Hello, my name is Andrew Einstein. I'm a cardiologist at Columbia University Irving Medical Center, and it's my pleasure to speak to you today as part of ASNIC's hybrid imaging workshop. I'm gonna be speaking about how to perform and interpret coronary artery calcium scores, quantitative and visual. These are my disclosures. The CT scanner was invented by Sir Godfrey Newbold Hounsfield in 1967 at EMI Central Research Laboratories in the United Kingdom. The first production CT machine was called the EMI scanner and it was limited to brain imaging. You can see here pictures of Dr. Hounsfield of his initial scanner and the initial patient images from it. It was installed in 1972 at a hospital in Wimbledon, England and we've never looked back. This had a small gantry and didn't enable body imaging. However, a larger scanner was begun a few years later. The first images from the body CT scanner called the Emerald scanner were taken of Dr. Hounsfield himself. And reputedly, he recognized the images of his lunch of crisps and beer eaten at the Blue Anchor Pub in Bermuda on that day. You can see the images here. By October of 1975, body CT scanners had been installed in three hospitals worldwide, the Northwick Park Hospital and at Washington University in St. Louis and at the Mayo Clinic in the United States. Dr. Hounsfield won the Nobel Prize in 1979 and in his acceptance speech, he stated, a further promising field may be the detection of the coronary arteries. It may be possible to detect these under special conditions of scanner. In fact, research in this area had already begun. The first scanner which imaged the coronary arteries was the Dynamic Spatial Reconstructor, also at the Mayo Clinic. However, this was really a massive scanner. It had a 13-ton gantry which rotated at 15 rotations per minute. It had 14 x-ray tubes and supposedly it caused a brownout in the city of Rochester, Minnesota because it used so much electricity to perform. So it was never really used in humans. But also in the late 1970s and early 1980s, the Electron Beam CT scanner was developed. This was commonly referred to at the time as an ultrafast CT scanner or EBT or EBCT. They all refer to the same thing. And this used quite a different technology from CT scanners before and after it. Really, there were few moving parts in this and it was based on an electron gun which generated an electron beam. That electron beam got magnetically focused and deflected. It hit tungsten target rings and generated x-rays there. And those x-ray beams projected through a patient and were detected on the other side. Because of the few moving parts here, the temporal resolution of this scanner was really good. And despite its limited spatial resolution, it was sufficient to enable imaging of the coronary arteries. This was captured by Dr. Arthur Agatston and Dr. Warren Janowitz at Mount Sinai Hospital in Miami Beach, who developed coronary artery calcium scoring, which is today still generally referred to by Dr. Agatston's name. You can see this classic paper from Jack in 1990 on quantification of coronary artery calcium using ultrafast computed tomography. And you can see the images, which are not so different from the images which we're going to look at together. We're talking about performance and interpretation of these scans. So let's talk about the performance of an EBCT scanner. Why? This is an ancient technology, but really as we've moved to the next generation of scanners to multi-detector row CT scanners, what's happened is that these parameters, which initially were used in electron beam CT, have largely been duplicated in the multi-detector row CT scanners, which we use. So this was scanned in Dr. Agatston's paper and subsequent work in the so-called high resolution volume mode, which meant at that time, three millimeter slices. The gantry, or rather the scan time for a single slice was just 100 milliseconds. We used an X-ray tube potential of 130 kilovolts. 20 slices were acquired beginning at the carina and additional slices were acquired as needed so as to cover the entire heart. You can see here from Dr. Agatston and Janowitz's classic paper, four consecutive slices, three millimeters apart without any inter-slice gaps. You can see calcium here in the left anterior descending coronary artery, for example, you can see subsequent slices in the same heart. There were no inter-slice gaps. ECG gating was performed with each image being triggered at 80% of the RR interval. And it required about 40 heartbeats for 20 slices because the scanner would image one beat, skip the next beat, image the next beat. So it required a total acquisition time for a calcium score of about 30 to 45 seconds. Interpretation again is quite similar to what we do today. Dr. Agatston and colleagues developed a system where they looked at the threshold for any calcific lesion in the coronary artery. If the brightness of a calcified lesion was greater than 130 Hounsfield units, these units of brightness are named after Dr. Hounsfield, and it had an area of at least one millimeter squared, then it's counted towards the Agatston score or the calcium score of that patient. If there's a CT number, which is greater than 130 Hounsfield units, the patient gets one point for that lesion. If the CT number is between 200 and 299, then it gets two points. Between 300 and 399, it gets three points. And if it's 400 or greater, then it gets four points. So each region of interest was selected. You'd identify an area of calcification in the coronary arteries and on a single slice, take that number denoting the brightness, multiply it by the area of that lesion. So long