0:02

Hello and welcome to Noon Conference hosted by MRI Online.

0:03

3 00:00:06,600 --> 00:00:08,400 In response to the changes happening around

0:08

the world right now and the shutting down of

0:10

in-person events, we have decided to provide free

0:13

Noon Conferences to all radiologists worldwide.

0:16

Today we are joined by Dr. Javad Azadi.

0:19

Dr. Azadi is an academic radiologist

0:21

at the Johns Hopkins Hospital.

0:24

He is interested in hepatic

0:26

steatosis and the metabolic syndrome.

0:29

He is the Assistant Course Director for the

0:31

Diagnostic Radiology medical student elective,

0:34

and is a member of the RSNA COVID-19 Task Force.

0:39

A reminder that there will be a Q and A session

0:41

at the end of the lecture, so please use the

0:43

Q and A feature to ask your questions and we will

0:45

get to as many as we can before our time is up.

0:48

Apologies for butchering those medical terms.

0:50

But with that being said, thank you

0:51

all for joining us today. Dr. Azadi,

0:53

I will let you take it from here.

0:56

All right, thank you.

0:57

And yeah, no, you didn't butcher it

0:58

too bad.

0:58

It's hepatic steatosis.

1:00

Um, and so again, uh, welcome everyone.

1:03

My name is Javad Azadi.

1:04

Um, and since my bio, I've also been promoted to

1:07

the Chief of Ultrasound here at Johns Hopkins.

1:09

Um, still an Assistant Professor in the

1:11

Department of Radiology, but, and I'm

1:13

still interested in hepatic steatosis.

1:14

And this, uh, Noon Conference will be about

1:19

MRI Elastography to assess hepatic fibrosis.

1:22

Um.

1:23

And, you know, this is important to, uh,

1:26

understand and if it's something that you're

1:28

interested, uh, in doing at your own institution

1:31

to learn more about, I have no disclosures.

1:35

Our learning objectives for today's talk, we're

1:37

gonna review the indications for MRI Elastography.

1:40

We'll review the techniques that we use

1:41

to measure liver fibrosis using MRI.

1:44

We'll review the techniques used to measure liver

1:46

fat content using MRI, and then we'll review some

1:48

techniques to measure liver iron content using MRI.

1:51

And the reason why we're gonna cover all

1:53

three of these topics is for our protocol

1:55

and many Elastography protocols out there.

1:58

All three are done, uh, as part of the

2:00

protocol, and that gives it a leg up

2:03

compared to ultrasound elastography.

2:05

I'll look at that.

2:05

It has its own timing, so lemme try to pause that.

2:08

Uh, okay.

2:11

Hopefully this doesn't do it itself.

2:14

Um.

2:15

So remember, liver fibrosis is gonna exist

2:18

on a spectrum of disease and it's going

2:20

to be their minimal hepatic injury and

2:23

it's gonna extend to end-stage cirrhosis.

2:26

When, uh, I was more junior, you know, I called

2:28

everything cirrhosis, and that was a misnomer

2:30

'cause it definitely rings, uh, alarm bells for the.

2:33

Hepatologists, even the primary care

2:35

physicians taking care of these patients.

2:37

'Cause cirrhosis is, uh, a much bigger deal

2:40

than hepatic inflammation or, um, the grades

2:43

of fibrosis that we see in the United States.

2:46

Common etiologies of liver fibrosis include alcohol

2:49

abuse, non-alcoholic fatty liver disease, chronic

2:52

viral infections, so with non-alcoholic fatty liver

2:55

disease that's seen in the metabolic syndrome.

2:57

And patients who consume.

2:59

Excess fructose, excess glucose,

3:02

they build up fat in their liver.

3:03

Then that ends up, uh, leading to inflammation

3:06

and that inflammation leads to fibrosis.

3:11

Some accepted clinical indications for assessing

3:14

hepatic fibrosis include evaluating the grade of

3:16

fibrosis prior to initiation of antiviral therapy,

3:19

such as for patients that have chronic hepatitis.

3:22

It's, it's really wonderful that we had these

3:24

cures for chronic hepatitis C, um, which

3:27

we really didn't exist 10, 15 years ago.

3:29

Um, and now there are multiple, uh,

3:31

regimens available for these patients.

3:34

It's also gonna be helpful to evaluate for.

3:35

Patients who present with symptoms of

3:37

portal hypertension, but there's no

3:38

unknown or there's no known etiology.

3:40

So you know, if there's kind of subclinical

3:42

fibrosis, this will show you it.

3:45

It can also be helpful to monitor for hepatic

3:48

fibrosis in response to treatment, whether

3:50

you're on a medication or your patient's

3:52

on a medication that can cause fibrosis.

3:54

Or if you're, uh, treating and you wanna see

3:56

does the fibrosis halt or reverse, you know.

3:59

Elastography would be really

4:01

helpful for those patients.

4:02

I will say anecdotally, um, the indications are a

4:06

little bit more broad, um, but it's, it really is,

4:11

um, you know, those are the accepted indications.

4:16

So for invasive grading, um, we

4:20

do it here with an 18-gauge core

4:22

uh.

4:23

Biopsy needle.

4:24

Uh, we select the kind of the safest window.

4:26

We take a two-centimeter path.

4:28

We send that, uh, sample to our pathologist,

4:30

our pathologist, and looks at the sample in our

4:31

microscope, and they're able to grade the fibrosis

4:34

based on the elements of fibrosis present in the core.

4:37

Um, the issue with this is fibrosis isn't a.

4:44

Uniform process, it can happen, um, geographically.

4:48

And so you risk under sampling.

4:50

There's inter-observer variability

4:52

with, uh, the pathologist themselves.

