Transcription
Hello everybody and welcome to the second talk in a series of talks where we're examining the signal within blood vessels and MRI imaging. In the first talk, we looked at a time of flight effect known as high velocity signal loss, where high velocity blood exited a slice in a spin echo pulse sequence prior to the 180° RF pulse, and as a result, the blood didn't generate a spin echo. The blood in that image appeared dark.
Now I want to look at a time of flight effect where blood vessels appear bright in the image, and that effect is what's known as flow related enhancement. And it's this concept that allows us to generate time of flight MRA images. We can generate an MRA without actually needing to inject contrast. The contrast comes from blood entering the slice. And I want to pair this concept with a concept known as saturation bands. Saturation bands allow us to prevent signal coming from vessels in a certain orientation and only have signal coming from vessels in the opposite orientation contributing to the Mr angiography that we are creating with this time of flight effect.
Now, in order to understand flow related enhancement, we first need to understand a concept that we've touched on previously known as partial saturation. I've drawn a pulse sequence here that we can see is a gradient echo pulse sequence. We apply an RF pulse, we measure the signal at TE, we wait a period of time, and then we apply another RF pulse. And this period of time between the RF pulses is what's known as the TR.
If we were to look at a graph here and see what the longitudinal magnetization vector is for our tissues within this specific slice, we could plot the change of longitudinal magnetization over time. We have a longitudinal magnetization vector in the longitudinal plane prior to the RF pulse. Our RF pulse will then take that longitudinal magnetization vector and flip it into the transverse plane. Our longitudinal magnetization vector now is zero. All of that vector is in the transverse plane if that flip angle is 90°. And we're going to use 90° in this example. Importantly, the degree of transverse magnetization, the magnitude of that signal, is proportional to the magnitude of the longitudinal magnetization vector. That's important. However big our longitudinal magnetization vector is prior to the RF pulse will determine the magnitude of that transverse vector. That's a really important concept to understand.
So, prior to this RF pulse, our longitudinal magnetization vector is high. We then allow this transverse magnetization vector to dephase, to lose signal at free induction decay, and at a point in time, we are going to sample that image at TE. We sample that image, remember, during the frequency encoding gradient, and we're going to generate an MRI image. Now, that waiting for that MRI imaging will depend on the period of time between the RF pulse and the TE, TR, as well as the TR. Remember, the TE and the TR determine the levels of T1 contrast versus T2 contrast within our image.
Now, we've got no longitudinal magnetization at this point because we flipped the net magnetization vector into the transverse plane. There's no longitudinal magnetization over time. Once those spins have dephased in the transverse plane, they're still in the transverse plane. It takes much longer, longer for T1 relaxation, longitudinal recovery to occur. Longitudinal recovery happens much slower than T2 relaxation or free induction decay, and eventually, we will get full regaining of that longitudinal magnetization vector, and we can plot that over time. The regaining of longitudinal magnetization until we repeat the TR again. We repeat that RF pulse and repeat the process. We will generate another image that will be identical to the initial image if we allow for full recovery of longitudinal magnetization.
Now, remember gradient echo sequences, the benefit of that is that we can acquire images much more rapidly because we can have a very short TR. We don't wait those long TRs like we do in spin echo imaging. So what effectively we do is we reduce the TR, and it's that reduction of the TR that allows us to fill k-space much more rapidly. Now, when we reduce the TR, what we are preventing is full recovery of longitudinal magnetization. We can see now that the longitudinal magnetization, that recovery of longitudinal magnetization represented by this graph here, won't be allowed to occur with the shorter TR.
If we were to flip spins into the transverse plane before we get full recovery of longitudinal magnetization, our transverse magnetization vector at this point is going to be much lower. Think about it. With our initial RF pulse, we flip that longitudinal magnetization vector into the transverse plane. We've dephased, we've allowed free induction decay and created our initial image. We've slowly started to regain the longitudinal magnetization vector component of this vector here, and that's represented by this regaining of longitudinal magnetization here. However, at that point, we haven't fully regained longitudinal magnetization. We've only got a small vector here, but we apply another RF pulse. And remember what I've said, the magnitude of the transverse magnetization vector is a function of how much longitudinal recovery we have. And at very short TRs, we haven't fully regained that longitudinal magnetization. So at this second RF pulse, our transverse magnetization vector is going to be much smaller. We are going to generate an image that has far less signal. Remember, signal is a function of the transverse magnetization vector and how in-phase those transverse magnetization vectors are.