as the lesion is at least one millimeter squared, and add that up throughout the coronary arteries in each slice, and that's the calcium score. That's the Agatston score for that patient. Very similar to what we do today. After calcium scoring was developed, there was great interest in developing coronary CT angiography, but this old generation of scanners, the electron beam CT scanners, simply didn't have the spatial resolution that's necessary for accurate coronary CT angiography. For that reason, and for other applications as well, the electron beam CT scanner, notwithstanding its great temporal resolution, simply didn't have the spatial resolution, which at best in the last generation of these scanners was one and a half millimeters. So really technology moved on, but there was some tug back and forth between these technologies. With the move from single slice CT scanners to multi-detector row CT scanners, the temporal resolution, as well as the spatial resolution of this more conventional geometry of CT scanners became better, and electron beam CT scanners went by the wayside. In fact, in 2023, when I'm recording this video, we're going to be having the American College of Cardiology Scientific Sessions in New Orleans, turned back the clock 19 years ago at the ACC Scientific Sessions in New Orleans, and there were raging debates about EBCT versus multi-slice CT. Ultimately, notwithstanding each technology having certain advantages for non-cardiac applications, there seem to be significant benefits for multi-detector CT, and there are a few sites today which still use electron beam CT. However, there've been multiple comparison studies, and in general, we find calcium scores, coronary artery calcium, or CAC scores, agreed quite closely between these two technologies. You can see in this one study, an R of 0.98, so really good correlation between the calcium score performed on this initial EBCT scanner and the current generation of multi-detector row scanners. So today, all of the hybrid scanners, be they PET CT scanners or SPECT CT scanners, from all of the major manufacturers only would have a multi-detector row CT scanner. So you're going to be, the calcium score which you're performing on your hybrid scanner is going to be performed on a multi-detector row CT scanner. That being the case, in general, these multi-detector CT scan parameters are pretty closely mimicking what had initially been done on electron beam CT. For example, reconstructed slices should be three millimeters for some scanners that's not possible, it's only possible to do 2.5 millimeters, but really you should only be analyzing calcium scores quantitatively using two and a half or three millimeter slices. If you use thinner slices, it's going to give you an inaccurate calcium score. The temporal resolution for multi-detector CT is actually not as good as it had been on electron beam CT, but in general, the gantry rotation time is going to be somewhere between 200 and 500 milliseconds. And as for all of these parameters, when you get a new hybrid scanner, work with your vendor to help them set up a calcium score in an optimal manner. Whereas for the electron beam CT scanner, the X-ray tube potential was 130 kilovolts. That's not available on most multi-detector Rho-CT scanners. So we use a tube potential of 120 kilovolts. And in fact, all of the epidemiologic data, which we have is at 120 kilovolts. Changing the tube potential changes signal brightness and hence it's going to change your calcium score. So in general, since our epidemiologic data is calibrated at this 120 kilovolts, we're going to want to stick to that and not adjust that. Images again, like on EBCT scanners are acquired from the carina to the bottom of the heart. There should be no inter-slice gaps. ECG gating should be performed, generally only triggered during diastole. But here it only takes a few heartbeats, anywhere between one and say seven seconds to acquire a full scan using contemporary equipment. Some important considerations, as we mentioned, really should use a tube potential of 120 kilovolts. There's a little bit of a literature on using reduced tube potentials or otherwise known as tube voltages. This paper from Rio Nakazato from Dan Berman's group addresses using 100 kilovolts peak and compared that to 120 kilovolts peak. And in general, there's close agreement, but you have to use a different threshold. Rather than using 130 Hounsfield units, they used 147 Hounsfield units. This in general is not built into conventional software. My colleagues at Columbia and I have developed a method for calculating scanner kilovolt peak and patient size specific Hounsfield unit scale thresholds for Agatston. So that's tailored, not just to the KVP, which can be 80, 100, 120, or any other value, but also for the specific scanner. So I think this is really in the research phase still and currently one should only perform calcium scoring using 120 KV. Your tube current should be selected so that you get a noise level that will give adequate images for calcium score assessment while keeping radiation dose reasonable. And as you use your scanner more, you'll get a better sense of that. And again, the vendor can help you select an initial value there, but there should be a continuous process to make sure that that's optimally selected. The scan length should be selected on the scout to ensure full coronary coverage, and it should be done with ECG gating and an inspiratory breath hold. Now, one consideration, which I think we're not entirely sure about right now is which image reconstruction algorithm should be used. Traditionally, we had used filtered