Transcription
Hello everybody and welcome back. So let's pick up from where we left off in the previous talk. Previously, we looked at how we go about generating a gradient echo, and we looked at why we need to generate a gradient echo. We saw that when we apply a gradient magnetic field along the x-axis of our slice, we get rapid loss of transverse magnetization, loss of signal at a rate that's much faster than free induction decay or T2 star.
Now, we have to apply a gradient when we're reading out signal. That gradient allows us to perform a one-dimensional Fourier transformation to frequency encode and find out where that signal is coming from on the x-axis location. Now, because this gradient applied along the x-axis location causes spins to process at differential rates, those differential frequencies of those spins rapidly accelerates the dephasing in the transverse plane, and that's what causes this rapid loss of signal. Now, we need to generate a gradient echo to recover or regain, gain some of that signal back up to free induction decay or T2 star levels. This gradient echo allows us to recover signal while still applying a gradient along the slice that we are trying to record from. So, we can still get spatial localization, but we have recovered some of the signal that we've lost from the gradient that we've applied.
Now, importantly, a gradient echo doesn't account for all the local magnetic field inhomogeneities that we saw in spin echo imaging. Spin echo imaging brings signal back up to T2 levels, not T2 star levels. Why then would we go about using a gradient echo over a spin echo? Well, you'll notice on our gradient echo sequence here, we only require one radio frequency pulse. We don't need that second radio frequency pulse that generates the echo that we see in spin echo imaging. Having only one radio frequency pulse, and often using a radio frequency pulse that doesn't flip the spins 90 degrees, uses smaller angles, also allows us to bring that TE much closer to that single radio frequency pulse. Remember here, we're reading T2 star signal is lost rapidly. We need to bring that TE close to the RF. So, this initial part of our sequence happens much faster in gradient echo imaging than it does in spin echo imaging.
The question we had at the end of the last talk was, where do we place this second radio frequency pulse? Where along this gradient sequence do we place it? Now, there's certain sequences, if you think about it, that need a long TR. If we want to give a T2 weighted image, if we want to create a T2 weighted image, we need to wait for these spins to relax back into the longitudinal plane. If we take two spins that are lying in the longitudinal plane, flip them, say 90 degrees with an RF pulse, they are going to gain longitudinal magnetization at different rates. Those rates are known as the T1 values, the T1 constants, and that's different for different tissues. A long TR will allow these spins to both regain fully their longitudinal magnetization before they flip again.
Now, this is a very key point, and this is where people often get confused. When we're talking about sequences, any type of pulse sequence, we need to remember that it is a sequence. These events don't happen in isolation. It's repeated over and over again. The TR, the time to repetition, is repeated over and over again, and the degree of transverse magnetization when a spin is flipped is proportional to how much longitudinal magnetization has been recovered from those spins. If we have a short TR and only a small amount of longitudinal magnetization has been recovered, if we were to flip, make our TR short, flip those spins early, the transverse magnetization will be much reduced. So, a long TR allows us to get large signal, as well as negate the T1 differences in tissues.
Now, in a gradient echo, our advantage was that we have done this part of the sequence very quickly. If then we were to wait a long TR, our acquisition time would still be very long, despite having done this very quickly. Now, the TE doesn't affect acquisition time. It's the time to repetition that determines how long it takes to image a single slice or to get data from a single slice. Now, in spin echo imaging, we could fill up that time with multiple different echoes, or we could do multi-slice imaging. That isn't on offer as much with gradient echo. We're going to look at some gradient echo sequences, and we see that because gradient echo is imaging T2 star decay that happens much quicker than T2 decay, spin echo, we get a lot of signal for a long period of time that we can sample over and over again, filling up a lot of K space between our first TR, between our first RF pulse and our second RF pulse. That can't happen in gradient echo imaging, and that means that we need to bring the second RF pulse close to the first RF pulse, make our TR short. And that is one of the features of gradient echo imaging is that we have short TRs. Short TR allows us to image rapidly, and the faster we can acquire an image, the less motion there's going to be within that image, motion artifact, and the quicker our scan time is going to be.
