years of author expertise embedded into this document. It's probably the best single summary I've seen in terms of helping you start with your PET practice. Myocardial blood flow comes as an added benefit of using PET and happily does not necessitate any increase in radiation dose to the patient nor any increase in time within the imaging facility and no additional staffing is required. Quantitative flow data improves both the diagnostic and prognostic assessment and allows for interrogation of not only the epicardial coronary arteries but also the mycrovascular bed. Traditional SPECT or PET MPI detects obstructive CAD but potentially under-assesses CAD extent and severity. Myocardial blood flow measurement improves assessment of epicardial CAD and also improves stratification of risk for major adverse cardiac events. Myocardial blood flow measurement improves the selection of patients for coronary interventions and or medical therapy alone. Myocardial blood flow can, as we've suggested, also identify coronary microvascular disease and importantly measurement of myocardial blood flow provides assurance that vasodilator stress has been effective. Myocardial blood flow is dynamic, responding to changes in heart rate and systolic blood pressure, the two main determinants of myocardial work. As such, there will be variability in a patient's myocardial blood flow from a physiologic perspective. Another source of variability is measurement variability and this can be controlled for to some degree by adherence to patient preparation and imaging protocols. The following tips help to minimize procedural sources of variability. Number one, use a free-flowing intravenous line for both stressor and tracer. Optimally, this will be a large 18 or 20 gauge needle in a forearm vein. Hand injection should be avoided. Number two, the degree of hyperemia varies in relation to the time after stressor injection. Time to maximal hyperemia is not well defined for adenosine dipyridamole or regadenosone. Regadenosone is the stressor most often used in the United States. Peak hyperemia likely occurs about one or two minutes following its injection, so a concomitant time delay between its injection and beginning the tracer infusion would seem prudent. Number three, be consistent in the time duration between the stressor injection infusion and the tracer injection infusion. A stopwatch is helpful for this purpose. Number four, utilize the same pharmacologic stressor as much as possible for myocardial blood flow measurements. This is because the degree of hyperemia is known to vary between different stressors. The alignment of the PET rest and stress emission images and the transmission attenuation correction images need to be assessed in the transaxial, coronal, and sagittal views, preferably using a fused display. Misalignment may be due to patient and or respiratory motion and result in image artifacts on the relative perfusion images, as well as corresponding regional decrease in myocardial blood flow. Any required correction of misalignment must be performed on the PET scanner console with repeat reconstruction after proper alignment. This is not something that can be done at the workstation. The placement of the region of interest for blood pool sampling, usually left atrium or left ventricular cavity, needs to be reviewed. The blood pool ROIs should be in the same location for rest and stress, and they should not touch the walls of either left atrium or the left ventricle to avoid spillover. The myocardial ROIs also must accurately encompass the myocardium during all frames that will be used for determining the tracer uptake. This ROI needs to be inspected to ensure that it is accurately tracing the myocardium and excludes adjacent non-cardiac structures that may contain tracer. The time activity curve must begin prior to the infusion or injection of the tracer. It must demonstrate an initial increase in blood pool counts followed by a precipitous drop in counts because of tracer uptake. The myocardial count time activity curve must demonstrate a steady uptake of tracer followed by a plateau of counts. For longer acquisitions, this plateau could gently increase or decrease depending on tracer kinetics. However, there should not be any abrupt changes in myocardial counts during the time-activity curve. And here is an example taken from the publication that I mentioned, which illustrates these points. There is good delineation of the boundaries of the myocardium at both stress and rest shown on the left. On the top right, you can see images where there is a small dot which acts as the arterial input function. This is nicely positioned in the left atrium in this case. On the bottom right, you can see an example of a time-activity curve from a dynamic perfusion study with rapid initial uptake followed by a descent and then a plateau phase. Here is a good example of a dynamic PET study where although the contours are originally well placed, as the study commences, they drift off substantially due to patient motion. The temporal sequence is from the top to the bottom. At the beginning of the acquisition, the blood pool and myocardial ROIs are in the correct place as we can see here. As the scan progresses, the ROIs are no longer positioned over their respective regions. To obtain accurate blood flow measurements, this motion during the dynamic phase must be corrected. This can only be done at the PET scanner console and not using the reporting workstation. This is why it is important to retain the raw data until the flow calculations have been verified as accurate. Myocardial blood flow reserve is the ratio of stress myocardial blood flow divided by rest myocardial blood flow. If heart rate or blood pressure is significantly increased, for example, due to the holding of medications on the day of study, the myocardial blood flow reserve ratio may appear abnormally low because the resting flow is unusually elevated, thus increasing the denominator in the flow ratio. In such cases, one of two approaches can be used. The rest myocardial blood flow can be adjusted using the rate pressure product and a reference value such as 9000. A common way of adjusting is to divide the resting myocardial blood flow by the rate pressure product and