back projection for coronary artery calcium scoring, but more recent studies have attempted to compare filtered back projection with iterative image reconstruction and even newer deep learning-based or machine learning-based iterative reconstruction algorithms. You can see three such studies here. In the first study, in comparison to filtered back projection, an iterative reconstruction method may decrease Agaton scores in calcium volumes, but in other studies, and I think this seems to be a consensus in the field now, is that in comparison to filtered back projection, iterative reconstruction techniques do not have a profound effect on the reproducible quantification of coronary artery calcium, according to the Agaton score and subsequent cardiac risk classification, although risk reclassification may occur in a small subset of patients. So my colleagues and I are comfortable using iterative reconstruction for coronary artery calcium scoring, but again, I think it's worthwhile to understand your own equipment and see what reconstruction algorithm is the best. Now, there are different types of CT scans which can be performed on a hybrid scanner, be it PET-CT or a SPECT-CT scanner. The best way to perform a coronary artery calcium score is to use a dedicated coronary artery calcium scoring scan. That'll include ECG gating. Generally, it'll be performed at an end-inspiratory breath hold, and the diagnostic quality for that scan should be adequate for coronary artery calcium scoring. It'll be done at 120 kV at some medium MA level and a slice thickness of 2 1⁄2 to 3 millimeters and a scan length, which is typically going to be about 14 centimeters. However, there are other CT scans which we acquire on our hybrid PET-CT or SPECT-CT scanners. For example, CT attenuation correction scans. Those will provide on some scanners information about coronary artery calcium. However, they're not optimized for calcium scoring. For example, they're acquired typically during free breathing without ECG gating. The diagnostic goal of these scanners may vary. In some cases, you may want to use the lowest MA that's acceptable for CT attenuation correction. However, other labs may choose to use a higher MA, a higher tube current, so they can acquire more diagnostic data, such as better quality images of the lungs, better quality images of the coronary arteries for that coronary calcium scoring assessment. That's going to vary from lab to lab. The kV here is going to vary. So remember, if you're using a lower kV, in general, things are going to be brighter, maybe overestimating your calcium score. The MA could be the lowest possible or higher. Slice thickness is going to vary, and the scan length in general is going to be a bit longer than that on a calcium scoring scan. And similarly, for an image registration scan, for example, ASNIC is really encouraging the use of SPECT-CT for the assessment of patients in nuclear scintigraphic imaging for cardiac amyloidosis, for example, a PYP or an HMDP scan. And for those registration scans, which are not performed so much for attenuation correction, but rather to register our nuclear images together with a CT for anatomic localization so that we can see the difference, for example, between blood pool uptake of pyrophosphate and myocardial uptake, those also will be free breathing without ECG gating. In general, there'll be a good enough image quality for image registration. The KV and the MA, again, can vary. And these, again, it's another type of scan which we can assess calcium score on, or at least assess some sort of burden of calcium on without having a dedicated high-quality coronary artery calcium scoring scan. We'll go over illustrations of both of these types of scans. So there's different types of calcium scores which we can acquire as well. Really, what we've talked about are Agatston scores, but there's other quantitative measures of coronary artery calcium burden, which one can obtain from a coronary artery calcium scoring scan, for example, the volume score or the mass score. And one can acquire both of those using a dedicated ECG gated 120 KV breath hold scan for coronary artery calcium scoring. However, on an attenuation correction scan on your SPECT CT scanner or PET CT scanner, you can't really acquire a proper Agatston score because those parameters necessary, for example, the breath hold and the ECG gating generally are not the case for the scan that's obtained there, unless one does a second scan. One can try, in some cases, try to roughly estimate an Agatston score by inputting the data which one acquires from the attenuation correction scan into the Agatston scoring program on your workstation. But more commonly, people are going to use these scans to visually estimate a coronary artery calcium score. We'll give illustrations as well of that visual estimation process. And then similarly, for these CT nuclear image registration scans, can't do a proper Agatston score, but you can try to approximate an Agatston score or visually estimate an Agatston score. I was privileged to do some research with a number of ASNIC leaders assessing the agreement of visual estimation of coronary artery calcium from low-dose CT attenuation correction scans with standard Agatston scans. And in general, we found very strong agreement between the assessment of coronary artery calcium scoring in experienced readers between that dedicated Agatston scoring scan and a VCAC, or visually estimated coronary artery calcium, from a CT attenuation correction scan. In general, 93% of cases, the readers estimated within one category, looking at six different levels of calcium