Now, we need to be able to scan quickly, but we still need good signal. We still need good and appropriate contrast within the image, and we want to reduce the amount of noise in our image while still maintaining spatial localization. Now, in today's talk, I want to focus on flip angles, specifically how much we flip those spins out of the longitudinal plane into the transverse plane. And I want to show you how differing the flip angle will have different effects on the amount of signal that's going to come from our slice. After we've looked at that, I want to show you how we can use different flip angles to allow us to make very short TRs while still getting T2, T1, and proton density weighting within our image. And that's really important. How can we have a short TR but still have a T2 weighted image? And hopefully, by the end of this talk, you'll appreciate how the flip angle allows us to do that.
So, let's have a look at two examples. I want to show you two different types of flip angles and show you how that affects signal within a pulse sequence. So, we've got a net magnetization vector here along the longitudinal plane. We're applying our main magnetic field. Right, at this point in our pulse sequence, we then apply a 90-degree radio frequency pulse, a pulse that we've seen over and over again, that flips our magnetization vector into the transverse plane, giving us the signal that we're measuring. Now, this red arrow here represents the degree or the magnitude of that signal in the transverse plane. Now, I've shown you these types of spins before. We're going to look at how the transverse vectors dephase over time as we stop our RF pulse, both from the side and front on. As these spins dephase, watch what happens to this net magnetization vector. As the spins now dephase like this, they lose their net magnetization. Our net magnetization vector in the transverse plane gets lost because these spins are all out of phase from one another. If we were to add up these vector values, they would equal zero. Notice how these vectors still have magnitude. Each one of these has magnitude and is in the transverse plane, but it's the dephasing that's caused us to lose our transverse magnetization. That process there is free induction decay or T2 star.
Now, this is a great illustration for how T2 star happens much more rapidly than T1 relaxation or longitudinal recovery. These spins are still in the transverse plane. We need to wait now for those spins to gain longitudinal magnetization, and that's what that long TR does before we flip the spins again. Now, what happens if we've dephased these spins and we only allow them to gain a very small amount of transverse magnetization before applying our next 90-degree RF pulse? Well, if we were to make this TR short here, we only recover some of the longitudinal magnetization. We don't fully recover that full longitudinal magnetization along the Z axis here. Now, if we apply the 90-degree RF pulse at a short TR, our transverse magnetization is going to be much smaller than it previously was, and that's what a short TR does. That's why I say we need to think about pulse sequences as sequences. They don't happen in isolation. The second RF pulse determines the degree of transverse magnetization based on the degree of longitudinal recovery that happened after our first RF pulse. And if we make that very short, we don't give that time for the longitudinal recovery to occur.
Now, you can see that if we repeat this at continuing short TRs, we can lose quite a lot of that signal, and we'll eventually get to a steady state where the degree of recovery and the degree of transverse magnetization will even out, and we will get a set signal for this specific pulse sequence at this TR. And say that signal is this, this magnitude here. Notice how a 90-degree flip angle doesn't always mean we get the full amount of transverse magnetization when we use short TRs.
Now, let's have a look at a second example here where we use a 15-degree radio frequency pulse. We are only flipping that spin 15 degrees out of the longitudinal plane into the transverse plane. So, as we apply that radio frequency pulse, notice how we have still got some transverse magnetization. When we apply a radio frequency pulse, those spins process in phase, and because they're processing in phase, they will have a transverse magnetization vector. All of those spins have the same net transverse magnetization vector. However, we haven't fully lost our longitudinal magnetization here. When we flipped a spin 90 degrees, we fully lost that longitudinal magnetization, and it took a long time to regain that longitudinal magnetization. Here, when we use a 15-degree radio frequency pulse, after the first radio frequency pulse, we seemingly have very little signal, much less than we had after our first 90-degree RF pulse. But remember, these things happen in sequences. This is a gradient echo sequence. So, if we apply a second radio frequency pulse here at a set TR, the same TR that we used in our previous example, what will happen? Well, this magnetization vector, because it has kept most of its longitudinal magnetization in this short period of time, the short TR, the spin will be able to fully recover that longitudinal magnetization. The amount of longitudinal magnetization that needs to be recovered here is very small, and that can happen quickly, as opposed to the full longitudinal magnetization or longitudinal recovery that needed to happen in our 90-degree example. So, we allow that spin to relax in this short period. See how easily that longitudinal magnetization is recovered, and we apply this second 15-degree RF pulse. We get the same signal again. And if we were to repeat this over and over again at the same TR, each time we would be getting these same transverse magnetization vector. And this transverse magnetization vector in this random example that I've chosen, the 15-degree example compared to the 90-degree example, shows us that the 15-degree example can actually give us just as much, if not more, transverse magnetization.
