October 2017 Webinar Transcript

Hello everyone, this is Stephen Buttars. I'm the product manager for preclinical ultrasound here at Fujifilm VisualSonics headquarters in Toronto. Thank you all for joining us today for this webinar, which will be on the Vevo Strain analysis software; which is an excellent piece of software for digging into subtle cardiac changes – I mean your animal models, and will be presented by Kelly O'Connell, one of our application scientists. 

Before we get started, just some housekeeping notes about the webinar. We are recording the webinar, and a recording of the webinar will be made within a few days after completion. All the phone lines and audio lines are muted for the duration of the webinar. If you have questions to send us, please use the questions panel that's on your webinar window. We will be monitoring that during the webinar, and we'll have some time at the end for answering those questions and any other discussion probably around hopefully around 10 or 15 minutes, once we're finished with the presentation. 

So, I'd like to introduce Kelly O'Connell. Kelly O'Connell has been with us here at Fujifilm VisualSonics for quite a few years now, gaining lots of experience as a field application scientist. She did her PhD work in molecular medicine at the University of Maryland in Baltimore; studying manipulation of mitochondrial function and influencing heart failure and various animal models. Afterwards moving to Brown for her postdoctoral work, working on pulmonary hypertension and RV dysfunction. And throughout her time at both the places she had the good fortune to be working with VisualSonics’ Vevo imaging systems - both the Vevo 770 and the Vevo 2100 system. And she wrapped up her postdoc; we snapped her up for her skills and hired her as an application scientist. 

So, with that, I will pass the webinar presenter rights over to Kelly, and she can take us through her presentation. Take it away Kelly, thank you. 

Good morning, everyone. Thanks Stephen for the introduction. So today I'll be talking about the Vevo Strain software and understanding subtle cardiac changes. We'll start with a discussion of what is strain and using strain, and as a sensitive early marker of dysfunction, and getting into some of the fundamentals of the calculations that we used in our strain software. Then I'll move to interpreting the strain results including functional information, the synchronicity parameters, and global regional changes. Finally, I'll end with some information about diseases for which strain is abnormal; and I do reference papers throughout this presentation. So, at the end there's a final slide that contains all the references that I use so you don't have to worry about writing anything down while I'm speaking here; it's all summed up at the end. 

So as we all know from using ultrasound, it is a in Vevo non-invasive tool to measure real-time changes in function, especially when it comes to cardiology. We can see the anatomy very clearly because of the high-resolution ultrasound transducers that we use, and we can get a lot of functional and data. 

In terms of echocardiography, we have four basic views that we use. So apical 4 chamber view in ‘a’, that is mainly used for diastolic function. Then we have parasternal long axis and short axis imaging that you can do Vevo Strain on. And then the aortic arch view in 'd' for the other parameters. And this talk is mainly going to focus on strain in the long axis, but I do touch on a short axis strain as well. 

So normal 2D-ultrasound and M-mode imaging can give you a lot of important functional information. These are mostly systolic functional parameters, like ejection fraction, and fractional shortening changes in volume; things like that, and really easy ways of measuring using a tool called LV Trace. And that's great for really over changes and function, like ejection fraction. But if you want early subtle changes, that's when we use to start to use strain imaging. 

So how is it useful? It's useful in clinical decision making, because it is more sensitive than LV ejection fraction as a measure of systolic function. And it's also useful preclinically as a sensitive indicator as well. So this was demonstrated a number of times, but this particular paper shows a diabetic model. And when you look at diastolic function, which tends to also be earlier than systolic function in certain models, we can see that normal parameters of diastolic function like the E/A ratio, and the E-Wave. Those you see changes in there about at six weeks after induction of diabetes. But if you start entering that same animal model into strain software, now you can see that the changes that are happening actually occur as early as one week after induction. So this is really a an early way of detecting subtle changes. And longitudinal strain in particular is a sensitive marker of dysfunction, which we'll get into in a little bit. 

So, let's move right into it – Vevo Strain software. I want to mention that this talk is mostly focused on some of the advanced quantification of the parameters, but there is a workflow guide that we've done a webinar for in April 2015. Actually Stephen, who introduced me did this webinar for you. So if you have questions about how to use the Vevo Strain software itself, this is a great tool and resource for you to perform strain. I also want to make sure that you have your Vevo LAB software up-to-date. 

All of the work that I'll be talking about is using Vevo Strain 2.0. The current version that's on our website is version 3.1.0, so just go to our website, we always keep that updated with the current version and make sure that you've got that Strain 2.0 on your software. 

Alright, so what is Strain? Strain uses speckle. Speckle is an interference pattern of natural acoustic markers in ultrasound. So, you can kind of think of it as if you're looking at a starry sky. You've got all of these stars. But if you want to look at just very specific ones, then that would be the speckles of interest that you'd be tracing. And if you know where those speckles are, then you can follow how those particular points move over time. And that's what we see when we're looking at strain. So instead of looking at you know seasonal changes of the starry sky, we can look at three different planes of movement of those little speckles in a frame-by-frame manner. And what that gives us is displacement.  