scoring. That is, for a visual estimation, one can look at the scan and say, hmm, this looks like a calcium score, probably about zero or one to nine, 10 to 99, 100 to 399, 400 to 999, or 1,000 plus. And 93% of cases here were estimated within one category of the true score, with that true score being determined based on a dedicated Agatston scoring scan. And indeed, other work from Paco Bravo's group at the University of Pennsylvania shows important prognostic information in this visual estimated coronary artery calcium scoring. So I think having some assessment of calcium scoring, even on our attenuation correction and image registration scans, is an important additional piece of information which we can provide to our patients and their providers. So now, why don't we switch and take a look at some actual cases, both to look at visual coronary artery calcium scores and actual measured Agatston coronary artery calcium scores. Okay, let's start with a case and determine an Agatston score for a patient. First thing which I generally do when I review a calcium scoring case is to take a look at the scan parameters, make sure that I have the right parameters. One thing which you can see on the screen here is it says slice pos, that's slice position. And by scrolling up and down, I can see what the difference in position between one slice and the next slice is. As you can see here, we're at 81.1 millimeters, here 83.6 millimeters. So the difference between those two numbers is two and a half millimeters. That's good. That's either two and a half or three millimeters, that's standard for calcium scoring. If, for example, I had 0.5 millimeter slice distance or 0.6 or 0.625, that would mean I'm using the thinnest possible slices. Those images will, in general, be noisier and not give us an accurate calcium score. So you want to make sure that you've pulled the right data set into your workstation. Second thing which I see on the screen here is it says calcium score ACER 30. It tells me that I'm using an image reconstruction algorithm called ACER, which happens to be one of the vendor's software programs which uses iterative reconstruction. That's a choice which we've made in our laboratory, and you just want to make sure that you're consistent and using a consistent image reconstruction algorithm for every patient. So you're not switching from filtered back projection to one patient, iterative reconstruction on the next, and deep learning-based iterative reconstruction on yet a third patient. Now I'm at the top of the scan, and what I'm going to start doing is scrolling up and down, and there are different ways that you can scroll using different programs. I'm starting with one program now. This program is called Heartbeat Calcium Score, which is part of our Intellispace portal package which we commonly use at Columbia. I'll illustrate another program later on. There's many of these programs, and they all work a little bit differently, but in general you're going to want to scroll up and down and get a look at the lay of the land, where you have coronary artery calcium here. One tool which we have here is this highlight button. I can toggle highlight on and off, and you can see pink appears and disappears here. So what is pink? Pink basically is pseudocoloration. It's going to color every voxel which has a CT number greater than 130 Hounsfield units in pink. So that's telling me, well, that voxel is a candidate to be part of a calcium score. Remember, again, you're going to want the area to be at least one millimeter square for it to be counted as part of a lesion in a coronary artery, and it's got to be at least 130 Hounsfield units. Now, not everything that's pink is going to be coronary artery calcium. You can see, for example, here is the patient's sternum. Here is the patient's spine. That's colored in pink. What the computer leaves me to do is to say what specifically is a coronary artery and what is which coronary artery. So let's scroll up and down. When I take a look here, and I can turn off the highlighting once again, turn off the pseudocoloration, looks like here is my ascending aorta. Here's my descending aorta. Here's the left atrium, and here is the left main coronary artery coming off of the ascending aorta, and it's bifurcating into the LAD and into the circumflex. So this first lesion, which I'm seeing, seems to be in the distal portion of the left main coronary artery, and what I can do is specify which coronary artery it is. Here on the left side of the screen, I select left main, and this program, Heartbeat CS in the Intellispace portal, gives me two ways of selecting things. I can either manually or use this one-click seed. I'm going to choose this one-click seed tool. So I'll click on this, and then I'll click on this lesion. I've already selected left main. Clicking on here, and it turns it from pink to yellow. That means it's selected, and what you can see at the bottom of the screen here is telling me already that the score in the left main, and this is an Agustin score, is 42 already. In fact, because I have this one-click seed, it lets me know in adjacent slices if there's anything that's contiguous with the slice and the area which I've selected, and it's automatically going to include that. So if I scroll up a level here, you can see there's a little bit of calcium which I identified here, and if I scroll down a level, there's just a little bit, and all of that the computer is going to assign to the left main coronary artery. And when I take a look here, that's already the left anterior descending coronary artery. So I'm going to no longer select left main. Go