0:01

Hello and welcome to Noon Conference, hosted by MRI Online

0:06

Noon Conference connects the global radiology community

0:08

through free live educational webinars that are accessible

0:12

for all and is an opportunity

0:13

to learn alongside top radiologists from around the world.

0:17

You can access the recording of today's conference

0:19

and previous noon conferences

0:21

by creating a free MRI online account.

0:25

Today we are honored to welcome Dr.

0:27

Fel Diarco for a lecture entitled Pediatric Epilepsy.

0:30

What the Radiologist Needs to Know, Dr.

0:33

Diarco completed his undergraduate medical training

0:36

and radiology training at the University Federico of Naples,

0:40

Italy, an internship at the University Hospital

0:43

of Luve Belgium, and a Head

0:45

and Neck imaging internship at the University Hospital

0:49

Hospitali Sevo in Brescia, Italy.

0:51

Dr. Diarco completed a fellowship in pediatric

0:54

neuroradiology at the Hospital for Sick Children Toronto,

0:58

Canada, and subsequently joined Great Ormond Street Hospital

1:02

as a consultant pediatric neuroradiologist in 2015.

1:06

He serves on the editorial boards of neuroradiology

1:09

and child's nervous system, is involved with the head

1:12

and neck subcommittee of the A SNR

1:14

and manages a YouTube channel dedicated

1:16

to pediatric neuroradiology education

1:19

for residents and fellows.

1:22

At the end of the lecture, please join Dr.

1:24

Daro in a q and a session

1:25

where he will address questions you

1:27

may have on today's topic.

1:29

Please remember to use the q

1:30

and a feature to submit your questions so we can get to

1:33

as many as we can before our time is up.

1:35

With that, we are ready to begin today's lecture. Dr.

1:38

Daro, please take it from here.

1:41

Um, thank you very much

1:43

for the kind introduction and the invitation.

1:45

Um, uh, my name is Felicia.

1:48

Uh, as was said, I, I work

1:49

as pediatric neurologist in Greater Ormond Street Hospital,

1:52

which is very big, uh, pediatric hospital in London, uk.

1:57

And my main field of interest are pediatric head and neck

2:00

and, um, pediatric epilepsy.

2:03

Um, so in, in this lecture, I will just, um, uh, show you,

2:08

uh, some cases, uh, some technical advances

2:11

and from basic protocol

2:12

to technical advances in the images of epilepsy.

2:16

And if I have time, I would like to have a last, um, um,

2:20

part dedicated to gray moderator Utopia in particular, uh,

2:24

how we use pattern recognition to, uh, give our clinician,

2:27

um, clinicians, uh, um, genetic

2:31

or non-genetic diagnosis in these cases

2:34

of gray matter Heterotopia, uh, this is the summary.

2:38

So we speak about the technical aspect

2:41

of modern neuro imaging in epilepsy, uh,

2:44

some fascinating cases, uh, preoperative planning.

2:47

And as I say, gray matter heterotopia, if I have time, uh,

2:50

let me start with this quote from the Little Prince,

2:53

a very beautiful book.

2:55

And, uh, uh, uh, they, they say in this book, it, it is only

2:59

with the heart that one can see rightly,

3:01

what is essential is invisible to the eye.

3:03

And something similar applied to epilepsy

3:06

for years and years.

3:08

But now something has changed in the last, uh,

3:11

25 years, I will say.

3:13

And now with the MRI, what is essential, which are these

3:18

small lesions, uh,

3:20

that are epileptogenic can be operated on,

3:24

are no longer invisible,

3:26

but we, we need to approach this patient, right?

3:29

So, uh, I'll show you

3:31

what is our approach grade demonstrate,

3:33

and in general, big center, uh, that are dealing

3:36

with epilepsy, uh, patient.

3:39

So first of all, keep in mind that

3:42

what we call invisible is in children.

3:44

Most of the time, small focal cortical dysplasia

3:48

or sometimes areas of gray matter, heterotopia and polyuria.

3:53

Um, but what we need to see this invisible is power.

3:58

And in MRI power means, um, three Tesla.

4:02

And if you look at the literature, most of the um, um,

4:07

uh, paper say the three Tesla is definitely better

4:10

and should be used for focal epilepsy, uh,

4:15

except for motion.

4:17

Uh, so remember, the more powerful is the magnet,

4:22

the more sensitive is to flow artifact to patient motion.

4:25

And this is especially relevant in children.

4:28

Um, and, uh, so you need to act, um,

4:32

accordingly to have a perfect scan

4:34

because you will need a perfect scan in order

4:37

to find this very small lesion.

4:40

This is an example, 1.5 entry Tesla.

4:43

Uh, you may have heard that there is a lot

4:45

of push in using seven Tesla images for this indication,

4:50

in particular for, for epilepsy.

4:52

But this is still, uh, not used widely on,

4:56

on a clinical basis.

4:57

There are some center, uh, in, in us, some in Europe,

5:02

but it's very difficult to have optimized sequences.

5:05

So I will stick with the three Tesla as far as this, uh,

5:09

clinical and practical lecture.

5:10

Le lecture is concerned.

5:12

Uh, you can see the difference here

5:14

and you can see here this small area of hyperintensity

5:18

is much better seen in three Tesla.

5:20

Remember, like Spider-Man say, with great power

5:23

and great responsibility.

5:25

And with three Tesla, I great sensitive to motion and

5:28

because we are acquiring a lot

5:30

of 3D sequences in these patients,

5:33

remember if you have a bit of motion in the original

5:36

sequence, like in this case

5:38

where you are wondering if there's something here,

5:40

you will have motion of course in the, in the reform.

5:43

So on a practical standpoint, if we are approaching a child

5:48

with focal epilepsy,

5:50

maybe they already had a previous scan which was reported

5:54

negative, so we are looking for something very small,

5:57

namely focal cortical dysplasia.

5:59

What we have to do differently from the other, um,

6:03

clinical indication from tumor inflammation

6:05

and so on, we need to achieve a perfect scan.

6:08

And this means that we need to have, uh, an expert setup

6:12

and most importantly, low threshold

6:15

for general anesthesia or deep sedation in this patient.

6:19

So basically we start speaking with our nurses

6:22

and try to ex that they select the patient

6:24

that needs general anesthesia or sedation, uh,

6:28

and we started to explain them

6:29

that they cannot apply the same, uh, reasoning for children

6:34

with focal epilepsy, that they apply for children even

6:37

with tumor, that if they move a bit,

6:39

it doesn't mean much in terms of our diagnostic power.

