Transcription
Hello and welcome to the third and final part in this introduction to gradient Echo sequences. We left off in the previous talk by looking at how changing the flip angle could actually manipulate the contrast weighting within our image. And we saw that small flip angles allowed us to still generate a T2 weighted image despite having these very short TR times. We saw that the small flip angle allowed for recovery of longitudinal magnetization prior to the next RF pulse. We could then choose a slightly longer TE time that will allow for the differences in free induction decay between two tissues to occur prior to sampling that signal at TE. And those differences in free induction decay would generate their T2 weighting within our image.
Now, the major benefit of gradient sequences is that we can image a large region of anatomy in a very short period of time. And that's due to these short TR times. Now, you'll see when looking at gradient echo images that we actually use TR times that are even shorter than the example that we used here. We often use TR times that are in the region of 50 milliseconds.
Now, there are two major issues that occur when we reduce LTR times to such short periods of time. The first that you'll notice here is that we reduce our potential TE time. Here, we can't have Tes that are longer than our TR. That when you have very short TE times, you don't allow for much difference in free induction decay to occur between tissues. It's very hard to generate a T2 weighted image. We know that TE times need to be slightly longer for T2 weighted images. We can't do that with very short TR times.
The second issue is that when you place two RF pulses very close to one another, at the time of the second RF files, there is still going to be some transverse magnetization within our sample. We weren't a fully lost that transverse signal prior to flipping those spins again. Doing this, putting two RF pulses close together will generate what's known as a stimulated echo, which we're going to be looking at today.
Now, fortunately, the stimulated echo can be used to compensate for this issue of short TE times. We can actually generate T2 weighting in our image by using the stimulated echo despite having short TE times. So, let's go through this pulse sequence and see what actually happens to the signal that's been generated from our slice. So, let's make a graph of the signal. We've got our transverse magnetization vector here, our magnetization in the XY plane over time. Now, this matches up with our pulse sequence below here. The first thing we do is flip our spins into the transverse plane using a slice-selected RF pulse. Those spins will then be in phase in the transverse plane until we switch off our RF pulse. Those spins will then dephase due to spin-spin interaction as well as local magnetic field inhomogeneity. So, we know that the loss of transverse magnetization is free induction decay, or T2 star.
So, let's look at what happens to that transverse magnetization signal within our slice. That signal is being lost at free induction decay because of that dephasing in the transverse plane. We then apply the dephasing frequency-encoding gradient that we looked at in the first part of this lecture series. That dephasing frequency-encoding gradient causes rapid loss of transverse signal. We can then apply our readout frequency-encoding gradient when we are actually measuring that analog signal, converting it into digital signal and placing it into k-space. That readout gradient will then allow for re-accumulation of signal back up to levels of free induction decay and allow us to get analog signal while still applying a frequency-encoding gradient, allowing us to spatially localize that signal along the x-axis. And we know that the loss of signal here is free induction decay in that tissue. It's T2 star.
Now, in this pulse sequence, we are measuring the signal here during the gradient echo at TE. And we are then applying the next RF pulse prior to full loss of this free induction decay, prior to loss of transverse magnetization. We have still got some residual transverse magnetization at this next RF pulse.
Now, in order to illustrate what effect this will have further down the line in our pulse sequence, I want to go back to the Cartesian plane that we've been looking at. Here is the longitudinal plane, and here is our XY plane, our transverse plane. Now, what we do at the first RF files is we flip our net magnetization vector to a specific angle. That vector can then be broken down into a longitudinal component and a transverse component. It's the transverse signal that we are measuring and plotting on this graph here. We wait a period of time until TE. Now, when we measure that signal at TE, there will be some transverse magnetization here. We see that transverse magnetization is back at levels of free induction decay. The transverse magnetization has been lost because of this dephasing. That's what I've tried to represent with these two arrows, this dephasing of that transverse magnetization. All the while, we're getting re-accumulation of longitudinal magnetization.
Now, this vector can be broken down into a longitudinal magnetization vector and a transverse magnetization vector. Now, at this period of time, there hasn't been complete loss of transverse magnetization. There is still some phase coherence in the transverse plane. That's why we have transverse signal that is measurable.
Now, notice how close our TR is here in this pulse sequence. This TR occurs so rapidly that we haven't yet lost transverse magnetization during this period of time. What's going to happen is we are going to gain some longitudinal magnetization, still lose some transverse magnetization, but it's not completely gone at the period of time in our next RF pulse. The TR is short enough that we've got what's called residual transverse magnetization. There is still some phase coherence in the XY plane.
