0:00

So the second half of our ischemia evaluation

0:03

talk is gonna be focused more on our approach

0:05

to actually acquiring the images

0:07

and then how we interpret the images.

0:10

So first thing to do with

0:11

for this test is you've gotta select the patients correctly

0:14

and you've gotta make sure they undergo, you know,

0:16

appropriate preparation.

0:17

Luckily for us in the cardiac MR community, uh, the recent

0:22

update to the chest pain guidelines in 2021 really showed

0:27

that many, if not most patients are likely good candidates

0:32

for cardiac MR evaluation, including patients with acute

0:35

and stable chest pain.

0:36

And so that kind of makes patient selection

0:39

a little bit more straightforward in terms of

0:41

who would is likely to benefit.

0:43

And we've got a summary slide here again from the site

0:46

for cardiovascular magnetic resonance, uh, about kind of

0:49

where cardiac MR fit into those guidelines.

0:51

And it's helpful to review those guidelines

0:53

for anyone who's doing this, uh, on a regular basis.

0:56

But then we get into other basic screening protocols

1:00

to we need to make sure we cover this MRI safety screening,

1:04

including things like do they have a pacemaker?

1:06

Now you can do stress profusion in patients with pacemakers,

1:09

uh, at least some sites can do that,

1:12

but you know, you have to have a whole pacemaker dedicated

1:15

protocol for making sure they can

1:16

be imaged with the pacemaker.

1:17

And then also your image quality is impacted

1:20

by the pacemaker ICD, so more difficult in those patients.

1:24

And then other things basically in our safety, things like,

1:26

you know, metallic form bodies including old pacemaker

1:29

wires, et cetera, especially in these cardiac patients,

1:32

Stress agent considerations.

1:33

So we are gonna be doing pharmacologic stress.

1:36

Uh, and so, uh, we'll go through some

1:38

of the specific contraindications,

1:39

but you know, we do an informed consent of every patient

1:42

who comes through our institution

1:44

before we do this procedure.

1:46

And it's specifically to make sure

1:47

that we're discussing all the contraindications related

1:50

to some of their medical history

1:52

and how our pharmacologic agents can impact those things

1:55

like asthma history of AV block.

1:58

We do get an ECG

1:59

before every exam that's evaluated by a cardiologist

2:02

to make sure there's no signs of AV block

2:04

that hadn't been detected before sinus bradycardia.

2:07

And then recent caffeine intake can actually caffeines a

2:09

competitive agonist with some of the

2:13

vasodilating drugs that we use.

2:14

So we ask people to not have caffeine

2:16

for 24 hours before their test.

2:18

Often that's difficult for people to adhere to,

2:21

but we try to push them in that way

2:23

and I can't really stress enough

2:25

how important the team is in this context.

2:27

So you really need an expert technologist,

2:29

a nurse who's comfortable delivering cardiac

2:32

medications and then a physician.

2:34

We have a cardiologist at the scanner at every one

2:37

of these exams just in case something goes bad,

2:39

we have someone onsite ready to sort of intervene

2:42

and monitor any adverse physiology in real time.

2:47

And we have MR compatible real-time,

2:49

physiologic monitoring equipment of course in the room

2:52

with the patient throughout the exam.

2:55

So when we talk about pharmacologic stress, uh, there's kind

2:57

of two main classes of drugs that can perform this.

3:01

One is vasodilators

3:02

and we'll talk about the other here in just a second.

3:06

We tend to do adenosine in most of our patients there.

3:08

The top row

3:10

and a couple of things I want you to focus on here,

3:13

one is the duration of action in this

3:17

and ADINE is, you know, very short duration of action.

3:20

So it's kind of on the image and

3:22

Then within, you know, seconds it's sort of off

3:25

and that's kind of a nice feature of this particular drug.

3:29

The key contraindications of all

3:31

of these generally are bronchospasm.

3:34

So patients with asthma, contraindication, AV block

3:37

and sinus bradycardia we've mentioned

3:39

you can give an antidote to these usually amino

3:42

and that can reverse the effects.

3:44

And I mentioned caffeine actually is a competitive

3:47

agonist with these agents.

3:48

And so if worse comes to worse,

3:50

you could give some one caffeine

3:51

and it may help 'em feel better.

3:53

The other one that we use occasionally is rone.

3:58

The bottom row there. One thing

3:59

that's nice about it is it's got less

4:01

impact on airways at least.

4:02

So sometimes in patients with asthma, uh, you can use it.

4:05

It can also be injected through a single iv,

4:08

which at Dene we usually have two IVs that we

4:11

inject one on each arm.