6:43

So if you have a child,

6:44

you are not sure you can stay completely still consider

6:46

general stage or deep sedation when it's not, uh, possible.

6:51

Consider training. We bought this inflatable MRI scanner.

6:54

It doesn't cost much. It's 2000, 3000, um, uh,

6:59

euros, and you can train the child.

7:02

The flow

7:03

of the air pumping into this inflatable simulate the

7:07

noise of the MRI.

7:08

So this is quite, um, useful, uh,

7:12

and cost effective.

7:13

But of course you need a team. We have a play team here.

7:16

This is our radiographer, Jess,

7:18

that train the children when deemed necessary.

7:22

Uh, we said about volumetric acquisition.

7:24

We need a lot of, uh, spatial resolution,

7:28

but look at this sequence.

7:29

This is, uh, of course 1.52, uh, dimensional,

7:33

not the best sequence.

7:34

We are not sure what's going on here in the three Tesla,

7:39

um, uh, scanner.

7:40

We did a 3D, T one.

7:42

So we have special resolution,

7:45

but also good contrast resolution, not only

7:48

because it's a three Tesla,

7:49

but remember you need the gradients such as MP range, uh,

7:54

or equivalent because the MP range

7:56

of the equivalent is very good

7:57

for gray white matter differentiation.

7:59

If you use Siemens, for instance, the 3D space, good

8:02

for post contrast,

8:03

but not as good for gray white matter differentiation.

8:07

So know your scanner,

8:09

but to keep it, uh, simple, the MP range is, is very good.

8:14

And in this case, there was this area of blurring

8:16

between the cortex and the white matter.

8:18

So this is a fogal cortal dysplasia same apply for T two.

8:21

The three DT two is not easy to achieve.

8:24

We still using two, uh, DDI dimensional,

8:28

uh, T two weighted images.

8:29

But of course, in three Tesla, the signal

8:31

to noise rash is much better.

8:33

You can see much better this area

8:36

of cortical thickening corresponding

8:38

to a focal cortical displacer.

8:39

Type two beep. Uh,

8:43

one important practical point I guys, I I try

8:46

to keep it practical for you.

8:48

If you have a small lesion, like in this case,

8:52

remember the neurosurgeon, uh,

8:54

do not have cross-sectional even with neuro navigation.

8:58

Uh, they, they, they are much more familiar

9:01

with surface anatomy, not with cross-sectional anatomy,

9:04

but you can reformat the surface of the brain

9:07

and you can tell them, look, this correspond

9:10

to this fogal cordial dysplasia correspond to this

9:14

area of the brain.

9:15

So when they open up the skull,

9:17

they can see the gy look at the anatomy that you, uh,

9:21

send them on a screenshot like,

9:23

like a screenshot, like this one.

9:25

And they found this very helpful.

9:26

And this is something that my, uh, neurosurgeons taught me.

9:31

So remember, um, you need a lot of technical aspect.

9:35

You need the three Tesla, but remember,

9:37

you need the multidisciplinary approach,

9:39

multiparametric approach.

9:41

And this is a teamwork. And I will go through some of this.

9:45

Remember, regarding theological differentiation

9:47

of fogal cardio space.

9:49

This is actually all, there is a newer, um, uh, paper,

9:54

uh, that was, um, that was, uh, uh, published with, uh,

9:58

also some mild malformation

10:00

of cortical development, moga and so on.

10:02

Uh, but, uh, as far as the, the,

10:04

the main classification is concerned, remember that

10:08

the type one is usually not easily visualized on MRI

10:13

look for indirect sign atrophy, hyperplasia,

10:16

abnormal ation of duration.

10:18

If you have a young child with a mild malformation

10:21

of cortical development like mga, what we call mga,

10:25

M-O-G-H-E look also for some subtle bound of hyperintensity

10:30

below the cortex.

10:32

Um, but remember the type two are much better seen,

10:36

especially the type two B.

10:38

So let's see some example.

10:40

Look at these cases, a lot

10:41

of things going on intractable seizure.

10:43

And you see this area of hyperintensity below

10:48

and the, the cortex.

10:49

So cortical subcortical area of hyperintensity with some

10:54

also line that we call transman sign

10:56

that we can track down to the ventricle.

10:58

But there are multiple, very rare, uh, to have multiple, uh,

11:03

areas of focal cortile dysplasia

11:05

that we call tumor in this case,

11:07

but very rare outside this condition.

11:09

Okay, so this is the first thing.

11:11

And then you have a lot of subependymal nodules

11:14

that are actually Ammar Thomas.

11:17

And in another similar case, you also have

11:20

a large tumor in the foram of morron.

11:22

So this we call Sega sub giant cell astrocytoma.

11:26

Most of you already did the diagnosis.

11:28

This is tuberculosis complex.

11:30

But I want to show you this case

11:32

because tub buro sclerosis complex is characterized

11:35

by several focal cortical dys,

11:37

paal type two B histologically.

11:40

Sometimes they have calcification.

11:42

It's very difficult to manage this, uh, in, um, uh,

11:46

uh, like in surgery.

11:48

But, uh, remember these are actually, um, a lot, lot

11:53

of focal cort phenotype two B,

11:54

they look exactly the same when they're in insulation.

11:57

But look at this is you, it is much more difficult to find

12:01

because there are not like 40 of these.

12:03

This is one only, uh, very close to the area of the brain

12:07

that control the movement of the feet.

12:09

But same radiological characteristic, a bit

12:11

of hyperintensity in T two, uh, blaring

12:14

with the gray white matter junction, a bit

12:16

of hyper intensity in T one.

12:18

This is a fogal cordial speco type two B.

12:21

And I show you here in, in, you know, um, a larger,

12:25

um, uh, picture.

12:26

But remember you need to know

12:30

where the seizure are coming from.

12:32

So another aspect, apart from the sedation I told,

12:34

or general anesthesia, the three Tesla, uh,

12:37

technical aspect, 3D volume.

12:40

Remember, you need to ask precisely

12:42

where the seizure are coming from.

12:44

I still receiving from other hospital a lot of,

12:46

uh, clinical indication.

12:48

I say focal seizure from where, uh,

12:50

and also the type of seizure helps you a lot.

