3D Ultrasound of Craniofacial Abnormalities - HD

Introduction

The reason why Dolores asked me to give this talk is that we were at a meeting together several years ago, and she asked me to, I, I was asked to give my own talk about 3D ultrasound, many of cranial facial abnormalities.

And she was there, she heard that talk, and she liked it a lot.

Dolores is having her knee operated on and couldn't be here.

So she asked me to give this talk.

Now, she asked, said I could give my own talk or substitute slides or do whatever I wanted, but I thought it was only fair to really give her talk because she's doing some things that nobody else is doing, things that we're certainly not doing.

So I went over the slides with her and I'll do the best they can to show you what she wanted to tell you guys.

Dolores had no financial disclosures.

I added this, I do manage a grant for Toshiba, and I thought it was only fair to mention that.

The objectives of the talk here are to describe what views we take for routine screening in the first and second trimester.

Demonstrate where 3D views are helpful in scanning patients, particularly obese patients, and then show clinical applications where 3D is useful.

Clarification on 3D Imaging

Now, on my own talk, I clarified one thing, and I'm gonna do that here before we start.

3D means different things to different people.

There's 3D surface rendering, which is what I think most people think of when they talk about 3D.

However, a lot of what Dolores has here is what I would call tomographic imaging.

So she's actually taking tomographic images through the fetus, and then there's always the multiplanar reconstructions as well.

And all of these seem to fall under the umbrella of 3D imaging.

First Trimester Applications

The way she does these at UCSD in the first trimester, she's really looking at the palate and she's looking at the embryo as well.

And these are multiplanar reconstructions.

And so this would be the axial image here.

Okay? And this is the palate right here, okay? And kind of where the nose is gonna be, they do coronal reconstructions as well.

So you have frontal bones, nose, a little bit of chin, and then the sagittal view as well.

And this is the head right here.

This is probably the most useful image.

Here's the nasal bone.

And this is the palate here.

And this is part of the mandible in chin.

And this is a 12 weeks, one day.

And looking at the palate is the key thing.

And just to compare a normal, this is the normal one I just showed you.

Okay? And then this is one with a midline cleft palate.

And you can see on this view, the axial view that there's a big defect running right up the middle of the palate.

Okay? And that's to me, a beautiful image.

We've not been obtaining that image.

This is compared to normal.

On this view, I think you can see there's a defect in the middle here.

Okay? This should look like this and continuous, and instead there's a big defect in it.

I think the sagittal view, you can just see there's no palate.

See, you should be able to see the bone here, the palate extending posteriorly into the fetus, and here you just don't see it.

Okay? In this case was a Trisomy 13, diagnosed at 12 weeks one day.

So very early, and a very good use of multiplanar imaging.

And this is another one, cleft palate here, no palate seen there.

Big defect in the middle.

And this is the normal for comparison.

Alright, unilateral cleft lip, same idea.

And she's put the normals here on each one of these.

So you can kinda see that.

Here's the defect here, right on the axial, I guess this is really the coronal image.

I think she switched these.

I think this looks axial to me.

There's the defect there.

And then on the sagittal image, it's a little harder to tell because this is not in the midline, it's unilateral, and I think this is how you tell it's unilateral.

So I think this is a very good use of multiplanar imaging to make the diagnosis in the first trimester of cleft palate.

And then she's been doing 3D surface rendering of embryos in the first trimester.

And I think there's great utility to this.

I think you can tell if there's arms, if there's legs, if the umbilical cord is in the right place, if the face looks reasonably normal and you can pick up abnormalities doing this.

We've really not been doing this, but we probably ought to be now 12 weeks.

We're doing a lot of nuchal translucencies, and I think we know that picking up neural tube defects can be difficult early in the first trimester.

When we do these, here's some ultrasound images, a coronal image and a sagittal image.

And I think you can see that something's abnormal here on these.

And I think you can see how abnormally angled this looks on the 3D reconstructions, you can see the ribs, the neck, the head here, and here's the defect.

And even though these aren't really even completely ossified, there's some value to the 3D rendered image.

When we look at these, here's another one.

This is a sagittal view, doing a nuchal translucency.

And there's no fourth ventricle.

We're all used to looking at this, but there's actually a space here that's supposed to be here.

And we look at this one, there's no cisterna magna here, and maybe the thalami, or the tectum is pointed, if you will.

And we get the idea that there's probably a neural tube problem on the axial view.

The skin overlying the thoracic spine has a defect in it.

And you can see that these laminae are protruding out laterally rather than closing like the triangle that they're supposed to form.

And there's a defect there.

And this is really the first trimester uses for what I would call multiplanar or tomographic imaging.

