 So the first thing I'm going to do is I'm going to actually go back and take a step back and show where 3D has come and what the technological impact happened in terms of pro-technology and computer technology in order to get to where we are right now. And then I'm going to talk and I'm going to start, this is going to overlap a little bit with Max's talk on how to acquire 3D and how to optimize, okay? And what we're going to do is try and get you sort of an understanding what the technology is and then how to actually do it. One of the things that we have to understand is that the technology of echocardiography has actually advanced a lot since we initially started with the 1D probe and all the way to the real-time 3D that we actually have now. What I want to take us back to the basics is that you've got to remember that for ultrasound, the basics of ultrasound is that you have a transmit beam and a receive beam. You have an electrical signal that in the head of the transducer gets converted into an ultrasound signal that goes out to your structure of interest and then gets reverberated back or echoed back to your transducer which then translates that ultrasound signal back into an electrical signal, okay? That's the basics of ultrasound. So what you've got to remember is that no matter what happens to technology you're always limited by the basics of the speed of light, okay? Or, sorry, speed of ultrasound, sorry. Sorry. Too much Spider-Man yesterday. So you're limited by ultrasound, okay? Now, original transducers just had a single element, okay? We got to 2D transducers because you could then put a row of elements together, okay? So we went from the original ultrasound which was one element to a row of elements. Now, when you have a row of elements you can now steer and direct your beam. This is how you can create a 2D image. This is a pretty complex image but what you want to know is that how you do is you can delay your activation of each of your elements and that gives you a beam that you can actually direct, okay? And then the same thing is when the signals come back after the different delays that they've been sent out you just reorganize that and you get your image, alright? So because you can take a 2D image, you can actually take multiple 2D images and fuse those together to create a 3D image. So that was the concept in the original 3D images. So what you would have is you would have a 2D probe, you have a localizer outside either be it a physical localizer or you can have a transducer that rotates automatically and you would do a series of 2D images which would then fuse to create the 3D image. So this is the original 3D, okay? Now how do we get from taking multiple slices of 2D images and fusing them together to our current technology? Well there had to be transducer changes. Now you'll hear the term matrix array, okay? That's what our current transducers look like. And initially they were called sparse arrays and the reason they're called sparse arrays is because not every element, so we've gone from one element to a row of elements to now a matrix of elements but originally not every element could be active, okay? And part of that was because every element in these original transducers had to be connected and processed in the transducer head. Now when you do that, if you have 4 to 8 elements that's doable but once you start talking about 400 elements being active that becomes a very large probe that's a chord for every single element. You couldn't do that. And so in order to go from the sparse array to a fully active matrix array there had to be a lot of changes in technology and I'm going to show you a slide of that. So this is to show you what these original sparse arrays were. The systems, like this is the ultrasound system. This is huge. It's a size of, it's bigger than what we have right now. It's like two carts together, okay? I don't know why my computer is doing this but ever since I updated it. So this is the original sparse array. You can see the picture is not playing properly and I'm not sure why it's doing that. But you can see that this is the original 3D. It looks a lot like how our current 3D looks but if you can see this is actually quite impressive because the date on that is 1998, okay? So how did they actually move from having these sparse arrays to fully active matrix arrays? Well one of the things they did was they broke up beamforming and the image forming sort of parts of the transducer. So they actually created these patches where they would take a group of the elements and patch them or group them together and they needed fewer cords or connectors to the ultrasound system. So they broke up the process of the data from part of it being in the transducer head and part of it in the cart itself, okay? This is the most simplistic way of doing that. So by doing that you can actually reduce the number of cords or connectors that you need between the transducers to the computers inside the machines, okay? So they broke up some of the tasks that the ultrasound machines had to do. So some of it's still in the transducer head and some of it's in the cart itself. And this is how you can get the matrix fully active transducer, okay? So this is where we were in 2002 with a surface image of a mitral valve. Now one of the things they also had to do when they were doing with these transducers, this is a TE probe broken down, is they also had to deal with the heat issues because when you