12:53

So they ideally should tell you on EG

12:55

or clinically where to look.

12:58

Remember that we are dealing with children,

13:01

and children dynamically change in terms

13:04

of brain marination.

13:06

And this is important.

13:07

So if you have a neo

13:09

or an infant very early on, you have a background

13:12

of unmyelinated brain,

13:14

which is bright in T two and dark in T one.

13:17

So the, the the, um, the white matter looks like water.

13:21

And in this interval, so very early on is very easy

13:25

or easier to pick up this malformation.

13:27

Look at the blurring, the thickening here,

13:29

but if you wait 4, 5, 6, 8 months, the myelination start

13:33

to kick in faster.

13:35

In T one, let's say visualize faster in T one than T two,

13:39

but your window opportunity to delineate this

13:44

abnormality is, uh, less.

13:47

So. Remember guys, either you image very early or

13:51

after two years in between, if you are forced, you have to,

13:54

but your di your diagnostic yield will be less.

13:57

Yeah. So remember that, look for this abnormality.

14:00

But the timing of scanning is important.

14:05

And remember, the transman sign, transman sign is abnormal.

14:09

Um, uh, the dysplastic tissue going from the abnormal cortex

14:13

to the vent because sometimes this is the only thing you see

14:16

and you track it down to probably abnormal cortex.

14:18

Very helpful, not always so striking.

14:21

But this is another characteristic

14:23

of foal cordial dysplasia, in particular the type two B.

14:28

This is another situation, uh,

14:31

where you don't have subcortical hyper intensity,

14:34

you don't have transplant sign,

14:35

but you do have the blurring.

14:37

If I zoom it here,

14:39

and of course we had the clinical indication,

14:41

an EG indication, you will have blurring

14:45

of the gray white matter junction,

14:47

and this was a type two A.

14:50

Um, some, some people say that it's not easy

14:54

to distinguish type two B

14:55

and type two A without the transplant sign.

14:57

They can overlap a lot.

14:58

It doesn't change your management,

15:00

but remember, sometimes you just have the blurring.

15:04

Sometimes you have something like that.

15:06

Look, you, we couldn't see very well the

15:09

blurring maybe a bit here.

15:10

We know the seizures were coming from the right frontal

15:13

lobe, but the right frontal lobe was slightly larger.

15:17

Interestingly enough, we had also a low grade glioma

15:21

with this typical cotton like out focus enhancement hyper

15:25

intensity on the same size.

15:27

So remember that there are genetic mutation involved

15:32

in the fogal cordial dysplasia.

15:34

Um, uh, and this can be an mTOR mutation.

15:38

You know, there, there are pathways

15:39

that include the tubals sclerosis genes as well.

15:43

But remember the differently from, uh, bros scs,

15:47

this is a somatic mutation.

15:49

And, um, uh, uh, this can be, uh, it can explain

15:54

why these, uh,

15:55

two abnormalities are in the same side of the brain.

15:59

So remember, look for the, um, um,

16:02

for the associated finding

16:03

and remember that this, uh,

16:05

if you do the genetic analysis on the biopsy specimen,

16:08

you can have a somatic mutation important

16:12

after the, uh, the, the focal cortical dysplasia,

16:15

the mesiotemporal sclerosis.

16:17

Mesiotemporal sclerosis can be secondary to seizure,

16:22

uh, can be, um, a primary cause of, of seizure.

16:27

It's not as frequent in children as is in adult,

16:30

but remember, you have a reduction of the volume,

16:33

bright signal and loss of the internal structure.

16:36

Uh, and, uh, uh, remember again,

16:40

a technical practical point.

16:42

Uh, you need to orient the coronal perpendicular

16:46

to the long axis of the hippocampus.

16:49

Uh, some people still orient the axial

16:52

perpendicular, uh, to, to this.

16:54

But to be honest with you,

16:56

this really makes very complicated to find the gy,

16:59

the anatomy of the gi.

17:00

So if you have fogal epilepsy, you are looking

17:02

for either hippocampus sclerosis of,

17:05

of fogal cortel dys patient.

17:06

You don't want an axial like that.

17:08

Uh, so I personally changed the protocol when I started the

17:12

gosh to have only a coronal, um,

17:16

inclination along the perpendicular to the axis

17:19

of the hippocampus in order to better see the, the, um,

17:23

hippocampus sclerosis.

17:25

You can see hippocampus sclerosis also in a 1.5 Tesla.

17:28

But the internal structure with this ilum nucle,

17:32

this eye point intense area, which is stratum radicalis

17:36

and the granulosis

17:37

and moleculars, so so called the KoSA mon is.

17:41

Um, and then, um, the, the, uh, this other bundle

17:45

that is the, the suum, basically this,

17:49

this laminar appearance is better visualized in three Tesla,

17:54

you can see here is preserved here, is lost,

17:59

interestingly enough, especially with the seven Tesla,

18:01

but also with a good, um,

18:03

three Tesla you can now distinguish.

18:06

And this is a beautiful publication from, uh, um,

18:08

pro Professor Middlebrook.

18:10

Uh, you can distinguish the different type

18:13

of hippocampus sclerosis.

18:15

So this is the normal appearance.

18:17

You can see Soum ave,

18:20

fia co carsoon, one, two, and three.

18:23

Uh, the stratum, uh, radium gran

18:26

and molecular is altogether here, ipo, intense bundle.

18:30

And the helium that is CF four, uh,

18:33

and Denate, uh, you cannot distinguish between the two,

18:36

but this 1, 2, 3,

18:37

and four alternating band of hypo hypo intensity needs

18:41

to be preserved.

18:43

If they, you have a global atrophy,

18:46

you can have a type one hippocampus sclerosis,

18:49

if mostly involves the corso is one, uh,

18:53

you have a type two.

18:54

And if you, if you have mainly the ilum,

18:57

but the rest of the corso is

18:59

preserved, you have a type three.

19:00

And there may be some, uh, implication, uh, um,

19:05

in terms of patient management, uh,

19:08

and, um, uh, so it's very important, uh, that you try

19:11

to recognize these, uh, small changes.

19:16

Remember also tumors can give you focal seizure.

19:20

This is a dnet.

19:21

We recognize the dnet for the band of Hyperintensity, um,

19:26

uh, the hyperintensity rim.