here and choose left anterior descending, and instead of using the one-click seed here, let's use manual seed. So I'm going to go here and choose manual seed, and what this requires me to do is draw a circle around it. It's a different tool. Both ways are good. There are certain advantages of each, and it's assigned this to the left anterior descending. I can't focus on all the coronary arteries at the same time, so I'm going to focus just on one vessel at a time. So I'm going to look up and down the left anterior descending before moving on to the circumflex in the right coronary artery. Let's move down one slice. Ah, I see there's a little bit of pink here, so let me circle that, and it's assigning that to the left anterior descending artery. You can see here the aorta, the pulmonary artery, and as I'm scrolling down, I'm getting to the left ventricle and the right ventricle. Here's the anterior interventricular groove where the left anterior descending coronary artery should be sitting, and I see that artery, and I'm not seeing any pink, so it doesn't seem like there's any calcification there. Let's scroll down another level. Same thing, same thing. Going down one level after the next, just looking for pink there, that I can assign to the LAD if it's greater than 130 hf units. I'm just not seeing anything more in that LAD here. So I've covered the LAD. Let's scroll backwards. We'll go back to the left main here, and now let's start looking for the circumflex, left circumflex artery. That's coming off here. Here's the left atrium. Here's going to be the left ventricle, and the circumflex in general travels in the left atrioventricular groove. So let's take a look in that groove and see if we see anything that's colored pink or pseudo-colored pink here. Nothing, nothing. I see a little bit of pink here. It's very small, smaller than one millimeter square, and it looks like it's on the wall of the left atrium. Often you'll see that. It doesn't look like it's in the groove. That's not going to be a coronary artery, and even if I selected it, it would be too small probably to register. Let's keep going here. I see pink here, but that's pink in the aortic root. That's not pink in a coronary artery. So let's scroll down and look at the circumflex distribution. Not seeing anything else that's pseudo-colored pink here. Let's scroll backwards and look at the right coronary artery. Here we are. I would say this is not part of the right coronary artery, but it's still in the wall of the aorta, but when I scroll here, I see that there's this this pink area in the right coronary artery. Let's select the right coronary artery on the left side of the screen here, and let's go back to our one-click seed. Click on it. It selected it, and it's because we're using the tool which attaches contiguous areas of calcification in adjacent coronary arteries. It's telling me that in the subsequent slice there's also this. In fact, it's telling me also that the average CT number here is 314. So remember, one point for 130 to 199, two points for 199 to 200. It's going to give me three points for each unit of area because it's between 300 and 399. So it's counting that towards the right coronary artery. Let's keep scrolling in that right coronary artery, see if we find anything else. Here we're in the right atrial ventricular groove. Here's your right atrium, your right ventricle, and in the right atrial ventricular groove. Looking for pink, not seeing anything. So that's all, and that's really it for this study. What is it telling me? It's telling me that my total calcium score here is 84, and that's broken down by vessel, the left main 42, the LAD 7, and the right coronary artery 35. So we've performed our first calcium score. Let's move on to another case. Once again, I'm going to look at the metadata, make sure that I'm looking at two and a half or three millimeter slices, make sure that I'm using the same reconstruction algorithm, and once again, 61.7 to 64.2, again, two and a half millimeter slices, and the same iterative reconstruction algorithm I had used before. So I'm looking at acceptable data. Let's scroll up and down, look at the lay of the land here. I'm seeing more calcium here than we had seen on the previous patient, although maybe I'm seeing more ribs and spine, et cetera. When I'm looking in the coronary arteries, I'm seeing a fair amount of calcification here. In the aorta, I'm seeing it as well. Here's our descending aorta. Here's our ascending aorta. Here's our pulmonary artery, main pulmonary artery, right pulmonary artery. So the anatomy is a little bit skewed in this case, but as I'm scrolling up and down, I'm seeing calcification in coronary arteries. See if we have anything in the left main. I don't think I'm seeing anything in the left main here. In fact, probably right around here, I'm seeing the left main bifurcate into the left anterior descending and left circumflex. So let's start off with our left anterior descending artery here. So not going to choose anything for the left main. Let's select LAD. Let's click on one click seed, and I need to choose multiple areas. Each of these is associated with LAD calcification. Now let's scroll down a level, scroll up a level, and it looks like this as well should be associated with the LAD. This I'm glad it didn't select because that's part of the circumflex. I'm going to scroll down further, and again, because I'm using this one click seed mode, adjacent things in consecutive slices are automatically going to be selected together. So it looks like it picked up all of that calcification in the LAD, but it didn't pick up this. There's a small