Second Trimester Applications

Now, in the second trimester, obviously there's more use for this.

And Dolores wanted to share with you images of the brain, face, spine and extremities that they get at UCSD.

And this is available and you can see all kinds of pictures of the face.

And I'm gonna make some comments about this work towards the end of this lecture here.

Brain Imaging: Cavum Septum Pellucidum

So one thing Doris really, really wanted to emphasize, and to me this was really new stuff, was imaging the cavum septum pellucidum using tomographic imaging.

So this is the normal, okay, 22 weeks.

And the thing that's really, really important to understand is that the frontal horns are supposed to be next to the cavum septum pellucidum.

And the cavum septum pellucidum should be really a rectangular shape.

Either a square or rectangular shape should have smooth margins, okay?

It's not a trapezoid, it's not any other geometric shape.

It's really rectangular.

And there it is.

Okay? Now here's a case.

See, this is a 22 weeks.

All right? Now, what's wrong with this case?

Well, first of all, the frontal horns are not next to the cavum septum pellucidum, okay?

And that makes this abnormal.

And then secondly, this is not exactly a rectangular shape.

This is a little oddly shaped, okay?

And part of the reason Dolores wanted me to give this talk is I made the comment in my own talk that it's not just about the cavum septum pellucidum being present, it has to look normal as well.

And abnormalities of the cavum septum have kind of been under reported, actually, in the literature.

And even clinically, here's a postnatal MRI.

Okay? And you can see the cavum isn't quite normal.

Maybe there's colpocephaly here.

Maybe this is agenesis of the corpus callosum.

But regardless, the cavum septum pellucidum is abnormal.

And that was at 32 weeks, by the way.

So this is a multiplanar display of the cavum septum pellucidum, and this is in the midline.

And you can see what this looks like.

This is on the coronal view.

Now the dot corresponds to the same location in each view.

You can see that the frontal horns are supposed to be next to each other here.

And you get that 20 weeks, five days, you get this multiplanar projection of the cavum.

Okay?

Alright. So normal, normal 3D views of these.

And you notice that the frontal horns are next to the cavum septum pellucidum.

Alright? And then this is the corpus callosum, okay?

And this is the cavum, okay?

This is on a coronal view.

This is on an axial view, and I think I'm not sure we're really seeing the frontal horns here, so I'm not entirely sure this one is normal.

And I think maybe it was agenesis of the corpus callosum, but the point Dolores wanted to share is that looking at this as extremely important and it's much, much better with tomographic imaging.

And I'm gonna show you some examples that are abnormal towards the end of this.

This is a technique called thick slice display that she's been using.

So this is really a multiplanar image, but it's reconstructed at a thicker interval than the normal display.

And normally, I think these are two millimeters thick, and these particular ones are four to five millimeters thick, 25 weeks.

One day, cavum septum pellucidum has been measured, and it's measured transversely across the head.

And like any other parameter, there's a distribution of normals that follow a curve.

And this is normal.

It increases with gestational age.

There are normal variants.

This is a cavum septum pellucidum at vermis.

And we see this sometimes, almost inevitably, whenever we see one of these, the sonographer will come out and think that this is something else.

This is either an aneurysm or this is the third ventricle being dilated.

But normally you can tell because it's not between the thalami, it's a little more superior and it just continues with the cavum septum pellucidum.

So it's really a normal variant.

Alright? Another thing Dolores has been working on is imaging of the posterior fossa.

And this is a normal vermis, okay?

The vermis is something a lot of times it's difficult to tell.

Is there really inferior vermian agenesis, is the vermis normal?

I think when there's really a Dandy-Walker variant, it's a little easier.

But sometimes it's not so easy to tell.

So this is the vermis here, okay?

Cisterna magna vermis cerebellar hemisphere is here 22 weeks, four days.

And Dolores says, this looks like an orange slice.

So you can think of these as being an orange slice.

Here it is here.

Okay? And we make a measurement usually from top to bottom.

And if the inferior vermis isn't gonna form, it's gonna be smaller than normal.

And these are all normal cerebellar hemispheres vermis on a coronal view, here's the axial view, it's here.

Little harder to see, but if you're concerned about it, it's a good thing to look at.

Fourth ventricle here is the cisterna magna, 18 weeks, four days.

And again, the orange slice.

The Brits named all these abnormalities after fruit, lemon sign, strawberry sign, banana sign.

Now we have the orange slice sign as well in honor of Dolores.

And here's the vermis with the thick slice.

And she uses the thick slice in patients that are of high BMI, big patients.

And this helps delineate the anatomy.

And you can see the vermis here.