4:54

One, one pathologist may say this is

4:56

mild, another may say it's moderate.

4:57

You know, that radiology is true

4:59

as well when we create something.

5:00

Um, and then for us, increasingly our core biopsies,

5:05

they'll fragment and then they become non-diagnostic.

5:07

So, um, if you were to.

5:10

Can I offer a patient saying, Hey, you

5:12

know, I want to grade your hepatic fibrosis.

5:15

Uh, I can offer you a biopsy

5:17

or a non-invasive procedure.

5:18

I would sure most would opt

5:20

for the non-invasive procedure.

5:22

And with any invasive procedure, the

5:23

complications would be bleeding and pain.

5:28

So the two, uh, in-use methods are.

5:32

MRI Elastography, which we'll talk about today.

5:34

And then ultrasound elastography,

5:35

which we've talked about before.

5:37

So the pros of MRI Elastography: we're able to

5:39

evaluate for fibrosis, steatosis, and iron overload.

5:43

And it has a much higher agreement

5:44

with biopsy fibrosis grades.

5:46

Um, so it really does alleviate the

5:49

need for biopsy in most patients.

5:52

The downsides are that it's very expensive, right?

5:54

You need to have an MRI machine,

5:55

you have to have MRI staffing.

5:57

Um, but in addition to all that, you also

5:59

need special MRI Elastography equipment.

6:01

So it's not something that you

6:02

can just do as an extra sequence.

6:05

There actually is extra hardware required,

6:07

which we'll cover a little bit. For ultrasound,

6:10

the pros would be that you can evaluate the

6:12

fibrosis, and to a lesser degree, the steatosis.

6:14

It's not as.

6:15

Reliable, in my opinion, as MRI.

6:17

Um, but what you do get with every ultrasound, as

6:20

long as your protocol spells it out, is that you

6:23

can look at the portal vein velocity, direction,

6:25

and that's also a marker for portal hypertension.

6:27

Ultrasound's gonna be a lot cheaper than MRI,

6:29

and it's gonna be more readily available.

6:31

Um, in this case it.

6:32

It generally is a package update for a specific probe.

6:35

The downsides of ultrasound elastography is that it

6:38

can overestimate fibrosis, especially in patients

6:40

who have elevated liver function tests, who aren't

6:42

fasting, or who have, um, hepatic congestion.

6:45

So they have, you know, right heart dysfunction.

6:48

The elastography may, uh, the ultrasound version may

6:52

say they have more fibrosis than is actually present.

6:55

Another downside of ultrasound is that it's more

6:57

user dependent than MRI. You know, MRI is, um,

7:00

the patient's there, the machine's working, but

7:03

with ultrasound you do have a technologist or

7:05

radiologist or potentially, you know, personnel

7:07

in the community doing the scan themselves.

7:14

So Elastography is a technique

7:18

where we'll distort a target tissue.

7:21

Uh, in this case we're gonna

7:22

apply external vibrations.

7:23

They're gonna start at the surface of the

7:24

skin, and those vibrations are gonna travel

7:26

through the patient's body to the target tissue.

7:28

And then we're gonna observe

7:29

changes to the distorted tissue.

7:32

We're able then to calculate the liver stiffness

7:35

from these external vibrations using a shear modulus.

7:38

And so this is actually a different conversion

7:40

technique than what we use with ultrasound.

7:42

And so the, uh, pressures and velocities

7:45

that we get with ultrasound are comparable

7:48

to MRI, but they're not one-to-one.

7:51

And that's important to just kinda

7:52

keep in the back of your head.

7:53

So while both can tell you if a patient

7:55

has grade 1, 2, 3, 4 cirrhosis, um.

7:59

They're not gonna be the same values

8:01

between MRI and ultrasound.

8:04

So what's some of the evidence for MRI Elastography?

8:07

Uh, Yin et al. published a series in 2016 looking at

8:11

1,377, uh, patients, and they were able to show that

8:15

there was, uh, you know, good reproducibility, uh.

8:19

A lot of the patients had biopsy so that you were able

8:21

to show concordance and they were able to, you know,

8:25

demonstrate that liver stiffness was significantly

8:27

higher in patients who had advanced fibrosis.

8:28

So that's important.

8:30

Um, they're actually showing that

8:31

their findings have clinical meaning.

8:35

Uh, again, Yin et al. was able to show that

8:39

MRI Elastography can discriminate between patients

8:41

with moderate and severe fibrosis, so grades two to

8:43

four, and those with mild fibrosis, so one or less.

8:47

And it's fairly sensitive, fairly specific.

8:49

And then, uh, these numbers are a little bit

8:51

higher than what we would see with ultrasound.

8:53

Um, so if cost was no issue,

8:55

patients would get elastography.

9:00

The way that we're able to transmit our

9:02

external vibrations is using a passive driver.

9:05

This is actually a mechanical device.

9:07

It's gonna be attached to the patient.

9:09

It's a, it's a paddle, you know, and

9:11

it, it, it acts like a drum almost.

9:14

Um, and it's gonna be placed kind of

9:16

at the xiphoid midclavicular line.

9:19

So about a third of it's gonna be over the ribs.

9:20

Two thirds of it's gonna be over

9:21

the liver and it's gonna, um.

9:24

Passively transmit external vibrations

9:26

at about 60 hertz, uh, continuously.

9:29

And those vibrations are gonna

9:32

be transmitted to the liver.

9:33

And then we're able to measure those

9:34

vibrations delivered to quantify stiffness.

9:36

This is a schematic of what that looks like.

9:39

And so you have a patient, they have the pad on.

9:43

It's an ideal location.

9:45

Um, you know, it is important

9:46

that it is over the liver.

