0:02

Hello and welcome to noon conference hosted by MRI online. Noon

0:05

conference was created when the pandemic hit as a

0:08

way to connect the global Radiology community Through free live educational

0:11

conferences that are accessible for all.

0:13

It has become an amazing weekly opportunity to

0:16

learn alongside Radiologists from around the world and we encourage

0:19

you to ask questions and share ideas to help the community Learn and Grow.

0:22

You can access the recording of today's conference and

0:25

previous new conferences by creating a free MRI online account.

0:28

The link will be provided in the chat box. You can also sign

0:31

up for a free trial of MRI online premium membership to

0:34

get access to hundreds of case-based microlearning courses

0:37

across all key Radiology. So Specialties

0:40

learn more at MRI online.com

0:43

Today we are honored to welcome Dr. Jeremy height for a

0:46

lecture on acute ischemic stroke Imaging and thrombectomy

0:49

treatment Dr. Hite is an associate

0:52

professor of radiology and neurosurgery in the

0:55

interim chief of neuroimaging and neuro intervention at

0:58

Stanford University. He is a practicing Diagnostic and

1:01

Interventional neuroradiologist who specializes

1:04

in the diagnosis and minimally invasive treatment of

1:07

ischemic and himrochix stroke

1:10

Dr. Heitz research seeks to understand the patho physiological

1:13

basis of cerebral vascular

1:16

disease in addition. His group is developing new minimally,

1:19

invasive image guided treatments at the end

1:22

of the lecture join Dr. Height in a Q&A session where he will

1:25

address any questions you may have on today's topic. Please remember to

1:28

use the Q&A feature to submit your questions we can get to as many

1:31

as we can before our time is up with that being said we

1:34

are ready to begin today's lecture Dr. Height. Please take

1:37

it from here.

1:38

Well, this should be a lot of fun. So thanks again to the

1:41

folks at MRI online for for inviting me to talk today.

1:44

I want to talk about acute stroke

1:47

Imaging which I think is of interest to the most of the audience here,

1:50

but also a little bit about how we treat patients with

1:53

ischemic stroke by thrombectomy. And I think you'll pretty quickly

1:56

appreciate that the Imaging is very intimately tied to

1:59

the procedure. And in fact, I would argue it

2:02

really is the first step of the procedure. So with that

2:05

let's go ahead and get started here are my disclosures.

2:10

And I'm going to briefly touch on sort of the underpinnings of

2:13

acute stroke physiology. And then how we image patients with

2:16

acute ischemic stroke, then we'll talk about the thrombectomy treatment

2:19

of stroke and some of the data the underpins US doing

2:22

these procedures and then I want to talk about stroke systems

2:25

of care and how the Radiology practices integrate into

2:28

the the transfer and evaluation of these patients and

2:31

then we'll really spend the end of it how to

2:34

reviewing some of the general Imaging triage of patients for thrombectomy treatment,

2:37

but let's start by talking about ischemic stroke.

2:40

So, you know, very briefly stroke happens when

2:43

you interrupt blood flow to the brain and that's what causes an ischemic stroke

2:46

and that can be due to blockage of an artery

2:49

in the brain such as this patient. We see here that there's the arrows

2:52

pointing to blockage of the first part of the middle cerebral artery,

2:55

which is a large artery that supplies the brain and

2:58

you can see that there's a lack of blood flow here to the MCA

3:01

territory in this patient and it could be due

3:04

to blockage of an artery in the neck. But what happens when you

3:07

do have a blockage of blood flow to the brain is it?

3:10

Result in permanent injury to the brain and that's shown here on

3:13

this ischemic stroke here on a diffusion weighted image. Now, if

3:16

you think about it, you don't necessarily have

3:19

all of the brain tissue that you've got

3:22

suddenly die. When you block an artery, right? You can have some tissue

3:25

that really is highly dependent on the nutrients that

3:28

are in the blood that die because you cut off that blood

3:31

supply and and that is kind of schematized here in red but very

3:34

often you can have a larger region of brain sort of

3:37

surrounding this area where you know, there's not enough blood flow

3:40

for the neurons here to maintain their ion gradients. They

3:43

don't function quite normally but there is enough collateral blood flow

3:46

that this tissue is still alive. It's just not functioning and

3:49

if you kind of fundamentally think about these two compartments you

3:52

have dead tissue and tissue at risk and these

3:55

are terms that we are also commonly assigned to

3:58

these two compartments which is the ischemic core and that

4:01

really refers to tissue. We think is likely to be irreversibly injured

4:04

and then the penumbra which refers to

4:07

the potentially salvageable tissue.

4:10

That if we're able to restore blood flow can go

4:13

back to normal and not go on to permanently dive.

4:16

So let's do a little thought experiment. So a patient comes into your

4:19

Ed and they're having an ischemic stroke and you're

4:22

looking at their symptoms and you know that they may look something kind

4:25

of like this. If you could take a picture of their brain, they've got some amount

4:28

of ischemic core that's dead. They've got some hopefully larger

4:31

amount of salvageable penumbra and what do

4:34

you want to do with this patient? Well, you want to reperfuse the brain you

4:37

want to open up that artery that is blocked because if

4:40

you restore blood flow to the brain, all of this blue penumbra

4:43

goes back to normal and the patients left with

4:46

the same amount of permanently and your tissue that they

4:49

had when they came in through your front door and that's really the

4:52

goal of what we're trying to do a stroke patient. So you could achieve this

4:55

reperfusion either with intravenous thrombolysis with

4:58

TPA or connect to place or with an Indo

5:01

vascular thrombectomy procedure to pull the blood clot out

5:04

that's blocking the artery to the brain. Now, if you're

5:07

not successful say the patient is not a candidate

5:10

for TPA or tnk or say that they're you

5:13

know, you're not able to to promptly

5:16

An endovascular thrombectomy. Well, what can happen in those situations

5:19

is that this penumbra will go on to permanently

5:22

die. And so now you can see this patient has

5:25

a much larger area of permanent infarction. And

5:28

obviously this patient is not going to do as well

5:31

as this patient. And so the really the

5:34

goal of all stroke treatment is to restore blood flow to the

5:37

brain and as timely a manner as possible.

5:40

So just a couple of real-world examples of this. This is

5:43

a gentleman who was treated at Stanford. He

5:46

was raking leaves in his front yard when he collapsed and

5:49

that was witnessed by his family. He was transferred initially

5:52

to an outside hospital and quickly transferred over to Stanford

5:55

given the severity of the symptoms and here's his diffusion

5:58

way to image from his arrival at Stanford. You can

6:01

see just the subtlest subtle hyper-intensity that's

6:04

in the caudate head on this diffusion weighted image.

6:07

He had a Mr. Profusion study that

6:10

shows you know, prolong blood flow into the

6:13

entirety of the left MCA territory. This is

6:16

someone that should be you know, a really good thrown back to me candidate and

6:19

here's his angiogram that shows you that M1 occlusion

6:22

in the cath lab and fortunately this was

6:25

revascular eyes. And so all the blood flow is very

6:28

quickly restored to his brain and so he had all

6:31

of this, you know large area of under perfusion

6:34

that could have gone on to die. But because blood

6:37

flow was restored he ends up with an

6:40

Parked that's basically the exact same size as what

6:43

he came in with and he did extremely well and was

6:46

able to leave the hospital a couple days later with minimal symptoms. So

6:49

that's that's what we want to have happen. Let's look

6:52

at this patient. So again, we can see a patient coming this

6:55

time with the right side at stroke and there's you know, a little bit larger

6:58

but still relatively small ischemic core that we're seeing

7:01

and the lentiform nucleus and in the caudate head on

7:04

the right side again, we see that there's Mr. Profusion

7:07

obtain so much larger penumbra here with prolonged blood

7:10

flow on this time to maximum math. This is another

7:13

good patient to take up for treatment. So here's the

7:16

angiogram

7:17

Again, we see an M1 occlusion this time on the right side

7:20

in this patient. And this was one of these very difficult clots

7:23

to remove. So we're pretty good at removing these

7:26

clots, but there is a good 10 to 15% of patients

7:29

where we just cannot open up the artery and

7:32

despite our best efforts. We could not open up this artery it

7:35

remain blocked and because of that all of this

7:38

tissue that the patient had that was at risk of infarcting did

7:41

in fact go on to permanently and fart and

7:44

so you can see that this is someone who's not going to do as well

7:47

in this patient did not do as well compared to the first patient

7:50

that I showed you in which blood flow to the brain was restored.

