Buzzwords: Contemporary Ultrasound Technology and the Future - SD

Introduction

I am Fred Crem, call professor of radiology
and director of the program
for medical ultrasound at Wake Forest University.

In this discussion, we are going to look at some
of the more recent developments in sonographic technology,
the things that have become available
to us in the last few years,
and we'll also take a brief look into the future
and where it looks like we're going
with diagnostic ultrasound and with therapeutic ultrasound.

Title of this lecture is Buzzwords,
contemporary Ultrasound Technology,
and a look into the Future.

We will look at some of the technological advancements
that have come along in diagnostic sonography in the last
few years, and we will be looking inside
and seeing what's going on in here.

Now these are the topics
that we will be discussing, coded excitation,
harmonic imaging, compound imaging, panoramic imaging,
volume imaging, and elastography.

Coded Excitation

This is an image of the carotid bifurcation.
We can see that inside the vessel where there is blood,
typically we do not image that
and it commonly appears black as it does here
because the echoes from blood are so weak
because the blood cells are so small.
But in this case, we are imaging the blood
and we are seeing it flowing.
It is not a doppler technique.
You can see that we're not looking at color doppler.
We are looking at a gray scale presentation of the blood
and we can see its flow.

The manufacturer that introduced this
years ago called it B flow And
that implies brightness flow
or B mode ultrasound, which we commonly call gray scale.
Ultrasound now is being used to image the blood as well
as the surrounding tissues.

But this is an example of the application
of pulse coating, which is something that has been applied
to radar for decades,
but more recently to ultrasound imaging.
B flow is just one example of how pulse coding can
improve our imaging.
It can produce improvements in imaging in many ways,
but let's look at what pulse coding is.

Typically in the conventional approach
to sonographic imaging, we will drive the transducer
with a single cycle of voltage that will have a frequency
that we want the outgoing pulse to have.
But in pulse coating, we will drive the transducer
with a more complicated voltage arrangement,
which is actually a sequence of voltages.

And this is an example of a pulse code sequence
where we have one cycle of voltage followed by
none and followed again by none,
and then followed by another cycle
and another cycle to describe this mathematically,
this would be called a sequence 1 0 0 1 1.
It's just one example of an unlimited number of sequences
that you could design.

Here's another example
where we have a 1, 1, 0 1
and minus one minus one, which simply means that these two
cycles are inverted.

Let's use this sequence to describe
how this works and why it gives us an advantage.
We're going to use the sequence of 1, 1, 1 and minus one.
That will be our coded pulse,
and the returning echoes will have the same characteristic
as the pulse that goes out into the tissue.
A returning echo will look like this then in the voltage
form as it comes back into the electronics,
and it will go through a matched filter in the electronics,
which has the same arrangement.
1, 1 1 minus one.

The result of that echo passing through this matched filter
as shown on the right,
and we'll see why that is advantageous.
But first we'll see how it is that this is the result
as this voltage that now represents that echo passes
through the matched filter.
There is a multiplication
and addition process that is carried out.
You can see that this is
how far the voltage has come into the matched
filter at this point.
And we have a minus one in the echo
and we have a plus one in the matched filter.
We multiply those two
and we get a minus one,
which is indicated here a little bit later.
It has come to this point
and now we multiply one times one, which is one
minus one times one is minus one.
We add those two results, which is zero,
and that's the result here.

Now we have a one times one, one times one minus one,
and so that's one and that's two
minus one gives us a plus one, which is shown here.
And now we have
one times one again again,
and a minus one times minus one is also a one.
So you add all those and you get four,
and that gives us this large one here,
which represents four.
Well, if we continue through, this is what we get,
when the process is complete.

Now there's good news and bad news here.
The good news is that we have a very strong result.
In fact, four times
what we had would have gotten had we taken the conventional
approach with just, one cycle of driving voltage,
we are getting a result that's four times
that which means we have a much more sensitive
system operating for us.

The bad news is that we have these other things hanging on
here, which have lengthened the pulse
and that's not good for resolution.
So we've improved sensitivity,
but we've degraded resolution.

Well, there's a solution to the problem of the resolution,
degradation also.
And that is that if we have a second pulse
that inverts the last two numbers as you see there,
also in the matched filter, of course, then
what will happen is that second,
echo from the second pulse from the same structure comes
through, the result is shown on the right
And we then add those two results from the two pulses
and the two echoes that have come from those two pulses.
When you add those together, you get a wonderful result.
Now it is eight
and the extra things
that were hanging on there are now gone.
So we have improved the strength of the result
by a large amount and still have good resolution.

This is called a gole code pair.
Of course, it takes two pulses then
to generate a scan line rather than one, so
that would slow down the frame rate some,
but we've improved the sensitivity dramatically.

This shows an image of blood flow within the tissues
that, as a result of that improved sensitivity
from pulse coating.
Another example and a third one.

Harmonic Imaging

Moving on to the second topic, and that is harmonics.

In harmonics, what we do is we send out
a pulse into the body of an appropriate frequency
for whatever depth it is that we're imaging,
but we will look for echoes
of twice that frequency.

Now, there are two questions to answer here.
One is why would we even get echoes of twice the frequency?
And secondly, why would that image
be better than just looking at the fundamental frequency
image that we have been using for decades?