So if we can follow the same point and how that moves between frames, then we end up with the displacement of that speckle. So in terms of cardiac analysis, what you do in the Strain software is you trace the endocardium and the epicardium. And this particular side on the right, is just looking at the endocardium, and these yellow dots follow a specific speckle on the endocardium and how that moves frame-by-frame. 

As I mentioned, what we get is, we get a distance that the speckle moves between two consecutive imaging frames. And if we know the distance it moves between the frames and the framerate, then we can get our velocity. This all occurs in three different planes. The first one is the radial plane - this is going to be perpendicular to the reference border that you're using; in this case - it's the endocardium. Then we have a longitudinal direction; this is tangential to the border that you trace. And thirdly, we have circumferential strain. So that's more of a twisting motion that's also tangential to the curve when you do strain in a short axis plane. 

What are the fundamentals of myocardial strain? Strain is a dimensionless quantity, and it's a fractional change in your dimension. So, this equation is what we use throughout all our strain imaging. Essentially, strain is the length at a given time, minus the original length, divided by that original length. 

So, it's a measure of contractile function. If you have a positive strain, then that must mean that your length at a given time is greater than your original length. In the original length in our software, and for most strain softwares is always taken at the end of diastole. If you have a negative strain, that must mean that your original length, or your length at a given time, is smaller than your original length. So a negative strain is a shortening of the myocardium, and a positive strain is an elongation or a stretching of the myocardium. And there are two types of strain - there's Lagrangian strain, and or Eulerian strain - also called Natural strain. So Vevo Strain uses Lagrangian strain and the only difference between these two is the idea of a reference frame. So Lagrangian strain, which is this equation here, shows that we have an original length. That original length is the same point, no matter when we look - and that's always end diastole. Eulerian strain has an always changing reference frame. So, most strain use Lagrangian, but if you choose, or you are interested in doing Eulerian strain, there are equations that can convert Lagrangian to Eulerian. So, you're not limited to only using Lagrangian - but that's what Vevo Strain uses. 

So, as I mentioned, you have negative and positive strain, based on the length at a given time. When we look at longitudinal, if you're doing parasternal long axis, and you do strain on that image, you will have a negative number because during contraction, the entire length of the heart is shrinking. So that means, your length at a given time is smaller than your original length; so, you're going to have a negative value. Normal negative values are roughly around -18%. Radial strain is a positive number because during systole, the myocardium gets thicker. So that thicker means that your length at a given time is now greater than your original length - so you have a positive number in strain. Circumferential is different, because by convention, a counterclockwise motion in a twisting circumferential plane is considered positive. 

Interpreting Strain Results: Functional Information 

Here, we'll talk about the functional information that the strain software gives you. When you first take your data and you analyze it in Vevo strain, on the very first page that you open up, you'll get a panel that looks like this. So, what we're seeing here is, in blue, we have change in volume over change in time. And in this reddish orange color, we have our volume. And at the top, you'll see a lot of the systolic functional parameters including ejection fraction, fractional shortening, cardiac output, and then this global longitudinal strain number; which I'll get into a little bit later. What I wanted to point out here is not only do you get your ejection fraction, and that kind of information. But this change in volume, over a change in time, that is actually your ‘E’ and ‘A’ ratio that you would get if you were doing diastolic apical four chamber view for inflow. Instead of velocity, you are getting a change in volume over change in time. We know this because if you take a look at the ECG in light green, that's the P-wave. Your ‘E’ and ‘A’ is always separated by that P-wave. This is still dependent on heart rate conditions; orientation of your image. So if you have a slow enough heart rate, and usually that's around the 400 beats per minute frame, then you can get a separation of your ‘E’ and ‘A’ on you the strain software. 

The parasitotic functional parameters of strain have been shown to be at least as good, if not better, in terms of a correlation to a gold standard cardiac function test which is cardiac MRI. So in this paper, we see that two-dimensional echo, which would be your sort of long axis or short axis M-mode picture, and if you look at end diastolic and systolic volume and ejection fraction, and you compare that to speckle tracking, which is using strain software, the correlation to cardiac MRI is often better with a speckle tracking software than it is with just simple 2d echo. 