9:47

You don't want to be transmitting

9:49

vibrations to the bowel or to the lung.

9:51

Um, and the active driver itself

9:54

will be outside of the room.

9:55

It'll be running.

9:57

And so really you, uh, if you're ever troubleshooting,

9:59

I'm sorry, I'm gonna have to pause that

10:01

lecture for a second 'cause it keeps advancing.

10:03

So I was just doing a little

10:05

troubleshooting on this, uh, side.

10:06

Dr. Ti, I think if you go to your, um, you know, uh,

10:10

slideshow manager, you can clear the timing off of it.

10:14

Yeah, no, I think that's what it is.

10:16

I think there's timings, um, pop out of this.

10:22

My apologies, everyone.

10:33

There we go.

10:33

I think that should have fixed it.

10:45

Okay, so hopefully now everyone can see my screen.

10:49

Um, hopefully now it will kind of advance on its own.

10:53

And so when you're troubleshooting, you know, things

10:55

that can go wrong would be with the device itself,

10:57

the tubing, the paddle, the patient, and those

11:00

are important considerations when you're actually,

11:02

uh, doing these studies in your own institution.

11:05

Our protocol, um, is, you know, it's been

11:09

standardized and we do it at, uh, multiple

11:11

sites in our institution, but, you know, always

11:13

gotta start with your scout, uh, sequences

11:15

to make sure you're imaging the right place.

11:17

Um, you wanna make sure there's not any ferromagnetic

11:19

materials that could cause susceptibility.

11:23

You wanna make sure there isn't, um, you know,

11:25

significant amount of ascites, things that

11:26

wouldn't be readily visible to the naked eye,

11:29

but would affect the quality of the study or

11:30

potentially make it unsafe for the patient.

11:34

Sorry, that was not the auto advance.

11:36

That was me.

11:36

Um, our elastography sequences are then

11:39

performed after that. The paddle could be removed.

11:42

We then perform fat quantification, um, sequences.

11:46

So we do both in and out of phase

11:47

sequences as well as Dixon sequences.

11:50

We follow that up with a T2 single

11:52

shot turbo spin echo with fat sat.

11:54

Uh, and that's a very important sequence,

11:56

uh, to review when you're reviewing

11:57

elastography because that's where you're

11:59

gonna see your incidental findings.

12:01

The rest of the, the data you're obtaining is

12:04

more, uh, quantitative, but this is actually

12:07

anatomic and qualitative, and so you can find

12:09

potential abnormalities you otherwise would miss.

12:13

We also obtain, uh, T2* relaxometry

12:15

and R2* relaxometry,

12:17

and that's used for iron quantification.

12:19

One question...

12:20

One question that we received in the past is, well,

12:23

can you combine this with a contrast-enhanced study?

12:26

And technically you can.

12:27

The reason why we don't is because

12:29

of the way these studies are billed.

12:31

This is billed as an MRI abdomen study,

12:34

just as a contrast-enhanced study would

12:36

be, so effectively you're giving the

12:37

patient two studies for the cost of one.

12:39

So we would, uh, ideally do an elastography

12:42

at a different visit than a contrast-enhanced

12:45

MRI.

12:48

So our elastography sequences, to just kind of

12:51

dig down a little deeper, um, you can use multiple

12:55

different sequences to track the shear waves.

12:57

Uh, we use gradient echoes, um, but really

13:00

it's what your manufacturer specifies.

13:03

The data's initially gonna be

13:04

obtained, it's phase contrast,

13:08

and then we're gonna process that

13:09

data to generate wave images and

13:11

stiffness maps using the shear modulus.

13:14

And then over that we're gonna.

13:16

Generate confidence maps to show the reader where

13:20

the liver stiffness being reported is considered

13:24

valid because you're gonna get lots of variation.

13:28

You're potentially gonna be outside of the liver

13:30

and that's gonna be considered invalid data.

13:31

So you really only wanna be reporting a

13:33

liver stiffness of the liver, where the

13:35

confidence maps say that it's valid data.

13:38

And so this is just an example of that process.

13:41

So you start with, um, some mag images.

13:44

You start with some phase contrast.

13:47

Then they're, they're converted

13:48

into an actual, uh, wave series.

13:55

Uh, it's not playing.

14:02

There we go.

14:03

And so you can actually see the

14:04

waves tracking through the liver.

14:05

So the alternating bands of red and

14:07

blue are the waves propagating from the.

14:11

Paddle through the liver, and the distance

14:14

between the waves corresponds to the

14:15

stiffness. It's gonna play again, my apologies.

14:20

We then take this map and we can color code it into

14:25

liver stiffness, um, with the more purple colors of

14:30

that side of the rainbow being low stiffness,

14:33

and then red being high stiffness, fibrosis.

14:36

Um.

14:38

And then over that you can apply the confidence map.

14:40

So we use hashing.

14:43

So anywhere there aren't hashing, that's

14:45

considered an area of valid measurements,

14:49

but you as the radiologists should be

14:50

aware of where the liver actually is.

14:52

So I would recommend that you have, um, on one of

14:54

your screens, you have kinda that T2 sequence or

14:57

even any of your other sequences that are anatomic

15:01

to make sure where you're measuring stiffness.

15:03

Um.

15:04

Is within the liver as well as within the

15:06

boundaries of this hash mark, and then we'll

15:09

go over some of that in the cases as well.

15:13

For our fat quantification.

15:14

Uh, we do both in and out of phase and Dixon.

15:16

So if you guys remember, in and

15:18

out of phase is a GRE technique.

15:20

And when it's in phase, fat signal's

15:22

gonna be added at the water signal.

