0:00

So now, we're going to be talking about CT perfusion.

0:04

And in order to talk about CT perfusion,

0:07

I just wanted to show you this cartoon first.

0:10

So this is a cartoon of a patient

0:12

with an MCA embolus.

0:13

And there's a core of infarction,

0:15

that's tissue that is so ischemic that it's already dead.

0:18

And penumbra is tissue that's at risk of infarcting.

0:21

And I'm going to show you how you identify those

0:25

two areas on CT perfusion imaging and the pitfalls,

0:29

and how it's used in decision making.

0:33

CT perfusion, basically,

0:36

you image one location and you watch bolus of contrast,

0:40

go through that and back to baseline,

0:43

and you image, you know, way past baseline

0:46

so you get the whole bolus.

0:48

And this is Hounsfield unit, so contrast.

0:51

And then, you use an arterial input function

0:53

and the output function and mathematics,

0:56

and you can construct cerebral blood volume,

0:59

cerebral blood flow and tissue transit time maps.

1:04

And we're going to be talking most about CBF

1:08

and another tissue transit time map called the T-Map.

1:11

And this is just a little bit about physiology

1:13

of ischemic stroke.

1:15

So, normal perfusion, this is CBF,

1:18

is over 50 ml per 100 grams per minute.

1:21

Hypoperfusion is under 50 ml

1:23

per 100 grams per minute.

1:24

There's oligemia,

1:25

where you have mild hypoperfusion,

1:27

but you're not really at risk of infarcting tissue.

1:31

And then below 20, you have ischemia.

1:34

Infarction is approximately 10 ml

1:36

per 100 gram a minute.

1:37

And penumbra is somewhere between 10 and 20.

1:41

So, this is the whole basis.

1:43

If you can threshold perfusion maps to below 10 ml

1:47

per 100 gram per minute,

1:48

you can figure out infarcted tissue.

1:50

So, this is the theory behind it.

1:53

This, after much studying and many research articles,

1:58

this is what the consensus is,

2:01

that CT perfusion can increase the sensitivity

2:03

of infarct detection

2:05

over the non-contrast CT and CTA source images,

2:08

the threshold for infarct core

2:11

is less than 30% houndsfield units

2:14

compared to the contralateral side,

2:15

or less than 30% CBF

2:17

compared to the contralateral side.

2:19

So basically,

2:20

people measured with thresholded CBF

2:22

and the final infarct volume,

2:24

and this has the best correlation,

2:27

again, over multiple studies.

2:29

A core infarct of less than 70 CCs

2:31

is the target for intraarterial therapy

2:34

based on multiple studies.

2:35

So, you're basically looking

2:37

at the thresholded CBF,

2:39

less than 70 CCs for the core.

2:42

The tissue at risk of infarction

2:44

is a T-max transit time map.

2:48

So, the threshold for penumbra is

2:51

anything that's greater than 6 seconds.

2:54

And basically, the optimal penumbra core ratio,

2:59

so the total ischemic tissue divided

3:01

by the core is greater than 1.8.

3:04

So, you have a substantial amount of tissue at risk

3:07

that suggests you should have,

3:08

perhaps have IA therapy.

3:10

So, core infarct less than 70 CCs,

3:12

penumbra core ratio greater than 1.8

3:15

amongst other factors.

3:17

CTP has poor contrast to noise ratio,

3:19

so there's a lot of measurement variability.

3:22

It may overestimate the infarct core,

3:25

and it's relatively insensitive

3:27

for identifying tiny strokes.

3:29

So, here's a case of a patient who had some

3:32

right sided weakness.

3:34

In the non-contrast CT,

3:35

hard to see anything.

3:37

CTA source images,

3:38

you can see some subtle abnormality in the left

3:41

putamen, but once you get the perfusion maps,

3:44

you can see an obvious abnormality on the CBV maps

3:48

and big abnormality on the

3:50

CVF and transit time maps.

3:52

This is a follow up CT that

3:54

shows the infarction.

3:56

So, this is just showing you

3:57

CTP does increase the detection of ischemia

4:00

versus non-contrast CT and CTA source images.

4:03

And so typically what happens is, again,

4:07

they create the perfusion maps,

4:09

and they threshold CBF less than 30%

4:12

compared to the contralateral side.

4:14

And so, this patient had a small

4:15

core infarct of 4.2 ml.

4:18

For the penumbra,

4:20

they use a T-max greater than 6 seconds.

4:22

This is the tissue that had a T-max

4:24

greater than 6 seconds.

4:26

There's a huge core penumbra mismatch,

4:29

small core.

4:30

So this would be,

4:31

you know, if it's approximal vessel occlusion,

4:32

it would be a good candidate for thrombolysis.

4:35

And the mismatch ratio was 18.4,

4:37

so it's ten times over the 1.8.

4:40

So, that's what these thresholded maps look like.

4:44

This patient has a lot of tissue at risk of infarction.

4:47

And I just wanted to reiterate,

4:49

though, that the CTP,

4:50

there's a lot of measurement error,

4:52

and it has a poor contrast to noise ratio.

4:54

So, here's a DWI.

4:56

This was taken, like, within 1 hour of the CTP,

4:58

and you can see that there's the infarct.

5:01

And if you threshold this

5:02

so that everything that's dark black

5:04

is lower than threshold,

5:06

and everything that's white is above the threshold,

5:08

it kind of matches the DWI,

5:10

but it's probably bigger.

5:11

There's also another area over here

5:13

that doesn't even match with the DWI.

5:16

If you look at this patient,

5:18

another left MCA stroke on DWI,

5:21

you can see that the CTP is overestimating

5:24

the infarct core, so it's a great technique,

5:27

but you have to be careful.