Interpreting Strain Results: Qualitative Synchronicity 

When we look at synchronicity on Vevo Strain, what we're seeing is essentially a heat map of how the walls are moving. So in this picture, what you're seeing is we're looking at radial velocity, and we've chosen three points along the endocardium. The blue dot is on the anterior wall, the red at the apex, and the green dot is on the posterior wall. We've chosen only those three to look at throughout each cardiac cycle - that's represented in this heat map below. So if it's red, that's essentially a positive velocity; so that's the velocity in in this case - the radial direction during systole. Then when it's blue, it's a negative velocity; which is the velocity that's happening during diastole. This is qualitative because it allows you to see the level of synchronicity, and because this is radial directions, all the walls are moving pretty much together. You can see a nice solid plane of red and then blue. And if you compare that to say an infarcted animal, you can see visually that there is dyssynchrony; we don't have these planes of red and blue that we do in a normal animal. 

This can also be represented in three dimensions using the same sort of heat map in a normal versus infarcted animal and you can see very clearly the amount of dyssynchrony that’s present. 

Interpreting Strain Results: Global and Regional Changes 

Let's get into global and regional changes. So, there's a second page on the strain software called Quantitative Segmental Synchronicity. What this does is, it breaks down the heart into six different segments, and allows you to look at four parameters: the velocity and displacement, as well as strain and strain rate, and this is a long axis picture. So, what we get is velocity on the top, and longitudinal velocity on the bottom - so radial, and then longitudinal. Here's where I want to get into some common definitions for echocardiography. 

So, when I referred to a mid-base in the heart, and this is about the segmental synchronicity page, what we mean is the midpoint between the endocardial points of the anterior and posterior wall. So, this cartoon on the top, it would really be right here at the middle at the red line. So that would be the mid-base. And that's important because the way the software defines the apex is the most distant point from that mid-base. So when you trace a long axis strain tracing yourself, it's defining that apical point based on where that mid-base is, which is the distance between the endocardium and the in the posterior and anterior walls, the mid-base point there. And then that apex is an important area because the anterior wall is going to be from that apex to the top of the heart. And that's divided into three different segments. And then from the apex below is the posterior wall, which is also divided into eight different segments, and each segment contains eight equally spaced points, resulting in 48 points total. So that's how the segmentation happens in the heart. 

All right. Some of the parameters that are in the Segmental Synchronicity page are peak % strain. In this panel that you'll see on that page, this is longitudinal strain, but you'll see it is also present in radial strain. And peak percent strain is the highest strain value in that segment. 

 So if we look at this read out here, this little bullet point on each curve is the peak percent; that's the highest strain value in that segment. There's also the average percent peak strain, which is not this down here, this is average curve, and there's a value there. Average percent peak strain is this black line that appears on the curve. That black line is an average of all the individual full values within each segment; it's not an average of the peak percentage. So, it does not always equal the average of this and if you take these values right here and you'll see that they don't average to -19, but the -19 is an average of all the raw data essentially of each one of those segments. 

There's also maximum opposing wall delay. So what that is, is a difference in the time to peak, which is here between the opposing walls. So if you look at how those colors are defined, the green and the blue would be opposing walls, yellow and light pink are opposing, and then pink and light blue are opposing. So, this maximum wall delay of 11 is the difference between the anterior mid and the posterior mid. So that's how we get 11. It's just the maximum difference between opposing walls and that can be a measure of dyssynchrony, and we'll get into that in a little bit. And then, another value we want to pay attention to is global longitudinal strain. This is on the first page of the Vevo Strain output parameters. And essentially the way this number is calculated is different from the segmental strain values. What it's doing is it's applying the strain formula to the length of the actual curve that you measure. So the length of this inside curve or the endocardium curve and how that changes in systole and diastole, it's plugged into this equation to give you that GLS number. So that's how that number is calculated. 

Identifying Landmarks in Longitudinal Strain 

All right. Let's identify landmarks in longitudinal strain. Say we have a long axis picture and you plug it into the strain software, and you go to the segmental page, and you see a curve that looks like this. You know, what does this mean? So, we're looking at longitudinal strain, not radial, longitudinal in this case. And we're only looking at the endocardium, but you can do this for epicardium too. 

So right off the bat, what we see is our peak systolic strain (PSS). That's what we just talked about on the previous slide, which is the maximum strain for that segment. But we also have a couple other landmarks that I want to identify. The first one is you have this little bump at the beginning of your strain curves. See how it rises - so it becomes positive, which means that you have lengthening, and then it contracts. So, there's a small area where it's positive here and that's normal. This is a normal heart, but it becomes more pronounced in a diseased animal. You can also see where the mitral valve closes, and that's based off of where your ECG R-wave is. 

And then another point is that we have the aortic valve closure - this is a really important timing component of longitudinal strain and it indicates the end of systole. And how do I know that that's where the aortic valve actually closes? Well, if we take that beginning change in volume over change in time and the volume graph from the from the first page, and we can compare that to the Wiggers Diagram, which is a great diagram to show how ventricular pressure volume and atrial pressure change corresponding to the EKG. If you look at the Wiggers Diagram, and then you look at the changes in volume and EKG on our output parameters, then you can see very clearly where the atrial ventricular valve closes, because that's the R-wave, where it opens just at the end of the GRS and where the beginning of Isovolumic relaxation is. And this is based off of your change in volume. So because this is E&A, I know that where this area is flat, that must be the beginning of Isovolumic relaxation. And because this is one of the lowest points in the ventricular volume based on the Wiggers Diagram, I know that that must be where Isovolumic relaxation ends. And then as we mentioned earlier, we have E, then the P-wave, and then an A-wave. 

