3D Sonography in the Abdomen - SD

Introduction and Disclosures

Hi, I am Franklin Tesler from University
of Alabama at Birmingham.
I'm going to be talking about 3D sonography in the abdomen.
I have two disclosures.
The first is that my department has a research relationship
with Phillips Healthcare,
and the second is that I promise
that no transducers were harmed in the
making of this presentation.
The reason for the second disclosure
will be evident shortly.

Apology for Jargon and Explanation of "Mainstreaming"

I'm going to start off my presentation with an apology,
which is unusual for me,
but I'm apologizing
for all the jargon in the title of my talk.
We see jargon in every facet of our lives these days.
It's impossible to escape it,
but I felt that the use of some jargon in my topic title was
warranted and I'm going to explain that to you.
The first jargon word
or phrase that I used was derived from the noun mainstream,
which dictionary.com defines as the principle
or dominant course.
Of course, if you looked at the title, you saw
that I used the verb form of this mainstreaming, which means
to send into the principle or dominant course.
What does this actually mean for ultrasound?
What's something I think we're all familiar with,
especially if we've been practicing ultrasound for a while
and that's putting a technique into common use.
If you've been practicing ultrasound for as long as I have,
you may recognize the image on the left, which is an old
black on white static image of an aorta
that was taken in the early 1980s.
At that time, static imaging was the mainstream
and real time imaging was not.
But the clip on the right which shows a gallbladder in real
time eventually became mainstream.
And now clearly anybody would consider real-time imaging
as part of mainstream ultrasound.
So you may wonder isn't 3D already there?
Why am I bothering to talk about mainstreaming something
that is already in the mainstream?
Well, certainly if look at 3D ultrasound
and obstetrics, you'd conclude that it is in the mainstream.
This is a Google image search from about two years ago for
3D ultrasound in ob,
and you can see from the response I got,
and this is one of many thousands of pages
and images that clearly 3D ultrasound has made it in
obstetric sonography.
The same is true to a possibly a somewhat lesser extent in
3D ultrasound applications in gynecologic imaging
as in this coronal reconstruction
of a uterus showing very clearly the endometrium
and surrounding myometrium.
In my practice at UAB, we do far less
OB ultrasound than we used to.
We do very little second
and third trimester work, mostly first trimester,
so we don't really use 3D there.
We do less gynecologic imaging than we used to
and we are using 3D there.
So to summarize,
I think 3D really is in the mainstream in both OB
and GYN sonography, but not in the abdomen.

Reasons for Limited Use of 3D in the Abdomen

Why not? I think there are a couple of reasons for this.
First is the potential lack of a killer app in the abdomen.
That is, there's no application that's so obvious
that everybody has to do it, as is true in OB and GYN.
However, there are other reasons as well
and it's those reasons I want to delve into in the remainder
of this presentation.

Transducer Design Issues: Size and Weight

The first issue is design of transducers
and the biggest problem for me has been the size
and weight of the first generation of 3D
transducers in ultrasound.
Here's an image of one of our older 3D transducers,
and you can tell even though there's no scale on this image
that this is a fairly large transducer.
In fact, after working with this for a day, you may think
that you've been lifting weights now less.
You think that I'm being wimpy in saying I don't wanna lift
a transducer like this and use it for a long time.
This is a real ongoing issue for sonographers
who scan all day long
and the SDMS has taken a lead in looking at the risks
of ergonomic related risks including size
of transducers and weight
of transducers on sonographer work-related injuries.

Mechanical Design Limitations

The second issue is the mechanical design
of these transducers
and that's illustrated in this animation here.
What these transducers are is really an incorporation
of a mechanism that sweeps through a volume
and acquires the 3D volume, but it takes time to do it
because it's a mechanical process.
Furthermore, the mechanism
of the transducer is bathed in fluid
and you can see that
because of that the face of the transducer
is not completely rigid as it is with most
of the probes that we're used to using.
And as I activate this clip, if you look carefully,
I was pressing on this actually fairly lightly.
You can see the face of the transducer deforming slightly
with each press of my finger,
and I promised I wasn't harming this transducer.
Again, my second disclosure, it was working just as well
after I finished this clip as it was before.
These transducers, because they contain fluid and
because they contain mechanical mechanisms tend
to be more fragile than other transducers.
And over the years we've dropped several of these.
I won't say that I've, I've ever dropped one,
at least I won't admit to it,
but they have been dropped in our lab
and the transducer housing will crack
and the fluid leaks out
and it's very expensive to have these repaired
Because of the mechanism of these transducers
resolution is limited as well.
Again, it's a mechanical process of sweeping the beam
through a volume and it takes time to do that.
And because it's mechanical, there is some
change in resolution
because it's very, very difficult to create a mechanism
that follows exactly the same path each time.
And it also takes time to do each sweep.
That doesn't mean that it's not possible with
non-moving objects like this kidney to get decent 3D
reconstructions in the clip you see at the lower left,
the reconstructed C plane,
actually the resolution is not bad,
but in this case it was possible to do this
because the kidney wasn't moving during the acquisition.