These are all normals.

Facial Structures

Now the palate.

And I find these images pretty exciting.

This is at 18 weeks, 24 weeks.

And a rendered image of this.

And you can see how clearly you can see the tooth buds here.

Okay? And this is very, very important when looking for cleft palate, actually.

And these are tomographic images and she goes up and down through the face and axial and produces these images.

And I'm gonna show you some cleft lips, and you can see, you can start at the orbits and you can really see the lenses of the eyes very well.

So the orbits are really there.

You can see the palate beautifully to know whether there's a cleft or not.

You can see the tooth buds, right?

You can even see the mandibles, the mandible right here going through tooth buds again.

So really beautiful images.

And I thought these were fairly novel.

I've not seen images that look like this.

Some clarity to visualization.

There's a lot of reverb artifact when looking at the normal image, looking at these eyes and looking at this face.

But this thick slice rendered image in this patient with a BMI of 58, alright?

And really does clear up the image quite a bit.

So there is some utility for using this tomographic, thick slice imaging in patients of large BMI.

Here's what the 2D image looks like.

A lot of reverb, a lot of dropout, thick slice of the palate.

And I think you can see that pretty well.

Okay? So it's definitely an improvement in the appearance of things.

Again, BMI of 35, here's the spine sagittal view of the spine.

And it's an okay image.

We can't really resolve all the vertebral bodies.

We don't see the laminae all that well.

But on the thick slice reconstructions, I think we get a little better resolution.

And then this is actually a pretty nice image.

I mean, this would be perfectly reasonable, I think diagnostically.

So quite a bit of improvement doing this 18 weeks, one day, normal spines, okay?

Fairly useful.

Sacrum, sacrum is always a problem.

Sometimes you can't see the sacrum.

The bigger the patient, the harder it is to see, is this some kind of sacral agenesis or caudal regression.

And some improvement in clarity doing the thick slice imaging here.

Cervical spine, pretty difficult to see.

There's some drop out here due to overlying structures, even at the lower cervical, thoracic spine.

But doing the thick slice, you can resolve the vertebral bodies.

And I think you can see this looks a little more normal.

It's a little more reassuring to the patients.

Some extremities, this is a hand, I think this hand looks strange, actually.

I told Dolores that I thought this looks strange, but it was normal.

Club foot or a normal foot, I'm sorry, just a normal foot there.

Normal hand, normal foot ears, okay?

We don't usually look at ears routinely, but obviously ultrasound doesn't do a great job at looking at the surface of structures.

We see through things better, but this is a volume rendered image of an ear and position shape.

There's various things that can be detected by looking at ears.

These are artifacts.

That's one of the problem with rendered images.

There's really no skin tags here, so I asked her about that.

Abnormal Brain Findings

So some abnormalities, an abnormal cavum septum pellucidum.

Alright? And this is agenesis of the corpus callosum on this coronal.

The coronal view is always the best view really to look for this.

But we don't see the corpus callosum.

The frontal horns are not down next to the cavum septum pellucidum.

And on the sagittal view, we see that the third ventricle is coming up somewhat superiorly here and here between the thalami.

And that indicates agenesis of the corpus callosum.

Now this is another one where the cavum septum is abnormal in shape.

Okay? This doesn't really look quite like a rectangle, right?

This looks like a flying squirrel to me.

And you can see the legs and the arms here.

The frontal horns are not next to it, reconstructed view.

These are 3D acquired images, by the way.

And then she's rendering them in the different planes.

And on the coronal view, you can see there is a corpus callosum present.

But there's fusion of the frontal horns of the lateral ventricles, and they're a little superior to the level of the cavum septum pellucidum.

Here it is on the sagittal view, a beautiful image of the corpus callosum.

And this turned out to be septo-optic dysplasia, and that's important to know.

Now, this is where I think this is really novel and this is really interesting stuff.

But these are tomographic images going from anterior to posterior through the cavum septum pellucidum frontal horns, cavum septum pellucidum looks completely normal anteriorly.

And yet as we run posteriorly, we can see that this isn't normal.

That there's incomplete septation here.

And here this is normal here, but here, there isn't.

Alright? And we see a little bit of corpus callosum here, right?

And this turns out to be septo-optic dysplasia.

And she has many examples of this demonstrating that if you do tomographic imaging through the cavum septum pellucidum, you can pick up really subtle abnormalities.

And I think if I had just seen this septum, this image is really the one we usually look at.

I probably would've cleared this.

And how does Dolores know this?

Well, they have a very good QA program there.

And they learned how to recognize these things.