So, knowing where those three points are is really important. And if you look at the strain parameters on the first page and we extend those points into the heatmap, now we can try to make sense of why the heatmap looks the way it looks. So, in longitudinal strain, you'll see that during Isovolumic contraction, the entire heart is not necessarily all red and all blue like radial was. And instead, if I bring up this map of where our points are, again, this blue line is the anterior wall, the green line is an apex, and then the red dot here is the posterior wall. So, what we see is the basal segments on the anterior and posterior are relatively similar, but they differ from what the apex looks like at certain time points, and this all comes down to myocardial mechanics. And what I mean by that is the Ventricular Torsion of the heart. 

So the myofibrils of the heart are differently oriented from epicardium, transmural and the endocardium. And we can see that just by the gross pathology here on the left, when we separate out the different layers of fibers, you can see that they're oriented completely differently. And then we can also, if we do 4D imaging, you can see that there is a difference in how the heart moves at the base and the apex. And this is the ringing motion of the heart during systole. Every time it contracts, you have the opposing motion. And if I go back to the heatmap, that's what we're seeing here. So during systole, you have a change in direction of the basal segments compared to the apical segments. And so that's why that heatmap in longitude looks the way it does. 

Recognizing Strain Patterns 

All right. So what is a what a strain supposed to look like for us? So if we look at normal longitudinal strain, so what we mentioned earlier is you have a slight area of positive strain before contraction. Then you have a negative strain value and then it goes back to zero at the end of your R-wave. So typically the values are, you know, less than or equal to -15%. It tends to peak at the aortic valve closure. So this is why that aortic valve closure time point is important because if it peaks too far left or right of that closure, that can indicate certain degrees of dysfunction. So normal longitudinal strain is less than or equal to –15, it peaks at the aortic valve closure and the walls are pretty much synchronous. What about dysfunctional longitudinal strain? What does that look like? So, let's put our normal strain values on there - with the aortic valve closure marked. The first common type of dysfunctional strain is something that looks like this green line. It pretty much looks like the same pattern as the red one, but it's smaller. So that indicates that you have globally reduced strain or less contractility during systole. You could have a strain value that looks like this, which is completely opposite of what a normal wall should be doing. Instead of shrinking during systole, it's actually expanding during systole - so that would be defined as Dyskinesis, or abnormal motion of the heart. And then lastly, we have Dyssynchrony, so to dyssynchrony, you need to at least be looking at two segments; you can't just look at one. And you can think of dyssynchrony as opposites walls doing opposite things. So that's what we're seeing here in orange and blue; we have opposite walls doing opposite things. While one is shrinking during contraction, the other one is elongating during contractions, and that would be indicative of dyssynchrony. 

So, here's the strain picture we've been looking at. If I see this for the first time and I want to know what's going on with this particular heart just by looking at longitudinal strain, I'll look at the peak strain percent values first. And knowing that the GLS, or the global longitudinal strain for this heart is -19.6, then I can see that it peaks at the aortic valve closure, which is about right here. I know that this must be normal. 

So, if I take a look at this tracing, then what I can see is, well, this looks awfully abnormal here. Instead of contracting during systole, it's elongating. So then combined with the fact that I know that this is -15.5, which isn't so very abnormal, it's sort of within the range. But seeing this yellow line that lets me know that this is a dyskinesis in this segment. 

If I look at this value, well, they all look like they're peeking at the aortic valve closure so they're pretty similar here. But then when I look at the actual average, this black line, and then I combine that with the idea that GLS is -13. Well, what do you think that would be? This must be globally reduced strain. 

And lastly, I take a look at this tracing. Well, there seems to be kind of a lot going on. I've got this pink curve on top that's positive, so that sort of looks like dyskinesias. But the other ones seem to be peaking around the aortic valve closure. But if I look at how those those segments are defined, again, I can see that, you know, the pink and the blue are actually doing opposite things and they're opposite walls. So opposite walls doing opposite things, knowing that the GLS is -10, which is an abnormally low number. I know that this has dyssynchrony. 

Interpreting Strain Results: Dyssynchrony 

And that will lead me into talking about what is dyssynchrony, and how do we measure it? So, it’s one thing to see that the walls are dyssynchronous and visually know that there is dyssynchrony there, but there are methods of actually measuring the magnitude of that dyssynchrony. The first one would be to look at the absolute maximum time to peak delay between the earliest and the latest segments. So here, you really only need two segments that you're looking at. And you just take the time that it takes to get to its peak strain value. And this is all longitudinal strain as a reminder. And so, you can get these values from your TPk (time to peak) strain values that you get in that panel on the software. 