15:24

And remember, it's, they're, you know, they're

15:25

spinning, um, in like a timewise fashion.

15:29

So at one time point they're in phase.

15:31

Another time point they're out

15:32

of phase, another time they're in phase.

15:34

So in this case, you know, we do out of phase then

15:36

in phase, um, based on the strength of a magnet.

15:39

Um.

15:41

In phase the fat signal's additive and out of phase

15:44

the fat signal's subtractive from the water signal.

15:46

And then you can use those two signals

15:48

then to determine what a fat fraction is.

15:49

And that's an estimate of, uh, or

15:52

how much fat is in the liver cell.

15:54

So fat fraction can be expressed as the signal

15:57

of in phase minus the signal of, uh, out of

16:00

phase divided by twice the signal of in phase.

16:05

And this is just examples of in and out of phase.

16:08

Um.

16:08

You know, I always find it helpful to

16:11

look at the, uh, TR, um, but you know, just

16:13

visually it can always kind of quickly

16:15

determine which one's in phase and out of phase.

16:17

Whereas the out of phase, you're gonna have that

16:19

India-ink artifact, and that's due to the

16:22

interface of the fat and the water molecules.

16:24

Um, and you can use these two

16:25

images then to determine, you know,

16:28

percentage of fat within the liver.

16:30

And similarly, just qualitatively,

16:32

looking at these two images, you can

16:33

see there is a lot of fat in the liver.

16:35

Um.

16:36

But the benefit is, you can quantify that

16:40

the, uh, newer method, I say newer,

16:43

it's, it's, you know, 10 plus years old.

16:46

But the method that we use is Dixon.

16:48

So this is a chemical shift fat suppression technique.

16:50

And so we're gonna generate four sets of images.

16:53

So we'll still have the in and out of phase sequences,

16:56

but then we'll have water-only and fat-only images.

16:58

So...

17:00

And the water-only images, only,

17:02

uh, water signal's gonna be shown.

17:03

And then fat-only images, only

17:06

the fat signal's gonna be shown.

17:07

And then we can use a fat fraction

17:09

based on that Dixon fat set.

17:11

So you have the Dixon fat divided by the sum

17:13

of the Dixon water and Dixon fat, times 100%.

17:16

And that tells you what the actual fat fraction is.

17:20

And this is an example of that.

17:22

You can see this image on the left.

17:23

All the water signal's being suppressed,

17:24

everything's fat bright, and then on the right,

17:27

it's fat suppressed, all the fat's dark, all the

17:28

water signal's preserved, and you can draw ROIs.

17:32

I would recommend, um, if you're doing this

17:35

regularly, to actually just create a spreadsheet,

17:38

um, in Excel or Google Sheets, uh, whatever

17:42

your preferred spreadsheet software is.

17:44

Where you can just put these measurements in and then

17:47

they'll automatically do these calculations for you.

17:49

So it makes it really fast,

17:50

um, and it is also available for whenever

17:52

you're reading MRI elsewhere and you're

17:54

just trying to decide, hey, this looks

17:55

like it's fatty, but is it actually fatty?

17:59

For our iron quantification, uh, we use T2 star.

18:03

So remember that.

18:05

T2 times aren't, uh, what we actually

18:07

observe with MRI, it's actually T2 star,

18:10

and that's just due to local field inhomogeneity.

18:12

Uh, in this case, uh, we are concerned about

18:15

iron causing inhomogeneity, and so it's

18:17

gonna cause a reduction in T2 times.

18:20

So what we'll do is we'll actually

18:22

obtain multiple single-shot GRE, and we

18:24

will derive a T2 star relaxometry map.

18:27

And that relaxometry map is

18:30

then further, uh, kind of validated

18:33

against degrees of iron overload.

18:34

So we can measure T2 star times, uh, to estimate

18:38

what the degree of iron overload is for a patient.

18:41

And this is just an example of what would be

18:43

obtained once you've done the post-processing.

18:46

This is a color representation

18:47

of the time in milliseconds.

18:49

And then this is, um, a grayscale image,

18:51

which we would actually do measurements off of.

18:53

And just like you would for.

18:55

Um, the liver stiffness or for fat,

18:58

um, quantification, you really only

19:00

want to be measuring on the parenchyma.

19:02

So I wanna be measuring outside of the

19:03

parenchyma because that's gonna be less accurate.

19:06

R2 star is, it's interesting.

19:09

So R2 times are just the inverse of T2 times.

19:11

So R2 star is the inversion of T2 star.

19:14

Um, and just as iron causes decreased T2

19:17

star, it causes increases in R2 star times.

19:22

We're again gonna use multiple single-shot,

19:24

uh, echoes to obtain this, to derive an R2

19:26

relaxometry map, and just as it was for

19:28

T2 star, uh, these types could be measured

19:31

and correlated to degrees of iron overload.

19:35

And the units are different because

19:37

they're not measuring time necessarily.

19:38

It's an inversion time.

19:39

Um, but you know, a nice color map here showing.

19:44

Kind of where it falls.

19:45

And this is windowed automatically

19:47

by the machine itself.

19:48

So I unfortunately couldn't make it look prettier.

19:50

The image on the right is a representative

19:53

image of what R2 star looks like.

19:55

And just as with T2 star,

19:56

measure on the liver parenchyma.

19:59

Uh, the nice thing about, uh, T2 star and R2

20:02

star relaxometry, they're fast acquisitions, they can

20:06

be obtained, uh, with single or multiple breath holds.

20:10

They're usable both on 3 and 1.5 Tesla.

20:13

Uh, the downside is that there needs

20:17

to be post-processing that may require

20:18

additional fees for your MRI machine.

20:20

So these are things you would have to discuss

20:22

with your vendor if you wanted to implement it.

