0:01

Okay, this next topic is phase contrast cardiac MRI.

0:06

So here's an example of phase contrast imaging.

0:10

Whenever you perform phase contrast, you get both

0:13

what's called a magnitude set of images, the ones on

0:15

the left, and the phase images, the one on the right.

0:17

The phase images have all the important information.

0:20

Phase contrast is also known as flow

0:22

velocity mapping or flow mapping.

0:25

And what we see here is that on the right-hand image,

0:28

each pixel is coded with extra information from

0:33

your standard 2D MRI, and that extra information

0:36

is the directionality of flow and a rate of flow.

0:40

The brighter the pixel, either on the far

0:42

end of the white scale or the really, really

0:44

darker pixels on the far end of the black

0:46

scale, those are the fastest moving pixels.

0:50

And then the direction is

0:52

coded by either white or black.

0:54

So white can be decided by the scanner,

0:57

but oftentimes, the white is towards you, as it

0:58

is in this case, and black is away from you.

1:02

So, like I said, in phase contrast imaging,

1:04

the brightness corresponds to velocity

1:06

and the color to the direction of flow.

1:08

It's either in or out of plane.

1:10

In this case, you'll see the upper

1:11

right-hand corner here. The black is

1:13

towards us, and the white is away from us.

1:15

So, that sort of polarity there can be decided

1:18

at the scanner, and it doesn't really matter.

1:19

You just need to make sure you

1:20

know what you're looking at.

1:22

You must select the VENC, the highest velocity

1:24

encoding, and VENC stands for velocity encoding.

1:26

So, what does that mean?

1:27

So if you're performing phase contrast,

1:29

and you want to measure the flow or the peak

1:32

velocity, you have to determine what is the...

1:34

basically, what are the bounds of your pixel encoding.

1:37

So for instance, if I say my

1:39

blackest black is going to be

1:41

my highest velocity pixels,

1:44

well, what is the top limit of that velocity?

1:46

Is it one meter per second?

1:47

Is it one and a half meters per second?

1:49

Is it two meters per second?

1:50

Usually, we're going to use around one and a half meters

1:53

per second for the aorta, and maybe one and a half,

1:56

or even one meter per second, for the pulmonary artery,

1:58

because that blood moves a little bit slower.

2:00

In the setting of some sort of stenosis, you may

2:03

need to go up to two, three, even four meters

2:05

per second, in order to measure that stenosis

2:08

because of the flow acceleration

2:10

that can happen in the setting of a stenosis.

2:13

The VENC is expressed in centimeters per second.

2:16

So, you know, I mentioned one meter per

2:18

second, so a hundred centimeters per second.

2:20

Typically, we use around 150.

2:23

You get aliasing when the peak

2:24

velocity is greater than the VENC.

2:26

So, for instance, if you're scanning a patient and you

2:29

set your VENC at 150, and it turns out they have aortic

2:32

stenosis and the blood's shooting out of there at 400

2:35

centimeters per second or four meters per

2:37

second, then you're going to get an area of

2:38

aliasing, very much analogous to what we see on

2:42

patients who have ultrasound and Doppler imaging

2:44

when the frequency encoding is

2:46

not set correctly, same thing.

2:48

And the only way to correct for that is

2:49

to actually rescan with a higher VENC.

2:52

You can take these data sets and use them in

2:56

special post-processing software to calculate flow.

3:01

So, basically, you create a contour

3:03

surrounding the blood vessel and propagate

3:06

that throughout the entire cardiac cycle.

3:08

The software will then measure that area and measure

3:11

the signal over time and integrate it over the entire

3:14

cardiac cycle to give you the flow calculation.

3:16

So, how much blood is passing through

3:19

that vessel over the cardiac cycle.

3:21

And from those calculations, you can

3:23

calculate shunts, which is an important

3:25

thing for congenital heart disease.

3:26

So the relative flow between the pulmonary

3:29

side of the circulation and the systemic side

3:31

of the circulation. If the pulmonary side is

3:34

greater than the systemic side, that's not normal.

3:36

One and a half

3:37

is usually somewhere on the borderline of

3:39

maybe we'll treat the patient, maybe we won't.

3:41

And, certainly, over two is a very severe shunt.

3:44

And those patients will always go to surgery,

3:45

let's say if you have an ASD or VSD, for instance.

3:49

Okay.

3:49

I just wanted to include an example of aliasing.

3:52

This is what it looks like.

3:53

And the way you identify it is that you

3:55

have the really, really dense pixels.

3:57

So, whatever is the upper limit of

4:00

your encoding, you'll see kind of a rim

4:03

of those really dark pixels that ends up

4:06

surrounding an area in the center of that rim.

4:09

That is the brighter pixels of the opposite color.

4:12

So, it's a halo of the really extreme high

4:15

end of your range, um, with the central area,

4:18

which is actually not even being encoded

4:20

and it's flipped over to the wrong side.

4:22

So that's the aliasing artifact.

4:23

You look for; if you see this, then you should

4:25

increase your venc and rescan the patient.