Another way of measuring dyssynchrony is to look at the standard deviation of the time to peak over all the segments. So essentially what you're looking at here is, you know, how spread out are these segments? Are they all peaking at the aortic valve closure or are they very different from each other? And if you just look at the standard deviation of that time to peak, that'll give you an idea of the spread of your data. 

This was shown as a really nice example in a paper by Satsuki Yamada at the Mayo Clinic. So, what she did here is at the top of this panel is she had a wild type mouse, and subjected that to the transverse aortic constriction model, and followed that animal over time. And what she sees here in just one cardiac cycle that she overlaid on top of M-mode picture, is that, and this is long axis radial strain not longitudinal, is that she's getting about a 60, maybe 50 on average peak radial strain. 

But as the animal has greater and greater hypertrophy and cardiac dysfunction, we see a globally reduced strain. So, it just starts decreasing. But you can see they're all pretty much in sync. In the same paper, she has a knockout animal that shows a mechanical dyssynchrony. And so, same thing, in the animal that's not subjected to this type of stress, looking at radial strain, it has very synchronous, good movement. And then after you give the animal the TAC surgery, the strain decreases. So, we're seeing globally reduced strain. But not only that, you can see this change in how the segments are peaking. Some are starting to peak really late, and then especially after three months, they're almost all doing opposite things. So, this is an idea of how to display dyssynchrony in your model. 

And she further quantified that using those methods that I mentioned, so she did intraventricular delay in the time to peak strain compared pre and post, and she also did the standard deviation of the time to peak strain in pre and post. So, it's great to show both the qualitative picture of dyssynchrony, but then you can also quantify that I'm using those two methods that I outlined. 

So, there's a variety of diseases for which strain and strain rate is abnormal. I took this from a table from Feigenbaum’s Echocardiography book and it just shows you these big classes of cardiomyopathies and some subcategories of where they're changed because this is clinical data. But we can do the same thing pre clinically. There's tons of information out there about diseases and what is different in those. And I get asked a lot about, you know, I am doing X or Y model and what's different in my model. So, I put together this little list for you which gives you an idea of what might be changing in your model if you're doing something on this list. And this is by no means an exhaustive list of all the different pathologies that have changes. So this is just a nice little overview of some things that could be changed. 

So, if you're doing something like these diseases, you can take a look here and see that. Oh, okay. Well, if I'm doing an MI, that radial strain might be different. Maybe look at radial strain. So, use this as a resource for you about what might be changing in your model. 

And to summarize, what can strain do for us again? Well, it can lead to early diagnosis because we're not waiting for the global change of ejection fraction to go down. We can see an early change of dysfunction, which can tell us an early diagnosis of dysfunction, and that can lead to better timing of therapy. So, if you are doing something that is supposed to ameliorate a decrease in ejection fraction, well, if your EF already has gone down by 20 points or 20%, then perhaps it's too late for the therapy to be effective, so you can have better timing of therapy. And it also helps inform your decision making in your experimental design. So because you can start things earlier or later, it's just an early indicator of dysfunction which can help you design your experiments. And hopefully this all can lead to better outcomes and better data. And also, the key parameters that you would want to look at would be your global longitudinal strain, that’s that GLS number that shows up at the beginning first page of your strain software, you'd also want to see the peak percent strain. That's the maximum strain in every segment. So not the average, but the actual strain in every segment. And that will help give you some information about that dyssynchrony. So, you want to use or one of the two ways of calculating dyssynchrony as one of your key parameters to look at. 

Here is the table of references that I wanted to mention. So, any paper that I referenced in this presentation is listed here for your convenience. And with that we can enter the Q&A session. 

Q&A Session 

All right, great. Thanks, Kelly. That's an excellent presentation, a great overview of strain and all the possibilities we had some questions coming in while you've been even talking. One of them actually, I can answer it off at the top. There is a question about 4D on the Vevo 3100 and 4D strain. This is not possible with any of our software yet, but we are actively researching that. 

Another question came in, in the literature, have you seen I haven't, have you seen Kelly, any correlation of the global longitudinal strain numbers with the ejection fraction published in most data with such high heart rates? 

In mouse data? So ejection fraction, the only that pops into my mind is that the GLS number often goes down before ejection fraction goes down, but then when ejection fraction does decrease, your GLS number is already low. So, there should be, I don't know if anyone has done a specific review of that comparison, but there is definitely a correlation as perhaps like a side data, you know, like a side output parameter that you can look at. But yes, when EF decreases, your GLS is already low. 