7:53

And so this is a very simple concept but

7:56

it really is something that we see every day and

7:59

simple is good. And if we just remember

8:02

we're trying to restore the blood flow to minimize the

8:05

amount of permanent ischemic injury to the brain. That's our

8:08

goal.

8:09

So with that you're already

8:12

getting a sense that Imaging is going to be really important

8:15

right? I showed you a bunch of examples already here of

8:18

where Imaging different compartments in the brain, but when

8:21

we think about the evaluation of a stroke patient, what are our

8:24

goals and I would argue that there's really kind of four goals that we're thinking about.

8:27

So first off we want to confirm that we

8:30

have the right diagnosis. We want to make sure that we're dealing within a patient

8:33

with an ischemic stroke. It could be a hemorrhagic stroke. It

8:36

could also be a stroke mimic and so we'd like to ideally rule

8:39

out stroke mimics. If we are able to we want

8:42

to make sure that we're identifying any contraindications to

8:45

revascularization and that could again be

8:48

by intravenous thrombolysis or by a throne back to

8:51

me procedure and then increasingly we're really thinking a lot

8:54

about identifying thrombectomy candidates and how we almost optimally

8:57

do that.

8:58

So when we talk about confirming the diagnosis, I

9:01

think that's something that's on all of our minds. That's Radiologists. We're

9:04

thinking about that all the time when we're looking at a

9:07

scan and it says, you know evaluate for a stroke and you know, we have two

9:10

modalities that are disposal. We have CT and MRI.

9:13

They both certainly work. I like this picture because it's

9:16

very easy to see the obvious stroke. It's quite large

9:19

and this patient on the diffusion weight at MRI, but on

9:22

the CT scan, it's more challenging. If you look though,

9:25

you can see that we have this hypodensity here that's in

9:28

the left insula. It's extending into the left lentiform nucleus,

9:31

but that relative drop in density compared to

9:34

the side. I think is more challenging to see and that's the advantage

9:37

of MRIs that the signal to noise is just much higher. It's much

9:40

easier to make the stroke diagnosis, but regardless, these

9:43

are two modalities and the Radiologists needs to be Adept and

9:46

interpretation of both of them.

9:48

So why do we typically do CT

9:51

in this country? Well, it's fast it's really

9:54

widely available. Pretty much every emergency department has a

9:57

CT scanner and it's relatively inexpensive compared to

10:00

to MRI and by Contrast MRI has

10:03

a reputation of being slower. It's certainly

10:06

not as always available and it is more expensive. The

10:09

machines are more expensive and they're they're more expensive to

10:12

to operate but I would make a slight

10:15

argument that these perceptions may not

10:18

quite be true and I would challenge you all to start to think

10:21

about that a little bit so I'm right can actually be done quite quickly. It is

10:24

increasingly available. You have to make a little

10:27

bit of effort to make it available for stroke patients all

10:30

times and it may actually be cost assist

10:33

efficient because a lot of times when you have a CT scan that doesn't

10:36

reveal a diagnosis. What do they do? Well, you go on and you

10:39

get an MRI, so if you can just go straight to the MRI, you're

10:42

gonna be able to answer more of those clinical questions and potentially there's

10:45

room for cost savings and that so this

10:48

is the

10:48

Most streamlined MRI protocol we use at Stanford

10:51

for triaging patients for thrombectomy. And I put

10:54

this up to show you that you can do this pretty quickly. What are

10:57

we doing here? Well, we have a diffusion weighted image to evaluate for how

11:00

significant the ischemic chorus. We have a

11:03

gradient Echo sequence where we're looking for the presence of any Hemorrhage

11:06

which could be a contraindication to treatment. It also can help you see

11:09

the thrombus in many cases. We have an MR angiogram

11:12

to evaluate where a large vessel occlusion is and

11:15

in this case, we have an M1 occlusion here on the left side. And then

11:18

we do perfusion imaging to identify the penumbra

11:21

and make sure that there is salvageable tissue. So these

11:24

are the scan times for each sequence. And if you add up the

11:27

scan time, it's just a little over five minutes and obviously there's some transition

11:30

time in between the two, you know, you've got a look image

11:33

but really you can do an MR pretty quickly and

11:36

around 10 minutes or less if you're really dedicated to

11:39

getting it done quickly and I would argue that that's pretty comparable

11:42

to CT in most patients

11:45

with an ischemic stroke. So Mr. Can certainly

11:48

be

11:48

quickly, and it's something to at least think about

11:51

So we also want to try to exclude stroke

11:54

mimics. And so these are three different patients who

11:57

were brought in for thought that they had an

12:00

ischemic stroke, but none of them does and so

12:03

let's look at each of these in turn. So this first one we've got a flare

12:06

image, you know, really? It doesn't look horribly remarkable.

12:09

I will tell you this patient is hemiplegic and

12:12

has a gaze preference so they certainly look like someone who has

12:15

a large vessel occlusion stroke. Maybe you

12:18

can see that there's some increased, you

12:21

know, T2 signal in the cortex in this patient.

12:24

But if you add an ASL to look at blood flow, we

12:27

see that there's extremely high blood flow to the cortex in the

12:30

right Hemisphere and this patient is actively seizing right. So

12:33

that's a stroke mimic. This is not someone that you want to give TPA to

12:37

This patient was brought in with what was thought to be a cute

12:40

weakness and then a little bit later you find out that it may have been

12:43

more of a slowly Progressive weakness, but here's their non-con CT

12:46

and we can certainly see areas of hypodensity in

12:49

the left MCA territory. There's some

12:52

Mass Effect on The ventricle, you might worry that you're looking at a

12:55

Subacute stroke that's you know already kind of the cats out

12:58

of the bag but it looks a little bit funny right? We've got some sparing of

13:01

tissue in the midst of the areas of hypodensity. Then

13:04

you want to look at the whole study and contrast can be

13:07

helpful because this is a patient with a secretory meningioma

13:10

where all of this edema and the patients

13:13

weakness are really due to this tumor that

13:16

just needs to be removed.

13:18

This last patient here came in with concern for

13:21

a posterior circulation stroke. They were having visual disturbances and

13:24

on the non-con CT. We see hypodensity and

13:27

the occipital lobes. It looks a little bit symmetric but

13:30

there is a little bit of sparing of the cortex over

13:33

the area of hypodensity and both of these regions and this

13:36

is actually a patient who is quite hypertensive. And

13:39

this is a case of press posterior reversible encephalopathy

13:42

syndrome. And so, you know, all of

13:45

these are stroke mimics and if you're not careful and you don't

13:48

look at the Imaging closely, you potentially could miss these diagnoses but

13:51

none of these patients is gonna benefit from intravenous

13:54

thrombolysis or thrombectomy. So if

13:57

we're able to identify the mimics that is good for the patients.

14:01

So what about contraindications to treatment? Well, there are a few.

14:04

So here's a patient who came in again with visual symptoms

14:07

concern for ischemic stroke in the Ed. Here's the

14:10

non-con CT and we're seeing a sort of heterogeneous

14:13

hypodense lesion here, you know,

14:16

you're getting the sense that doesn't really look like a stroke

14:19

and sure enough this ends up being a primary brain neoplasm.

14:22

This is a GBM certainly not something you

14:25

want to be doing giving TPA to as the presence

14:28

of an intra a parenchymal tumor is a contraindication this

14:31

middle patient here with someone who is altered came

14:34

in with some isolated weakness was thought to

14:37

maybe have a deep stroke. It's a non-con CT and we're

14:40

seeing little Foci of hyper density that certainly look

14:43

concerning for small areas of hemorrhage. And these are

14:46

Hemorrhage is related to mycotic aneurysms. So certainly

14:49

not a patient. You want to be giving TPA to important to

14:52

identify this contraindication this last

14:55

patient. This was a I can't make this up and investment banker

14:58

who tripped on the cobblestones in nantuck.

15:01

And actually had weakness on the left side of the body got

15:04

this head CT and you can get sensors a

15:07

little bit of trauma here and there is some hyper density

15:10

in the central sulcus here. And this is traumatic

15:13

subarachnoid hemorrhage. That was probably causing some

15:16

irritation may be a focal seizure to account for

15:19

the patient's weakness. So three contraindications to

15:22

treatment that you want to pick up and certainly are non-invasive

15:25

Imaging is helpful and identifying these contraindications.