Now, what will happen when we apply this next RF pulse? Well, we know this longitudinal magnetization vector is going to be flipped by the flip angle. The transverse magnetization vector is also going to be flipped. Now, I want to stress what I'm showing you here is a gross oversimplification of what actually is happening. These vectors are occurring in the 3D plane, and flipping these vectors causes complex overlapping of these vectors. What I'm showing you here is to get you an idea of how these vectors are generating what's called the stimulated echo.
So, let's look at what happens when we apply this RF pulse. Well, the longitudinal magnetization vector gets flipped by our flip angle, as we saw in our first RF files. We can think of that residual transverse magnetization vector as also being flipped by that flip angle. We can also see that we're getting an increase in the magnitude of that transverse or that residual transverse magnetization vector. And that's because the RF pulse not only flips spin into the transverse plane, but it also phases spin. We know that this net magnetization vector gives transverse magnetization because the net magnetization vectors are in phase, and in phase increases the level of that transverse magnetization vector.
Now, if we were to wait a given period of time until the next RF pulse, what would be happening at this RF pulse? Well, we saw what happened to this net magnetization vector. It gained longitudinal magnetization as well as losing transverse magnetization. And the same thing is happening to our residual transverse magnetization vector. It is realigning into the longitudinal plane, and it is dephasing in that period of time. It's losing some transverse magnetization. You'll notice that at this period of time, though, this magnetization vector will be aligned in the transverse plane. It will be giving signal. That signal that we are reading out during this longitudinal recovery and transverse decay is what is called our stimulated echo. And I'm going to show you what that looks like now.
This is very similar to the spin echo sequences that we saw before. We applied two RF pulses that generated an echo. Now, the difference in spin echo was one, the RF pulses had different magnitude. Generally, we looked at an example of 90 degrees and 180 degrees, and they generated an echo in a period of time where there wasn't an RF pulse. That echo was when we were reading the signal. That echo was at TE. In gradient echo imaging, our RF pulses are equally spaced out. In spin echo imaging, we had one RF pulse, a second RF pulse, and waited until our next TR before we repeated that sequence. Gradient echo imaging, our RF pulses are equally spaced out. So, the generation of the stimulated echo is going to be superimposed over this third RF pulse. This third RF pulse is going to generate a free induction decay curve as well. And what we get is those two signals, both stimulated echo and the free induction decay, being superimposed over one another, both contributing to the signal that we measure in our gradient echo imaging.
So, let's have a look at what is exactly happening to the signal after this third RF pulse. This third RF pulse is going to flip the longitudinal magnetization vector into the transverse plane, and we'll get free induction decay, just as we looked at for each one of these RF pulses. But now we've mentioned that we've got what's called a stimulated echo. The combination of these first two RF forces that is causing an echo to be occurring in the transverse plane at the period of time of our third RF pulse.
Now, what is causing the signal? What is contributing to the signal in the stimulated echo? If you think about it, tissues that have a very short T2 time, a T2 time that is much shorter than this TR interval, they're going to fully lose their transverse magnetization prior to the second RF pulse. There's going to be no residual transverse magnetization in those tissues. They are not going to generate a stimulated echo. The only tissues that are going to generate a stimulated echo is tissues that have a longer T2 time than this TR interval, that still have transverse magnetization. Those tissues that have residual magnetization at this second RF pulse have longer T2 time constants. They're the ones that are going to be generating this stimulated echo at the third RF pulse. You can see how this stimulated echo then is going to be giving us T2 weighting. Tissues with long T2 time constants are going to give signal, and those with short T2 time constants are not going to give a signal.
The free induction decay curve that we mentioned here is going to give a signal that is basically T1 weighted. Now, why is that the case? Well, when we look at these sequences, actually our Tes are very short, as I mentioned previously. Our Tes are shorter than 50 milliseconds. And if we were to accurately draw those on our free induction decay curves, that TE is happening very early on in this free induction decay. There hasn't been time for differences in free induction decay to be generated in this signal here. We saw that short Tes give us proton density and T1 weighted images. So, we're getting high signal, high transverse magnetization, but very little differences in T2 weighting here. So, you can see that this sequence that we've done here is going to give us a combination of both T2 weighting from our stimulated echo as well as T1 weighting from our free induction decay or our gradient echo here.
