0:00

So, today we're going to talk a little bit

0:02

about diffusion and FLAIR and gradient echo

0:06

in the analysis of acute ischemic stroke.

0:09

So, diffusion-weighted imaging is one of

0:11

the best biomarkers in all of radiology.

0:13

It provides level one evidence

0:15

for identifying infarct core.

0:17

It's the most accurate method.

0:18

Its sensitivity and specificity

0:20

is 95 to 100 percent.

0:23

And detection depends on processes

0:25

that happen as early as 15 minutes.

0:27

What happens, there are decreased

0:29

sodium-potassium ATPase activity

0:31

in the glial and neuronal cells.

0:33

Intracellular organelles break down.

0:35

There's increased intracellular viscosity.

0:37

So that leads to decreased diffusion.

0:39

The cells swell.

0:40

There's decreased size of

0:41

the extracellular space.

0:43

So that leads to decreased diffusion as well.

0:46

Again, it happens really early on.

0:48

There are some false negatives,

0:50

usually they're little tiny lacunes in

0:51

the brainstem and deep brain nuclei.

0:54

And, um, if you have a posterior fossa,

0:56

you're suspicious for a posterior fossa

0:58

stroke, and you don't see one, you can

1:00

get high-resolution coronal DWI images.

1:03

False positives are stroke mimics,

1:05

which I've covered in another section.

1:08

These are images of a patient who had acute

1:11

ischemic symptoms approximately 45 minutes

1:14

before these images were obtained, and you

1:17

can see these tiny little strokes in the

1:19

left corona radiata, and then this more

1:21

confluent stroke in the left corona radiata.

1:25

This is a two-minute DWI.

1:27

There's also a faster experimental

1:29

DWI that we're using, and even on that

1:31

you can see this very early stroke.

1:35

So best method of imaging infarct

1:38

core, highly sensitive, very early on.

1:42

Reversibility is very uncommon, meaning what

1:45

you see is going to infarct, it is the core.

1:49

So a definition of reversibility is DWI

1:51

of normal tissue that appears normal

1:53

on follow-up FLAIR T2 or follow-up CT.

1:56

And basically some pooled data

1:59

from trials a while ago, out of 119

2:02

strokes, only eight strokes had

2:04

lesion reversal, meaning the final infarct

2:07

was a little bit smaller and the median lesion

2:10

volume reversal was 1.5 mils, so it was trivial.

2:12

So the bottom line is what

2:14

you see is infarct core.

2:15

This is a patient who has a left MCA stroke

2:18

involving the basal ganglion, Corona radiata.

2:21

It's bright on DWI, dark on a, b, c.

2:25

On follow-up you can see there's this

2:26

tiny little area that didn't embalm.

2:28

So tiny little area of reversibility.

2:30

70 cc's is an important threshold.

2:35

This is data from a couple papers,

2:37

a number of people looked at this.

2:38

These were patients with MCA strokes.

2:41

This is DWI lesion volume, um, and

2:43

then good outcome versus poor outcome.

2:46

And you can see that on this initial study, all

2:48

the patients with over 70 cc's had poor outcome.

2:51

And then this was a validation study.

2:53

Almost all the patients with

2:54

over 70 cc's had poor outcome.

2:56

These are young patients and

2:57

patients with right-sided strokes.

2:59

So, 70 cc's has been good.

3:01

a relative cutoff for patients for IA therapy.

3:04

So it's over 70 to 100 ccs.

3:06

Usually, they're going to have poor

3:07

outcome and you really have to decide

3:09

whether you want to thrombolysis or not.

3:11

As far as measuring lesion volume on

3:14

automatic software, they use a threshold

3:16

of less than 620 times 10 to the negative

3:19

3 millimeters per second to threshold.

3:21

And it's very accurate

3:23

with what you see visually.

3:24

We can also measure Length times width times

3:27

depth divided by two, and that has a pretty

3:29

high correlation with automatic thresholding.