15:28

All right. So increasingly we're

15:31

really concerned about identifying patients for

15:34

thrombectomy because these are patients who can really benefit from

15:37

that procedure and it's up to the radiologist to

15:40

work with the stroke teams to make sure that we're identifying these patients

15:43

correctly. And so generally, who are we looking

15:46

for? I would say that patients who are good. Thrombectomy candidates

15:49

fit these three criteria first off. They should

15:52

have a relatively small ischemic core.

15:55

All right, right. Now the evidence is that if patients

15:58

have less permanent injury.

16:01

More likely to do. Well that does not mean thrombectomy is

16:04

not going to benefit a patient who has a larger core it

16:07

probably does and we're we have

16:10

ongoing trials that may change this first point and the

16:13

next year or so the patient should also have some

16:16

salvageable tissue. All right, and that's evidenced by

16:19

a larger penumbra and that makes sense. Right? If

16:22

you're going to open up a blood vessel. You want to restore blood flow to

16:25

some amount of brain that's not dead because if

16:28

it's all dead, I don't see how restoring blood flow is

16:31

going to help that patient. And so the patient should have a

16:34

small core a larger penumbra and they should also have a

16:37

blood vessel that is amenable to end of vascular thrombectomy.

16:40

This is defined right now as a large

16:43

vessel occlusion, and that means a occlusion of

16:46

the internal crowded artery the first part of the middle cerebral artery

16:49

the first part of the anterior cerebral artery and

16:52

now we finally have evidence for the basil or intra

16:55

cranial vertebral arteries. This is

16:58

another area where I can probably give you this talk next year and the

17:01

Data may be new and different and we are increasingly going out into

17:04

M2 vessels and to more distal vessels within

17:07

the intracranial circulation. But right now

17:10

at least we should really be looking for patients who have a relatively

17:13

small amount of ischemic injury a larger penumbra

17:16

and a vessel that is amenable to thrombectomy meaning

17:19

a large vessel occlusion.

17:21

You can obviously identify those characteristics on

17:24

MRI or CT and so by CT. We

17:27

can you CTP to find the core we can use a non-con we'll

17:30

get into that in a minute. We can identify the number using

17:33

perfusion or patient symptoms and we can certainly

17:36

do a CTA to look for an lvo.

17:40

Okay. So let's take a second and

17:43

talk about Precision Imaging and I think this is

17:46

really important because there's really been an evolution and how we

17:49

use Imaging to guide stroke patient treatment. And this

17:52

is a relatively old church and I

17:55

kind of like colored the picture here to make it look old but

17:58

it is actually old and I put this up here because of the

18:01

clock and so for many many years the treatment of patients

18:04

with stroke was really governed by the clock how long has it

18:07

been since their symptoms started and basically you

18:10

would have a really short time window to offer

18:13

an active treatment to stroke patients. If you

18:16

were a patient who shows up within zero to three hours of your symptom

18:19

onset, you could get intravenous TPA if

18:22

you qualified based on your Imaging and then that

18:25

got extended out to four and a half hours with some additional

18:28

data all the FDA on label approval for

18:31

for TPA remain only at three hours. Most people will still give

18:34

it up to four and a half hours from symptom onset though, but

18:37

what's the problem with using the clock and let's talk?

18:40

About that for a second and this is a Nantucket. By the

18:43

way, if you haven't been it's a really lovely place to go. Well, here's the problem with

18:46

the clock. These are three different patients and they were

18:49

all last seen normal within 30 minutes of each

18:52

other and they're getting evaluated at six and a half to seven hours

18:55

from symptom onset.

18:56

So here's the MRA in each of these patients and

18:59

so we can see the each of them has an M1 occlusion

19:02

on the left side. They each have a very similar Stroke

19:05

Scale as well.

19:07

Here's their perfusion in it. And so each of them has a

19:10

pretty similar sized ischemic penumbra, right?

19:13

We're seeing that there may be some heterogeneity and

19:16

you know how severe the blood flow is on these T-Max Maps,

19:19

but we can see each of them has a perfusion delay into the

19:22

left MCA territory so far. They're looking really quite similar.

19:26

But when we look at their DWI and look at

19:29

the size of each of these patients, ischemic core, look at

19:32

the variability that we're seeing these first two patients have

19:35

pretty small ischemic cores, whereas this last patient has a

19:38

pretty large ischemic core. And so even though

19:41

all three of these patients had their symptoms starting at essentially the

19:44

same time. There's a market difference in

19:47

the size of their ischemic core when they're being evaluated for

19:50

potential treatment. And so if you're really relying on

19:53

the clock, you're not going to be able to sort out these three individuals

19:56

and you would either have to treat none of

19:59

them or treat all of them knowing that at least this patient is probably unlikely

20:02

to benefit from your treatment and this

20:05

was looked at in the diffuse 2 cohort and so

20:08

these are patients with known ICA or MCA collusions, and

20:11

they have a known symptom onset time. And so if

20:14

you graph their core based on when their symptoms

20:17

started and on the x-axis to the time of

20:20

Imaging evaluation, and then they're core volume by

20:23

DWI on the y-axis you can see

20:26

There's tremendous variability, right? You could be this patient

20:29

you're at 11 hours and you've got a super small core

20:32

or you could be this patient at a little bit more than two hours

20:35

with a Holo hemispheric infarction. And so there's

20:38

a lot of physiologic variability people have very

20:41

different abilities to tolerate an ischemic

20:44

Challenge and you always hope you're one of these patients but

20:47

the reality is when you're just looking at a patient without Imaging

20:50

you don't know where any individual patient lands in

20:53

the Spectrum and that's why Imaging is so so critical and

20:56

it really is a form of precision medicine. So what

20:59

we're really looking for is a favorable patient who's

21:02

got a small core a larger penumbra and

21:05

there should be some quantifiable difference that you're

21:08

seeing and of course the large vessel occlusion patients who

21:11

are less favorable really have what we call a matched

21:14

profile their core is the same size as their potentially

21:17

salvageable tissue. And so the treatment of these patients is

21:20

unlikely to offer a benefit and there's very good data to

21:23

support this concept.

21:26

So even though historically stroke treatment is

21:29

based on the clock. I think we're really moving into this era of

21:32

precision Imaging where we're using physiologic Imaging and

21:35

other Imaging techniques to really triage

21:38

our patients for treatment and that's really great for radiology. And

21:41

it's really important that we all understand that so that our

21:44

role in helping take care of these patients as well understood. Okay.

21:47

So let's pivot and talk now about thrombectomy. This

21:50

is what a thrombectomy looks like.

21:53

This is one of my patients. There is a right M1

21:56

occlusion shown here by the arrow. This is the thrombectomy

21:59

procedure being performed and then after a

22:02

successful thrombectomy, the blue arrow is showing you restoration of

22:05

flow into that that right MCA territory.

22:08

And so, you know, we do this procedure quite a

22:11

lot at Stanford and we do it because it is so tremendously impactful

22:14

and here's an slide that

22:17

just shows the evidence from the largest trials that have validated the

22:20

efficacy of thrombectomy. So this is again for internal credit

22:23

artery or M1 occlusions of The Middle.

22:26

Little artery and these are all randomized trials. All

22:29

of them are published in the New England Journal of Medicine. This dashed

22:32

line separates the early window trials from

22:35

the late window trials and we Define early as now zero

22:38

to six hours and then late window as 6 to 16 or

22:41

24 hours. So the trial names are here. The

22:44

medical arm is the blue and the red arm

22:47

is the throne back to me and the y-axis is what percentage

22:50

of patients had a good functional outcome. Meaning that they

22:53

are living independently three months after their procedure. And

22:56

if you see in each one of these trials there is

22:59

a market benefit of thrombectomy over

23:02

our standard medical therapy and the publication of these trials

23:05

is really led to a sea change in the care of these patients. And

23:08

in fact, the number needed to treat is only

23:11

two to three for these procedures and that's really quite impressive. So

23:14

for every two to three patients you treat one of

23:17

them will do well and just to contrast that with something you

23:20

might be more familiar with if you look at patients with an st

23:23

elevation Mi who go to you know, get

23:26

Into the cath lab for a coronary artery intervention with

23:29

an angioplastier stent placement. The number needed to treat for

23:32

that procedure is somewhere around 17. So this

23:35

is much much more effective than even coronary intervention

23:38

for St. Elevation Mi and we have certainly

23:41

built whole systems of our health care around the

23:44

care of those cardiac patients and that starts to make you

23:47

appreciate the importance of doing something similar that works really

23:50

well for our stroke patients. So how do we

23:53

triage these patients for thrombectomy and we've talked about this a

23:56

little bit more but let's delve into a little bit more detail. Okay. So

23:59

what do we need? Well we talked about right now. We're

24:02

treating patients with a large vessel occlusion. And so

24:05

that we need to have some sort of vessel Imaging in

24:08

most patients and that's really can be done by an MRN

24:11

angiogram or a CT angiogram. You can see the left M1 occlusion

24:14

in both of these patients. There should be some sort

24:17

of assessment for how much injury there

24:20

is to the brain at Baseline. All right. How big do you think the ischemic core

24:23

is so if you've got a diffusion way to

24:26

Great. We think that's a great estimate of you know, how big the core

24:29

is if you just have a non-contrast CT

24:32

great. We can look for hypodensity and make sure that we're

24:35

not seeing, you know, the whole entirety of the MCA territory being

24:38

hypodens, which would be consistent with probably an

24:41

irreversible infarct and then we can

24:44

use perfusion imaging to identify the presence of

24:47

a salvageable penumbra.