20:24

Um, how much would it be and is it a

20:27

continual license or is it a one-time upgrade?

20:30

So,

20:32

you've decided to purchase that

20:33

MRI Elastography, uh, application.

20:36

You've started to do it and your

20:39

technologist calls and says, Hey, I'm,

20:41

I'm not getting, um, the expected images.

20:44

Um, that's when troubleshooting comes into effect.

20:47

And there are a couple kind of easy fixes.

20:50

Uh, the first one would be you have poor

20:51

quality stiffness maps, meaning most of the

20:55

liver's not in that, um, high confidence area.

20:58

It, it's covered up by hash marks.

21:01

So in that case, check the driver, check the

21:03

attachment, check the settings, make sure that the

21:05

external vibrations are getting to the patient.

21:08

Recheck the scouts for any susceptibility

21:09

artifacts, something that's gonna, uh,

21:11

affect your elastography sequences.

21:14

And then maybe easiest of all, just make

21:16

sure the patient's not moving too.

21:17

So, um, most of you are aware patients, uh, have

21:21

to be coached to stay still during these exams.

21:24

They're not short exams.

21:25

And you know, some people are claustrophobic in the

21:26

MRI machine, so it's, you know, proper coaching.

21:29

If the patient can't be compliant, it

21:30

doesn't matter how perfect your hardware

21:32

is, you're not gonna get good images.

21:34

There's, um, a couple special categories

21:37

if you have non-diagnostic stiffness maps.

21:40

So let's say you did the study and none of

21:43

it is confident, it's, it's all hashed out.

21:46

Uh, the first thing you'd think

21:47

of would be iron overload.

21:48

Um, so you'd use a lower Tesla

21:51

magnet if iron overload's

21:53

suspected, and again, that's because the T2 star

21:56

times are less affected with a 1.5 than a 3.

21:59

Um, and you'd want to avoid gradient

22:01

recall elastography techniques,

22:03

again, that's because of the iron.

22:07

Ascites can affect, uh, the study.

22:10

That's just because the waves aren't gonna

22:11

transmit through the ascites to reach the

22:13

liver, so you can't measure liver stiffness.

22:15

And then finally, obesity.

22:16

And that should make sense

22:17

because the patient's bigger.

22:19

The waves, the mechanical waves,

22:21

have to travel through more tissue.

22:23

They're gonna, um, be dampened as

22:26

they travel through the tissue.

22:27

And so there's gonna be less effect on

22:28

the liver as you get deeper and deeper.

22:31

Overall, if you had to choose, you would do

22:32

a 1.5 Tesla for this over, uh, 3 Tesla,

22:36

'cause you're gonna get more diagnostic studies.

22:40

And with that, uh, we'll switch over to some cases.

22:43

Um, our first case, this is a real case.

22:47

And I will say the majority of the cases that

22:49

I've read over the last couple years were normal.

22:52

So good for the patients, but not

22:53

necessarily good for teaching.

22:55

But in this case, this first one was an

22:56

indication of fatty liver and they wanted

22:58

to see how much fibrosis was present.

23:02

So first thing we do is we obtain our Dixon sequences.

23:05

So you have the.

23:07

The fat Dixon, water Dixon. We'll draw some ROIs

23:10

and just visually looking at it, you can already

23:12

appreciate there's a lot more signal in this liver

23:15

than there were in the prior, um, fat Dixon sequences.

23:20

So, you know, 23, 44, 66.

23:23

So you've got high signal here and then you

23:25

also take the signal of the water Dixon.

23:29

And then you compute it.

23:31

In this case, we were able to determine the

23:32

patient had an average percentage of about 20%.

23:36

So we would say that's compatible with fatty liver.

23:38

Um, we then will.

23:40

Include our reference value.

23:41

So we consider 6% or less normal. Patients who have

23:46

6 to 26% would be considered mild steatosis, 26

23:49

to 37% moderate steatosis, 37 or greater severe steatosis.

23:53

And it may be helpful to describe geographic patterns,

23:56

but ultimately, you know, the treatment would be

23:59

reverse the cause of the fatty liver in this patient.

24:02

Next technique would be to look

24:03

at the iron quantification.

24:04

So here I have.

24:06

Relaxometry maps. On the left would be T2

24:08

star, and on the right is R2 star.

24:11

Um, again, kind of just drawing, um, ROIs in the

24:15

liver, you know, 2–3 centimeter diameter.

24:19

You don't want it to be too small.

24:20

You don't want it to be too big.

24:21

If it's too small, it's underrepresented.

24:22

If it's too big, it might have overlapping structures.

24:25

You don't want to get the hepatic veins or

24:27

the, um, portal vein if you can avoid it.

24:30

You don't want dilated ducts if you can avoid it,

24:33

um, and in this case, you know, we did it.

24:35

Um, and so, you know, you're getting

24:39

averages for the T2 star about 26, and

24:44

these are about, they say like a thousand.

24:45

I don't necessarily believe

24:46

that word reported, but um.

24:50

I think 50 is what we actually got.

24:52

So you know, in this case we're gonna

24:54

report the range of the T2 star values,

24:56

we'll do three measurements and then we

24:59

will look at our internal references.

25:01

And here, uh, 11 milliseconds or 11.4

25:05

milliseconds or greater, consider normal.

25:07

And as that T2 star time decreases below that, it

25:11

corresponds to mild, moderate, severe iron deposition.

25:13

I've never seen severe iron deposition on this study.

25:17

Usually when it's severe, I think they clinically

25:19

suspect it, and they'll do a formal, uh, FerriScan,

25:22

and that study could actually be, uh,

25:26

processed to determine how much iron content is.