4:12

And so if you can't get two IVs, sometimes we use rezy,

4:16

same antidote, but note the longer duration.

4:18

So we don't often have to reverse

4:20

or do anything to reverse adenosine.

4:22

If we use rezy, often we give a reversal agent

4:25

or at least consider giving a reversal agent.

4:28

The other pharmacologic agent that we can use is domine,

4:32

which is chronotropic and ionotropic agent.

4:35

And this has been historically used mostly to look

4:37

for regional low motion abnormalities.

4:39

So we don't often use that.

4:41

We opt for one of the vasodilators instead. Here's a basic

4:45

Imaging protocol.

4:46

Uh, you can see compared to our viability, you know,

4:49

we've kind of been building over the course

4:50

of the course from a very basic viability protocol now to a,

4:54

you know, more advanced stress perfusion protocol.

4:56

The big changes here are

4:59

The fact that we do stress perfusion first

5:03

and we do rest perfusion last.

5:05

So what that ends up looking like is

5:07

after we do our long axis syn nase,

5:10

we then deliver the stress agent.

5:12

We use a half dose of gadolinium.

5:15

So whatever weight-based dosing we determine

5:19

we end up using half of it as the stress first pass imaging.

5:24

Then once we complete that, we

5:26

do our short axis syn a images to both give time for some

5:31

of this contrast to wash out

5:32

and also the patient's heart rate to come down, you know,

5:35

after we turn off the stress agent.

5:37

And then we do the resting first-pass perfusion

5:40

after that, this is followed by a short weight period

5:44

and ultimately then we end up doing late gal limb enhanced

5:46

imaging at the end.

5:48

And ideally, you know, it's ends up being, you know,

5:51

probably about 10 minutes or five to seven minutes

5:54

after the rest and 10 to 12 minutes after the stress.

5:56

So hopefully we've given more than enough time for a lot of

5:59

that contrast to wash into any areas of scar.

6:02

One note, the only FDA approved indication

6:05

of gadolinium in cardiac imaging is for stress profusion.

6:09

These gaca studies were recently completed

6:11

and that was what was used to get FDA approval

6:15

for gadolinium and cardiac mr.

6:17

Then it's at a single dose quote,

6:19

single dose 0.1 millimeters per kilogram.

6:21

Gata butal is the agent that was approved

6:24

and that's again approved

6:25

for divided doses in stress imaging.

6:28

I won't get too much into the technique here,

6:30

the technical parameters of the technique,

6:32

but just note that these first pass perfusion techniques

6:36

have a little bit higher spatial resolution

6:38

and we have a temporal resolution of every requiring image.

6:42

We'd like to reconstruct an image about once every

6:44

heartbeat, there's trade-offs to this of course signal

6:46

to noise ratio, like everything in cardiac mr,

6:49

there's always trade-offs.

6:50

But part of the issue with the spatial resolution

6:53

and signal to noise is that

6:54

because you're trying to acquire a lot of data

6:57

relatively rapidly, you kind of have to make decisions about

7:00

how much spatial resolution you need.

7:02

This is can, can be improved

7:05

by imaging acceleration approaches like parallel imaging

7:08

or some novel case-based sampling strategies.

7:10

So you know, over time we anticipate

7:13

and it has happened, that image quality has improved a lot

7:16

in these and and it, I think we'll continue

7:18

to go in that direction.

7:20

We always acquire at least three axial slices.

7:22

So basement and apical slices

7:24

and use a saturation recovery preparation

7:26

pulse before imaging.

7:28

And then you can choose to have AGRE readout

7:31

or an SSFP readout might depend on the

7:33

field strength that you're working at.

7:34

For example, three T probably would necessitate AGRE

7:38

readout, whereas we do a lot of our cases at 1.5 T.

7:41

So we tend to use an SSFP readout in those patients.

7:45

And then, uh, here's the result.

7:46

This is what we have worked to acquire.

7:49

Bottom row is stressed, top row is rest

7:53

and uh, you can see contrast coming into the right side

7:56

of the heart, then filling the left ventricle,

7:58

then perfusing the myocardium.

7:59

And this is exactly what we wanna look at.

8:02

How about interpretation?

8:04

So probably the most important part of any of this

8:06

for people is how do we actually interpret these images?

8:09

So I will say that imaging setup is very important

8:12

and normally still I think the mainstay is a qualitative

8:15

evaluation as opposed to something more quantitative,

8:17

which we'll look at a little bit.

8:18

But we always try to review the rest

8:20

and stress slices together.