multiply the result by the reference value. An alternative approach is to explain in the report that myocardial blood flow reserve is artificially low because of high resting myocardial blood flow and then default to the peak myocardial blood flow. Either approach is reasonable for clinical reporting. PET is generally much more sensitive than SPECT for CAD, but may be slightly less specific since all the measured blood flow parameters may also be affected by both diffuse non-obstructive and microvascular disease if present. Nonetheless, it is universally accepted that a myocardial blood flow ratio greater than 2 means that there is a very low chance of the patient having high risk CAD and that the patient has an excellent prognosis. Of course, this additional data needs to be integrated with the relative perfusion images and the functional data that we have already discussed. It makes sense to evaluate rest and stress MBFR and MBF in three complementary ways. Globally, i.e. for the whole LV, per vascular territory, and then also segmentally within each vascular territory. Bear in mind that a segment may be abnormal, but the integrated flow values for the artery may still be normal as there may be enough normal segments to average out a single abnormal one. At the same time, the noise in the data increases and the precision of flow measurement decreases in individual segments, and therefore care is needed in order not to over-interpret small regional changes in flow. On the right-hand side are rest and stress values for a case showing myocardial blood flow as well as the ratio of stress to rest flows, which, as we have said, is myocardial blood flow reserve. By the way, the exact threshold for normal stress flow is debated. It's often regarded as being around 1.7 mgs per minute, but this varies a little from lab to lab and will also vary according to the stress agent that is used. Normal rest flow is generally below 1 and maybe below 0.5 mgs per minute. So with that in mind, if we review this chart, we can see that the rest flow is normal in each artery as well, and therefore so is global flow. Now look at the far right column for flow reserve, and you'll see that the flow reserve for each vessel, and therefore the global reserve, is also normal. If we stopped there, we would be tempted to call this a normal study, if the relative perfusion images were also normal. However, let's look at the middle column for stress flow. Here we see that global flow is above our magic value of 1.7 and therefore is normal. However, when we look at stress flow on a per vessel basis, both the LAD and the right coronary artery have stress flows greater than 1.7 and are also normal. But, what about the circumflex? Well, that has a stress flow value of 1.65 and so you might be inclined to round that up to 1.7 and say it's normal as well, and in some cases that might be a reasonable thing to do. However, this is an example where averaging of the segments that comprise an arterial territory can lead to obscuration of real findings. So if we look in more detail at segmental flow reserve values in the bullseye plot on the left of the screen, we can see that the flow reserve is less than 2 in several of the distal circumflex segments, suggesting that there may be distal circumflex or side branch disease. Some software will also provide plots of segmental stress flow values as well as flow reserve, and a similar reduction in stress flow would have been visible in the same segments. So the lesson here is to think global first, most prognostic data is based upon global flows, and then think local second. It sounds confusing and on occasion it can be, but mostly it gets easier the more cases you do and the more you correlate with subsequent angiography, which is an absolutely essential part of ongoing quality control and learning for your lab. The added value of myocardial blood flow reserve varies depending upon context and is usually easier to interpret in the context of those with no known prior CAD. In this group it has a high negative predictive value, especially when the relative perfusion images are also normal. The ability to review MBF and MBFR on a per vessel basis also makes it possible to assess the relative extent as well as severity of disease. Similarly, balance 3 vessel disease which may not be obvious on relative perfusion images can be readily identified upon review of the quantitative data. Finally, if there is little or no augmentation of flow, that is to say if the MBFR is less than 1.2 and there is little change in rate pressure product, then we can suggest probable failure of vasodilator stress with reasonable confidence. By now it should be obvious that consistency is important. Measurement of myocardial blood flow may result in differing results depending on the software, the radiotracer, and the protocols used for imaging and analysis. Serial patient MBF measurements are best done using an identical approach for imaging and analysis on both studies. When moving to incorporate blood flow measurements into a reporting template, one thing that needs careful consideration is how to do this in a way that conveys your expertise to the referrer whose own knowledge of PET is likely to be rudimentary. Referring physicians will not initially know how to incorporate MBF measurements into management decisions. Therefore, the wording needs to assist with how the MBF measurements affect the overall study results with respect to diagnosis and risk, as in the table that we reviewed a few slides earlier. CASNIC has also provided a number of examples of clinically useful statements that convey clear meaning to non-nuclear physicians. Now I've never seen this done before by any other imaging society, but I really like it because it places MBF and MBFR in a clinical context and conveys a clear message. There are circumstances where MBFR values may be confusing or misleading, some of which are shown on the slide here. In these circumstances, careful wording of the report will be required in order to convey a useful interpretation to the referrer. Some centres may choose not to quote MBFR in these instances. This is reasonable, but perhaps even better is to quote but explain clearly, as we discussed on the prior slide. As we all know, there