So she's doing tomographic imaging through the cavum septum pellucidum in almost every case, prominent cavum of the cavum septum pellucidum is associated with various trisomies, 22q11 deletion, all kinds of things.

So if it's enlarged without any other anomalies, it isn't entirely clear.

And this is features, this is in trisomy 18, 92% of 'em are gonna have a prominent cavum septum pellucidum down syndrome.

Trisomy 13.

And this was published.

So this is a cavum septum pellucidum that's normal, about two millimeters or so.

And I showed you the table for this.

Here's the big one.

This is nine millimeters, almost a centimeter here.

And they've had 15 cases of this at UCSD.

And five of them were thought to have chromosome abnormalities ultimately, and here's one that was trisomy 21.

Okay? So in large cavum septum pellucidum, it's not something I had really thought about.

I mean, we've noted it before and seen it, but I hadn't just really thought about it that much.

So that the 15 cases at UCSD 13 had other congenital anomalies.

10 had adverse outcomes, adverse outcomes, brain anomalies, heart disease, myotonic, pye in utero, fetal demise, and abnormal karyotypes.

And this was published in 2017 recently.

So an enlarged cavum septum pellucidum really has prognostic implications here.

Prominent cisterna magna, okay?

Something we see all the time.

We used to think it was a normal variant.

Alright? But now the question is, is this Dandy-Walker malformation?

Is this a Blake's pouch cyst with just rotation of the vermis?

Is it just the big cisterna magna?

The big thing is, is it vermian hypoplasia?

I think that's what everybody wants to know.

Here's some schematics showing that this is the normal cisterna magna, Blake's pouch.

It connects basically to the fourth ventricle here and a mega cisterna magna, okay?

Just the markedly enlarged.

These are MRIs okay to correlate with these, okay?

And I think you can see that this is kind of an abnormal position of this cerebellar vermis and on the tomographic images, here's the cyst there.

These get a little hard to see here.

We come up, here's the Blake's pouch connecting to where the fourth ventricle would be.

And this is the enlarged cisterna magna.

Here's the vermis here.

Alright? So this is useful for telling inferior cerebellar agenesis, if you will, from other abnormalities.

And we do this every now and then when it comes up, but we don't do it in every case.

And neither does Dolores.

Sutures and Skull

Sutures.

Okay? She's using it to look at sutures.

And you can see the various sutures, you can look this up.

And 3D rendering of the skull.

Now this is a situation where you really do need a surface rendering to see these sutures.

18 weeks, 21 weeks.

Okay? Sagittal suture.

Oh, this is a coronal suture, actually.

Craniosynostosis, okay, no sutures here.

Normal fused here.

We really haven't been doing a lot of this.

This is the lambdoid suture, and this is fairly normal, okay?

This is wide open.

It hasn't really fused, okay?

To acquire these, generally you'll get a volume acquisition, and she usually gets this in the axial plane, and then gets coronal reconstructions as well of individual surface reconstructions.

It's very important to know that with these structures, you need to do multiple planes and multiple acquisitions to see these.

You can't see the sagittal suture if you're acquiring the images in sagittal.

You have to be coronal or axial to see them.

And the same is true with the other sutures.

You have to do multiplanar imaging.

And you can see these beautiful images here.

And this is abnormal, okay?

This is wide open.

The sagittal suture just hasn't really fused there.

And she's just showing the location just so that you can line up your images and know where you are.

Nasal Bone and Facial Anomalies

Nasal bone is a very good use of this.

And personally, I think our sonographers are really, really good and they can get this view most of the time.

But sometimes you get this maxillary process here, the malar bone next to the nose, and you think you're seeing the nose and you're really not all right?

And you can get fooled by that.

So there's 3D multiplanar imaging is a very good way to make sure you're really in the midline and that you're not tilted some way and not really in the nose.

Here's an absent nasal bone.

So we're doing tomographic images, somewhat axially through the nose.

And we can see as we come down orbits that we should be seeing a bone here.

The nasal bone should be here and we really don't see it.

We're starting to get into palate a little bit with the bottom of the eyes.

So it's a very good way to really make sure whether you're seeing nasal bone or not, since that's the best predictor of trisomy 21.

We have really other anomalies that you can see.

Cleft lip orbits, abnormal facial profiles.

So here's a unilateral cleft lip.

Dolores gave me 85 slides to show you, by the way.

So I'm going the best I can.

Bilateral cleft lip.

Okay? A little hard to tell that this is bilateral and this is one that they missed.

Okay? Now, I don't know that we would ever know we've missed the cleft lip, frankly, but this is one on a surface rendering.

It's kind of hard to see, but you can see the premaxillary protrusion here.

And then on the rendered axial image, you can see that there's actually a cleft there and a bigger cleft here.