So we've now generated an image that has different weighting. We've got a much shorter TR, and it has much less signal. As we repeat this process here, we flip now with the second RF pulse this smaller magnetization vector into the transverse plane. That's going to slowly recover its longitudinal magnetization, and at our next TR, we are going to get recovery of this longitudinal magnetization at the same rate that we had it initially. This process is repeated out, and you can see how the signal from our tissues here is drastically reduced compared to the initial image that we created. And our final image here, after the same amount of time in these two examples, you can see the difference here. This difference in signal is what's known as partial saturation. It's the short TRs preventing that full recovery of longitudinal magnetization. And because we are repeating these RF pulses rapidly through the same slice, we are going to get a reduction in signal that's known as partial saturation. This concept is incredibly important when we're looking at flow related enhancement in time of flight MRA imaging.
So, we've generated this image here, and we get partial saturation because these tissues aren't leaving the slice. They're constantly getting this barrage of RF pulses here, unable to fully recover their longitudinal magnetization. And what ends up happening is we reach a steady state where the TR and the T1 of the tissues means there will be a steady state where we get the same amount of longitudinal recovery between each TR. And we can draw a graph here that shows that signal reaching a steady state here. That steady state, the difference between full recovery and this steady state, is what's known as the partial saturation. We will look later at saturation bands where we get full saturation, and that happens through a different process to partial saturation.
So, we've partially saturated our slice that we're looking at. How then do we get bright vessels showing up within that slice? Hopefully, you've realized that it's the blood entering the slice that's not receiving multiple RF pulses in a row that is going to give us bright blood here. Let's look at our schematic that we looked at in the previous talk here. We've selected a slice. We've got blood flowing from left to right in our image here. We know that tissues that are remaining in the slice here are experiencing multiple sequential RF pulses and becoming partially saturated. We're losing signal from those tissues. And as we continue to expose that slice to RF pulses, stationary tissue is going to lose signal.
Blood entering the slice, however, hasn't received multiple RF pulses. Remember, the RF pulse is slice selected. The frequency of this radio frequency pulse matches this specific slice. Blood that's not within that slice will have its full longitudinal magnetization vector. As this blood now enters the slice at the next RF pulse, it has its full longitudinal magnetization vector. So when it experiences the RF pulse for the first time now, because it's within the specific slice, it will have its full transverse magnetization vector proportional to that fully recovered longitudinal magnetization vector. So the signal coming from blood entering the slice will be high. It hasn't been partially saturated, and it's that that allows us to get bright signal from fresh blood entering the slice. We call that fresh blood, blood that has been tagged. It has got its full longitudinal magnetization vector, and we're going to look at other concepts later, other ways to tag blood. But in this example, fresh blood is unsaturated. It's coming in with its full longitudinal magnetization vector, and subsequently, when experiencing that RF pulse, gets a large transverse magnetization vector. It will then leave the slice, and new fresh blood will enter. During these sequential RF pulses, the tissue remaining stationary will experience the same RF pulses over and over again and get a reduction in that longitudinal magnetization vector.
Now, you may be wondering, how do we know which direction this blood is traveling? Well, we don't at this stage know which direction. We could have blood vessels coming in from left to right in this image, or from right to left. Most commonly, it's arteries and veins that are opposing directions. This blood here is also unsaturated and is going to enter the slice at the same time that the arterial blood had entered the slice. The new fresh unsaturated blood, in blood traveling in this direction, is also going to provide a bright signal.
Now, what if we're only interested in say, arteries in our example? We're going to need a mechanism to null signal coming from right to left in our image here, and that is where we look at what's known as saturation bands. Now, previously, we've looked at the concept of saturation when we were looking at fat saturation. We could apply an RF pulse that had a very narrow bandwidth that was specific for the frequencies of fat. That would mean that only fat longitudinal magnetization vector would be flipped into the transverse plane. That RF pulse was very specific for fat precessional frequencies within our slice. Once that fat had been flipped into the transverse plane, we provided what's known as spoiler gradients that caused the fat in the transverse plane to be completely dephased from one another. It no longer had a transverse net magnetization vector because it was completely out of phase, and it no longer had a longitudinal magnetization vector because fat had been flipped into the transverse plane. We then ran, in this example, a spin echo sequence with our normal RF pulses that selected for both water and fat, but only water had a longitudinal magnetization vector, and therefore only water contributed to that signal, at least in theory. Fat had been what's known as saturated. It had been flipped into the transverse plane and completely dephased.
Now, we selected fat based on its frequency here. In this example, with saturation bands, we need to select tissue to saturate based on its location. So let's have a look at our diagram here. We want to saturate tissue coming from this side and moving into our slice, blood that's traveling from right to left in our diagram here. So we want to be able to select this region here and saturate these tissues. Now, the way we do that is instead of having an RF pulse that matches the precessional frequency of a specific specific tissue like that, we have an RF pulse that matches the precessional frequencies of this region here. Remember, we've applied a slice selection gradient along the longitudinal axis of our patient here. So spins in this region are going to precess at a different frequency than spins in this region here. That's how we slice select our radio frequency pulse matches the frequency of that slice. So we can do the same here, provide a different frequency radio frequency pulse prior to our gradient echo sequence here. We apply that RF pulse that is specific for this region, this band on our imaging. All the spins then are flipped into the transverse plane, and we apply spoiler gradients. We lose the longitudinal magnetization and lose the transverse magnetization of any spin that is within this region, and we can make this band much larger, covering entire sections of our patients.