Now, this small flip angle in combination with a short TR is giving us similar or equivalent transverse magnetization vectors as opposed to that large 90-degree RF pulse. Now, we can represent this graphically as we change the flip angle, like we looked at in those two examples, at a set TR for a specific sequence and at a set T1 for a specific tissue. We can then plot the steady state signal that we would get if we were to sequentially repeat our gradient echo sequence. Now, this graph here, importantly, is for one specific tissue. If you think about what determines the rate at which a spin will recover its longitudinal magnetization, that rate is defined by the T1 time constant for that tissue. Different tissues have different T1 values, and the rate at which that longitudinal relaxation will occur will be different for different tissues. So, these graphs are looking at one tissue in particular, and they're looking at one set TR. If we were to wait a longer TR, it would give spins longer time to regain that longitudinal magnetization before we flip it again. So, here in our example, see, see how the 90-degree, say, if we were looking at our previous example, gives us a specific signal value, and the 15-degree gives us a signal value that is the same, that's equivalent here. And we can connect these dots here and get a graph that represents the signal change as we change flip angle within our sequence.
Now, if we were to look at the point for this specific tissue at a specific TR that gives us the greatest transverse magnetization, that value there is what's known as the Ernst angle. This angle for this specific pulse sequence will give us the maximum signal. Now, if we were to look at a second tissue that had a longer T1, it took longer to recover the longitudinal magnetization, but still had the same TR, it was still part of the same pulse sequence. We would see that those signal values are different. Our Ernst angle would be slightly lower. Our flip angle that will give us maximum signal will be lower because our recovery happens slower in a tissue that has longer T1s. We can then also join those data points there and get a different Ernst angle for a different tissue, a tissue that has a different T1 value.
Now, when going about looking at a single slice, we know that that's made up of lots of tissues, and we're not going to have one specific Ernst angle that's best. If we were comparing these two tissues, we may say that perhaps an Ernst angle of about 40 degrees is going to give us on average the most signal within our tissues. Now, remember, when we're imaging, it's not only signal that we want. If everything gave maximum signal, we would just have a white blob on the screen. What helps us in radiology is that we need contrast. We need to be able to differentiate tissues in normal anatomy, and we need to be able to differentiate pathology from normal. And it's contrast that gives us a lot of accuracy. We need signal, we need contrast, and we need good spatial localization in combination to give us an adequate diagnostic image. We may want to choose an angle that gives us the most contrast between the two tissues, the most signal difference between the tissues. So, choosing a flip angle doesn't always come down to what gives us the most signal. Sometimes we want to think about what gives us the most contrast within an image.
Now, the take-home point that I want you to get here is, flip angles, differing flip angles can give us differing signals, and using maximum flip angles in a sequence that has a short TR is not always beneficial. We can use smaller angles and actually get better signal because that TR is short, and that short TR doesn't allow these larger flip angles to fully regain their longitudinal magnetization before we flip them again, and that's responsible for this decrease in signal here. Now, importantly, today we're looking at gradient echo imaging, which requires a short TR. A gradient echo sequence with a long TR is not really beneficial to us. We may as well then be doing a spin echo pulse sequence that will have less chemical shift, will be less susceptible to magnetic susceptibility artifacts, and will give us more signal, T2 signal. The benefit of a gradient echo is that we can acquire signal quickly with that short TR.
Now, let's look at what changing TR does to these signal values using different flip angles here. These TR values of 2,000 milliseconds are long TR values. These are the kind of TR values that we'll be seeing in T2 weighted imaging or proton density weighted imaging in our spin echo pulse sequences. In gradient echo pulse sequences, we are using very short TRs, as we looked at in our sequence here. And this is going to be the area of interest in gradient echo imaging. Obviously, if we have long TRs, those spins will be able to regain fully their longitudinal magnetization before the next radio frequency pulse, and the degree of signal is going to be based on the flip angle here. The higher the flip angle, the higher the signal. Now, that's not the case when we look at our short TRs here. If we zoom in onto these short TR pulse sequences, notice how the order of the angles has now changed here. If we use very short TRs, there's actually going to be a period of time when a 10-degree flip angle is going to give us the best signal. And as we increase the TR, we can see that 20 degrees at one point will give us the best signal, 30 degrees, 60 degrees is going to give us better signal than the 90-degree flip angle. And this graph here actually shows us for specific TRs and a specific tissue what the rough or estimated Ernst angle is going to be at that specific TR.