amount here. Let's click on that. Maybe too small to include. No, it picked it up, and then in the next level, it's picking up this additional piece of calcification, yet another piece of calcification. Again, here's the left ventricle, right ventricle, anterior interventricular groove, that's the LAD. Parenthetically, if I saw something in a diagonal distribution, in general, when we're assigning vessels, branches of a vessel will count towards that vessel. So calcium in the diagonal branches would count towards LAD. Calcium in the obtuse marginal branches would count towards the circumflex, etc. Here's LAD calcification again. Let's see if there's anything else, and then I'm going to need to make a determination here. Do I think this is calcium in a coronary artery? It's looking to me, and why don't I turn off the highlight here, that there may be just a, this doesn't, I don't think this is in the, in the epicardial fat where the coronary arteries sit. Rather, I think this is the pericardium, so there's a little bit of calcification of the pericardium here. I'm not going to count that towards the coronary artery calcium score. Let's keep scrolling backwards and forwards. So again, we won't count this. It looks like that's it for the LAD. So now let's move on to the circumflex. I'm seeing a lot right here. I have to ask myself, where is this? And this is something which can fool you, because mitral annular calcification can sometimes be confused for circumflex calcification. Here's my left ventricle. Here's my left atrium. This is the location of the mitral valve. So what we're seeing here is probably mitral annular calcification, not coronary calcification. You can scroll up and down, and sometimes they can appear quite close to each other. So I would venture to say that probably this is calcification in the left circumflex. This is calcification in the left circumflex, but this is all mitral annular calcification. I'm not going to use that. I'm going to turn the highlight back on. I'm going to select the circumflex. I'm not going to use one-click seed. I don't want to get fooled. I'm going to just manually select things. I'm going to assign this to the circumflex. I'm going to assign this to the circumflex, and the rest of the calcification we're going to leave as mitral annular calcification, and then just look on further. Make sure we're not missing anything here, and I think that's it. And then let's look in the distribution of the right coronary artery here. Maybe a little bit of calcium in the over here, and move this out of the way. In the right coronary distribution, often you can see blurry coronary arteries in the middle of the right coronary artery. That's the area which is most susceptible to motion artifact, as we see on coronary CT angiography as well. You can see that there's calcium here. Let's continue, and it looks like I would venture to say that this is in the right coronary artery, as is this. We'll assign those, and that's all. So what do we have? A total calcium score of 387, mostly in the LAD with a local calcium score in that vessel of 346, 29 in the circumflex, and 12 or 13 in the right coronary artery. So that's the second case. Okay, let's return to the first case, which we had looked at, but look at it using a different program just to illustrate the subtle differences between programs. Again, I want to take a look at the slices and make sure that I have two and a half or three millimeter slices. In fact, here it tells me that the slice thickness is two and a half millimeters. In fact, it tells me what the KVP was, that I ensure that it's 120 KV, tells me what the MA was, and here it doesn't tell me directly what the image reconstruction software was. However, on my second monitor, which you're not able to see here, tells me what the image reconstruction was. So why don't we, once again, here we'll use the middle mouse wheel to scroll up and down. This program is called vScore. It's part of the Vital Images Vitrea workstation package, and you can see it works quite similarly to the other system. And this we had assigned to the left main coronary artery. So here let's select left main, and then we can use either a select tool to identify a contiguous cluster of calcium, or we can choose the free tool to draw freehand shapes to identify calcium clusters, or use an ellipse tool to draw a set of ellipses to identify calcium clusters. So let's try each of these. I can select this, that's the left main, 42, which I think is very similar to what we had previously had using the other software. This program will pseudocolor lesions differently depending on which vessel you assign it to. Again, this had been the left anterior descending, the LAD. Let's use the free tool here. That is a freehand tool, and I'll draw a circle around that, and it's assigning that to the LAD. I'll draw another circle here, that's the LAD. Sometimes you need to use this freehand tool if you've got one lesion which was partially in one vessel and partially in another vessel. So for example, if this lesion were not all in the left main, but partially in the left main, but spilling over into the LAD, you'd want to draw a freehand lesion, part of it in the left main, and part of it in the LAD. That's not the case here, but it's a useful feature of that tool. Again, I'm going to scroll up and down, that's all we see in the LAD. Let's look in the circumflex distribution, and this can sometimes get in the way, but I can pick it up and drag it away. And not seeing anything in the circumflex, but in the RCA we have something as well, so let's choose the RCA. We can use the ellipse tool here and draw