Parallels with CT Evolution

It's instructive to look at the parallels between CT
and volumetric ultrasound over the years.
This is an old article
as you see it's from A JR October, 1979
by federally etal talking about three actually not 3D,
but more correctly reformatted CT images.
And you can see the acquisition plane on the top axial
and then on the bottom you see two orthogonal coronal
and sagittal planes.
And if you look at those images,
you could see the stair step effect, which shows
that the resolution of those images
was considerably less than that of the acquisition
or axial plane.
Compare that to what we have now.
This is a coronal reconstruction from a CT of the abdomen
and it looks like the scan was acquired in this plane.
Another way to say that is
that this represents near isotropic imaging
where the resolution,
the spatial resolution is the same in all directions.
So really what you're trying to do with any type
of volumetric imaging as shown in this animation is
to get an ideal volumetric data set where each element,
they're called voxels for volume elements
are the same in each direction.
So far that's been impossible to do
with the older style transducers,
but again, that doesn't mean
that in ideal situations we couldn't get good volumetric
images with those probes.
Here, for example, in a 5-year-old
sequence is a 3D reconstruction of the aorta and IVC,
and as I turn this volume around, you can see
that there is a dissection flap in the aorta
and the SMA, this is looking at from the opposite direction.
But generally speaking, all other things being equal,
the older types of probes
had less resolution than we would like.
They were certainly not isotropic.

Enabling Technologies for 3D Ultrasound

So the second jargon word
or phrase that I used in my title is Enabling technology
and my other source of truth wikipedia.com defines that
as an invention or innovation that can be applied
to drive radical change in the capabilities of a user.
The new iPhone four s from Apple, for example,
that you see in the image on the right its enabling
technology is the A five dual core processor
that lets the iPhone do what it can do.

Fully Electronic Probes

The first enabling technology that I'm going
to talk about here is fully electronic
probes or transducers.
Because of their construction being all electronic
with no moving parts, they can be lighter,
they acquire volumes faster
and they produce higher resolution images.
Here's an example of one such probe that we have in our lab
and although it's not quite as small as I would like,
I'd like it to be a little bit smaller than this.
It's certainly a big improvement over the large, bulky,
heavy mechanical probe I showed you earlier
and it has no moving parts in it.
In this animation from Phillips, you can see
that the beam produced by this type of transducer,
which they call MI a matrix array,
is produced electronically and
because it's done electronically,
it can be moved a lot more quickly than a beam
that is swept mechanically.
In order to show you the difference in speed
between these two types of transducers, I immersed both
of them in a water bath
and then did volume acquisitions
with each one using the same volume size
and that's going to be illustrated in the next slide.
Mechanical on the left, electronic on the right,
and when I hit the begin button, watch
to see which one finishes first.
Here goes and you see the electronic one is finished far
earlier than the mechanical one.
It works that much more quickly
because it's electronic and not mechanical.