8:22

So here you can see my setup,

8:23

stress on the bottom, rest on the top.

8:25

And then we try to sync all these images together

8:28

with contrast timing so

8:30

that we're really watching the myocardium be perfused at

8:33

stress and at rest at the same time.

8:35

And we'll see some examples of that in the cases later.

8:38

And what we're looking for here is perfusion defects.

8:40

So areas where there's low perfusion,

8:42

we can see a nice example here in sort of basal interseptal

8:46

and then kind of mid anterior infra septum

8:49

and circumferential of the apex.

8:51

We see these sub endocardial areas of hypoperfusion

8:54

at stress, we don't see them at rest.

8:56

So these are kind of stress rest mismatch

8:58

or reversible profusion defects.

9:00

That's what we're looking for.

9:02

And then we want to correlate those with late gadolinium

9:04

and enhanced images

9:06

or areas of regional wall motion abnormality.

9:09

So here's the same case now correlating with areas of LGE

9:12

and there's no real LGE in this patient.

9:14

So this is gonna be consistent with ischemia.

9:17

And what do I mean when I talk about a perfusion

9:20

defect? So that just means

9:21

There's no signal in the myocardial segment

9:24

during first pass of contrast

9:25

and they're usually sub endocardial, they should conform

9:28

to a vascular territory most of the time.

9:30

And just a brief overview of the physiology.

9:33

As we know as we discussed earlier,

9:35

the mean arterial pressure is actually low

9:38

during vasodilation due to that epicardial stenosis.

9:42

So we've already kind of maximally dilated our arterials.

9:45

We give a vasodilator they can't dilate anymore

9:48

and so they actually drop their overall pressure and

9:51

because they drop their pressure, they're actually trying

9:53

to maintain an elevated pressure

9:56

so they end up vasoconstricting at that point.

9:58

They're the pressure's too low, they vasoconstrict

10:00

and they vasoconstrict so much

10:02

that they actually block contrast from flowing

10:04

in to those segments.

10:06

So that's the reason that we get hypoperfusion,

10:09

low perfusion or low signal

10:11

or main inject contrast in these areas of ischemia.

10:16

If you get a circumferential area of hypoperfusion,

10:19

there is a differential diagnosis for

10:21

that Dark REM artifact is one, we'll talk a bit about that.

10:24

Balanced ischemia is the other three vessel disease

10:27

and then microvascular disease can manifest is that,

10:31

and you know, if there's one takeaway from this part

10:33

of the talk, this is probably at the slide

10:35

for interpreting stress fusion MR.

10:37

So you've really gotta interpret in context

10:39

of the stress rest and LG E images.

10:42

And obviously if they're all normal then

10:43

that's a normal result.

10:45

If you have a perfusion at stress that isn't present at rest

10:50

and there's no associated LGE in that territory

10:53

and that's consistent with ischemia, if you have uh,

10:56

stress profusion defect and a defect at rest

10:59

and LGE in that territory, then

11:01

that's most consistent with infarction.

11:04

There is a caveat to that piece though

11:07

because if you look at the next row,

11:09

I have their stress defect positive rest negative

11:13

and LGE positive, that's also can be an infarct.

11:16

But the reason that that happens is if you remember we give

11:19

the stress injection of contrast first

11:23

and if you have an infarct even five minutes or so

11:27

after when we start to perform the rest imaging the contrast

11:30

that you injected during stress,

11:32

if there's an infarct there, it will still be taken up

11:35

and sit in that area of infarcted scar.

11:38

And so essentially what your imaging at rest is contrast

11:41

that's in scar kind of the equivalent

11:44

of a late gadolinium enhanced image.

11:46

And that can cause actually a false negative on those rust

11:49

images where it looks like there's

11:50

enhancement but there really isn't.

11:52

And so, so that's the one caveat

11:53

to be aware if you have a stress defect

11:56

and an LGE defect that's probably still an infarct with sort

11:59

of this quote early enhancement phenomenon.

12:02

Then the last category is when you have defects that stress

12:05

and rust, but there's nothing on LGE most often

12:08

that's an artifact most often dark growth.

12:11

So that's, that's the really important takeaways.

12:14

A couple of quality assurance issues

12:15

that you need to be aware of.

12:17

When we give stress, we like

12:18

to see some physiologic response obviously.

12:21

And so increase in heart rate by 10 beats per minute.

12:24

We like the patient to have some symptoms flushing,

12:26

experiencing their heart racing, that type of stuff

12:29

to make sure that they're actually getting

12:31

adequate stress response.