All blood entering the slice now is fully saturated. You see how this is a different mechanism to partial saturation where we have multiple RF bands. Here we are having one RF pulse and then spoiling or dephasing all of those spins in the transverse plane. As a spin now enters the slice, we can see in our artery, the spin entering the slice has full longitudinal magnetization. However, this spin that's entering the slice, or this tissue in the blood vessel that's entering the slice, has no longitudinal magnetization and no transverse magnetization. It is fully saturated because of this saturation sequence here. And this is what's known as a saturation band, where we saturate tissues based on their location. Blood entering the slice from a particular direction now won't provide signal because it has no longitudinal magnetization vector that can be flipped into the transverse plane with our RF pulse within our gradient echo sequence here. So you can see how blood would be bright right in vessels in this orientation, and there would be no signal from blood coming from regions that have been affected by the saturation band.
We can then take all of these slices that we've created where our blood vessels are bright and the background stationary tissue has been partially saturated. We can take the brightest regions in these images, stack the images on top of one another, and then stitch together those images, creating a 3D image using only the brightest regions on these slices. And that's what's known as maximum intensity projection or MIP images. We've created what's known as a 2D time of flight MR. We can actually then rotate this image here and examine the vessels more closely.
Now, I've shown you the mechanism for acquiring individual slices just so that you know there's a mechanism where we can acquire thick slices within the patient and still generate these time of flight images using those thicker slices. We can generate a more accurate MIP image here, which is known as a 3D time of flight MRA, which is outside of the scope of this talk, but it's something that I will cover in a future talk.
Now, there are multiple problems that come with time of flight angiography. This only works if fresh blood is entering our slice between the individual TRs. If blood flow is very slow, the blood itself is going to become partially saturated if it doesn't leave the slice quick enough. As we've seen, the time of flight effects work best when the blood flow is perpendicular to the slice that we're imaging. That allows for blood to enter or exit the slice in time. We can also run into problems when blood is running along the slice instead of through the slice. If you look at the MCA here, you can imagine how blood will be running along the slice for a longer period of time and can also become partially saturated.
Now, we're going to look at a concept in the next talk where we look at gradient moment nulling, which allows us to compensate for blood that is flowing through a slice. Now, flow related enhancement that allows us to generate these images is also vulnerable to turbulence. Turbulence is going to cause rapid dephasing. It's going to prevent bright blood vessels. Here we've seen that turbulence causes darkness within blood vessels. It causes a loss of signal because of the physical mechanisms causing dephasing of those spins in that region. And when we create these images here, we might have gaps between our blood vessels, and we're not sure if those gaps represent stenosis, where they represent occlusion, or whether they just represent turbulence within the vessel. And our 3D acquisitions allow us to get around some of those problems.
Now, when we were looking at partial saturation, we saw that the rate of longitudinal recovery, how quickly tissues regain their longitudinal recovery, is going to determine how much transverse signal they have at the next TR. If we have tissues with a very fast T1 time constant that gain their longitudinal recovery very quickly prior to the next TR, they're going to show up as bright on these images as well. These images have brightness because fresh blood is entering the slice, but if there's stationary tissue within that slice that has a very quick T1, it's also going to show up as bright because it doesn't have the same degree of partial saturation. Our flip angles also determine how much partial saturation we get. With very small flip angles, they will regain their longitudinal recovery much quicker than with large flip angles. So larger flip angles are going to lead to more partial saturation of background tissue, and that's why in these image acquisitions, we see that we use much larger flip angles than our normal gradient echo imaging where we use small flip angles to allow us to compensate for those short TR times.
So now we've looked at how we get dark blood vessels based on time of flight effects, and we've looked at how we get bright blood vessels based on time of flight effect, and we've seen how we can compensate for the direction of flow using saturation bands. In the next talk, I want to look at a concept known as spin phase effects, which becomes really important when we look at phase contrast MRA. Spin phase effects show how the phase of spin changes as blood travels along a slice instead of through a slice, and we can compensate for these effects that cause dark blood using a concept known as gradient moment nulling, which is incredibly important if we're trying to generate accurate MRA images. Gradient moment nulling allows us to get bright signal or compensate for this loss of signal within blood vessels.
Now, remember, if you're studying for an exam, these questions come up over and over again, and as always, I've got a question bank link below. I can't wait to get into the next topic. We're moving rapidly through Mr angiography. I hope you're finding this helpful. So until then, goodbye everybody.