24:50

So we can also do perfusion imaging

24:53

to look at the core. Okay, so that's

24:56

a kind of a variation on the CT technique. And if

24:59

you're getting the penumbral Imaging you would also be getting the same

25:02

Imaging of the core from that same technique. However, you

25:05

really do not have to do perfusion in

25:08

early time Windows. It is perfectly acceptable to

25:11

do so and two of the randomized trials

25:14

did use perfusion to treat patients. So there's good

25:17

data for it, but you don't need to do it to

25:20

make a good treatment decision that's important to understand. Once you

25:23

get into those late time window patients Beyond six hours

25:26

both of the trials that suggested a benefit in

25:29

those patients did use perfusion imaging to identify at least

25:32

the ischemic core and you can discern whether there's a

25:35

salvageable penumbra either from profusion or from symptoms severity

25:38

that is mismatch from the size of their core.

25:42

All right. So that's our Imaging goals. That's kind of

25:45

what we're looking at. How do we start to integrate all of this into the

25:48

care and the workflow of patients with ischemic

25:51

stroke? So I'm going to kind of talk a little bit about some of the systems of

25:54

care that we have in this country and there's certainly variability worldwide. But

25:57

you know since I practice in the US, that's what

26:00

I thought I would talk about. So first off this slides a couple years old

26:03

now, but I put this up to show that Throne back

26:06

to me really remains a limited resource. There are

26:09

not neuron eventualists practicing absolutely everywhere. There

26:12

are still relatively few of us. And I

26:15

think we're getting closer to having the number be appropriately matched

26:18

to the number of patients that need us, but there's certainly a geographic difference.

26:21

If you live in a major metropolitan area

26:24

on the coast. There's likely a neuron

26:27

eventualist near you. But if you live in the center of the country, there are

26:30

a large areas where you could be quite a ways

26:33

away from a neuron evangelist that could offer your throne back

26:36

to me that's shown here by driving or flying time is what these

26:39

color overlays mean.

26:41

So the way patients tend to flow to neuron

26:44

eventualists and stroke care teams in this country is

26:47

they will have a stroke and they will be brought

26:50

into a hospital and usually their first brought to a primary

26:53

stroke center primary stroke centers are fabulous. They're

26:56

able to offer TBA or connect to

26:59

place and they're sort of the Frontline hospital system for taking

27:02

care of these patients, but in general they don't have

27:05

more comprehensive level of stroke care, which

27:08

is something you're going to get at a comprehensive stroke center. And

27:11

so if you are a patient with a large vessel occlusion who needs

27:14

a throne back to me, you're going to get transferred to a comprehensive

27:17

stroke center or a thrombectomy capable center

27:20

with Staffing who can perform the throne back to me

27:23

procedure. There's some very interesting work going on right now

27:26

a Rhode Island's really doing a lot of great work at Brown on potentially

27:29

by passing a primary stroke center for

27:32

patients who are found to have a large vessel occlusion in the field and

27:35

going direct to a comprehensive stroke center. And that seems

27:38

to actually potentially be very beneficial.

27:41

For the care of these patients and lead to improved outcomes. So

27:44

this is a very active area of research that I think

27:47

will lead to some changes in our practice in the coming years.

27:51

So at Stanford most of our patients here are transferred in

27:54

so about 85% of our patients are transferred from

27:57

primary stroke centers the number of patients who walk in

28:00

through our Ed with a stroke or who have a stroke while

28:03

they're in our hospital is a smaller proportion. It's around 15% last

28:06

time we looked at our numbers. So what I

28:09

want to do is kind of walk you through the different workflows for

28:12

each of these patients to give you flavor of how we do things that Stanford and

28:15

maybe that'll help you think about what you're doing your hospital

28:18

things you could do differently certainly probably do

28:21

things some things better than us, too. There's a whole bunch of ideas and

28:24

that's why we talk about it. So let's start if you present to an

28:27

outside hospital with the stroke what happens? Well, you go

28:30

into an Ed and you're found to have, you know

28:33

symptoms consistent with a large vessel inclusion you get

28:36

your valuation and that Ed thinks that you need

28:39

to throw back to me, but they don't have a neuron eventualist.

28:42

What they do is they call our Stanford Transfer Center and I

28:45

get paged you can see this happens at all kinds of times a

28:48

day and the Stanford's Transfer

28:51

Center will activate me our stroke neurology

28:54

attending and our stroke neurology fellow and the three

28:57

of us will get on the phone within five minutes and be connected to

29:00

the Ed doc at the referring hospital.

29:03

Do you hear about the patient at the same time that

29:06

we're taking that phone call? We actually launch our helicopter. We

29:09

just go ahead and say let's get the the ball rolling. Let's

29:12

start to arrange the patient transfer and so our life

29:15

flight will lift while we're on the phone and if

29:18

we accept the patient, they keep flying if we change our mind for

29:21

any reason they come back so they go and they pick up the

29:24

patient and here's now the helicopter returning to Stanford. This

29:27

was a stroke transfer where I was with three

29:30

of my four kids and the patients about to land and I'm doing a patient

29:33

transfer for the helicopter and a child

29:36

transfer with my life. So sometimes you got a multi-test but

29:39

it works really well and the helicopter.

29:42

Lands on the roof. And then what do you do? Well, you've got to really

29:45

expedite things we go straight from the roof of the hospital to

29:48

the Imaging suite and we want to go straight down

29:51

and we want to image our patient and that's what we will

29:54

do and here we can see we've got an image that shows no ischemic

29:57

core on this patient, but a large salvageable penumbra

30:00

and perfusion imaging this is a great stroke treatment

30:04

patient. They're clearly highly impaired and here is

30:07

the patient's angiogram in the

30:10

cath lab. So this is an anterior posterior view. So

30:13

you're looking right at the patient and the black is the contrast diet that

30:16

we've injected into the left internal credit artery and right

30:19

here. We see blockage of the first part of the middle cerebral

30:22

artery. So that's our Target. And what I'm going to do is show

30:25

you a few videos about how we do this storm back to me procedure. So

30:28

the first thing we do is we actually cross the clot with our equipment. Here's

30:31

a micro wire and a micro catheter. That's Traverse

30:34

the area of occlusion inside the middle cerebral artery and

30:37

this gives us a conduit to introduce them back to

30:40

me devices. So now we

30:42

Of the wired here. We are introducing a stent retriever device inside that

30:45

micro catheter and you can see that we're across

30:48

the clot with a little bit of device that will be past the

30:51

clot to make sure it helps to drag it out.

30:53

Next thing we do is we unsheath our stent

30:56

retriever device here. It is kind of opening up as we pull the micro catheter

30:59

out and this will pin the clot between the

31:02

outside of the stent Retriever and the inner wall of the artery and

31:05

the stent kind of pushes into the clot to

31:08

grip it.

31:09

Then what we can do is we could Advance this aspiration catheter

31:12

right up to the clot interface. And once

31:15

we get up to the claw interface, we can hook on suction

31:18

to really kind of suck the clot into this catheter and

31:21

pin it with the stent helping to drag it out.

31:24

Now. This is something that we've started doing

31:27

now with this is a balloon guide catheter that

31:30

arrests flow proximately so that as you're pulling the

31:33

cloud out the blood pressure is not pushing it away from you and this

31:36

really seems to be something that's helpful for more complete removal

31:39

of the clot and there's a lot of variations that

31:42

are going on and trying to determine what the optimal thrombectomy

31:45

procedure should look like, but this is something that

31:48

we do pretty commonly here after the balloons inflated. We

31:51

can remove the clot in this manner here. It is

31:54

being pulled out of the patient and you can see do a

31:57

nice slow pull the clot should be pinned right here if you're

32:00

lucky and then you can defeat the balloon and you

32:03

can do an angiographic run and

32:06

you can see that you go from a blocked vessel.