an ellipse around this lesion, draw an ellipse around this lesion, that's in the RCA. So different set of tools, quite similar, but each software program you're going to want to learn the ins and outs of it, whatever you're using. That's it for this patient. Why don't we take a look at one more patient in terms of a regular calcium score using this program. Let's look at one final case using the V-score program within the Vital Images Vitrea workstation. Let's scroll up and down. Once again, it's 120 kV. These images were actually reconstructed using filtered back projection. They look a bit noisier. For whatever reason, the iterative reconstruction images, which I have on my workstation, I couldn't open immediately. So we'll take a look at the filtered back projection images here. And I'm going to choose the select tool. The question is, where's the beginning of the LAD and circumflex, and what's in the left main? Still, I would say this is the beginning of the circumflex. This is the beginning of the LAD around here, and maybe something like this is in the left main. So why don't I choose my free tool. We're going to assign all of this to the LAD, except for that one last piece, which we're going to, oh, looks like I assigned it to the left main. So what I'm going to do is choose the remove tool, and I can go back, correct my error, and it's removed those. So we're going to assign to the left main simply this little piece, and the rest of it we're going to assign to the LAD. So I'll choose the LAD here, and draw this free region of interest around all of this calcification in the LAD. This is LAD as well. LAD, left ventricle, right ventricle, anterior interventricular groove, that's LAD. You can see the LAD here. However, there's nothing pseudo-colored pink. That's it for our LAD. That's it for our LAD. And this we said is circumflex, so let's choose the circumflex, draw a circle around here. This is getting in my way a little bit, so I can always move it. Maybe noise, maybe real. In any event, it's not going to make too much of a difference in terms of the calcium score, and it didn't change my Agatston score at all. Here's the circumflex again. Circumflex. Scrolling up and down. And finally, we have the right coronary artery here. Let's choose the RCA, draw a circle. And that's the end of the RCA and that's it's our total calcium score which we got here is 183. You can see how the breakdown is in the Agatston score between left main, LAD, circumflex and RCA. Here we get a volume score as well that's a different algorithm which is used. But in general what we report are Agatston scores, not volume scores. So once again the Agatston score is 183 here. Now we've talked about Agatston scores, let's go back and look at visual estimation of coronary calcium in another couple of cases. Let's move on to a few cases where we look at visual estimation of calcium scores. The so-called VCEC, Visually Estimated Coronary Calcium Score. So this is a PET stress myocardial perfusion imaging study. You can see the fusion images with PET images overlaid on CT images here. In general when we're assessing these CT scans for coronary artery calcium, we're going to move over to a different module in this program called CT viewer. This program is 4DM by Envia, and we run this as well through our Intellispace portal. For our PET CT scans in general, we will do our visual estimates of calcium score within that program, so we don't have to close the program and open another program. But you might equally view it in a different program in the PACS or in a different advanced visualization software. But here, let's take a look at a visual estimated calcium score within 4DM. So let's move over to CT viewer. This is going to show you just the CT images. In general, the images you're seeing are going to be quite small here. There's a lot of black space, which is not the most well-used real estate by default. So I'll simply click on the magnification here, and that'll give me a bigger image. I may want to make it even larger. So to zoom up more, I can click on this little arrow next to the magnifying glass and zoom up a little bit more. And to scroll up and down, I'm going to choose typically the coronal or sagittal, pick up the blue line here and drag it up and down. Now here, I'm not going to measure calcium scores. This attenuation correction scan was not performed ECG-gated. You can see how blurry the coronary arteries are here. But it's good enough to visually estimate the burden of coronary artery calcium. Moreover, because the PET acquisition is typically 10 minutes and the patient is going to be at various phases of the respiratory cycle, we'll acquire these images in a free breathing manner. But let's take a look here. Here, we're not going to draw any circles. We're just going to look up and down in the vessel. And I'm seeing here a lot of calcium in the LAD. Seeing a decent amount of calcium. And unfortunately, I can't scroll up and down and point at the same time, so I have to stop scrolling up and down to point. But you can see calcium in the circumflex here as well. And I'm seeing a decent amount of calcium in the right coronary artery too. So I would estimate that this calcium score is not in the highest level of those six levels which we talked about, but maybe in the second highest level. So I would estimate a visual calcium score here between 400 and 999. And I would include that in our reports. In Epic, and anyone who has access to Epic is welcome to borrow the templates which we have. I can simply click here on 400 to 999, and it'll generate a sentence that the visually estimated calcium score is