Now, let's look at a pulse sequence here that we've been looking at. You may notice that I've been using an alpha symbol here in many of our gradient echo image sequences, and that's to show we can use varying flip angles. We can set the flip angle. Now, in spin echo imaging, it also turns out that we don't always use 90-degree flip angles, but we do use much larger flip angles, generally speaking, than we do in gradient echo imaging, and that's because of this short TR. Now, let's have a look at an example here where we use a TR value of 150 to 200 milliseconds, somewhere in this range here, where we've got high signal coming from 20, 60, and 30-degree flip angles. Let's take one example where we use a 20-degree flip angle. A 20-degree flip angle is going to give us a signal value of about this much at our steady state precessional value, and I've represented this signal value here by this red arrow here. Now, with this short TR, what's going to happen when we stop this radio frequency pulse? That spin that's been flipped 20 degrees into the transverse plane is going to fully recover its longitudinal magnetization before being flipped again because of that small flip angle here. So, we're going to maintain a steady state of this signal here.
Now, if we were to look at a different flip angle, say a 60-degree flip angle, notice how we've lost more transverse magnetization because of that larger flip angle. That loss of transverse magnetization because of the larger flip angle means at this short TR that we're not going to fully recover the longitudinal magnetization, as we saw in the previous example. That means that our steady state transfer signal is actually going to be the same as our 20-degree flip angle. And you can see that from this graph here. 20 degrees, the blue line, is giving us the same signal as 60 degrees, the orange line, at this specific TR. And that's because this short or small flip angle gets full recovery of its longitudinal magnetization in this example, and the 60 degrees doesn't get full recovery of its longitudinal magnetization. And that's why when we flip it over and over again in this sequence at this specific TR, we will get a steady state signal that is identical, according to this graph here.
Now, what does that actually allow us to do? Well, which sequences do we want full recovery of longitudinal magnetization? We want full recovery in T2 weighted images. We don't want to see the T1 differences in tissue when we are trying to create an image that has T2 weighting. Now, this image here, created by the 60-degree flip angle, if we flip these spins, two different tissues, 60 degrees, they are going to recover their longitudinal magnetization at differing rates. And we've just said here that in this example, this net magnetization vector is not going to fully recover its longitudinal magnetization. So, there is going to be differences in longitudinal magnetization between these two tissues. If we were to flip at the next TR, this net magnetization vector here, the flip net magnetization vectors are going to have differences in their transverse magnetization, T1 differences. And if we sample that really quickly at a short TE, we're going to create an image that has T1 weighting.
If we look at the example with only a 20-degree flip into the transverse plane and use the same TE value, these will gain their longitudinal recovery at different rates, T1s. But because of that short flip angle, they are going to fully recover when we flip them again. There's going to be no T1 differences in these two spins. That short flip angle or small flip angle has allowed us to negate the T1 differences in our tissues. And this is the crux for how in gradient echo imaging we can still create a T2 weighted image despite having a very short TR. It's not the long TR that allows the two separate tissues to regain fully their longitudinal magnetization before we flip it again. It's the small flip angle that allows two two different tissues with differing T1 values to regain their longitudinal magnetization before the next flip. So, hopefully, you can see that small flip angles allow us to create a T2 weighted image, allowing those spins to regain their longitudinal magnetization before the next flip.
So, we can see that small flip angles allow us one, to create a shorter radio frequency pulse. They allow us to get more signal often than larger flip angles. And small flip angles allow us to create a T2 weighted image with a short TR. And it's the short TR that we need in gradient echo imaging to acquire our image rapidly. Our TR is what determines that acquisition time.
So, now we've looked at how a gradient echo occurs and why we need to create gradient echoes. We've also looked at flip angles and the Ernst angle. Let's move on now to looking at different types of gradient echo pulse sequences: coherent sequences, incoherent, and spoiled sequences, as well as steady state free precessional sequences. We're also going to look at a phenomenon that occurs when we create these short TRs here. We've now got two radio frequency pulses that are closed together, like we saw in spin echo imaging, and they in fact themselves also create an echo known as a stimulated echo. That's all going to be covered in the next talk. Again, link below is a question bank. If you are studying for a specific exam, go and check those out. And many people have passed their Radiology physics exam and said to me, those kind of questions have come up over and over again in the exam. So, until the next talk, I'll see you all there. Goodbye everybody.