are some patients in whom we suspect non-response to the stress agent, whether that be due to intrinsic physiology or more often due to recent consumption of caffeine-containing beverages or products. Very often we can get some sense of this from the lack of symptoms during stress, although remember diabetics and some post-bypass patients do not always exhibit much symptomatology even with adequate stress. So then we can assess change in the rate pressure product. In most cases we expect the heart rate to rise 10-15 beats per minute and the blood pressure to fall by around 10 mmHg. When we see this, we are fairly sure that stress is adequate, but what when we don't? Well more recently a number of groups have described the utility of a sign known as splenic switch-off. This was first noted in the stress CMR world and was quickly picked up by nuclear physicians. It transpires that the spleen has adenosine receptors that respond to the stress agent by local vasoconstriction and this means that far less tracer is uptaken by the organ at stress than in the resting state. This therefore gives the impression that the spleen shuts off with stress. This can be assessed both visually and with various semi-quantitative ratios and I have provided references for anyone interested. I would like to specifically credit this nice image which is in the Bacula paper and comes from the Zurich PET lab. There is one very specific use case for quantitative PET and this is in the setting of patients who have undergone cardiac transplantation and are being followed for the development of transplant vasculopathy. At the University of Ottawa Heart Institute, based on data published by Sharon Chee et al, we report the following parameters for cardiac transplant patients. 1. Myocardial blood flow reserve which should be greater than 2.9 2. Stress myocardial blood flow which should be greater than 2.3 mpg per minute 3. Coronary vascular resistance which should be less than 55 units This last one may not be familiar to you but the index is simply calculated by dividing the systolic blood pressure by the stress myocardial blood flow and I have included a reference here for anyone who is interested. So having talked for a bit about the interpretation of quantitative PET, let us turn back again to the report once more. As imagers we all know that many of our colleagues don't bother to read the entire thing and simply skip to the conclusions. So the overall impression is the most important part of the report and when the study is positive it is a good idea to also include a statement about the significance of the perfusion results. It is also no use having a study done and a report drafted but the final report not being read by the referrer so ASNIC recommends that interpretation of the study be completed within one working day and that the final report be transmitted to the referrer's office within two working days. There may also be some high risk studies where you would consider emailing or calling the referring physician to give them an immediate heads up. This certainly occurs in our lab multiple times per week and it is also not infrequent for us to admit patients with very high risk scans directly to the ward. Certainly every lab needs to have a well defined strategy for dealing with high risk results. This overall impression or conclusion needs to provide a concise summary of the most salient points regarding the ECG, the relative and quantitative perfusion results, the LV function together with some kind of statement of overall risk in most cases. Now a PET study doesn't always occur in a vacuum. Very often there will have been prior imaging by PET, SPECT, CMR or angiography and a good report provides a comparison to prior perfusion imaging and may make reference to recent anatomic imaging where CT or CAF have been done. So I've told you what a report should contain and what information it should attempt to transmit. What do you think of this report? I wonder how many things you can find missing or inadequate about this kind of report. Pause the talk here for a moment and see how many issues you can remember before moving on to the next slide. This is the same patient with what I would consider to be an optimum report. I have to say that because it comes from my centre. But seriously why do I think this is better? Well firstly it is well structured along the ASNIC principles that we have discussed. It remains concise but transmits a huge amount of information. It provides details about the stress agent and the radioisotope used, ECG findings are reviewed, technical adequacy is considered, function and volumes are quantitated. Comment is made about qualitative and quantitative transient ischemic dilatation. The opinion is not in free text but laid out point by point with, I would argue, the most important comment first, abnormal study, high risk. Even if the referrer reads no further, he or she has received 95% or more of the intended message. There then follows a clear description of relative perfusion findings followed by quantitative estimation of both ischemia and scar respectively. Not only is EF reported but special note is made of a greater than 5% drop in EF with stress which is known to be a high risk marker. Finally quantitative flow is described and a clear statement of abnormality in all three coronary territories is made. And at the very end mention is made that there are no prior studies with which to compare. Now perhaps you can find room for further improvement with this report. This may not be a perfect report but it is certainly a very good one. And so on behalf of myself and Dr. Peddarazadeh, I'd like to thank ASNIC for inviting us to prepare this talk and to thank you the viewer for following along. I hope that you'll come to enjoy PET imaging as much as I do. And for those who want to ramp up their experience quickly, I can thoroughly recommend the PET workshops that are put on by ASNIC several times a year. I've personally attended twice, found that there was just an enormous amount of valuable expertise available from the top people in the PET community.

PET myocardial perfusion studies

interpreting

reporting

quantitative assessment

ischemia

American Society of Nuclear Cardiology

clear communication