19:29

So, uh, this is, uh, um, uh, typical in children,

19:33

not in a of, of dnet in, in a way, you know,

19:36

of course in other to have the EDH

19:38

mutation with this things.

19:40

But in children, like, uh, two more close

19:42

to the cortex things a dnet.

19:45

Uh, so remember the flare hyperintense re uh,

19:50

if you have contrast, you have a gang Oma, uh,

19:54

but do not try

19:55

to distinguish tumor too much this low grade al tumor,

19:59

because of course now there are,

20:01

with the new classification, more

20:02

and more, uh, tumors like, uh, um, PLN

20:06

and TY, other, uh, there are several of them.

20:10

You don't need to distinguish them because the management,

20:13

or you can try, but the management is similar.

20:14

So, so if you're not sure in the context of epilepsy,

20:18

you see a peripherally located tumor, just say, uh,

20:21

there's a low-grade neuronal tumor

20:24

that is most likely responsible for the epilepsy,

20:27

but remember to give contrast, uh,

20:29

and remain that they can be associated for focal

20:31

with focal cortical dysplasia.

20:35

You can also have seizure related to bleeding, of course,

20:38

you know that, but remember that you need to keep looking,

20:42

especially in children.

20:43

Genetic diseases are a problem.

20:45

And you can have here not only this bleed with the,

20:49

with edema,

20:50

but look other areas of, um, susceptibility.

20:54

And then you do an SWI or a gradient,

20:56

and you see a lot of areas susceptibility.

20:59

So this is a carbon osis, a genetic form

21:02

of carbon smart formation.

21:04

So you need to tell your clinician to look for the genes.

21:09

That's very, very important.

21:10

And also they tend to, uh, bleed

21:14

the cortex is partially functioning, which again, I mean,

21:18

this is a diffuse poly micro jia in deafness.

21:20

This is a, um, CMV infection prenatally.

21:24

But remember, once you recognize this lumpy,

21:27

bumpy appearance because differently from the focal cortical

21:31

dysplasia, this is something that, um, the, the,

21:36

the cortex in this cases

21:38

maintain some function is more difficult,

21:40

the surgical management of the isolated form of polyuria.

21:45

But remember this lumpy, bumpy appearance,

21:48

prenatal infection or prenatal insult can give you that,

21:52

but also genetic disease.

21:54

Look at this lumpy bumpy cortex,

21:56

different from this very sharp

21:58

gray white matter differentiation.

21:59

There is also hypo myelination here

22:02

and some cyst along this, the, the ventricle in the context

22:07

of a neonate with hypotonia localized poly micro

22:10

jia hypo myelination.

22:12

Uh, and this cyst are typical of cell vector syndrome.

22:15

So again, pattern recognition is very important,

22:19

and pattern recognition can give you a lot of, uh,

22:22

differential diagnosis in the context of seizure.

22:26

But really I want to focus, uh, in this lecture on the,

22:30

the fo cordial dysplasia.

22:32

Now we found them. So I, um, of course I could speak

22:35

for hours about different neurogenetic insult, uh,

22:40

uh, and and so on.

22:42

This is a lesion that was highlighted by the, uh,

22:47

nuclear medicine examination.

22:48

So remember, you can use a pet FDG pet, very important.

22:52

Most of the time you will have high perfu.

22:54

We don't know why.

22:56

Um, um, I was, uh, uh, lecturing at the Eli, um,

23:02

neuroimaging course,

23:03

and there was a beautiful lecture on a pet expert.

23:06

Uh, and the way they look at the pet, the way they, they,

23:10

they changed the color mapping and so on.

23:12

But as far as you are concerned,

23:14

remember the 10 na hypometabolism in the,

23:16

in the inner critical, um, uh, moment is very,

23:21

very suggestive.

23:23

Uh, and this is another mapping of, of course,

23:26

you can have here clear asymmetry.

23:28

And then you come back and you see

23:30

that there is a cortical thickening.

23:32

So you suspect the cortical dysplasia ear,

23:35

but remember, this is the Interictal pec al pec

23:39

better the spec can give you area of metabolism

23:44

because of course, the cortef is seizing.

23:48

Remember that you have other, uh, chances

23:51

to, to find a lesion.

23:52

This was a lesion that was for some reason,

23:54

and we can discuss about the technical analysis of that,

23:57

but sometimes it, for some reason, it's better seen on 1.5

24:01

that Tesla and then was less, uh, visual look at different,

24:05

this is sharper than this,

24:07

and then it, it becomes more difficult

24:10

to differentiate the two.

24:11

But then you do a meg magneto ence biography,

24:14

and you have a lot of cluster coming from a,

24:17

so you have also Meg, who don't have

24:19

so much experience with Meg.

24:20

We send the children to other institution,

24:23

but remember that Meg is use electrical current horizon

24:27

inside the neurons of the brain.

24:29

Uh, and so the skull

24:31

and soft tissue affect less meg than eeg.

24:34

Uh, so you can combine like in this case MRI, uh,

24:38

EG meg pet,

24:40

and this is the same case showing also the

24:43

abnormality on pet.

24:46

Uh, I will go very fast on DTI tractography, you,

24:50

you all know that DTI shows that the, the bundles,

24:52

the white matter, and you can use to show the relationship

24:56

with the p cortical dysplasia.

24:58

Same you can do to functional with functional MRI, I have

25:02

to say OO of course, that this helped the surgeon,

25:05

but the surgeon will always use,

25:07

especially when the focal cor t space is

25:09

close to critical areas.

25:10

Uh, intraoperative stimulation is more important

25:13

for preoperative, uh, uh, planning to, to show

25:16

where the language is coming from.

25:19

Like in this case, from literature, you see, uh, uh,

25:22

of course the language is very close to this large tumor.

25:26

And when we have to do operation,

25:28

especially when we do stereotactic, SEG, it's very important

25:32

for them to have an, uh, to visualize the vessels

25:35

so they can, um, say, you know, to the, the, they can,

25:40

uh, figure out how

25:42

to a better approach reducing the complications.

25:44

So sometimes we do, uh, CTA.

25:48

What are the future direction?

25:50

Uh, first of all, uh, we are trying

25:52

to establish a specific focal cortical dysplasia MRI

25:56

protocol using more advanced technique

25:59

and artificial intelligence.

26:01

We try to evaluate this. We send a pilot to epilepsy.