have a lot of active elements, it created a lot of heat. And so they had to create ways of dealing with that heat dissipation. That was one of the challenging tasks for many of the transducer, for many of the companies that were doing that. Now this is your original TE probe and this is what the 3D TE probe looks like when it's broken down. And you can see the little elements going into the cord, into the body of the transducer head here. So when you look at the mitral valve in itself, that's one of the structures we've actually... This is what the mitral valve looked like in 1953. This is in 2003 with the multiple slices that are fused together and you can see each of those slices that some poor research fellow had to fuse together to create this image and you see there's a little bit of a prolapse over here in the mitral valve. This is where we were in 2012 and we are in 2019 in terms of the technology now. We are getting beyond just creating high volume rate images of the mitral valve but now we're actually getting these kind of post-processing imaging visualization things that can make the tissue look realistic. So here this is the same valve over these two on the bottom where's my mouse, is the same valve and this is a photorealistic image of the same valve and you can actually see the details better, you can see the chords better, you can see the prolapse better with this and also we have new visualization techniques for looking at color. Now, that's my brief introduction to the simplistic way of technology development. But I'm going to direct you back to the guidelines for what's going to happen in the next part of the talk and what Max will continue on. One of the things I do want you to remember is that terminology may differ between each of the companies but the concepts underlying them are the same. And that's what you have to understand and take out of these talks. They may call them a little bit something a little bit different but the underlying principles are the same and how they do it may be a little different but some of the adjustments for the images, acquisition images, they may be a little bit different on the cart but the actual underlying principles are the same and that's what you really need to take out because if you understand the principles it's like doing math, if you understand the principles you can do math. This is just a quick slide, it's one of my pet piece. I want people to realize that multi-plane mode and bi-plane mode are 3D modes so a lot of times I hear, oh, we don't have 3D on our probe but we do. You have this on there and that's just a little thing. Now these are the steps I kind of go through when I'm actually working with fells and resins in terms of how to teach 3D and I'm not going to go through all of the steps because some of that is going to be covered by Max but what you want to do is you want to go through these steps in your head as you create your 3D image and this will help you. You want to first focus on image optimization and then you want to decide what your acquisition mode is and it breaks down into whether you do zoom, narrow volume, wide volume, single meet multi-beat and we're going to talk about those terminology and then whether or not you're going to add color onto it but essentially you're going to optimize your image then you decide what acquisition modes you're going to use and then you're going to decide if you want to render a crop before or after you acquire your image and then you actually want to present your data and then do any further analysis. One of the things I'm just going to keep on emphasizing to everyone is before your 3D acquisition your 2D image has to be optimized. If you look at the image on the right you can't see that anterior wall over here but on the 3D and so once again on the 3D image that's not going to make it better. 3D is not going to improve something you can't see on 2D. If you see here this is a nice amyloid heart thick with pericardial fusion you can see the image is much nicer. You've just got to optimize that 2D image as much as you can in order to get a good 3D image. One of the things is a lot of times people don't use the other planes that you can see on acquiring 3D to ensure that they've optimized their 3D image and this is your reference image is what I always call the reference image these are your cross planes and these planes that are perpendicular to your image are actually what you really need to look at to ensure that you don't have stitch artifact. What you also want to do is you want to make sure that when you're acquiring that you've optimized the image on these other cross planes because these are the walls that you don't see on your main image and that's where you want to tweak a little bit to make your 3D picture better. Now I have this slide up because what I want to point out is in a 3D probe and you can pretend this is a TE rather than a surface probe is that your reference image is going to be usually your X plane and then the cross plane which is your Y plane is going to be your second best resolution image and then your elevation is going to be your worst. So these are the X, Y and Z planes and the one thing you want to know is that the spatial resolution in the X and the Y are going to be the best your Z is usually your worst. So if you have to do a measurement you want to make sure it's either in your reference image or the cross plane perpendicular to that and not the one that's