25:29

So it's even more sophisticated than

25:31

what we're doing here, but this is

25:32

kind of a nice technique nonetheless.

25:36

And, um, for R2 star, remember, they're inverse.

25:39

So 88 hertz or less is normal.

25:41

And then as these, uh, frequencies get

25:43

higher, it becomes mild, moderate, severe.

25:47

And this is just the elastography.

25:49

So this is the wave, you know, we QC it, you

25:52

have really nice wave propagation to the liver.

25:55

You know, visually you can

25:55

see it outside of the liver.

25:57

It's a little less, uh, you know, reliable.

25:59

But that's also 'cause it's not in the patient.

26:03

So here we just, we have the

26:06

elastography sequences themselves.

26:09

You do have color coding, um, you know, zero

26:12

being black, eight being white, and these

26:15

values correspond to liver stiffnesses.

26:18

So we there, there, so it's

26:21

a 145, but really it's 1.45.

26:24

So we take 10 measurements.

26:26

We ignore the standard deviation of the range.

26:28

Again, it's, you wanna do it within the liver.

26:31

And so going through the different segments in the

26:33

right lobe, the left lobe is a little less reliable

26:36

just because of, uh, the lack of confidence.

26:39

But you could draw ROIs in the left lobe as well.

26:42

But again, more, you can see they're all highest.

26:45

Might be 1.7 here.

26:46

And so 1.3 to 1.7 is what we're reporting

26:49

as the liver stiffness in kilopascal.

26:52

The average was 1.5.

26:54

Um, and we have reference values in our report, so

26:57

anything less than 2.5 kilopascal is considered

25:59

normal, and then it's gonna go up in a stepwise

27:02

fashion, greater than 5 kilopascal being cirrhosis.

27:07

We've started to report on, uh, focal fibrosis where

27:12

the stiffness values are above background liver, but

27:14

only if they're greater than one hepatic segment.

27:17

And this has come up a few times when discussing

27:19

elastography, are the values you get a

27:23

distribution and or sampling of the whole liver?

27:26

So if you get one single measurement

27:28

and it's, you know, two stages above

27:30

what you're measuring, is that real?

27:31

Maybe not.

27:32

But if you see an actual, uh,

27:36

an actual segment of the liver

27:38

that's, you know, stages higher,

27:40

that probably is real and that's worth

27:42

reporting because that may guide therapy.

27:46

That was, you know, a nice easy case for you.

27:48

Uh, case two, uh, this is a patient who

27:51

had elevated liver enzymes, reported

27:54

history of hepatitis type C, 16 years.

27:56

It wasn't specified what kind. Patient

27:58

had used methotrexate, so they really

28:00

are concerned about fibrosis and steatosis.

28:05

This is again, our fat quantification.

28:07

Just visually, the fat signal's really dark.

28:10

Remember that's, that can be related to window.

28:11

So you do gotta, um, do actual ROIs.

28:15

And the water signal looks nice.

28:18

Draw your ROIs, your, you know, your

28:20

fat signals between 2 and 3,

28:23

your water signals between 97 and 140.

28:27

So just without even writing anything

28:29

down, you know it's not gonna be fatty liver.

28:32

And in this case.

28:34

It'd actually end up being 2%,

28:35

which you'll see from time to time.

28:36

It's kind of bizarre.

28:38

I don't know what to make of it when it's

28:39

low or even 0%, but I will record it.

28:42

Um, and again, same grading

28:45

values that we've used before.

28:49

Quantify the iron in this case.

28:51

Uh, T2 star and R2 star maps, you

28:53

know, times between, you know, 17, 18, 27

28:59

milliseconds, and then 55, 50, 36 hertz.

29:05

So the range of T2 star times was

29:08

16 to 21. T2 star time was 18.6, so

29:11

inconsistent with iron deposition.

29:15

R2 star values are 36 to 54 hertz.

29:18

The average is 44 hertz.

29:20

Again, inconsistent with iron deposition of liver.

29:22

Um, but in this case they're, they're

29:25

more concerned about the fat and

29:26

they're more concerned about fibrosis.

29:27

So then we look at the fibrosis.

29:32

And, you know, the waves are okay in this case.

29:34

They're a little discontinuous, but good news is the

29:38

confidence maps were a little bit more forgiving.

29:41

And here you can see stiffness is 1.9, 1.3, 2.2.

29:48

Six.

29:49

You know, the reason why it's reported

29:50

this way is only it's due to, uh,

29:53

this was made from Carestream or PACS.

29:55

I wish I could make it prettier for you guys, but

29:57

again, you know, the highest looks like 2.1 on

29:59

this measurement, but this is just to kind of

30:03

highlight when you're producing this, the machine can.

30:06

Instead of it being black and white, can it

30:08

display it in color, the lookup table?

30:11

And so visually it becomes very obvious.

30:13

There's no fibrosis here.

30:14

This is all very low stiffness liver.

30:17

Um, it's all near that zero mark.

30:19

So 0, 2, 4, 6, 8, you know, and we're

30:22

really focusing on stuff that's below

30:24

here, uh, to be normal stiffness.

30:27

So it looks good.

30:28

Uh, and that's, that's good for the patient.

30:30

So this patient had an average stiffness of 1.7.

30:33

It's incompatible with fibrosis.

30:34

There's no areas of focal fibrosis.

30:37

Um, sometimes you will, and I don't have

30:39

a case of this, I wish I did, but, um,

30:42

you'll get it when it's 2.5 to 3.

30:44

And in that case it could still be normal.

30:47

Um, but it may also just be related to

30:51

kind of their underlying hepatitis, and

30:54

you're not able to distinguish that on the.