26:05

They didn't like it. I was sent somewhere else,

26:07

but just to say that all these advanced technique

26:10

that you read in literature needs to be, um, optimized

26:15

and, and, um, used in the real life context.

26:18

So you need to wait.

26:19

The, the cost, uh, in terms of time,

26:22

not all the economical cost per se, uh,

26:25

and the, the benefit.

26:27

Uh, so what are the problem where we should look?

26:30

I told you you need to have clear indication where to look.

26:33

So the discussion with clinician is the,

26:36

the most important thing.

26:38

And I told you that we need a floral scan,

26:42

but most important, we need

26:45

to keep advancing our sequences

26:48

and then determining what to use and what to not to use.

26:51

So this is our pattern.

26:52

When we have a patient with focal epilepsy

26:54

and previous MRI negative

26:56

or questionable result, we request this specific protocol

27:00

where we add AMP two age sequence

27:03

with the age announcement post-processing.

27:05

We had a SL, uh, sorry, I didn't put a SL here.

27:09

And we had post-processing with meld, which is one

27:12

of the artificial intelligence software available.

27:15

This is the one we use.

27:16

There are other out there

27:18

with slightly different characteristic,

27:19

and then we re discuss

27:22

and, uh, in case we do PET or spec,

27:26

and we also have a seven Tesla St.

27:27

Thomas Hospital where I work.

27:30

So sometimes where I, I work in both hospitals

27:32

and sometimes we use also the seven Tesla.

27:35

Um, and then we reevaluate again.

27:38

Um, this is our protocol just to show you.

27:42

This is the MP agent, you MP two agent.

27:45

You see how many different contracts we have.

27:48

This is the age announcement

27:49

that I will go through in a moment.

27:52

But MELD is, uh, open source.

27:54

So you can go on the MELD project website

27:57

and ask, uh, them how to do, uh, to use this software.

28:02

Uh, and of course we have the other standard

28:04

sequences and a SL.

28:07

So this is MELD report.

28:08

Basically, they tell you what parameters are abnormal

28:11

and when the cluster of abnormal comes from.

28:14

And, uh, uh, the MP two range

28:18

with the age announcement was originally used

28:22

to target this, the, the different nuclei of the OME

28:26

for deep brain stimulation.

28:27

For instance, children with lenon gastro, um, syndrome.

28:31

From there, uh, you know, we, we tried, uh,

28:35

to use the same approach for visualization of small

28:40

for focal cortical dysplasia.

28:43

And every time we scan with the amputee age,

28:46

and then we do this post-processing,

28:48

we have all these different kind of, um, different, um,

28:52

um, contrast.

28:55

But this agents is the one we use the most together

28:59

with the, um, original unit sequence.

29:02

So does it work? There are lot publication, uh,

29:06

some on artificial intelligence software, including meld,

29:09

uh, and some about the age added value, uh,

29:14

and age.

29:15

Um, I want to show you just some examples.

29:19

This was, uh, old scan.

29:21

You see also the flare is not optimized.

29:24

The seizure were coming from there,

29:25

but we were not sure about that.

29:27

This lesion becomes very clear, uh,

29:30

with the subsequent, uh, scan.

29:33

The flare is more optimized,

29:35

and this is the age of the ude agent.

29:37

You see this age of brightness between gray

29:40

and white matter is interrupted

29:44

where the lesion is.

29:46

So this can confirm suspicious on other sequences.

29:50

Uh, the same, in the same, um, uh, patients.

29:55

There, there was an IPO perfusion, uh, reduction of CBF,

30:01

uh, uh, with a SL

30:03

and meld, um, uh, picked up an hotspot

30:06

because they picked up the thickening, the blurring,

30:09

and the abnormal signal in the same area.

30:11

So everything was consistent. You can operate on it.

30:15

Another case of slight as C, this is a bit artifact.

30:19

The slight asymmetry, um, of the, of the temporal pole,

30:23

the seizure were coming from the, uh, striking hypoperfusion

30:27

and again, positivity on melt.

30:29

This was a case where we report a normal,

30:32

even though we did this,

30:33

the focal cordia dedicated protocol.

30:37

But then Mel picked up this area here,

30:41

we review flare was possible,

30:44

a bit blurred, very difficult.

30:47

Uh, but then you see the, the age showed this blurring

30:52

of the age between gray and white matter.

30:55

Um, other cases, frontal bilateral seizure.

30:59

Again, uh, the age in this case was negative.

31:01

So we need to look at all the sequences we,

31:03

but Mel picked up this, and this was thicken it on flare

31:07

and possibly there was a small transman sign

31:12

or we were, we are still not sure, but we need to operate.

31:15

But just to say, to show you how we reason,

31:17

and we come back to the scan, we look at the flare, we look

31:20

for small transman sign based on the age

31:23

or new EEG findings.

31:26

Uh, so, uh,

31:27

before I start, I, I go from the last, uh,

31:29

last 10 minutes on the pattern recognition,

31:32

gray moderator Radiotopia, just to summarize,

31:34

you need a perfect scan.

31:36

Try to use the three Tesla.

31:37

You need general anesthesia, sedation

31:40

or training for children.

31:42

That cannot be, that cannot stay very still.

31:45

You need to know where the seizure are coming from.

31:47

You need to speak with neurosurgeon

31:49

and keep an eye on artificial intelligence,

31:52

new intelligence, new sequences and so on.

31:55

But it has to be a critical eye

31:57

because, um, you, you,

32:00

I think the most important thing is still the experience

32:03

of the radiologists and the time we spend looking at the

32:06

scan and learning from, from each other.

32:07

I'm very lucky that I have amazing colleagues

32:10

that pick up on my misses and mistakes and, and, and misses

32:14

and, and, and, um, and I can learn from from this

32:18

gray moderator opia.

32:20

So, uh, these are the learning objective of this.

32:23

This second part, uh, what is the definition, how we assess,

32:27

but most important guys, I want to show you

32:29

how you can distinguish when you see gray moderator utopia

32:32

in association with poly micro gy,

32:34

especially if this is genetic or not.

32:37

So definition gray matter heterotopia are clusters

32:41

of normal looking neurons, let's say normal, uh,

32:44

in abnormal location.

32:47

And the most common is the periventricular nodular

32:50

heterotopia before we call sub heterotopia.

32:53

And there are subtypes.

32:55

The subtype of this is a laminar heterotopia, uh,

32:59

rem I will show you this in a in a second,

33:01

but remember, this is the most common.