So this is really a nice image demonstrating the abnormality.

And I think it's better than what we would do, just getting a single slice through there.

And here's normal just for comparison, okay?

Another bilateral cleft lip we look through here and as we come down, you can see there's a big cleft here coming through here.

I think the other cleft on the other side is smaller.

This is bilateral cleft lip and palate, okay?

On these axial images, which I think is an excellent way of seeing this.

And she had normal here.

And another bilateral cleft lip.

Here it is rendered.

Okay? You don't even see much of a palate here, frankly.

And this is bilateral cleft lip.

And I think this is better than just looking at the coronal view we usually get.

And here's the bilateral cleft lip here, the premaxillary protrusion.

This is a 3D rendered image, and I think it's pretty clear this is quite abnormal.

But the fact that this comes back into this does not really come back into the palate, I think is the point of this one.

And here's what this image without the 3D reconstructions and the rendering, this is what this looked like.

Alright? And they missed it.

This is the one she told me they missed.

So very useful for improving the quality.

A pitfall is the nostrils.

Okay? You look through these and it looks like there's a cleft.

We've seen that sometimes the nasolabial folds will be very prominent and that will fool you.

There's no premaxillary protrusion micrognathia.

This can be difficult sometimes if you draw a line along the forehead, it should be along the front of the chin as well.

And this is one that's clearly recessed in Treacher Collins syndrome.

And these are 3D rendered surface shaded images that demonstrate this frontal bossing.

Okay? This is normal, okay?

And this is the frontal bossing.

You notice how deeply inset the nose is right there.

This is a good use for this surface rendering, showing a scaphocephaly and the eyes very close to each other.

Hypertelorism.

Okay? There's not really hypotelorism in these.

It's the two eyes are right next to each other.

Spine and Skeletal Dysplasias

Spinal anatomy, okay?

I'm running out time here.

But this is she uses this to look for neural tube defects.

Okay? I think you can see these pretty well on the standard images.

But segmentation anomalies are much better seen on these surface reconstructed images than on regular images.

She doesn't like the term hemivertebrae, she told me to say spinal dysraphism defects and or spinal dysplasia type defects.

And that's 'cause there's so much variability in these.

But you can see this clearly very well, certainly easier than on this image where this is the abnormal set of vertebrae here, and it's subtle here where there's more obvious there.

Okay? And skeletal dysplasias, this is osteogenesis imperfecta.

And I think you can see this bone doesn't look that all that abnormal.

And yet on these rendered images, I think you can see that there's some bending of the bone here and there's some irregularity of the bones here.

Oh, that was a clubfoot.

But a lot of times the best way to see this is on the ribs.

You do a 3D reconstruction of the ribs and you can see the compression fractures very well.

Placenta Accreta

The last thing that they're using this for is accreta.

And this is a problem that comes up all the time.

And you can see on the standard sagittal image that there's this bulge in here.

I don't think any of us would have a problem recognizing it.

It is interesting how much worse this looks though on the reconstructed axial and coronal images.

You can see that this little lobe of placenta has extended quite far anteriorly into the myometrium.

Here's a percreta.

And the idea here, I guess, is that this placenta is really extending into the urinary bladder.

There's really no myometrium here at all.

And you can see this little bit of irregularity along here.

Okay? And this is really extending.

This is the coronal view and it's extending all the way to the bladder here.

And you can't even see anything on this particular image that isn't placenta.

And she just wanted to describe what they're using it for.

Conclusion

I think the real useful stuff is the palate in particular, and the face and the nose to some degree, limb defects, and especially in the spine and brain, especially the cavum septum pellucidum, it seems extremely useful.

Now, in my own talk, I had made the comment a few years ago that you could look at in utero behavior of fetuses with these, you could see this one's yawning here.

It's like one of our residents, this is for bonding.

This word here is bonding and entertainment.

Why am I showing you this?

Well, I really, in counseling, I wanna give Dolores a lot of credit because Dolores is offering something to her community that here's kept her in the game of obstetrical ultrasound.

And it isn't strictly diagnosis.

A lot of it is these other things that she can offer to patients, and especially the counseling, to be able to be a radiologist and be the one who's producing these images and telling the patients and showing them what they're seeing is something that is very worthwhile to the perinatologists and to the obstetricians.

And while a lot of us think, well, 3D and rendering and all this isn't necessary to make diagnoses, it probably isn't.

But in terms of offering other intangible qualities and tangible product to the patients and to the clinicians taking care of them, I think it's very useful.

So I really give Dolores a lot of credit for doing this kind of work.

And with that, I made it 85 slides and.