Now, why do these TEs have to be short? Well, one, our TR is short. Two, the shorter the TE, the higher the signal we get. And three, we have other things going on during this pulse sequence. We need time, for example, to put our phase encoding gradients here. Now, you'll notice that prior to TE, we have a specific phase encoding gradient, and then we have a rephasing gradient after our TE before our next RF pulse. That allows us to then sample the signal at the next TE for a different phase encoding gradient, a different line of k-space. Now, what this means is that at every single TR, the spins in the phase encoding gradients are completely in phase because these two phase encoding gradients allow for rephasing of those spins. This is what's known as a coherent gradient echo, coherent meaning that the spins are in phase at the next TR.
Now, what a coherent gradient echo does is it samples signal from both the stimulated echo as well as the gradient echo. And ultimately, the image that we generate has both T2 and T1 contrast within that image. Now, generally speaking, tissues with a long T1 have a long T2, and we generate images that have very little contrast here because of these superimposed signals from these two echoes here. Now, this becomes very useful later on when we look at MR angiography. I'm going to come back to this, why we might want little contrast within our tissues and actually have bright signal coming from blood within our vessels. So, this image I'll be generating has a mixture of T1 and T2 weighting. And how much T1 and T2 weighting is dependent on our TR, our TE, as well as our flip angle, as well as the innate T1 and T2 time constants of the tissue that we're imaging.
Now, what if we want to create an image that is purely T1 weighted? We need to get rid of this stimulated echo and only sample this gradient echo here. Now, what was causing that stimulated echo? The stimulated echo was a result of that residual transverse magnetization at the next RF pulse. How can we get rid of that residual transverse magnetization? Well, what we can do is instead of rephasing without phase encoding gradient after the TE, allowing us to get a coherent signal, what we could do is take away that rephasing phase encoding gradient and apply what's called a spoiler gradient. We can apply a phase encoding gradient that is incongruent with the initial phase encoding gradient. That spoiler gradient would cause the spins to completely dephase, completely lose transverse magnetization at the next TR. At the next RF pulse, there will be no residual transverse magnetization. The only signal then that we generate will be from the longitudinal magnetization vector that is left at this specific period of time. That will then generate a free induction decay. This spoiler gradient causes incoherence, and that's why this is called an incoherent gradient echo or a spoiled gradient echo because of this spoiler gradient here. The signal that we measure is purely the free induction decay. And because that TE is very short, there's very little differences in the free induction decay between tissues. What we're getting here is a T1 weighted image. We've got a short TR allowing for differences in T1 to be attributed here, and a short TE giving us a T1 weighted image.
Now, is there a way that we can generate an image that is purely T2 weighted? We know that the signal we're getting from this gradient echo has got mainly T1 contribution. Can we sample that stimulated echo in isolation? Well, it turns out there's a complex way to separate this stimulated echo from the free induction decay. We can decouple these two signals. Now, if we are able to decouple these signals, we could move our TE to sample the signal only at the stimulated echo here.
In order to decouple the stimulated echo from the free induction decay, we need to utilize what is called rewinder gradients. What rewinder gradients do is take the echo that was generated from these two RF pulses and accelerate the rephasing of that echo to occur prior to the next RF pulse. That residual transverse magnetization vector that was flipped by the second RF pulse will rephase naturally at the timing of the next RF pulse. We can use complex rewinder gradients to accelerate that process to cause that echo to occur prior to the next RF pulse. Now, the actual mechanisms for doing this is beyond the scope of this talk. But what we're able to do here is to separate that stimulated echo, and we can sample that echo independently of the free induction decay.
This is what's called steady state free precession or SSFP gradient echo. The signal that we are sampling and we're using to generate our image is predominantly T2 weighted. And because this rephasing is actually an echo that's being generated from our forces, we're getting a signal that is more than our free induction decay. It's less than true T2 imaging, but we are compensating ever so slightly for some local magnetic field inhomogeneities by sampling the stimulated echo here.
The three pulse sequences that we've been looking at here, coherent, incoherent, and steady state free precession, all give us different weighting within our image. Coherent gives us a mixture of T1 and T2 weighting. Incoherent gives us a predominantly T1 weighted image. And steady state free precession gives us a predominantly T2 weighted image. Now, these three gradient echo sequences are broad, overarching categories for gradient echoes. And in fact, there are multiple different, much more complicated sequences used by all of the different manufacturers in MRI imaging. What I want to give you here is an understanding of how these broad categories differ from one another.
Now, that brings us to the end of gradient echo pulse sequences. Now, we're going to move on to the final category of inversion recovery pulse sequences. After which, we would have covered the three main pulse sequences within this MRI module that I want to cover: spin echo, gradient echo, and inversion recovery. So, I'll see you all in that next talk. Until then, goodbye everybody.