33:04

And then we have subcortical heterotopia, basically islands

33:08

of, of white matter, of gray matter in the white matter

33:12

between the cortex and the lateral ventricle.

33:14

So no longer along, no, no longer, uh, attached

33:19

to the ventricular margin,

33:21

we have the transman heterotopia from the ventricle

33:24

to the cortex with these two subgroups

33:28

and sub cortical bind band heterotopia.

33:31

This is pattern recognition when you know how to pick up,

33:35

and I will show you in a second the

33:36

subcortical band heterotopia.

33:38

You know that this will be a genetic course,

33:40

and this changes a lot in terms of,

33:42

of management of the child.

33:45

So these are island of abnormal, um,

33:47

of gray matter in the white matter.

33:49

So let, let's look at that.

33:51

This is the typical perticular nodal opia attached

33:55

to the ventricular margin.

33:57

This is the subgroup of laminar heterotopia.

34:01

This is not the sub cortical band heterotopia smooth coline

34:04

heterotopic layer with associated, uh, um,

34:07

white matter abnormalities.

34:09

Uh, I forgot to mention,

34:11

please look at this beautiful paper from Zaino.

34:14

If you want to, um, look at the classification

34:18

of brain malformation sub band heterotopia.

34:22

It looks like that is a band

34:24

of heterotopia in the white matter can be anthrop anthrop

34:28

posterior like diffused leg in this case,

34:30

or can be only posterior or more posterior or more anterior.

34:33

But remember when you see something like that,

34:36

this will be genetic, especially 80% this two gene, TCX

34:41

and li one transman sub corate

34:46

utopia is a bundle of abnormal G matter from the,

34:50

from the ventricle to the cortex.

34:53

And you see also here the ventricle is abnormal.

34:55

There is probably abnormal corpus callosum

34:58

by the look of the ventricles.

34:59

This is most likely an insult, prenatal insult.

35:05

One subgroup of this is sub global dysplasia.

35:08

So-called brain, brain is s quite extensive

35:12

and it looks like a small brain into the bigger brain.

35:17

Another subgroup of that is the ribbon,

35:19

like heterotopia is diffuses look like a rebo.

35:23

Again, pattern recognition.

35:24

If you see something like that,

35:26

this ribbon like diffuse heterotopia,

35:29

think genetic cause.

35:31

And Elma el e ml one is the main gene.

35:38

Before I show you the pattern recognition mimics,

35:40

we have seen already tubo sclerosis.

35:43

Please don't call the sub penal nodo sub, uh, uh,

35:47

superal Amar Thomas, uh, in a con of tub sclerosis,

35:51

gray matter rot over, they are not,

35:54

and do not call the transplant sign gray matter heterotopia.

35:58

This is dysplastic tissue.

36:00

There are papers that shows how the, um, typical appearances

36:05

of the, um, type two B detail

36:07

or focal cordial dysplasia are also in this transplant.

36:12

We need to assess this with an epilepsy protocol

36:16

because this is isso intense to the cortex,

36:19

but can be very small.

36:21

Sometimes you have isolated noles,

36:23

so you need a high resolution.

36:25

Remember though that you can use the restriction

36:28

to show the epileptic network.

36:30

Look at these cases. There are nodular utopia right

36:34

and left, but there is also

36:37

abnormal cortex showing the future restriction here, um,

36:42

in the median nup posterior aspect of the brain,

36:45

probably this abnormal cortex

36:47

and the nodal utopia at

36:49

that in this side are seizing together

36:51

because a part of the same network, this is not showing

36:55

restriction during the peral phase.

36:58

So it's very interesting to know that there is some changes

37:01

in the diffusion restriction.

37:04

So now most importantly, can we distinguish

37:07

between genetic and non-genetic?

37:10

So if you have association with destructive phenomena that,

37:15

uh, can be, uh, uh, found together with no doula, with, uh,

37:20

with gray matter heterotopia, think a prenatal

37:24

destructive event, not a genetic cause.

37:27

But keep in mind that there is some,

37:29

there are some genes like called four a one

37:32

or call four A two that create a high rate

37:36

of prenatal ischemia.

37:38

Okay? So that's the first thing.

37:41

And more and more genes that comoso ide have been described,

37:45

uh, uh, are being described over time.

37:48

So first thing, if you see gray matter heterotopia with

37:53

dandy walker, oh, sorry, there is a problem here.

37:55

Uh, uh, I don't dunno what happened. Yeah.

37:59

Um, oh, strange, sorry.

38:04

Okay, so, uh,

38:06

so if you see Dandy Walker malformation in gray,

38:09

moderator utopia together,

38:10

think prenatal destruction like in this case, uh,

38:14

from professor o gamma Ddy worker,

38:16

big transplant heterotopia,

38:18

this was probably a prenatal event.

38:20

This is one case that we had at gosh in my hospital,

38:24

perticular Heterotopia, a cleft that is a destructive event.

38:28

Dendy worker mark, mal formation, loss of white matter bulk.

38:32

This is, um, probably a prenatal insult.

38:36

And we know from publication that a ma, a major contribution

38:40

for non-genetic prenatal factor are present in

38:45

children individual with, uh, dandy walker.

38:49

Um, and this is to probably to very early on

38:54

destructive event in the, the, the, the, um,

38:57

development of the cerebral.

38:58

And this is a beautiful paper from Dr.

39:01

Aldi, really brilliant paper showing the histopathological

39:05

changes in ddy worker.

39:07

So if you said in the worker

39:09

and you, you know, the Ddy work case,

39:10

a prenatal insult most likely,

39:12

and then you see, um, um, gray mi heterotopia,

39:16

you don't think too hard.

39:17

It's always, everything is possible,

39:19

most likely is a prenatal insult.

39:22

If you have poly microglia in association with heterotopia

39:26

and particular trans heterotopia

39:28

and perr heterotopia, like in this case think insult.

39:33

There is a, uh, quite, uh, strong theory

39:37

that large area of poly micro jia

39:40

or schizo celi are actually due to prenatal vascular, uh,

39:44

um, damage.

39:46

And this is, uh, this is, uh, um,

39:48

a paper from Professor Griffey.

39:49

They basically say when the ischemia happened too early on,

39:52

the, the, the brain cannot produce gliosis.

39:55

So it fill the gap producing, uh, uh, abnormal gray matter.