32:09

to Restoration of flow

32:11

All right, and then here's the lateral angiogram just to show

32:14

you that we've really revascularized the entirety of the

32:17

MCA territory in this patient. And that's exactly what

32:20

you want to see happen. And if you're successful, here's this

32:23

patient's initial perfusion. Imaging again. It

32:26

didn't call a core because of the thresholding but here's

32:29

where you would expect to see some infarct in the posterior insula the

32:32

fall of MRI confirms that that injury stays

32:35

about the same size.

32:37

This was the initial CTA where the M1 occlusion is located.

32:40

Here's the MRI after thrombectomy, you can see now the vessel

32:43

is wide open and then the profusion Imaging

32:46

goes in a normalizes on the arterial spin

32:49

label perfusion sequence. And so this is

32:52

the patient who did very well and this is an illustration of how

32:55

we would bring a patient in from an outside hospital to Stanford

32:58

for thrombectomy treatment. So in general a

33:01

stroke will be identified. Typically an outside hospital is going

33:04

to at least do a head CT make a decision about giving TPA

33:07

or TMK if they're eligible and then they're going

33:10

to do additional Imaging and clinical evaluation to determine

33:13

if there are neuro Interventional thrombectomy candidate if they

33:16

are you call a comprehensive stroke center and you request

33:19

transfer the patient gets transferred right now.

33:22

We do a lot of repeat Imaging because our

33:25

transfer times are often quite lengthy given the large geographic

33:28

area we cover this is something that other

33:31

places don't necessarily need to do if they can transfer patients quickly

33:34

and then the patient good and go for thrombectomy treatment.

33:38

All right. So how does that differ compared to a patient who

33:41

comes in through our emergency department? Well, here's the Stanford hospital

33:44

about a year ago or two years ago now, I guess from the

33:47

Christmas lights are up. So getting close to the holidays and a patient

33:50

will be brought by an ambulance into the Ed here

33:53

and ideally you get a heads up from the ambulance that this is

33:56

someone who may have a stroke and you can bypass the evaluation

33:59

area and go straight to the scanner and evaluate the

34:02

patient at the time your Expediting the

34:05

Imaging sometimes my fellows here

34:08

about these cases before I do and they shoot me a text. They look

34:11

like this and if you open up I just get the bat signal which is kind of

34:14

fun. And then other times we will just be activated through

34:17

our Transfer Center in a similar way and sometimes there

34:20

are a whole lot of patients with Strokes that we get activated for

34:23

and a single day but we're gonna do the same

34:26

thing. We want to keep our Imaging consistent. That's how we practice at

34:29

Stanford. So this is a 59 year old with afib who came

34:32

into our edu with the left facial group and was Hemi product

34:35

on the left side. And here's the patient.

34:38

Sign Imaging so a small estimated core on the perfusion here.

34:41

We can see that there's a larger penumbra surrounding

34:44

this small core and we can see that

34:47

there's a right M1 occlusion on the CTA. So this is

34:50

this is a go case we got a small core large penumbra and

34:53

a large vessel occlusion.

34:55

So here's the patient's first picture in

34:58

the cath lab where we're injecting the right internal credit artery and

35:01

you can see that sure enough on this patient there. Is

35:04

that right M1 occlusion. So we were

35:07

very quickly able to go up and aspirate

35:10

this clot out and here's the picture right after we

35:13

did that and now you can see that we've really restored all

35:16

the blood flow with that vessel being widely open

35:19

and that's what you're trying to do. And if you have a really

35:22

efficient system you can do this exceptionally fast

35:25

and the time in this patient from arterial puncture

35:28

to when we're able to completely restore the

35:31

blood flow here was four minutes. Okay, and not that's really really

35:34

quick that we're able to do that but with a highly trained

35:37

team and a very streamlined workflow, you can

35:40

potentially revasculars patients and really really

35:43

quick time frames. Sometimes it takes longer

35:46

and this is certainly I'm showing you the faster and of

35:49

things but I just want to point out that you can do this stuff very efficiently.

35:53

So this was the estimated core before we took the

35:56

patient and here is his follow PET CT the following

35:59

day and you could see there's no growth in the amount

36:02

of the core and this patient actually walked out of the hospital with only

36:05

the mildest official troops two days later and was

36:08

changed on and his anticoagulation regiment.

36:11

So here's a kind of a schematic

36:14

of our in-house stroke workflow. So again, we'd

36:17

like to get a heads up from the EMS if they are suspecting a

36:20

stroke they get put on the gurney and wield in straight

36:23

to the CT scanner. We're checking things like glucose. We do a non-con

36:26

to make sure there's no Hemorrhage if they are eligible. We

36:29

will you know, grab our IV send in lab work make it

36:32

just determination about TPA after

36:35

we've assigned a stroke scale and made sure we think it is a stroke while

36:38

the non-con is being interpreted and then if

36:41

they're severely impaired, you know, we will do a CTA and CTP to

36:44

determine whether they're potentially a throne back to

36:47

me Canada. So really you're kind of doing things overall relatively

36:50

similarly. It's just how they're getting to

36:53

The Imaging Suite that differs just slightly. So where

36:56

does that leave us? Well, we have a pretty good sense of

36:59

whom we should be treating. And again, this is a

37:02

moving Target. There are a lot of active studies going

37:05

on that are looking to continue to expand the indications for

37:08

thrombectomy. And I think we're going to continue to learn

37:11

more in the coming years, which is great because I

37:14

expect we'll be expanding our eligibility. We

37:17

know how to treat them pretty well. So we've got good

37:20

techniques. There's a lot of work on the Interventional side to sort

37:23

out. What's the best thing to do? But we've made a lot of progress and

37:26

we're getting better at having systems in place

37:29

to perform back to me quickly and there's still I think room to you know,

37:33

Interstate level and even perhaps a national level to really optimize

37:36

this

37:38

So what are the challenges and how can we leverage technology

37:41

to help us out? Well, after these late window Trials of

37:44

Dawn and diffuse 3. We suddenly had

37:47

a market increase in thrombectomy treatment

37:50

because now large vessel occlusions could be treated up to

37:53

16 to 24 hours since patients were last known

37:56

to be well, so that's suddenly opens up the floodgates of

37:59

people who could potentially be treated which is a great problem to

38:02

have and this is what happened afterwards to

38:05

our stroke volumes and just kept going up and up and

38:08

up but this creates some real workflow challenges, right? How do

38:11

we evaluate all these patients? Well, we're gonna

38:14

have to figure out a way to deal with this because you can't just

38:17

transfer everybody to these comprehensive centers without

38:20

having a better sense of whether they actually are potentially

38:23

a treatment Canada and that's really meant that

38:26

it's been a little bit more up to the Community Hospitals to help

38:29

identify who could be a treatment Canada. All

38:32

right, and that's at least going to be identifying the

38:35

presence of a large vessel Collision. We'll talk about this in a minute.

38:38

And then you'd be really great. If we have these long transfer times

38:41

because they're coming from far away if we can kind

38:44

of predict. Well, what's going to happen after we transfer it and then how

38:47

can we use automated Imaging and artificial intelligence to help

38:50

us manage this increasingly complex workflow. Well,

38:53

let's talk about the possibility of doing profusion Imaging

38:56

in the community. So you're still kind of

38:59

now asking Community Hospitals to ideally identify

39:02

a thrombectomy candidate, which means can they tell

39:05

us if there's a small core a salvageable penumbra and

39:08

a large vessel occlusion. Well, what used

39:11

to happen was an outside hospital with you

39:14

know, admittedly fewer resources in a place like Stanford,

39:17

they would be able to get a non-con CT and they would

39:20

be able to do Stroke Scale. And so you would be able to hear well, you

39:23

know, there's maybe some subtle hypodensity on the left

39:26

MCA territory and I've got a patient who's very severely impaired

39:29

symptomatically, so I'm inferring that. There's probably

39:32

a large vessel inclusion and that's okay, but

39:35

it turns out there are a lot of patients where you're getting fooled by.

39:38

Mimic or where you're looking at a chronic infarct

39:41

or where it just doesn't quite make sense. And so then

39:44

we started asking pay outside hospitals. Well, can you

39:47

at least get us a CTA so that you can really demonstrate that

39:50

they do have a large vessel inclusion because that's

39:53

the patient we want to bring over and potentially help. And so now this

39:56

is sort of the minimum that people should be doing

39:59

as a non-concy Tina CTA to help

40:02

us make a transfer decision along with of course a

40:05

clinical evaluation.