Hmm. Okay. Interesting. Yeah, I don’t think I run that anywhere yet. We've got some technical questions; one that I actually might take here. There is a technical question about why strain is restricted to the endo and epicardium and not the mid myocardium or any trends, merely average measurements? That's really just down to the software that we developed. We have seen other packages that do the mid myocardial strain measurements, but the software package that we worked on was really restricted to tracking across the endo and epi. It's something that we kind of thought about a little bit, but to add that function would be quite outside of the boundaries of what we have in our software right now. 

Another question, “Will radial strain in a long Axis View show how the wall thickness is locally change shifting?” Kelly, could you take that one? 

So, the wall thickness and radial strain? So theoretically, if the wall is thicker in the radial direction, it will be represented with your tracing. But strain itself is a fractional change. So, if it's thickened and then shortens to the same extent that some wall is thin and then shortens, then you won't notice that. So, what you will get with strain is on the first output parameter, you'll get a measurement of left ventricular mass. So, if the walls are thicker, you'll see that the LV mass is changing. But strain itself, because it's a fractional change, will be an absolute measure of wall thickness. 

Mm hmm. Mm hmm. Yeah, well, that sounds pretty reasonable. A question here. I'll take one, actually, and then another one for you. Kelly, there is one question here about the quality of the images required to use this double tracking. It's kind of. And Kelly, feel free to jump in. I often say this is kind of an old software developer's joke, but the truth of it is, garbage in, garbage out. If you have images that are over gained, that have a lot of ring down artifacts, or noise in them, or positioning of the heart is not great in the image, then the speckle tracking is not going to be that great. 

Whatever you can do with adjusting the gain, adjusting the dynamic range, getting the frame rate up, frame rate is a really key component of speckle tracking, working properly, 120 frames per second at a minimum with a large rat heart. I usually like to see, you know, 250 frames plus four for tracking your mouse heart. To do the analysis with the Vevo Strain package, you really need to do everything that you can to really optimize your images, the quality of your images. That's been your experience as well, Kelly? 

Yeah, absolutely. And you want to make sure that you don't have your shadows in your image. And when you create your tracing and you enter the strange software, you can review how well it did just visually. So, if you play through your video, you can see that perhaps your tracing fell off at a certain frame area. And so keep in mind that, you know, double check your strain tracing and make sure that actually did a reasonable job and if not, just go back and you know, redo it. 

Yeah. And there is a possibility to go back and edit the analysis. So, once you place the tracing and you, you run it through the analysis and it does the tracking, if it's if it goes off rails somewhere, you can go back and edit and reanalyze. There's also one of the big things that we added in the updated strain package that Kelly mentioned was actually an undo function that was never there. So when you redo the analysis, if it's still off the rails or off the rails, even worse, you can actually undo and step back and try it again. And in another spot. 

Another question here, actually, this is going to be for you, Kelly. “How could we approach segmental analysis using the segmental analysis and strain to look at regional compensatory contraction after MI?” 

So after MI, you would want to look at the peak percent strain of each one of those segments. And so that one that I showed, I can actually go back to right here, in the examples, this one, actually sorry, not this one. So, this was an MI animal, and you can tell because this area on the anterior wall, like apical anterior side is red, almost three times as long to get to a peak velocity in this view. 

And that's right where your LAD Occlusion for an MI would be. So to approach understanding your MI and the compensatory contraction, you would look at your peak strain values here. So your blue and green, these are going to probably have a higher percentage than something that's closer to your infarct. So, you can compare the peak strain in each segment to each other to see what the velocity is of that segment; it doesn’t have to be strain. But it could be the radial velocity; longitudinal velocity. But looking at the segmental analysis and seeing what those absolute values are, relative to where you expect your MI to be. 

Hmm. I guess a little extension to that question, too, since we were talking about MI models and the differences in how they behave. What about areas where post-MI, you're getting wall thinning? Things to look for in, not only the strain pattern, I suppose, but also in the velocity and displacement patterns. Any tips on that? 

So if I go into this area here, so on a specific model like the LCA Occlusion, there are papers out there that show some of the specific parameters that change in that model. So, I'd really direct the listener to go and look at some of these papers to see what specifically changes for an MI. Every disease is a little bit different in in what changes. And so, these become a great resource to look at the specifics for your model. 

Seeing a couple more questions coming in that are related a little bit more specifically to the actual use of the software, how to trace the borders manually automated, where to place points. Just to reiterate, Kelly did mention during her presentation that we did do a l how-to webinar a number of years ago, and that is on our website. Yeah, there's the link to it. Thanks for bringing that up. Kelly. So that webinar that I did back a number of years ago, I believe that might be with the older version of the strain software, but the fundamentals are the same. That one actually goes through the point-and-click workflow to actually get your image into strain and actually place the points and do the analysis. So, anybody that has questions about those sorts of things, go back and take a look at that webinar and it is quite informative. 