3:32

So if you're trying to do this immediately

3:35

at the scanner, and you don't have automatic

3:37

thresholding, you can do ABC over two.

3:39

And when we image somebody with acute stroke

3:42

with MRI, we also get gradient echo images.

3:45

So gradient echo, which is 70 to 90 percent

3:49

sensitive, if you do the highly sensitive

3:51

SWI, it's greater than 90% sensitive.

3:54

However, this takes longer, although there

3:57

are newer methods where you can image

3:59

in two minutes instead of five minutes.

4:02

And then for intracranial hemorrhage

4:03

gradient, echo and SWI are actually more

4:05

sensitive than CT microhemorrhages.

4:08

If you see microhemorrhages, they're

4:09

not a contraindication for intravenous

4:11

or IA therapy, but if you have greater

4:13

than 10 microbleeds, then probably

4:16

don't want to give IV and IA therapy.

4:18

There's, there's some data showing that

4:20

more than 10 microbleeds at increased

4:22

risk for a far larger hemorrhage.

4:24

You can also see prominent medullary

4:26

veins on SWI suggesting hypoxia.

4:30

So, this is an EpiMix sequence,

4:32

and these are SWI sequences.

4:35

And what you can see here is there's

4:37

susceptibility in the MCA stem.

4:38

You can see it on the SWAN sequence as well.

4:41

And if you look closely on this SWI sequence,

4:44

you can see these prominent medullary

4:46

veins confirming that there's hypoxia.

4:51

FLAIR, in general, measures

4:53

timing of infarction.

4:55

So in general, what happens is early

4:59

on, there's a lack of vasogenic edema.

5:00

It indicates that early on, you're

5:03

just looking at cytotoxic edema.

5:04

And then eventually there's

5:05

breakdown of the blood-brain

5:06

barrier and

5:09

formation of vasogenic edema.

5:11

So, typically, early on, you

5:15

don't see much FLAIR abnormality.

5:16

So, here's a stroke at 1.

5:17

5 hours, a large MCA stroke, and you're

5:20

not seeing much on the FLAIR image here.

5:22

And then, typically, by six hours, you see some

5:25

breakdown of blood-brain barrier, some vasogenic

5:27

edema, and you see FLAIR hyperintensity.

5:29

And the DWI hyperintensity,

5:31

however, it's variable.

5:33

Here's another patient with six hours where

5:34

you don't see much FLAIR hyperintensity at all.

5:37

So it does measure time, but it

5:39

is variable is the bottom line.

5:41

It has been used for wake-up stroke trials.

5:44

So basically if you have a patient who wakes

5:47

up with a stroke, you don't know what time

5:49

it occurred and there's little to no FLAIR

5:51

abnormality, it's safe to give IV TPA.

5:56

MR Witness and the Wake Up trials.

5:59

The other thing that you can see on FLAIR

6:01

imaging are vascular hyperintensities like you

6:04

can see in this case in the Sylvian fissure.

6:06

And they're thought to represent good

6:09

collaterals, there's some controversy about

6:10

that, but here's a patient who had known good

6:14

collaterals on CTA, has FLAIR hyperintensities

6:17

in the Sylvian fissure, suggesting some slow

6:20

flow in collaterals, a relatively small infarct,

6:23

a mismatch on the perfusion-weighted images.

6:27

This patient got thrombolyzed and on

6:29

follow-up, the infarct didn't grow at all.

6:31

So this patient had good collaterals.

6:33

And this patient is a patient

6:34

with an ICA occlusion, and you see

6:36

none of those FLAIR hyperintensities.

6:38

You don't really see collaterals

6:39

suggesting they might be poor.

6:41

There's a tiny little DWI

6:43

abnormality, big Tmax abnormality.

6:46

The patient got thrombolyzed,

6:49

but still had a huge infarct.

6:50

FLAIR images can also

6:52

help you assess collaterals.

6:54

Although honestly, CTA and to a

6:56

lesser extent, MRI are much better.