40:07

So what about if a community hospital is able

40:10

to do perfusion? Well, then you can say well we know what the

40:13

core looks like. We know if there's a salvageable penumbra.

40:16

So that's really helpful. And then if they're all doing

40:19

a CTA great, you can see that there's a large vessel occlusion. This

40:22

is a lot of Imaging for smaller hospitals

40:25

to do and and I think we all recognize that and

40:28

want to try to minimize that so what could you

40:31

do to simplify things? Well instead of using a

40:34

CTA to identify lvo. Could you use profusion to

40:37

identify the presence of an LDL and it

40:40

turns out that that may be a valid strategy. This is something that we're

40:43

interested in testing. And so if you look at these T-Max images,

40:46

I think anyone looking at this you can

40:49

probably discern that these are patients with large vessel occlusions,

40:52

or at least thrombectomy eligible occlusion. So

40:55

this first one, you know, the whole MCA territoring

40:58

the left side has a perfusion delay. So that's an M1

41:01

or potentially an internal credit artery occlusion.

41:04

This patient we can look at this and say well that's not the whole

41:07

MCA territory, but certainly disappear division

41:10

looks like it has a perfusion delay and that's often a good

41:13

Target for neuron eventualists, but we can discern that

41:16

there's an M2 Superior division occlusion. And in

41:19

this patient that perfusion delays most more posteriorly

41:22

located. So this is a patient with an inferior M2

41:25

occlusion. So all of these patients you can identify,

41:28

you know where their vessel occlusion is likely to be

41:31

and that's really helpful without even doing a CTA that

41:34

we can potentially discern the presence of that blockage. We

41:37

looked at this in a slightly different

41:40

form here, but we're able to identify with just perfusion

41:43

alone that we were extremely sensitive and specific for

41:46

making an accurate thrombectomy triage. You see

41:49

our sensitivity was 97% and our specificity was

41:52

almost 99% And so that was I think

41:55

good evidence at least in a retrospective manner that should

41:58

be validated prospectively.

42:01

Okay, so does this work in the real world? Well, we actually

42:04

have a lot of our partner hospitals that do perfusion and we're

42:07

able to see their Imaging via some of these remote

42:10

apps on our phone. And so but let's look at

42:13

this. How do we think about that? Well, what if you had this image here

42:16

and you're looking you're saying okay, they're telling you it's patient with

42:19

stroke symptoms. Well, it looks like there's a small core. It looks

42:22

like there's a salvageable penumbra so that certainly sounds

42:25

like thrombectomy maybe indicated but can we discern

42:28

that? There's an lvo. Well this perfusion pattern

42:31

here in green and the penumbra that really looks like

42:34

an M1 pattern and sure enough when a CTA is

42:37

obtained. This patient has a left and one occlusion so that

42:40

certainly was correct. How about this patient? You're

42:43

being described to patient with a modest Stroke

42:46

Scale around sex to eight and you

42:49

know, it sounds like they could have a large vessel occlusion. Here's their

42:52

perfusion imaging so there's no ischemic core that's being

42:55

identified and there's no penumbra.

42:58

So we would based on this say well, we don't

43:01

think they have a large vessel occlusion and sure enough. There's CTA was

43:04

normal. So you know that you don't need to activate your

43:07

cath lab team for this patient because there's not a large vessel

43:10

occlusion.

43:11

How about this patient? Well, number one. We know that we're

43:14

dealing with a patient with a quite large core and again

43:17

right now. I don't know that we have evidence that this patient would likely

43:20

benefit. I think that that'll be an area that we may

43:23

have evidence for soon. Even though there is a mismatch and

43:26

that's something we have to look into so, you know, they have

43:29

a large Baseline core. They may have some salvageable penumbra

43:32

and if we look at this perfusion pattern you see how medial and

43:35

hide this goes. That's something you often see in

43:38

a patient with an internal crowded artery occlusion and

43:41

sure enough. This patient has a right ICA occlusion with relatively

43:44

poor collaterals, which probably accounts for

43:47

the size of their infart. Here's another

43:50

patient you're being called about transferring and you

43:53

know, you can look at their Imaging and you say okay. I see that there's a core here.

43:56

I see that there's a penumbra but you know

43:59

what? I'm looking at these areas and they look like they're kind of

44:02

the same and even volumetrically, you

44:05

know, the core is being called as larger than the penumbra.

44:08

So even though you can look at this pattern and

44:11

It looks like there's you know, probably excuse me.

44:14

Probably a large vessel occlusion. This is someone with a match

44:17

deficit. That's probably not likely to benefit from a

44:20

thrombectomy procedure.

44:22

Okay, so can we use Imaging to predict the

44:25

future? This would be really great. Right? And how do we start to

44:28

think about that? Well, what if a patient you do the Imaging

44:31

at referring hospital and they're a candidate but are they still likely to be

44:34

a candidate after you transfer them?

44:36

Well collateral blood flow is really what determines

44:39

whether patient's ischemic core stays the

44:42

same size or whether it grows. If you have good collaterals, you

44:45

would expect that their core will stay similarly sized at

44:48

the outside hospital to what it would be after transfer to

44:51

a place like Stanford. Whereas patients with poor collaterals are

44:54

likely to experience core growth and may no longer be

44:57

a treatment candidate.

44:58

So we often think about collaterals based on CTA and

45:01

these are two different patients and you can see here that there are

45:04

very few collaterals. In fact, perhaps no

45:07

collaterals being identified in this patient. These are poor

45:10

collaterals, whereas this patient here has very

45:13

good collaterals past this M1 occlusion. And so

45:16

what you do is you say this is someone sorry.

45:19

This is someone that's likely to keep their core small

45:22

and still be a treatment candidate. Whereas this patient you'd be worried that

45:25

they're experiencing a very rapid growth in their ischemic core.

45:29

So if we have CTP, how do we judge whether

45:32

they have good collaterals? We don't

45:35

usually think about profusion and collaterals. So is there a way

45:38

to do that? Well one way to do that is to use something called

45:41

the hir or hypoperfusion intensity ratio. So

45:44

if you divide the volume of brain tissue with

45:47

a T-Max greater than 10 seconds divided by the volume of tissue

45:50

at the TMax greater than six seconds that gives you

45:53

hir what you want to have is less severe blood

45:56

flow delay. So you want to have a bigger denominator and so

45:59

patients with a lower HR are people

46:02

with presumably good blood flow delivery to the brain and

46:05

that implies the presence of good collaterals. And so

46:08

sure enough around an hir of 0.4 is

46:11

0.5 is a good threshold less than that.

46:14

Usually indicates favorable collaterals greater than that number

46:17

indicates relatively poor collaterals. So visually,

46:20

what does that look like? Well, you want to see less severe blood

46:23

flow. So a lot of red on a TMax map is bad

46:26

less red is good.

46:29

And so we've published several papers to really look at

46:32

this and it's now undergoing prospective validation. So

46:35

how do we leverage this in terms of predicting core

46:38

growth? Well, look at this patient 56 year old

46:41

male left hemiplegia right gaze preference Stroke Scale is 11.

46:44

We're not seeing a ischemic core being

46:47

measured by the automated software, but we're seeing a large right

46:50

MCA territory a number.

46:53

All right. This patient is located a hundred and

46:56

seventy five miles away from Stanford and Fresno. That's

46:59

a pretty long flight, but it's doable. So

47:02

let's delve into their penumbra a

47:05

little bit more and think about this hir. So here is sort of a you

47:08

know, looking in more detail at the severity of blood flow

47:11

delay and there is some red but there's a lot more tissue that's

47:14

not red. And this patient has an hir of 0.3. And

47:17

again, we're taking that team x 10

47:20

volume divided by the team X6 lime. So this divided by

47:23

that to get this 0.3 ratio. And so

47:26

if you look at this, can you predict core growth?

47:29

Well, we studied this. This is a paper on the Adrian Road

47:32

and annals of Neurology a few years ago. And this is

47:35

what we found in fact growth seem to be related

47:38

to what their hir is at the

47:41

referring hospital and right around 0.5. There really

47:44

is an inflection or patients after that experience a

47:47

pretty rapid growth in their scheme and core that often rendered them

47:50

not treatment candidates by the time they got to Standford.