and I have an example of that to show you. The other thing you have to realize is that if you are perpendicular to your beam such as the mitral valve on your TE probe that gives you a better image than if you're parallel. This is why we have dropout or issues with imaging the aortic valve versus the mitral valve on TE because the aortic valve tends to be parallel to your beam versus the mitral valve on TE tends to be perpendicular. So that's the other factor that will influence your image quality and you've got to realize that. Now most of the time when you're doing narrow volume or color Doppler so that's no longer an issue and with some of the probes or carts that are out there now you don't even need to Nestle go into multi-beat at this point because you can get such high quality data with the single-beat or high volume rate. Now let's go back to the individual mode so I've used the term zoom narrow sector and wide sector before. Zoom and I have some examples which we'll show you is usually the smallest pyramid that you have. Narrow sector is an intermediate-sized pyramid and full volume or wide sector is the largest. It used to be that these volumes were pretty fixed on the machines. Once you went into it you couldn't adjust them but now I think on most of the carts they're a starting point. They're a fast way to get a certain size of a pyramid and then you can adjust them bigger or smaller depending on what you're trying to acquire. So they're a starting point for you now and that's the difference between all of them. Okay. I think I touched upon this earlier when I said that the different companies have different ways of knowing this. This is kind of a rough guide to how they're kind of equivalent across the different systems. I don't have the Siemens up here but I do have sort of the Phillips and the GE systems listed here and this is sort of how you can adjust and remember these are just starting points in terms of the size of the pyramid you initially get started and then how you adjust further. Now zoom is the smallest size pyramid so it's actually very good for small structures like valves such as the aortic valve and you can see this is a zoom back in the day when you get the smallest size pyramid that you actually have and you're going to see that it rotates and you can see that it's a very small size pyramid or 3D volume within that larger set. Let's go ahead to the minute. See that? And see and all it has is the aortic valve in the status set. Especially for valves, ASDs, VSDs, small, fast moving structures the one thing I have to say is when you use zoom mode is please leave some of the external structures around it so if you're doing the mitral valve leave a little bit of aortic valve in there or left atrial bendage this helps you orient and it helps someone who's looking at the data set after you've acquired it orient to where they are too. One of the things people sometimes tend to do is they crop in too much you don't know what you're looking at or what the external structures are or how to orient so you want to be careful that you don't crop in too much. Narrow volume is a thin long pyramid and what we actually tend to use a lot is if we're doing procedures and stuff and we need a quick look at something it's also called live volume mode as you can see here it's rotated and you can see that it's just a very thin slice of that. Now wide angle is the biggest pyramid that you can get. We usually use it for acquiring large structures like the left ventricle, right ventricle if you want to get the atrium in the ventricle or the whole heart this is the size of the pyramid that you're going to start off with. Now why does pyramidal size matter I'm not sure why it's doing that but the size of the pyramid gives you about a 30 hertz you widen that a little bit and still you care about frame rate because most data sets are gated by ECG. So diastole is always captured in your data set so your volumes for LVN diastole volume is captured but if you have a frame rate that's skipping you may miss end systole. Your end systole volume is going to be bigger than you expect and your ejection fraction is going to be off. So this is where you care about volume rate because it's the end systole. Now if your system is not gated by ECG and it's a time based system that's a totally different question but most systems are gated by ECG and so this is where it matters okay. So you want to get the highest volume rate that you can especially if you're doing volumes or ejection fractions. All right. So this is just a summary to recover the same ground that I talked about smaller volumes that you want size of pyramids you want for valves, interatural septums, inter-particular septums, larger volumes you want for bigger structures such as the ventricles or the whole heart. Okay. Whether or not you can do single or multi-beat with the different volume rates will actually depend on the machine that you have and I'm not going to spend a lot of time on this. I want to spend more time on this slide here the trade-offs between volume rate just to reiterate the point. So let's look at this left-most image here. You have a set volume rate that you have you have a set pyramidal size and this is the resolution you get. Oh, sorry. Oh, that might be very competitive for me to try and talk about the drilling but now on the second slide I've made my pyramid bigger I want to keep the same volume rate I'm going to sacrifice visual resolution