40:00

This is the origin of gray, uh,

40:02

of poly microglia in some cases, of course,

40:07

not in a berg,

40:08

but if you have a unilateral area of poly micro Gaia, uh,

40:12

and then association with gray matter,

40:14

probably this was an ischemia very early on.

40:17

This is another case with bilateral schizos, open lips,

40:20

closed lips, um, uh, uh, heterotopia,

40:24

trans heterotopia poly microglia, destruction

40:27

of the corpus callosum.

40:29

Uh, this is probably, um, a destructive event,

40:33

but remember to test

40:35

for genes like called four A one, A two.

40:38

If you have a craniofacial abnormalities like fronton nasal

40:41

dysplasia, you can have, um, uh, look at the bones

40:45

of the nose here and antibody heterotopia transman

40:50

heterotopia here, polymeal gia.

40:52

You can actually have, um, uh, most likely this, uh, uh,

40:57

negative genetic result.

40:59

This has been published.

41:00

We had the case and all the panel was negative.

41:03

So think out of the box

41:06

and don't push too much towards expensive genetic test.

41:10

If you have enough, uh, data, uh, to, uh, to think, uh,

41:15

that this is most likely an insult.

41:18

What are the genetic we spoke about sub,

41:21

sub cortical band heterotopia.

41:24

We have several genes, mostly this one, DCX, uh, A CTB,

41:28

and a Ct G one also, uh, present with this.

41:32

But, uh, you know, there is an overlap.

41:36

Of course, the gradient can help you in distinguish

41:38

between least one and this six.

41:40

But just run the panel of malformation

41:44

and you will pick up the right gene.

41:46

The important thing is the radiologist is you recognize

41:50

the presence of a band here.

41:52

Look, the posterior grad in li one, uh,

41:56

and you call it a genetic per ventricular noal,

42:00

gray matter heterotopia, a lot

42:02

of nodus especi along the ventricle, especially superiorly,

42:06

uh, they are typical of Phin aid.

42:09

These are female average cognitive ability, uh,

42:13

and they can have other association,

42:16

but remember the same pattern in both female

42:20

and male in a bad, um, uh, like in a bad context,

42:25

uh, uh, neurologically of eth brain atrophy

42:29

and with scarring of the pina.

42:33

So same radiological pattern,

42:35

but with different association, different gene.

42:38

This is, um, uh, a FG, uh, EEF two.

42:42

Uh, and this is, um, actually, um, uh,

42:48

very bad, uh, uh, uh, caries very bad prognosis

42:51

if you have deafness per ventricular artery.

42:54

Utopia. Another abnormalities, corpus callosum

42:58

corpus cephalic, uh, typically, uh, posterior,

43:02

so inferior cerebellar dysplasia.

43:05

This is pattern recognition guys.

43:07

This is Charlie Mecho syndrome.

43:10

You have a specific genes gene.

43:13

So, uh, this is very important.

43:15

So look at the cerebella abnormal, the first in association

43:19

with pert utopia

43:21

and other abnormal, the poly micro abnormal corpus colo

43:25

in the clinical context of deafness.

43:27

Very specific for this,

43:29

and we, we see, we saw recent recently, uh, diagnosis

43:33

with this partner was missed.

43:36

So you need to know, and once you know, you'll pick it up

43:39

and save a lot of money and time to your clinician

43:42

and a lot of us for your, for the family of your patient

43:46

per heterotopia.

43:47

I need to thank my colleagues, um, Nia Sakhar

43:50

who show me this pattern.

43:52

Uh, when you see something like that almost wing like

43:57

peri, uh, uh, temporal per trigonal et heterotopia

44:01

with other association, pon of microcephaly.

44:04

This is either a chap, um, so it's one of the chap gene,

44:09

either chap 1, 3 5

44:11

or uh, also, uh, TNPJ mutation.

44:15

They look very similar,

44:16

but again, this is pattern recognition.

44:19

Once you see once

44:20

or a pattern like that, you can solve that a lot of, uh, uh,

44:25

problems and give them a bunch of genes to look for.

44:30

Like heterotopia, I told you already bilateral

44:34

very specific pattern.

44:36

Again, it's your eyes that give you the diagnosis.

44:38

You remember this pattern,

44:40

and you have gene ML one, uh, uh, which is also associated

44:44

with the brain overgrowth, uh, syndrome.

44:47

And this is the paper if you want to look for it.

44:51

So, uh, what are my take home messages? Right?

44:55

MRI protocol is essential, uh, but more

44:58

and more genes are associated.

44:59

So you need to look for pattern recognition,

45:02

but do not forget the prenatal insult.

45:04

And now you navigate, you remember your patterns

45:07

and you remember what abnormal is in the brain are

45:10

associated most likely with a prenatal insult.

45:14

You put the things together

45:15

and you navigate between the two, uh, categories.

45:20

Um, do not forget the mimics.

45:23

Uh, and, um, uh,

45:25

and again, uh, remember that this is like for the sequences,

45:30

new or new genes are being described every day.

45:33

Uh, you need to keep up with the literature. Okay?

45:36

Um, my lecture is finished.

45:39

Um, and, uh, I am, uh, ready for to go

45:43

through some questions.

45:45

Uh, I'll start with the q and a and then the chat.

45:47

Please let me know if it, I was clear or not clear.

45:50

You want me to come back if I say something wrong?

45:53

Uh, you always learn.

45:55

Um, um, so, okay, uh, Sarah,

46:00

ask when you when to schedule the patients.

46:02

Uh, very early when, so ideally if you have a, a neonate

46:06

with focal seizure, you do the scan, and if you have seizure

46:10

and they are focal, of course there are a lot of causes

46:13

for seizure are not focal,

46:14

but the, uh, your clinician needs to guide you.

46:17

But the earlier the better.

46:19

If you are in a patient older than four months

46:22

and younger than two years,

46:24

and they have fogal seizure, they want, they want the scan,

46:28

you just need to be aware that it's more difficult

46:32

because of the physiological blurring due to the myelination

46:35

that is kicking in and reaching the cortex.

46:37

Okay? Um, so sometimes we scan like that,

46:42

but the earlier the, the better.

46:44

Um, so, uh, marai

46:49

Solomon, I don't know what is the surname?

46:51

Um, the flat maps, uh,

46:56

no, I haven't, I don't have experience with the flat maps.