47:53

All right. So here's the helicopter. It's about to

47:56

land on the roof. It's you know more than five hours after we

47:59

had that outside Hospital. Imaging we go straight down to the MRI

48:02

suite. This is what the outside Hospital CTP looks

48:05

like and here's our DWI at Stanford. So the

48:08

core is now 28 and if you look at this, it looks

48:11

like it's pretty well confined to this red area where the

48:14

most severe hypoperfusion was located but still

48:17

fortunately there is a large mismatch in

48:20

this patient. So this is someone that's still thrown back

48:23

to me Canada. They were taken to the cath lab and this internal

48:26

crowded artery occlusion was opened and

48:29

the patient had restoration of blood flow to the right cerebral hemisphere.

48:33

So we think that this

48:36

could potentially be very useful. So HR

48:39

is really measuring tissue level collateral blood flow and it's

48:42

it seems that it's going to be a very good predictive core

48:45

growth and this is really going on undergoing perspective validation

48:48

at stand for right now with Martin landsberg's Chris

48:51

to study.

48:53

Okay, so I want to now pivot toward the end and

48:56

just review some general stroke Imaging algorithms that

48:59

you might consider using and this is just meant to be a general

49:02

overview and it's really important that whatever Imaging

49:05

algorithm you set up at your Center. It is

49:08

something that you can do within your resources. That's

49:11

really really important. So again, we're looking for patients

49:14

with a large vessel occlusion a small core and a

49:17

large pin number. All right, but we can use a whole host of strategies to

49:20

identify these patients. So let's start with what I

49:23

would call the minimalist approach and that is a head

49:26

CT and a CTA and this was the approach that

49:29

was used by the Mr. Clean trial which was the first randomized trial

49:32

showing a benefit of thrombectomy in 2015.

49:35

And so it would do you would do something like this.

49:38

There's an 86 year old with left side of paralysis a very

49:41

high Stroke Scale. We have a non-con CT

49:44

and we have a CTA. So our CTA is

49:47

showing us an internal crowded artery inclusion on the right. So there is a large

49:50

vessel occlusion and on the left we're saying well, you know.

49:53

We don't see a huge amount of ischemic injury, like maybe there's

49:56

some subtle like loss of density here because

49:59

the cell loss of density here and the basal ganglia on

50:02

the right, but it doesn't look that bad. Maybe the internal capsule posterior lens

50:05

a little bit more hypoense, but certainly there's no large

50:08

area of obvious injury that seems to be favorable and

50:11

their symptoms are clearly much

50:14

more impaired than any hypodensity that we're seeing here. So

50:17

that's a good treatment candidate. Here's their AP

50:20

and lateral angiogram showing you internal crowded

50:23

artery inclusion. Just past the abhalmic artery and this

50:26

patient gets successfully revascularized and that's

50:29

a very valid treatment strategy and it works.

50:33

What if you want to be a little bit more discriminatory and

50:36

you want to include a little bit more nuanced information

50:39

about how much ischemic injury there

50:42

is on a head CT. Well, then you could talk about using aspects

50:45

as well as a CTA. And this was the approach

50:48

that was used by the ravasket trial which was also published in

50:51

the New England Journal of Medicine. And so aspects is

50:54

an attempt to quantify the

50:57

degree of ischemic injury on non-con CT. And so

51:00

this is from the folks that Alberta and what you do is you look at

51:03

these 10 stereotyped regions in the MCA territory

51:06

and you identify whether there is

51:09

significant hypodensity in each region. And if you

51:12

do have hyponency in a region, you remove a point and

51:15

there are 10 regions. So the best score you're going to get is 10,

51:18

but if you haven't aspects of six to 10

51:21

that is favorable for thrombectomy treatment, and so

51:24

here's a patient who's got a right M1 occlusion and

51:27

you want to window this to help you really discern the

51:30

presence of hypodensy and when you do that we

51:33

See that we've got hypodensity in the insula the lentiform nucleus

51:36

the caudate head and up into the M4 region.

51:39

Well, that's four points off. This aspects is

51:42

6 that is still favorable for treatment. Okay. Here's

51:45

another patient and 86 year old with left-sided paralysis.

51:48

Again, the high Stroke Scale here is the non-con CT

51:51

and when we look at this we say well

51:54

the M2 region looks hypotens the insula, the lentiform

51:57

nucleus looks typodens, and we've got an M1 occlusion. That's

52:00

someone favorable for treatment in the patient

52:03

did get a successful from Back To Me.

52:06

Okay, what if you want to incorporate some information about

52:09

collaterals and you could do that on CTA

52:12

or you could do a little more advanced collateral assessment

52:15

using multi-phase CTA and that's the

52:18

strategy that was used by the Escape trial another one

52:21

of the randomized trials that was published. And so how

52:24

does this work? Well you first are going to look at the head CT

52:27

and you're going to say okay. Well how much hypodensity is that is

52:30

there and there's a two patients where we see some hypodensity here

52:33

in the lentiform nucleus, maybe into the coding on

52:36

each patient, but overall looks like a high aspect score.

52:40

So what about the presence of an lvo? Both of these patients have

52:43

M1 occlusions this patient on the left this patient

52:46

on the right, but what about collaterals? Well, let's assess

52:49

the collateral. So this patient has very poor collaterals.

52:52

There's not a lot of contrast in the arteries past

52:55

this blockage and this patient over here has really good

52:58

collaterals, like very good filling past this blockage. And

53:01

so which one would you offer treatment to the one

53:04

with good collaterals? Okay, and this

53:07

patient goes on and gets treated successfully and that

53:10

also is a valid treatment strategy that was

53:13

worked very well in the Escape trial.

53:16

So we're getting into the more complex or Advanced sort of approach

53:19

and that's the use of perfusion imaging as well as

53:22

CT and geography and perfusion was actually used in

53:25

the bulk of the randomized trials that have been published. And so this was

53:28

in the early time Windows the Swift Prime and extend IA as

53:31

well as both late time Windows Dawn and diffuse 3.

53:34

And so that that will work something like this. This

53:37

is a patient of mine. She's an 81 year

53:40

old female with Aphasia and right-sided paralysis. She held her

53:43

Coumadin for a dental procedure and woke up with severe

53:46

stroke symptoms went to an

53:49

outside hospital was found to have these symptoms transferred to Stanford. And

53:52

here's MRI on admission to our hospital and we

53:55

can see a relatively small core on the diffusion weighted

53:58

image clearly a large penumbra here

54:01

and a pattern that looks like an ICA occlusion and sure

54:04

enough on the MRA. We see in ICA occlusion, but

54:07

clearly a mismatch here between these two the

54:10

core and the number which we quantify the course

54:13

seven the numbers 128 milliliters.

54:16

That's someone who's a good treatment candidate. So here we are treating the

54:19

patient and we see that I see

54:22

occlusion again just passed the ophthalmic artery and just passed

54:25

the anterior corital artery.

54:27

This is the clot that we removed from the patient and after

54:30

doing so we were able to restore flow

54:33

into that left anterior circulation territory. So,

54:36

you know a very good treatment a good restoration of

54:39

flow and here is the patient in the

54:42

cafeteria on the day of discharge, which is three days later and this

54:45

is my fellow former fellow Eric assessment who did the procedure

54:48

with me and I just want to point out that here's the patient.

54:51

So she looks fabulous and arguably and

54:54

better shape than her husband, and she really did remarkably. Well.

54:58

So I just want to end with a few words on

55:01

automated Imaging because this is something that is really going

55:04

on now and everyone should be aware of so AI is

55:07

here. These are some of the automated software platforms that you might have

55:10

seen you are able to get all kinds of information on

55:13

your phone, which is how the neuron eventualists and

55:16

strict neurology folks are really following these patients and triaging them.

55:19

So the Radiologists need to be aware of that you can window

55:22

things you can scroll things on your phone now, which is pretty

55:25

incredible. You can see on email whether a patient

55:28

has a mismatch and an ldo and you can

55:31

see when a patient looks like they should not be transferred even though they have

55:34

an lvo they may have a matched deficit or

55:37

a very large core. But automation is here. It's

55:40

helping already and that makes us a very exciting

55:43

time and stroke care. So just to wrap up this is a really fun

55:46

time to be taking care of stroke patients and it's a

55:49

really fun time to be Imaging them thrown back

55:52

to me is a very highly effective therapy for ischemic stroke due

55:55

to a large vessel occlusion and our stroke systems of care are really already.

