0:01

Okay, next we're going to talk about

0:02

cardiac MRI protocols in the management

0:05

of non-ischemic cardiomyopathy.

0:07

First, some technical notes.

0:09

One question that comes up all the

0:10

time is whether one should use

0:12

1.5 Tesla or 3 Tesla MRI for imaging of the heart.

0:18

And it turns out, contrary to what one might think,

0:20

that the 1.5 Tesla is generally preferable

0:24

for cardiac MRI.

0:25

And this is actually really handy for a lot of

0:27

practices because, as we know, 3T MRI often gets

0:31

really busy with neuro and musculoskeletal

0:35

indications.

0:36

And so, having some studies that

0:38

can actually be performed on the

0:41

1.5 Tesla can be really helpful for some practices.

0:45

Certainly, it's very helpful in our practice.

0:47

So generally, the preferred and

0:49

kind of workhorse is the 1.5 Tesla.

0:52

Why would one ever want to use the 3 Tesla?

0:55

Well, in some situations, if you're really interested

0:58

in doing perfusion images, so those will be the

1:00

people who want to do a lot of stress cardiac MRI.

1:03

We're not going to talk about that today.

1:05

We'll talk about it in different courses, but in

1:07

general, if you're doing a lot of perfusion imaging

1:09

for stress MRI, 3 Tesla can be a little bit better.

1:12

If you're performing 3D late gadolinium

1:15

enhancement, which is a very sort of

1:17

specialized, almost research-type application,

1:20

but if you're in a place where they are really

1:23

interested in very, very highly detailed late

1:26

gadolinium enhancement images—and I can think

1:29

of maybe perhaps if you're in a practice that

1:31

does a lot of electrophysiology and they want

1:33

mapping of scar prior to any ablation procedures—

1:36

then maybe 3 Tesla might be helpful.

1:39

And then finally, 4D flow.

1:41

4D flow applications really are probably, in my mind,

1:45

the best in congenital heart disease, which again,

1:47

we won't be talking about today, but if you're

1:50

really interested in 4D flow and you're doing a

1:52

lot of congenital heart disease, then perhaps

1:54

3 Tesla would have advantages for your practice.

1:57

What about coils?

1:58

Well, ideally, you want as many coils as you

2:00

can get, but certainly you want at least

2:02

an eight-element or greater thoracic coil.

2:05

We use a 32-channel coil as our

2:07

workhorse, and that tends to work great.

2:10

Just the most amount of signal you can get.

2:12

Obviously, saves some time and allows

2:15

you to make higher resolution images.

2:17

You want to have a power injector.

2:20

Usually, we inject at roughly 5 mL

2:23

per second for any perfusion images that we're doing.

2:25

So certainly, a power injector

2:26

becomes helpful in that situation.

2:28

If you're not doing perfusion, then maybe you

2:30

can get away without a power injector, and you

2:32

can even just do a hand injection because for

2:35

late gadolinium enhancement, we're generally

2:36

waiting about 10 minutes after the injection.

2:38

So, you know, it doesn't really matter how

2:40

fast that contrast gets into the patient.

2:43

And then you do need some

2:44

specialized software for cardiac MRI.

2:46

In particular, you need some specialized post

2:48

processing software for function and flow measurements.

2:52

Generally, the PACS that you have is not going

2:54

to have the tools necessary to do ejection

2:57

fraction calculations, volume calculations,

3:00

and so, I would say it's almost universal that you're

3:03

going to need some sort of secondary software

3:06

package to do that kind of post-processing.

3:09

Okay, so let's talk a little bit about

3:11

cardiac MRI protocols and what are kind of

3:13

the key elements to performing a cardiac MRI.

3:16

We always start with an anatomic overview.

3:19

We get an axial steady-state free precession,

3:21

bright blood image just to get the anatomy.

3:24

This is helpful looking for things like pleural

3:27

effusions, adenopathy, possibly in somebody

3:30

with sarcoidosis, pericardial effusion, you can

3:33

look for any abnormalities in the upper abdomen,

3:35

perhaps like cirrhosis that can, you know,

3:37

be helpful for evaluating the patient overall.

3:40

You could also add dark blood HASTE.

3:43

We do that in some of our protocols, not all.

3:46

The dark blood HASTE images we'll talk about later.

3:48

They give you a little bit more anatomy.

3:51

You can get them as sort of a T1

3:53

weighted image, and they may be helpful,

3:55

for instance, at looking at the aorta.

3:57

We always include them in our MRA protocols

3:59

to look for any aortic thickening in

4:01

somebody with chest pain, for instance.

4:03

But, it's kind of plus or minus whether

4:06

or not you want to include it in your

4:07

cardiac MRI, standard cardiac MRI protocol.

4:10

We always include functional imaging.

4:12

You have to have functional imaging,

4:14

and this is done with Cine SSFP, steady-state

4:16

free precession, bright blood images.

4:18

The only situation where we might not use SSFP

4:22

images would be if there are a lot of artifacts.

4:25

In particular, if you're at a place where you're

4:27

doing scans in patients with cardiac devices,

4:30

those cardiac devices create a lot of artifact.