46:59

We are trying to, um, optimize this.

47:03

Uh, first there are also new sequences that are, you know,

47:07

my colleagues in Boston, they are checking some, uh,

47:10

different artificial intelligence.

47:13

Um, oh wait, you know, flat maps, you just need, uh,

47:17

mean the, the, the map when you open up or, or the,

47:21

because there is something called Flo flowers, maybe

47:25

that is a different sequence.

47:26

So if you, if you say, oh, when you open up the scan, yes,

47:31

I use this all the time.

47:33

Uh, if you intend some other artificial intelligence,

47:37

there are so many, and, uh, to be honest with you,

47:40

it's a bit difficult to navigate.

47:42

We use this because it's done at UCLH.

47:45

Okay, sorry, I'm not sure what what you mean.

47:48

Um, per ventricular nodularity heterotopia, uh, well, no,

47:53

you don't have all the time ventricular

47:55

enlargement, to be honest.

47:57

Uh, so I would not, uh, um,

48:00

sometimes you have localized periventricular

48:02

nodularity heterotopia.

48:03

There is actually the, the, if anything, the,

48:06

the ventricle is contracted.

48:08

Look very carefully in the posterior

48:11

and inferior aspect of the, of the ventricles,

48:15

because sometimes there, they, they go misdiagnosed, uh,

48:19

sub cordial band heterotopia.

48:21

Sometimes you're missing some of the sequences,

48:24

especially when, uh, there are these other genes, A, B, T, C

48:28

and, and where is a bit more blood.

48:31

Um, but yes, it,

48:33

it always have a genetic substrate when you see a clear

48:36

bond, um, uh, not a transplant, right?

48:40

Clear bond. Um, um, yes, the, the,

48:45

the, the current knowledge is that this is genetic.

48:52

So, uh, perfusion, so I show you the A SL,

48:55

we use only a SL In theory, perfusion helps in this,

49:00

but at the moment,

49:02

nuclear medicine is still more established.

49:06

Um, but, uh, in theory can help you.

49:10

Uh, I, I, I don't understand if it's our a SL that is not,

49:14

uh, optimized because maybe the child is moving

49:17

or the age is too early, uh,

49:19

but we don't find us helpful as the pet at the moment.

49:23

Um, uh, so for, for the age announcement smoothing,

49:28

I'm not sure about the values, but you can get in contact.

49:31

I can, uh, put you in contact with my thesises

49:35

because we create our own software of the MPH

49:38

that create this ages, not the age sequence of the Siemens

49:42

that created automatically.

49:44

Um, so yes, I,

49:49

I use a SL routinely for protocol.

49:52

I send a paper to epilepsy in only 14 patients

49:56

because we wanted this serological diagnosis.

49:59

Now, I may not agree with all due respect with the reviewer,

50:03

but most of the results were kind of saying the flare,

50:06

optimized flare is still the best sequence

50:09

and it was rejected.

50:11

Um, so I think it's easier

50:14

to say, uh, yeah.

50:16

So great result and so on.

50:18

Our experience is that a SL most of the cases confirm

50:23

the lesion, but was never the one that say, oh,

50:26

this is for sure a lesion.

50:29

But on the other side, uh, found some lesions that then we

50:34

looking back at the scan we found on the scanner.

50:37

So, uh, at the moment, this is my experience,

50:40

but of course, you know, I'm improving.

50:43

Uh, we are trying to to, to really to

50:45

to improve technically and to look better.

50:48

So it may change, but this is my, uh, experience.

50:52

Uh, so we use contrast for tumors, uh,

50:56

and development arm malformation.

50:58

It depends where if you need vascular malformation,

51:00

of course, search web, stuff like that.

51:02

We use tumor, uh, we use contrast

51:04

for foal, cortical dysplasia.

51:05

We never use contrast

51:07

unless we are requested to do an MPH post contrast

51:11

because they want to see the best cell

51:13

before they like CCTA.

51:15

Uh, it depends by their now neuro navigation software.

51:18

Uh, sometimes they ask for contrast.

51:23

We use meg. Yeah, I said, um, in, in my presentation

51:27

that we have, uh, used MEG in selected cases

51:30

when I was in Toronto.

51:31

They were much more stronger in meg.

51:34

Uh, but we use only in selected p uh, cases.

51:36

Now, there is a big push in using stereotactic ESEG,

51:42

um, more than, uh, than to send the patient, uh,

51:45

somewhere else with Meg, because we have to do Meg

51:47

because we don't have the meg, um, in-house.

51:50

Uh, so Quin are done. There is something in chat.

51:58

Uh, yeah, just, okay. Thank you for the compliment.

52:01

And, and, uh, uh, you know, again, just to say, um,

52:05

there are a lot of, not only my YouTube channel

52:08

and these, uh, beautiful resources here, there are a lot

52:11

of resources out there.

52:13

Uh, old people are very available for me as well.

52:15

I'm learning like constantly,

52:16

or people are very available if you have a doubt to share

52:21

the protocol sequences suggestions.

52:24

So, uh, get in touch with me, with, with other colleagues.

52:29

Uh, and, um, I mean, the lectures, uh,

52:33

I think will be recorded if you, if you want

52:36

similar lectures are also my YouTube channel, uh,

52:40

separated, but they are there.

52:41

So all, all the, all the things are, um,

52:44

are there. Okay. Awesome.

52:46

Thank you so much Dr. Dco.

52:47

That was a great lecture

52:49

and thank you so much for taking the time

52:50

to answer everyone's questions today.

52:53

Oh, my, my pleasure. I'm very happy to, to share also

52:56

because I learned a lot really, um, every day.

53:00

Thank you so much. And, uh, guys, have a nice, uh, evening

53:03

or day whenever you are.

53:05

Yes. And thank you to everyone

53:06

for participating in our noon conference

53:09

and asking such great questions.

53:10

You can access the recording of today's conference

53:13

and all our previous noon conferences

53:15

by creating a free MRI online account.

53:18

We'll also email out a link to the replay later today.

53:22

Be sure to join us next week on Thursday,

53:25

July 18th at 12:00 PM Eastern,

53:27

where Dr. Dennis Beki will deliver a lecture entitled

53:30

Introduction to Arthritis Part three.

53:33

You can register for@mrionline.com

53:36

and follow us on social media

53:37

for updates on future noon conferences.

53:40

Thanks again, and have a great day.