55:58

Drink to this artificial intelligence is

56:01

here. That's really kind of showing up first in neuroradiology and

56:04

stroke care and I think is going to continue to expand in use

56:07

and here's just a view of the Stanford campus

56:10

from the helicopter when I when I joined them for a quick flight. It's

56:13

really beautiful here. So come visit the Bay Area sometime and

56:16

with that I will go ahead and stop my sharing

56:19

and I'm happy to answer any questions

56:22

that you all might have and

56:26

I'm going to navigate here the Q&A box. So

56:29

there's one

56:32

question here about how do we Define a small core?

56:35

And that is a bit of a moving

56:38

Target. So right now in the early window trials 50 MLS

56:42

to 70 MLS was used in the late window trials.

56:45

There was a sliding scale of core size that

56:48

went up to 70 MLS if you want to be as inclusive as

56:51

possible. What we do at Stanford is we Define the

56:54

core of 70 or less as favorable in greater

56:57

than 70. We we don't know and don't always offer treatment.

57:00

I suspect next year. We're gonna have data about

57:03

larger courses already been one trial published

57:06

out of Japan suggesting that large course can

57:09

help similarly acapca Strokes.

57:12

We just don't have a lot of data right now many people are starting to

57:15

consider treating them, but we don't have data from

57:18

randomized trials yet.

57:20

Another question here patients who

57:23

have recurrent stroke. Can you still treat them

57:26

with thrombectomy procedures so you can I

57:29

think you have to understand what's going on with the patient

57:32

and each situation can be variable. And so

57:35

for example, we had a patient a few years ago who are

57:38

one of my former fellows his first

57:41

three thrombectomy procedures. He did with me. We're

57:44

in the same patient within a 36-hour period She

57:47

had a new diagnosis of cancer and is

57:50

very hypercoagable and she kept throwing M1 clots on

57:53

the left side. And so we did a repeat, you know thrown back

57:56

to me three times in a row because her Imaging triage

57:59

every time still looked favorable and so

58:02

you can certainly still help those patients. You just have

58:05

to make sure you're taking to into account the clinical situation

58:09

of what you're doing.

58:12

So a question about hyperfine for

58:15

wake up stroke. So hyperfinds pretty interesting. We don't have

58:18

it at Stanford. I think that it would be very interesting to

58:21

have these sort of In-House Mrs. To to make

58:24

kind of triage decisions. I've just not familiar with how well

58:27

they perform but certainly you can get a DWI and

58:30

gain some information. I don't know how well other you

58:33

know sequences work on that scanner, but that would

58:36

be a very interesting way to potentially evaluate a

58:39

patient in your neuro ICU. For example, who may need a

58:42

throwback to me treatment.

58:45

So a question about chronic vessel occlusion on

58:48

CTN MRA being an issue for thrombectomy. This

58:51

can sometimes be difficult. Right a patient comes in

58:54

with a stroke and there's a large vessel occlusion and you don't

58:57

always know whether that is an acute occlusion or

59:00

a chronic occlusion. If they are

59:03

newly presenting with symptoms, I would

59:06

argue it's probably a good idea to operate under the

59:09

assumption that it is an acute occlusion. But again, you

59:12

can be wrong so you can sometimes leverage trip

59:15

tricks on the profusion to help you and so,

59:18

you know, if you're looking at Rapids MTT,

59:21

for example, which is less sensitive to

59:24

delays symmetry on the MTT map can often

59:27

indicate a more chronic inclusion. We have a paper under review about that

59:30

right now. Whereas if there's asymmetry in

59:33

the MTT that may indicate that it's an acute occlusion. But

59:36

sometimes you just need to say, you know, we don't know and take

59:39

the patient to the neuronovascular suite

59:42

and and see if you can discern based on the pattern.

59:44

Or whether you can cross it

59:47

successfully and find out that maybe you were tricked on the non-invasive Imaging.

59:52

A very timely question on posterior circulation stroke

59:55

care. So we finally have a

59:58

randomized data that doing a Basler artery

60:01

thrombectomy is beneficial. So the motion

60:04

attention trials were published within the last month and

60:07

they were both randomized trials in China and showed

60:10

a benefit for thrombectomy for basala artery occlusion,

60:13

as was asked a little bit early. We don't have data on PCA

60:16

occlusions that could be certainly a reasonable

60:19

thing. But it's it's an area where we don't have a lot

60:22

of evidence right. Now. The Imaging evaluation of

60:25

posterior circulation Strokes needs a little

60:28

bit more work. And that's one of my grants's asking that that question

60:31

but both of those positive trials use something

60:34

called the posterior circulation aspects and you had to

60:37

have a PC aspects of six or greater as well as a

60:40

CTA showing that you had a Basler artery occlusion or

60:43

occlusion of both intercranial vertebral arteries.

60:46

And so that's sort of the inclusion criteria that I

60:49

think is reasonable to be operating under right now.

60:52

But the stroke guidelines have not been updated to address that

60:55

at this time.

60:57

So I will there's a couple more and then I'll let you all go

61:00

for time. Our neuron inventional team. We currently have four

61:03

neuron eventualists. We have a three fabulous

61:06

nurse coordinators who help

61:09

operate our outpatient service, but we have a super busy inpatient

61:12

stroke team that includes, you know,

61:15

many many other folks in terms of a very robust drug neurology

61:18

team a stroke neurology fellowship program coordinators. It

61:21

takes a village to really do this this well,

61:24

and that's a really great question. And then

61:27

why don't I

61:30

For that you've got a good question. Nice to hear from you. There's

61:33

a lot of different terms for salvageable tissue that

61:36

can cause confusion so penumbra and

61:39

mismatched perfusion deficit. Yeah. So the I

61:42

think the terminology does get a little bit sloppy. So in general what

61:45

I would say is tissue at risk and penumbra are synonymous those

61:48

mean the same thing mismatch refers

61:51

to a core that is smaller than

61:54

a penumbra. So those two volumes

61:57

are mismatched if the core and the

62:00

penumbra are the same size, they are matched. And again,

62:03

that's a situation where I don't think it makes a lot

62:06

of sense to open up a large vessel occlusion. If the

62:09

core is already the same size as the salvageable tissue

62:12

perfusion deficit can refer to

62:15

any form of the penumbra certainly is

62:18

a perfusion deficit being could also have more mild profusion delays

62:21

that may reflect us to stenosis and the neck for

62:24

example

62:25

And then why don't

62:28

we end with one more question from Steven Wong here. Do I

62:31

put Stenson for intracranial arteries stenosis? So this this

62:34

is another very active and kind of tricky area a

62:37

lot of the best data are coming out of Korea and

62:40

Japan for how to handle patients with intracranial atherosclerosis. The

62:43

medical therapy is really critical. So initiation

62:46

of anti-platelet therapy is really important

62:49

in keeping these arteries open. I try to avoid acute

62:52

stunting if I can because I'd prefer to

62:55

not put metal in if I can get away with

62:58

not doing it. So I usually will try angioplasty if the

63:01

vessel is is continuing to go down after

63:04

a throne back to me. But if it doesn't respond

63:07

to angioplasty or if there's you know a dissection that

63:10

results placement of a stent is a reasonable

63:13

consideration that we have a little bit less data to

63:16

help us to guide that decision.

63:19

So last question

63:22

can MRI CT predict a stroke that might happen in

63:25

days.

63:25

Months not yet. You can that's a great

63:28

question. We can't quite predict the future that accurately but look there's

63:31

a lot of smart people on this call on this

63:34

zoom and I'm sure someone will come up with a good idea that we can

63:37

test in the future.

63:38

So, thanks again everybody. This was a lot

63:41

of fun and thanks again

63:44

to the folks at MRI online for for supporting it.

63:48

Dr. Hai, thank you so much for your lecture today and thanks to

63:51

all for your participation in our new conference a reminder

63:54

that you can access the recording of today's conference and

63:57

all our other previous named conferences by creating a free MRI online

64:00

account.

64:03

MRI online has launched a new stroke Imaging Focus membership plan

64:06

to help you gain access to help you gain confidence with stroke and

64:09

neurovascular. Imaging. Learn more at MRI

64:12

andline.com slash neurostroke. Be sure

64:15

to join us next week on Wednesday, November 9th at 12pm Eastern

64:18

time for a lecture with Dr. Donald Resnick on

64:21

the articular disorders of the peripheral. Skeleton

64:24

emphasis on morphology and Target

64:27

sites. You can register for that lecture MRI online.com and

64:30

follow us on social media for updates and reminders on upcoming

64:33

new conferences. Thanks again, and have a wonderful day.