12:33

And then there's something we

12:34

Can look at on our images called the splenic switch off.

12:36

This only works if you're using adenosine,

12:38

it will not work for denin.

12:40

And if you see in columns A

12:41

and B here, so at rest we've given contrast

12:45

and the spleen is bright at stress, we've given contrast

12:48

and the spleen is dark.

12:49

That's a physiologic response related to

12:52

how adenosine affects the splenic vascular

12:55

bed actually reduces flow.

12:57

This is the response that we expect.

12:59

This is what we like to see in a normal case

13:01

where the adenosine is worked

13:03

and these particular columns in C

13:04

and D at rest, you see the spleen is bright at rest,

13:08

at stress when we've given adenosine,

13:10

the spleen is also bright.

13:11

So this is a lack

13:12

of splenic switch off which says the adenosine either did

13:16

not work as well as we hoped it would.

13:18

The patient had caffeine

13:20

and didn't allow the adenosine to work as well as it could

13:22

or the adenosine didn't run in the way that we thought.

13:24

So this is something that you have to interrogate

13:26

'cause it's gonna make your results not reliable.

13:30

We also do motion correction.

13:32

This is particularly important for quantitative perfusion

13:35

where you're trying to do

13:36

quantitative voxel wise comparison.

13:38

You can see the difference between corrected

13:40

and non corrected there.

13:41

It works quite well.

13:43

A lot of us tend to like

13:44

to read qualitatively off the non corrected,

13:47

but the motion corrected looks pretty nice.

13:50

Dark rim artifact, you've heard me mention a few times,

13:53

very important to be aware of.

13:54

That's what looks like a circumferential profusion defect,

13:57

but it's usually present at both stress and rust.

14:00

And a true perfusion defect will kind of persist longer

14:03

through the cardiac cycle.

14:04

Dark REM goes away after a couple of heartbeats

14:06

and there are a few speculation ideas on what the sources

14:09

of this are, but they're really, you know,

14:12

people aren't actually a hundred percent sure why we get

14:14

this degree of dark rem.

14:15

It's probably some type of either susceptibility artifact

14:18

or a kind of any, any type artifact related

14:21

to the blood pool and myocardium border.

14:24

But there are other ideas around what this could be as well.

14:27

But something just to be aware of.

14:29

Last couple of notes here.

14:31

There is a semi-quantitative evaluation where you can start

14:34

to put numbers to the profusion process here.

14:37

The semi-quantitative approach looks at the segmental up

14:40

slopes of various segments of the myocardium.

14:44

So here's for example, the blood pool steepest up slope,

14:48

and then at various myocardial segments you can kind

14:50

of see different UPS slopes here.

14:52

And if you can pair those ratios, the ratios of the segment

14:55

to the blood pool at stress

14:57

and rest, you can get what's called a myocardial perfusion

15:00

reserve index.

15:01

And that can be useful for sort

15:04

of a semi-quantitative approach to

15:05

what segments are actually hyper perfused.

15:08

That technique is probably going to go

15:10

by the wayside though, because increasingly there's programs

15:14

and tools for true quantitative myocardial perfusion.

15:17

And this is actually really exciting.

15:18

We're starting to use this in our practice on a daily basis.

15:23

But what quantitative myocardial perfusion actually does is

15:25

it allows you to have a voxel wise assessment

15:28

of blood flow in the myocardium.

15:31

I won't go into the technique or the math,

15:33

but essentially it looks at the arterial input in the blood

15:37

pool and then looks at how the myocardium is enhanced.

15:40

And it uses some math to sort of understand

15:44

what the actual flow into each of one of those voxels is.

15:47

So there's a few assumptions to be made,

15:49

but the bottom line is it works pretty well

15:52

and there's increasing data to show that.

15:54

And just examples here in single vessels,

15:55

coronary artery disease, you can see areas of

15:59

hypoperfusion at stress and normal at rest.

16:03

Here in sort of this mid infra lateral

16:06

and inferior segment, here's an example of three vessel

16:09

where you see pretty much circumferential hypoperfusion

16:12

here at stress.

16:14

So the stress and rest maps look pretty much the same.

16:16

And so this is three vessel disease.

16:18

And just to wrap up to show you, this is kind of

16:20

what you know when we do this clinically, this is kind

16:23

of the readouts that we get.

16:25

We get segmental information on, flows at stress,

16:29

flows at rest, and then when you divide the flow at stress

16:32

by the flow at rest you can get a perfusion reserve.

16:34

And this should be greater than two, at least for normal.

16:37

So this would be a normal case.