4:33

In particular, they create abnormalities of the local

4:35

magnetic field around the device, and that's going to

4:38

create a lot of artifact, particularly on SSFP images.

4:42

And so in those cases, you may want

4:43

to use cine gradient echo images.

4:45

These are kind of the older way to do cine gradient imaging.

4:48

Turns out they have less artifact than the SSFP images,

4:51

and so in that situation, you may use those

4:53

in place of SSFP.

4:55

The other situation where I've had to use

4:58

it would be sometimes patients with, say,

5:01

for instance, iron deposition in the liver.

5:03

If you're evaluating for iron deposition

5:05

in the heart, a lot of times those patients

5:07

will have a lot of iron in the liver.

5:09

That iron actually can sit really close

5:12

to the undersurface of the heart and

5:14

create a lot of artifact in some patients.

5:16

And so in those cases, again, gradient

5:18

echo may be a better choice.

5:20

Okay.

5:20

What else is part of the protocol?

5:22

Well, certainly tissue characterization is important.

5:24

In any of our non-ischemic cardiomyopathy evaluations,

5:28

we want to include some dark blood

5:30

T2 weighted images to look for edema, and we

5:33

want to include late gadolinium enhancement

5:35

images to look for the patterns of late

5:38

gadolinium enhancement if it's present.

5:40

And really of all the various types of acquisitions

5:43

that you get in cardiac MRI for this particular

5:46

indication, non-ischemic cardiomyopathy,

5:48

the most important one is going to be this

5:50

late gadolinium enhancement, far and away.

5:52

So really that's the one that you

5:53

want to put the most effort into.

5:56

Other optional types of acquisitions.

5:59

Flow, like I mentioned, so this is

6:01

not particularly important in the

6:02

setting of non-ischemic cardiomyopathy.

6:04

The only situation where I could see it useful would

6:07

be if you're not completely sure if it's a patient

6:10

with severe valvular heart disease that might have

6:13

the cardiomyopathy, and so you may want to do some

6:15

flow measurements to try and quantify, say, for

6:17

instance, aortic regurgitation or mitral regurgitation.

6:21

But otherwise, I don't see flow

6:22

playing a big role in these patients.

6:25

So, the T2 star mapping can be really helpful

6:27

in patients with sickle cell anemia, and then

6:29

also in patients who have beta thalassemia.

6:32

Um, those often have a lot of iron deposition as well.

6:35

T2 mapping and T1 mapping are slightly newer

6:40

applications.

6:41

They have been getting a lot of attention in

6:43

literature, and I would say that they're still

6:46

kind of on the cusp of routine clinical use.

6:49

Of the two, I think T2 mapping right now is probably

6:52

a little bit more useful and ready for prime time.

6:55

The T2 dark blood images that I mentioned, they tend to

6:59

have frequent artifacts and are sometimes less reliable

7:03

in people who can't hold their breath very well.

7:05

The T2 mapping sequences can overcome some of those

7:09

problems and give you a more uniform evaluation of T2.

7:12

And this is helpful in looking for edema in patients

7:15

who perhaps might have myocarditis as the sort of

7:17

the number one application, or maybe sarcoidosis.

7:22

T1 mapping is something that's gotten even

7:24

more attention really in the research space

7:27

than T2 mapping, but on an individual

7:30

per patient basis, I don't think

7:32

it's quite as useful for diagnosis.

7:35

I think the numbers that you get for T1 mapping,

7:38

you basically get an exact number of the T1 time

7:41

of the myocardium, and the idea is that T1 times

7:45

are abnormal in patients with a lot of fibrosis.

7:48

The problem is that there's a lot of overlap between

7:51

normal patients and patients with disease, so much

7:55

so that the T1 mapping often doesn't really help you

7:58

distinguish between patients who have a particular

8:01

cardiomyopathy versus a patient who's normal.

8:04

I would say the one exception to that...

8:06

There are really two exceptions to that.

8:07

First would be patients with amyloid.

8:10

The T1 mapping is really, really

8:11

abnormal in those patients.

8:13

And so, in that case, it can be useful.

8:15

And then the other is in the patients with the

8:17

glycogen storage diseases, particularly Fabry's.

8:19

They can have really abnormal T1 mapping as well.

8:22

So in those cases, you may have enough

8:24

distance between normal and abnormal

8:27

values that you can use them for diagnosis.

8:30

Perfusion is also optional.

8:32

In our practice, we always get perfusion images.

8:35

So you'll see when we go through the

8:36

cases that we always have perfusion.

8:38

Do we need to get it as a routine?

8:40

Not really.

8:41

Rest perfusion is not particularly helpful.

8:45

Sometimes we'll see in patients who

8:47

have a recent MI that there's areas of

8:49

hypoperfusion on the rest perfusion.

8:52

And then also in people who have really severe acute

8:54

sarcoidosis, we can see some hypoperfusion at rest.

8:58

But really, for nearly all other indications,

9:00

we're not going to see any abnormalities.

9:02

So it's, it's plus or minus whether you do it.

9:05

We tend to do the perfusion, um, just

9:07

because it doesn't really lose us any time.

9:11

And it helps get our technologists used to just

9:13

performing the perfusion so that if we ever do a stress

9:16

MRI, they're very comfortable with that sequence.