Post-Processing Advances

The second issue has to do with post-processing.
And again, if you're old enough to remember the days of film
back when we used to read ultrasound on film, you know
that the type of processing I'm talking about was
the film processor.
Those were not fun days in those days.
The sonographer would produce images, take them
to the film processor, wait for them to be produced,
bring them to you to look at,
and if you needed more images taken,
you'd go through the same cycle.
Again. It was not a good way to do ultrasound. It worked.
It was all we had, but definitely left a lot to be desired.
Fortunately, by the mid 1990s,
many labs were moving into the electronic arena, such
as on this early workstation where we were able to see
for the first time both clips
and static images brought to us
as the sonographer acquired them.
And to me this was a revolution In reading ultrasound,
I was able to, as I am now often to have a
near complete report
by the time the sonographer brought the images to me,
to discuss and I would look
and say, discuss the case with them
and decide if any more acquisitions needed to be done
or if I needed to scan myself.
But in many cases I was done
and ready to sign off the final report.
That was a great change for me.
It's again, instructive to look at the effects
of this technology comparing CT versus
Ultrasounded workflow.
As it regards post-processing, the first way
to do it is using a fully dedicated workstation.
We recognized that several years ago at UAB
and set up a dedicated 3D lab,
which is located in our outpatient clinic
and that's staffed by technologists
who spend their entire day reconstructing images,
mostly from ct, some from mr.
Now here's another view of one
of the technologists working on a CT reconstruction.
Now it's certainly possible
that these people could do ultrasound reconstructions
as well, but they tend to come from the CT or MR world
and are not really as familiar with ultrasound
as sonographers and radiologists to do ultrasound are.
So over the past few years in doing 3D ultrasound,
I've been using an independent workstation
that's shown here, and here's a picture from a few years ago
of me working at this workstation producing
multiplanar reformats
and producing volumetric shaded surface display images.
That's what I would use.
The problem is that it would take time to get the
raw data to this workstation.
I'd have to fire it up.
I'd go through the effort of creating the images I want
and then I'd be stuck.
There was no easy way
to get these images back into the patient's study along
with the originally acquired images.
The second way of post-processing volumetric data sets is on
the ultrasound machine itself.
And again, here's the CT parallel.
This is a picture of one
of our CT technologists working at the CT console.
Typically we do mill multiplanar reformatting on all our
scans, usually at least coronal,
sometimes coronal and sagittal.
And those are done automatically by the scanner.
We do them all the time, so it's part of our workflow.
However, if we want to do some non-standard reformats,
we can call the CT technologist who sits at the workstation
and produces, multiplanar reformats
or volumetric images as needed.
The corollary of that in ultrasound is shown here.
That's the ultrasound machine on the left
and at the top you see a shot of the work page
and you'll notice in the middle there's a
button called Q Lab.
Q Lab is Philips software for doing analysis including
slicing and dicing volume data sets
and produ producing shaded surface displays.
If you hit that button, what you see on top is this display
and you can then interact with Q Lab
to get whatever you want.
And this is how it looks in actual practice.
This is me manipulating a volumetric data set
on the machine itself.
Again, similar to the workflow
that we sometimes use in ct, the trouble
with doing it on the machine is that it's time consuming
and relatively inefficient.
That is when you're using the machine
to process a 3D volume, you're not using it
to scan another patient and you're wasting
time and time is money.
However, this workflow can still be used for
what I call protocol scans, where like in ct, you know,
you always wanna do the same type
of post-processing every time
and you could do it efficiently,
but for the type of scan where you have a volume
and you just want to explore it
and figure out what structure is connected to what structure
and figure out what's going on, it really doesn't work well.
It's inefficient.
So the third way of doing this is on the pacs,
and here's an example in the CT world of one
of our PACS workstations with one menu selection,
I can fire up the, post-processing software
and do all the reconstructions I want
and save them as part of the original acquisition.
So the second enabling technology is ultrasound packs
with integrated 3D
and we've had this at UAB for several months.
Now, here's a video of one
of our workstations showing you the diagnostic monitor on
this side and the 3D monitor
that I'll show you a little bit more of in a moment.
On the other side, this monitor, which is the same
that we use for all our diagnostic reads,
is actually divided into two virtual monitors.
It displays all our clips, our static images,
and it also displays our 3D volumes.
And here's what those look like.
This is the a shot of the left kidney,
just a static image showing some complex
cysts at the lower pole and on the workstation.
It's from a company called IM Morgan, which does,
create is the vendor of our ultrasound packs.
The volumes look like that
and they look a little bit different from the way clips
ordinarily do, and this mimics the way the
volume was acquired.
More importantly, if you look carefully here,
you can see on this workstation some of these cl
acquisitions have a little box next to them
and I'm going to magnify that up for you.
That box has 3D in it,
and if I click on 3D, what happens is
that Q lab opens up on this monitor here on the left,
I can then use Q Lab just
as I would on the ultrasound machine,
but in this case, I'm not wasting time on the machine,
I'm doing it on the PAX workstation
that I would also be using to read my regular studies
and I can do all the manipulations I want
when the time comes to export the images,
I click on this button down here
and I get this dialogue box,
which lets me export whether it's a volume shaded surface
display volume I've created
or a multiplanar reformatted image
or what have you, that gets exported to the PS
and is then visible on the Air Morgan workstation just
as if it had been acquired
and post-process on the ultrasound machine itself.
And this ability to do this post-processing on the
PS workstation has really been a major change for us.
It enables me to use volumes
much more frequently than I could before
because there's a lot less, work to do
to actually get the volume ready to post-process.
And once I've created the post-process images, it's easy
to get 'em back to the packs.
So all I do is do export to packs and save and it's there.