on your image. If I actually then shrink the pyramid size back down but not as small as my original one I can improve my resolution a little bit. You've got to keep those trade-offs in mind. Now single or multi-beat so when we talk about single-beat you take the same pyramid you acquire the exact same structure multiple times in a period of time. That's single-beat. When you do multi-beat what is happening is that you are acquiring multiple pyramids over different portions of a structure and then fusing them together to make one big pyramid because each smaller pyramid can get a higher quality data set and then by fusing multiple small high quality data sets you can get a large one that's high quality. This is what multi-beat is. There are methods to try and improve quality or volume rates or volume rates within the limitations that we have currently. I'm not going to do a deep dive into the science behind it but some of them that are built in are in the background that you aren't going to know about but parallel beam farming, frame reordering multi-line transmission, high pulse repetition you're going to see that in the literature but the things that you can control are something called there's an interpolation mode that's on the machines as well as a virtual array mode. So parallel beam farming this is just to make you aware of what's going there that you have one transmit beam so one beam comes out of the probe and you have multiple receiving beams and that's how it increases the volume rate because the number of receiving beams lets it form an image better. The problem with this is the center of your beam is going to have very high quality there's a lot of noise and artifact with this mode but you won't be able to control this this is part of the machines. Frame reordering is technology that's actually out there it's similar to what they do in MRI you actually sample around the different parts of the cardiac cycle and then you take the images and you reorder them and you can see here with this frame reordering now they've skipped you can see the cusp over here opens first before the other cusp open and this is how you can actually increase your volume rate. Multi-line transmission another technique once again you have two source generators and you set a beam and you can actually image two different spots at the same or get two different spots at the same time once again you speed up your transmission and then high pulse repetition this is where you can it's mainly used in color from what I can see in the literature and it improves your color volume now in terms of things that you can actually see on your cart the high volume rate tool the machines are interpolation methods so I'm going to spend a little bit of time on this so what happens in interpolation methods is you guess half the time so not all the matrices of the head of your transducer are active you only sample some of them and if you're only sampling some of them you can speed up how much you acquire but how do you get a full image if you're only sampling some of the spots you guess the machine has an algorithm and it comes up with a pink spot that's why in some of these modes it's a little bit fuzzy initially when you acquire that then in the second cardiac cycle what happens is it samples the ones where it didn't the first time and then it decides is the new image better than its guess image and then you get this fusion image where it's interpolated both guesses and real images together to give you that frame to give you that image and that's how it speeds up the acquisition last thing that you can see is not showing up very well but what I had is there's something called a virtual array or virtual APEX that you can see on your machines and if you look on the cartoon usually this is your head of your transducer with your elements and the beam starts right from the elements the other one is there's something called virtual source where they set up up in space in your probe so here's your element and your virtual array starts back there and because by the time it goes through the elements of the water you can actually get a narrower box with a wider box so if that makes sense in the same way and you can speed it up and I actually just had a picture showing a triangle with a triangle here and the one where it is not where you have that virtual array where the top is actually flat on the machines you can actually see that you can adjust it that's going from rest to spend you can see the top of the machine change now I'm getting towards the end I'm not going to talk a lot about these because Max is going to cover some of this in his talk but one of the things I do want to point out is that there are blooming artifacts that you can actually see so here's a clip going in and in terms of blooming you can see that this part of the wire is thicker than this part of the wire over here and that's some artifacts similarly when you look at the repaired you can actually see sometimes that these stitches are probably not that chunky but they appear that way and that's a blooming artifact and shadowing now shadowing used to be a thing we tried to get rid of now people are putting it on the cards deliberately so I'm not sure if we're going to be calling this an artifact in the future but that's what shadowing is so before I finish I'm just going to have put this back up and I'm going to let you move on to rendering, cropping and image analysis with Max thank you