The question here is about long axis versus short axis, doing both, doing one is, is it's sufficient to do one, do you need to do both? Thoughts on that? 

It really depends on your model. So, the most well studied image of doing strain is on the longitudinal side of things. Most of the data that we get for understanding the changes in strain all come from a clinical view, and so most of the data is on it. For clinically, it's really on an apical view, but the equivalent for us would be to do it in the long axis. 

So, the long axis is usually very sufficient for looking at changes. If you've only done the short axis, it doesn't mean that you have to go back and do long axis. Based on some of these papers, you can see that like here's a Doxorubicin cardiotoxicity: they looked at circumferential strain and that changed for them. So, here's an exercise-induced; that’s circumferential as well. So, it's going to be dependent on the model, but you really don't necessarily have to do both. People typically choose one or the other based on how expect there to be regional changes. So, in long axis, if you are looking at regional changes that span from the base to the apex, you'd want to use long. If you're looking at regional changes that are going to affect maybe septal and free wall areas, then you might want to look at circumferential or short axis strength. 

Mm hmm. Yeah. Yeah. That's certainly been my experience from use and from showing people and seeing things as well. People using the software, the majority do seem to go for the long axis image rather than the short axis. I think the general consensus maybe is that there's more information in the long axis than in the short axis. 

I think part of it is, you know, the GLS number (global longitudinal strain) number is a very robust number that's been studied a lot. We don't have that same background of data on a global circumferential strain number, but also the image you and the previous listener asked a question about image quality needed for strain. 

And as I'm sure most people know from doing their own imaging, that when you do a short axis image, it can be really hard to get a very clear septal wall and a very clear LV3 wall because of the sternum and the long shadows. So, if you don't have those walls very clear, then when you enter that that short axis picture into Vevo Strain, you're not going to get the best data out because it's not going to track very well under all those shadows. So, I think a lot of people choose long axis for that reason as well. 

Mm hmm. Mm hmm. Yeah. Yeah, I agree. And that actually leads to a little bit more of a technical question I can sort of how to answer, but then I'll toss to you, Kelly. There was a question that came in about the current version and provide the current version of strain software provides a global longitudinal and the global circumferential, but not a global radial strain measure. 

Now for myself as on the product management side, it I, I've actually never been asked that before and that's a good question. It's not in the software because we didn't put it in. We can do some investigation and see if that's possible. But Kelly, on your side with, you know, looking at papers and dealing with researchers, have you ever seen global radial strain (GRS) as a as a figure that comes up? 

You know, I have not seen any global radial strain. The GLS the longitudinal view is really again, it's just the most studied parameter and the most sensitive parameter that we have in strain for dysfunction. So definitely something to consider, and perhaps I haven't seen it because it just isn't used very frequently. But definitely something to consider adding to our strain package. 

Yeah. Yeah. And I would concur. It's something that we can look at on the development side and see what's involved and if it's even valid. Sometimes these things that come up, you think, oh, we should do that. And then you actually get looking at it, and mathematically it doesn't always necessarily turn out to be the correct thing to do but can happen. 

QUESTION Actually, an interesting question. Again, something that I hadn't kind of thought of, but it seems obvious when you read the question is asking about the dependence of strain on blood pressure and any other factors, any thoughts on normalizing some of the numbers to blood pressure? Taking blood pressure measurements? 

Yeah, that's a good question. And it's something that, you know, we deal with in ultrasound, in echocardiography, you know, any time we scan an animal. So, your normal parameters like ejection fraction, your E&A ratios, all of those things; they're all low dependent, almost everything we measure is load dependent. And so, there is no real way of normalizing your strain. I suppose if you know, if you just normalized it to a base parameter of your own, that would be something. But it almost is never normalized. It's always taken in the context of which model you're using. So, if you know that there's pressure overload in your model, like attack model, and you do strain and do a sham, well, that that overload is going to influence your strain value because strain is a measure of, you know, systolic function. So, there is no normalization value to it, just like there's no normalization value to ejection fraction. 

Yeah, yeah. I feel like sometimes that's one of those things that you almost just have to accept that there's some variation there when you get down to base levels of trying to normalize everything to everything else. Sometimes, you almost drive yourself crazy, trying to account for everything in the models and in the experiment. 

And I think that if you're having load changes and you see concurrent changes in your strain, you know, that's a finding that you don't want to normalize away. That's indicating that there is a load change that perhaps you didn't know about. So, that's the real value of your strain parameters. 

Yeah, absolutely. The question here, and I think you touched on it a bit, Kelly, and it's certainly a question that I've had before. I’ll give a sort of a 15% answer from product management, and then I’ll pass it over to you. A question about evaluating diastolic dysfunction from the strain image; from the strain analysis. Now I know sort of algorithmically, the way that the software works, the diastolic phase and any diastolic dysfunction can be a bit of a noisy measure because of the way that the relaxation state goes on. Having said that, though, certainly I've seen it and Kelly, I think you touched on it a bit during your presentation. 