Equipment Availability Challenges

The third issue is a little bit different from the others
and that relates to equipment availability.
And if you look at a timeline in ultrasound technology from
1975 through 2012, you look at real time
sometime in the seventies
and a vaginal ultrasound in the early to mid 1980s,
color Doppler in the late eighties, and then 3D
or volume imaging in the earlier part of the last decade.
There's something similar with all these advances
and I think color Doppler really exemplifies that.
Well, this image that I found online is a shot
of an ad for the first
color Doppler ultrasound machine from
a company called Quantum.
And you can see how they were trying to market this.
They called it angiography,
let you see blood flow for the first time.
It really was a great change in the way we looked at blood
vessels and looked at flow and solid tumors and so on.
The trouble is we, most labs didn't have a lot
of these machines, even when they were available from other
vendors as they ultimately were,
you often ran into a situation where you wanted
to examine somebody with the Color Doppler machine,
but you didn't know where it is.
So you'd yell, where's the color machine?
And typically it would be in another room
or somewhere where you couldn't use it on that patient at
that time and you would just forego doing the color.
In rare cases, you might make the patient wait
or bring the machine from somewhere else,
but it was difficult to do.
But eventually color Doppler ultrasound became ubiquitous
and we got to a point where every machine had color.
We're not there yet with 3D.
We have five inpatient scanning rooms at UAB shown here.
Unfortunately, we only have, despite having lots
of transducers of different types, we only have one
of these 3D transducers and only one machine it works on.
And Murphy's lobbying what it is guaranteed that most
of the time when I wanna do a 3D acquisition on a patient,
the machine or the transducer is in use somewhere else.
And then you have to go through the same thought process
that we used to with color Doppler
and decide, do I wanna wait or do I forego it?
And most often for practical,
practical purposes, we forego it.
So what's the solution to that?
Unfortunately, it's not as simple as an enabling technology.
The solution is really this,
and I'm going to paraphrase a line from one
of my favorite James Bond movies, Goldfinger,
where James Bond is lying bound to a table
and about to be, sliced in half by a laser beam
and or a Goldfinger.
Um, it responds to James Bond question,
James Bond's question, do you expect me to talk?
Says no, Mr. Bond, in this case I expect you to buy.
And really the only solution to this is going
to be acquiring additional equipment that can do, 3D
with the solid state matrix array transducers.
Uh, and that eventually will happen. We're not there yet.
Like many labs, we have a refresh cycle
that runs every few years.
So we may upgrade a machine or two every few years,
but eventually we will get there.

Workflow Strategies for 3D Ultrasound in the Meantime

So what do you do in the meantime?
Well, there are three types of workflow you can use in 3D.
One is to scan in conventional 2D to get what I call the lay
of the land and see what's going on with the patient.
And you can tell fairly quickly if it's going to be normal
or near normal and in most cases you could get by
with this 2D acquisition.
If you have a lot of 3D equipment, you can use 3D,
but you don't really need it.
For many cases.
If it's abnormal but straightforward, say gallstones
or hydronephrosis, again, in most cases, 2D plus
or minus 3D will suffice.
But if it's an abnormal complex case
where you really wanna acquire a volume
so you can sit down later
and interrogate that volume to decide what's going on,
3D is very helpful along with the basic 2D.
Again, that doesn't mean that you can't,
if you have the available equipment, use 2D in if
or rather 3D in other ways.
For example, you can acquire a 3D volume of a kidney
as shown here and slice it very quickly
to duplicate the images that you get
during a conventional 2D acquisition and do it more quickly.
You can do the same thing
as shown here with the gallbladder.
Again, taking the single acquisition
and then slicing it in, multiple parallel planes
or even as in the second use case I just showed
with a relatively straightforward abnormal case
of a liver cyst.
Just take the volume and then slice it multiple ways.
But I think that 3D, especially in the area
where it's limited in availability,
is best reserved for more complex cases as
this was actually part of a much more complex case,
but showing some clot in the IVC in two planes
in any example where you want to again, acquire a volume
and then interrogated later offline on the PACS 3D,
I think is the prime use.
Again, this is most useful in instances where you don't,
3D isn't ubiquitous in your department.
So the use case we're employing most often
is this third example, abnormal complex cases
where you want to acquire a volume
and then interact with it later where you do 2D
and 3D imaging.
The other two use cases normal or near normal
or abnormal, straightforward certainly can be,
useful.
It's can be useful to use 3D in those,
but if the equipment isn't as available as you'd like,
then I would reserve it as we are in our lab
for the more abnormal complex cases over time.
However, the new technology
that enables the further application
of 3D in the abdomen just as it did with color
and just with as with endo vaginal sonography in real time
before it'll become more available
and eventually will be available on all
or most machines in our lab and many other labs.
And that will drive further applications
of 3D sonography in the abdomen.

Conclusion

Thanks very much for your attention.