in the range of 400 to 999. I'm also going to want to comment on which vessels. And I see calcium in the LAD. I'm seeing calcium in the circumflex and the RCA. Do I see calcium in the left main? I think probably not. So I would comment on LAD, circumflex, and RCA, and a visually estimated calcium score of 400 to 999. Let's move on to another case. So here we are in another case. This is a sarcoidosis study actually. Just because it's a sarcoidosis study doesn't mean that we can't estimate a calcium score, even though the goal of the study isn't primarily to assess coronary artery disease. It's additional information that's there for the taking and may be useful for the patient. Let's go to the CT viewer again here. Let's zoom things up. We can scroll up and down again. Now, this is one of the challenges which we have sometimes. You can see here, again, everything that's bright doesn't necessarily translate to coronary calcium. It looks like this patient has a device, which is quite common in sarcoidosis patients who have implantable cardioverter defibrillator devices. I can also see coronary calcium here. You can see in the LAD distribution there's calcium here. Scroll up and down. Probably in the left main there's calcium here as well. Just after the takeoff of the left main there, you see LAD as well. It doesn't seem like there's as much calcium as there was in the previous patient. I take a look here and I often will put my finger on the monitor just to follow the vessels. I'm seeing minimal calcification in the circumflex here, very little. The RCA may be more difficult to interpret here because of the leads. I'm not really seeing anything in the RCA. We have a little bit in the left main. We've gotten the LAD. My question here is whether I should say it's greater than or less than 100. I could give this the 10 to 99 grade or I could give this to the 100 to 399 grade. I think I would assign this the 100 to 399 grade, although maybe I would study it a little bit longer if I were doing this clinically. I would talk about calcium in the left main, the LAD, and a very little bit in the circumflex here. Okay. That's all for this case. Let's look at another case. Let's use a different program this time. Here I'm using the CT Viewer program within the IntelliSpace portal. That's a program which is not going to pull in my PET images, but if I just want to look at the CT scanner, that's adequate. I can save time and it's going to load the CT study more quickly. I'm going to be looking to your side because I'm viewing this on a different monitor. Let's scroll up and down and look for coronary calcium here. I'm seeing areas which are bright. I see the leads. I see spine. Let's look for coronary calcium. I'm not seeing too much here. Let's zoom up a little bit more. I'll choose the zoom tool, get a closer look at the heart, and then go back to our scrolling tool. We're just not noticing calcium in the coronary arteries here. I would give this a visually estimated coronary calcium score of zero, VCAC zero. Simple case. Okay, let's look at one more case. Again, I'm looking at it in the CT Viewer package. And again, scroll up and down, get a sense of the lay of the land. I'm not seeing calcium right away in the RCA here, but I'm seeing other stuff close to it. I'm thinking that that's pericardial calcification. You can see the pericardium here, and that's calcified. Again, not everything that's bright is coronary calcification. I'm seeing something in the LAD here too. I'm seeing something here, but I think this is cranial to the coronary arteries. I'm not going to count that towards a visually estimated coronary calcium score. Seeing something here, but that looks like there's a little bit of calcification of the wall of the left atrium. Really, what are we seeing here? In the LAD, we've got a little bit of calcification here. In the RCA, we've got a little bit of calcification here. Again, that's pericardial, that's not coronary. It's really just that LAD and RCA calcification. I'm going to give this a visually estimated coronary calcium score, VCAC, of 10 to 99. I think you'll find as you read more of these studies and report them, you'll get quicker and quicker. At this point, it takes less than a minute in general for me to assess coronary artery calcification. In some cases, if the VCAC is zero, I can do it in a few seconds. It's really valuable information, which we're providing for the patient and their physician, which is complementary to the perfusion information, to the flow information, to the functional information, which we acquire in a PET study. We do this in our SPECT-CT studies as well. Really, we've gone through the basics now. We've discussed the history of CT, how we got to electron beam CT scanners, where calcium scoring was invented, basically translated that over to multi-detector row CT scanners. There are certain peculiar scan parameters, which are really designed just to mimic what we had done in the past. But the epidemiologic data is based on those scan parameters, so we shouldn't play with them too much currently. We've talked about determining Agatston scores from ECG-gated studies, and we've talked about, in hybrid imaging studies, not dedicated calcium scoring scans, but CT attenuation correction scans or image registration scans, determining a visually estimated coronary artery calcium score. With that, I think we can end our session today, and I thank you for your time and attention.

coronary artery calcium

CT scanners

electron beam CT

calcium scoring

Dr. Andrew Einstein

CT attenuation correction

coronary artery disease

calcified lesions

risk assessment