Yeah. So, in the Strain software, there is a feature where you can check reverse peaks, and it'll change all the math to looking at time-to-peak of the diastolic peak strain values, or I guess the least amount of strain. So as Stephen mentioned, it just re-applies the formula to looking at diastole, but when it comes to looking at strain in the first place, strain is a measure of systolic contractile function. So, everything that has been published about validating strain values has all been based on the contractile function. Just like, you know, ejection fraction is a measure of systolic function, not diastolic function. So, you could get the parameters for that, but the values of what those numbers are, they really haven't been explored. So I wouldn't be able to say what's normal, what's not normal. I mean, it could be a great project for someone to look at trying to validate those values based on something like your E&A. And, you know, one of the things I mentioned earlier during the talk is, you know, the normal progression or at least what we think is the normal progression of systolic dysfunction is that there is diastolic dysfunction first, then there is systolic dysfunction. 

So, everyone you know, we used to use, or we still use ejection fraction and all of those parameters. But then if you wanted to see quote unquote early, you would do your E&A, and your E-prime, A-prime information. But now, what we have is we have strain, which is even earlier than diastolic function, but it's still a systolic number. So, you have an option to look at the diastolic information. But it's sort of an unexplored territory when it comes to those values. 

Yeah, that's one my feeling as well. I mean, there's definitely information there. It's just that ‘a’, it wasn't really designed to look at that phase. And then the signal-to-noise on that side of things is poor because of not having been designed to do that in the first place. But there is definitely information there that could be teased out. 

So, we've got quite a number of questions still coming in. We will try to get to those via email or other follow ups. We're just coming up pretty close to the top of the hour here right now. So, I do just want to get wrapped up. So, I'm just a change presenter back to myself here. And I just wanted to mention a couple of things just to wrap up. 

Wrap-up Segment

Please do visit our website. Kelly mentioned and linked to a couple of things on our website during her presentation. Please do visit our website. Register for an account if you don't have one, there's a login button there and you can create an account. It doesn't take too long, it just has to come through to us here for a check when you do a request creating an account. 

Once you do that and you’re logged in, you've got access to all kinds of information such as educational resources, tips and tricks, documentation, we have training videos that are coming soon. Actually, both Kelly and myself have been working away, and will be starring in some training videos that are coming up, will be posted on our website fairly soon, along with some of the other folks around here. 

And another thing that's very useful on our website, too, is software downloads. Kelly did mention the Vevo Lab software, making sure that your Vevo Lab software is up to date, so that if you are a strain user or are going to be a strain user, you can access the most recent version. We always post the most recent versions of our software on the website, and those are available to download for Vevo customers at no charge. We, we don't we don't charge for software updates, so that information is all there. 

In addition, there is also training courses that you can sign up for on our website too. Cardiovascular Training, Imaging Training Courses, Image Acquisition, learning how to capture some of the quality images that we mentioned that you need to do to carry on and do your strain analysis with. So please do visit our website and reach out to us. 

We've got some fun stuff going on right now. We've got an image contest going on. So if you go to our website, you will see a poster there on the stroller on our website. If you've got a really great image that you've captured, or that you've analyzed with our system, one of our systems, any of our systems, go ahead and submit it and you've got the chance to win one of these cool FUJIFILM Instax instant cameras. 

Coming up very soon, in hopefully sunny California, the American Heart Association annual scientific sessions meeting in mid-November. I think it's coming up two weeks away now at this point, we will be there at Booth 1500. I'll be there and a number of our staff will be there. So, if you're at the conference, do drop by and ask us any questions. We will have scientific applications, people and also myself there to answer any scientific or product-related questions. We'll be doing our regular Guess the Image contest, which is fun. We're giving away another Instax camera for that. And we will actually have a special guest presentation a couple of times in the booth from Dr. Craig Goergen from Purdue University, who will give a short talk on an overview of some work that he's been doing with some of our 4D imaging and his preclinical research. So again, if you're at Scientific AHA Sessions in Anaheim, please do stop by and say hello. 

You can connect with us anytime through our website. There's forms available to submit support questions, both Technical support and Scientific support. There’s lots of our past webinars that are posted on the website there. Kelly mentioned the webinar that I had done a number of years ago on the Vevo Strain, that's there, along with lots of other useful webinars for image acquisition, tips and tricks and those sorts of things. 

We’re on social media and there's a conference list on our on our website. So, if you're not going to be at AHA, but you're going to another conference, check the list there and see we might be there. And again, lots of customer support possibilities that are there. 

So, with that, we will wrap up and thank you everyone for joining us, those that are looking forward to it, a link to the recording of the webinar will be sent out very shortly. Thank you again, bye.