30:56

Elastography, and if necessary,

30:58

they could get a biopsy.

31:01

And the equivalent of that for ultrasound would

31:03

be, you know, not a statistically significant,

31:06

like clinically advanced, uh, fibrosis.

31:11

Like it, it says a completely

31:12

different, um, nomenclature than that.

31:15

All right.

31:15

Case three.

31:16

This is a funny case because, uh, I read this

31:19

that went out with a fellow, uh, last year and I.

31:24

She's on the phone with the technologist and the

31:26

technologist is saying, Hey, I'm doing my best.

31:28

I can't get a diagnostic study and you know,

31:31

I'm not really paying that much attention.

31:33

I say, oh, it's probably this.

31:33

I might just go back to, you know, reading

31:35

out the resident who I'm with, and.

31:39

You know, I don't even pay that much attention to it.

31:41

And then the fellow gets to the spoiler and,

31:44

and you know, she was like, oh my God, you're right.

31:46

And I said, okay, thanks.

31:47

Um, it's because of this talk.

31:49

I was aware of it being a problem.

31:51

Um, but already off the bat,

31:53

look at the fat quantification.

31:55

You have these little areas of speckling.

31:57

It looks a little abnormal.

31:58

It doesn't look like the prior Dixon

32:00

fat test to fat up to this point.

32:02

Um, the signal of the liver is dark, so

32:06

it doesn't look like it's fatty liver.

32:08

And again, you're getting, you know,

32:09

2 to 5 for the fat signal.

32:12

You're getting 107 to 140 for the water signal.

32:15

So apparently I'm saying 22%.

32:19

I don't,

32:23

I don't think that's right.

32:26

Okay.

32:26

Well, apparently we reported 22%,

32:28

but um, visually it was not 22%.

32:32

So I'll have to QC this for

32:34

the next time I give this talk.

32:36

Um, but the interesting thing

32:38

was in the iron quantification.

32:40

So you look, and this doesn't necessarily

32:43

look the same as the other, and you've got

32:44

lots of areas where there's really no signal.

32:46

And so you gotta measure the areas where

32:47

there is signal, and here you're getting.

32:50

T2 star times about 7 milliseconds,

32:54

and you're getting R2 star times of

32:57

134, 144 hertz.

33:00

And this is actually a case of iron overload.

33:03

So the T2 star average was 7 milliseconds, so

33:06

it's consistent with iron deposition, so that would be mild

33:09

iron deposition based on our reference values.

33:12

Similarly, the R2 star values

33:14

are 150, mild iron deposition.

33:16

This is really quality control, right?

33:18

Because if.

33:18

I've had cases where T2 star was right and

33:20

then R2 star was, was vastly off, and it was

33:24

more, okay, we've gotta get the machine serviced,

33:26

'cause they are inverse values of each other.

33:28

They should be, um, congruent.

33:32

But this patient has iron overload.

33:35

And as we mentioned earlier in the troubleshooting

33:37

section, if you have non-diagnostic confidence

33:39

interval maps, the thing you should think

33:41

about is the patient has iron overload.

33:43

So this was unknown.

33:44

Um.

33:46

Or this wasn't known, but it was suspected,

33:49

and in this case proved to be true.

33:51

So we tried to get the elastography, we

33:53

looked at the waves, and you know, already you

33:55

know, if you can kind of hallucinate through

33:57

the Zoom video, the waves are broken up.

33:59

They're not, they're not

34:00

continuous, they're not smooth.

34:01

Like that second case or that

34:03

first case, they look really nice.

34:06

So we apply the, the confidence interval map.

34:08

And there's no liver that's really measurable.

34:11

You could potentially measure

34:12

that, but I wouldn't justify that.

34:14

Um, so it was considered a

34:15

non-diagnostic study for iron overload.

34:19

Um, and that's a known pitfall.

34:22

So this case is interesting because then

34:26

you've got the kind of a bonus here.

34:29

Remember I told you that you wanna look at the T2.

34:33

In this case, the T2 fat sat, you

34:35

can see really dark signal on the liver.

34:37

That makes sense.

34:37

They got iron, and I don't show you the pancreas,

34:40

but the pancreas signal is relatively preserved,

34:42

but then you have really dark signal in the

34:43

spleen, and so that supports a secondary cause

34:45

of hemochromatosis versus a primary cause.

34:49

Um, so there were other findings on the T2

34:53

that were, you know, potentially interesting.

34:55

Patient had IPNs.

34:56

But, uh, for hemochromatosis, I

34:59

think this is a helpful image.

35:02

And then finally we will kind of

35:03

round it out with a fourth case.

35:05

This is a patient who had obesity, abnormal

35:07

liver ultrasound, and they were concerned

35:09

for non-alcoholic steatohepatitis.

35:12

So I will say that this was probably not the right

35:15

study for non-alcoholic steatohepatitis, um,

35:19

but it is the right study to evaluate, um, fatty

35:24

liver and if there's any evidence of fibrosis.

35:28

So in this patient, uh, there is, it looks

35:32

like there is a little bit of fat in the liver,

35:34

although 5 to 6%, so normal.

35:39

Okay.

35:41

Again, you always want to calculate it out,

35:43

so the patient doesn't have fatty liver,

35:44

so they probably don't have non-alcoholic

35:46

hepatitis, so they probably just have regular

35:48

hepatitis. The T2 star and R2 star maps.

35:53

You know, we wanna draw our circles.

35:55

26, 27, 20, 42, 46, 37. Both are normal.

36:01

Uh, so there's no evidence of iron

36:03

overload in this patient, which is good.

36:06

And then our elastography maps.

36:12

So I try to convince you these waves are continuous.

36:15

They're just really tightly bound together.

36:20

And on the actual elastography sequences, we're

36:23

getting values much higher than we've gotten.

36:25

For the first two cases you're getting 4.4, 3.5,

36:29

3.1, and then there's even an area that's 9.3.

36:32

And then on the color map it becomes very obvious.

36:35

Lots of fibrosis throughout the liver,

36:38

lots of areas of green, yellow, red. Red

36:40

being the areas of the worst fibrosis.

36:43

And so this would be a case of, um.

36:48

So the range was between 2.6 and 8.9.

36:51

The average stiffness was 3.8, so

36:53

that's stage 2 to 3 fibrosis.

36:55

So it's, it's advanced fibrosis,

36:57

but it's not, uh, cirrhosis.

36:58

But there are, uh, areas, uh, segment

37:02

4A where you do have cirrhosis.

37:05

And so that potentially is, uh, a portion

37:08

of the liver that if you were to treat

37:09

the underlying issue may not recover.

37:13

And with that, another little bonus.

37:16

You got a T2.

37:17

The bonus here is you can look at

37:19

the spleen and see that it's big.

37:21

I don't have a measurement on it, but, um,

37:22

promise when I say it's 13 and a half centimeters.

37:25

So they're already starting to develop

37:26

signs of portal hypertension, and that

37:29

fits with their pattern of fibrosis.

37:31

So you're able to answer a question for

37:33

all four of your patients and give them

37:35

appropriate treatment down the road.

37:37

Uh, with that, we got about 20 minutes left.

37:40

I would just want to open it up

37:41

to questions if there are any.

37:45

Yeah, if any of our attendees do have

37:47

a question for Dr. Ti, please direct

37:49

it to the Q and A feature in Zoom.

37:55

It appears everybody's questions

37:56

are answered or they're all working.

37:59

It was the mo— it was, the talk was so comprehensive.

38:02

Yeah.

38:03

I mean, mean, it, it, and the, you

38:05

know, the attendees, you know, defense.

38:07

It, it, it's more of a how-to-do-it kind of thing.

38:10

And, uh, hopefully a lot of 'em know how to do it.

38:12

Well, I will keep my eye on the Q and A

38:14

feature just in case anything comes in, but

38:16

as we bring this to a close, I wanna thank Dr.

38:19

Ti for this lecture.

38:20

And thanks to all of you for

38:21

participating in our noon conference.

38:23

A reminder that this conference will be

38:25

available on demand on MRIONLINE.com in

38:29

addition to all of our previous noon conferences.

38:31

And be sure to join us again on Tuesday.

38:34

But it looks like we have one question.

38:40

All right, so, um, one of the questions

38:44

is the interest of T2 fat sat.

38:51

I'm not sure what the question is, but I

38:54

will say we use T2 fat sat in this, um, in

38:58

this protocol for, uh, anatomic evaluation.

39:01

So.

39:02

Remember that patients with hepatic fibrosis are

39:05

at risk for developing hepatocellular carcinomas

39:08

as well as other kind of incidental pathologies.

39:10

But it would be, uh.

39:12

It would be a disservice to the patient if you had

39:15

the technique available to you, especially because

39:16

the sequence itself isn’t that long to obtain, to

39:19

just make sure there aren't any, um, abnormalities

39:23

that would require further workups, like a liver mass.

39:25

So most common example would be

39:27

you're doing a fibrosis workup.

39:29

There's an indeterminate, mildly

39:31

hyperintense mass in the liver.

39:33

The patient subsequently.

39:34

Gets a multiphase liver MRI, uh, it has the imaging

39:38

characteristics of a HCC, and now the patient went

39:41

from possible fibrosis to, you know, a LI-RADS 5

39:46

lesion with definitive HCC so they can get treatment.

39:52

And then the next question is, um, is ultrasound

39:56

elastography as sensitive as MR elastography?

39:59

There's similar sensitivity and specificities.

40:02

Uh, if I recall correctly, the

40:04

sensitivity and specificities, while

40:05

comparable, MRI's a little bit higher.

40:07

It has a little bit higher

40:09

agreeability with, uh, biopsy.

40:11

Um, but it's not, I.

40:14

I, I, if, uh, if I were to, and again, uh,

40:17

this is off memory, it's like 5% difference.

40:20

Uh, 10% difference.

40:21

So, um, both ultrasound and MR perform really well.

40:25

It's just MR is a little bit

40:27

more sensitive and specific.

40:31

And then another question, what

40:33

are some indications for biopsy after

40:34

elastography with either MRI or ultrasound?

40:38

Um, one would be to kind of

40:41

document what the findings would be.

40:43

So say you said a patient had grade 2–3

40:46

fibrosis, um, that would be, you know, they, and if

40:50

the treatment protocol depended on it being stage

40:54

1 or stage 4, uh, they may proceed with biopsy.

40:58

I will say my experience, I haven't had any

41:00

referrals for, uh, liver biopsy after a diagnostic

41:05

elastography, but I have had cases where there

41:07

were non-diagnostic elastography, so there'd be

41:09

patients with ascites or obesity, um, who, or even

41:13

iron overload, who had non-diagnostic, um, studies

41:16

and they wanted to grade the elements of fibrosis.

41:20

All right, well thank you again to Dr. Ti and I

41:24

just wanna be sure to let everyone know to join us

41:26

again on Tuesday for a lecture from Dr. Mahesh on

41:31

managing radiation risks in medical X-ray imaging.

41:35

You can register for that at MRIONLINE.com and follow

41:38

us on social media at MRI Online for updates

41:42

and reminders on upcoming noon conferences.

41:45

Thanks again and have a great day.

41:47

Alright, thank you everyone.