The answer to the first question, why do we even get
harmonic frequencies?
And the answer to that is
because as a pulse travels out
through tissue, it changes shape.
This is because the speed
of the pulse depends on pressure.
The pulse itself is a pressure variation.
The higher pressure portions travel faster than the lower
pressure portions, and so the
pulse changes shape as it travels.
This also means that the frequency content of it will change
if this represents the frequency of that pulse,
we might call it a five megahertz pulse,
but it contains frequencies around
that which we call the bandwidth, which is indicated
by the width of, the bottom here.

As that pulse travels
and changes shape Harmonics are introduced
and the harmonics are multiples
of the fundamental frequency.
So this would be the second harmonic.
It's actually the first one,
but we call it the second harmonic
because it is two times the frequency of the fundamental,
and this is the third harmonic, which is three times
the frequency of the fundamental.
And so we, and we would get others also,
and we would have even harmonics and odd harmonics.
The first even harmonic, which we call the second harmonic,
is going to be the strongest one.
So that would be the one that we would want to work with.
And so as that pulse travels, the harmonics gets stronger.

What we will do then as the echo stream comes back
is filter out the echoes of the fundamental frequency
that we would've normally used.
And in our reception system, we will have a filter
that will accept the frequencies of the second harmonic
and the bandwidth around it,
and we will image then the echoes of the second harmonic.

And why would that be advantageous?
Well, first of all, the beam that,
is indicated here, which is the harmonic beam,
is not even generated for a while
until the pulse travel some distance
and the shape has changed enough to develop harmonics
that are strong enough to be detected.
Secondly,
that beam will be much narrower than the fundamental beam,
and that will be good for resolution.
The fact that the beam has not formed
for a while will be good for eliminating some things
that degrade the image.

So in conventional imaging, we would be using this beam,
the blue beam, and in harmonic imaging we would be using the
red beam, which you can see,
will give us some improvement in the imaging.
It's kind of like putting the transducer inside the patient
and avoiding some of the problems that would happen out here
in the initial tissue layers
and also having a narrower beam than we
otherwise would in cardiac imaging, for example.
This can,
reduce
or eliminate the clutter that results from reverberations,
also the distortion that results from refraction.

And here are some examples of the improvement.
Here we are imaging the gallbladder
with fundamental imaging.
When we go to harmonic, we see the improvement.
We have a reduction of the noise in the gallbladder.
We have an improvement in the imaging of the shadow
that has come from the stone.
Here we are imaging a stent.
Not very well, it's hard to tell where it is,
but with the harmonics, much improved
Noise in this aneurysm reduced significantly.

Panoramic Imaging

Moving on to the next one here, we have a larger view
of the anatomy than we normally would.
In fact, it is larger than the field of view
of the transducer that produced it.
So how did we get the increase in the field of view?

This is called panoramic imaging,
and what's done is
to physically mechanically slide the transducer
along the surface of the patient along the scan plane.
This will yield a panoramic image such as the one
that we see here of the liver.
This is an obstetric example, and another one

As the transducer is physically sliding along the surface
of the patient, along the scan plane,
new information is being added to the old,
and the instrument will be doing a spatial correlation
of the overlapping regions of old
and new frames so that the new frames are located correctly
relative to the old ones.

For example, here, a slight, motion to the right
would introduce this new scan line here.
How does the instrument know how to place that with respect
to the old information?
Well, there's overlap.
The vertical scan line number three in the new frame
corresponds to vertical scan line number
two in the old frame.
So it would overlap three with two here,
which would cause number one to be added here next
to the old number one.
So this is a spatial correlation of the overlapping regions.

This is an example that includes power doppler to show
how the,
almost the entire arm has been included in this image.
And this shows the generation of an image its breast.
In this case, this represents the overlapping region
that is correlating to add the new information to the old
to develop the image,
which will be much larger than what
that transducer can produce at any particular time.
And there is the resulting panoramic image.

Compound Imaging

Moving on now to compound imaging
or spatial compounding, traditionally
in sonography we send out a pulse,
we get a stream of echoes.
We write a scan line by just showing that stream
of echoes visually on the display.
It's the pulse echo principle that we use
in sonographic imaging.

We keep sending out pulses.
In this case, they all go in the same direction,
namely vertical from different starting points across
the surface of the patient and
therefore across the top of the image,
this yields vertical parallel scan lines
and would result in a rectangular image composed
of vertical parallel scan lines.
These scan lines would be independent.
They would not interact with each other or cross each other.
That's the conventional approach to sonographic imaging.

But if we, instead of
having independent scan lines, if we allow pulses
to go out in different directions and
therefore the scan lines will overlap,
we will then be doing spatial compounding.

Let's see how that works
and why that would give us an improvement in the image.

In a linear array, we have a couple of hundred elements.
We will energize them in groups, small groups,
to produce each individual scan line.
But what we will do in compound imaging,
rather than having each group sending a pulse straight down,
we will send out more than one pulse from each group.
It will be an odd number.
For example, here we would use phasing
as in a phased array.
We would use phasing to send that pulse out in
that direction and then this direction and that one
and that one and that one.
So in this case, we sent out five pulses in different
directions from that same group.
Then we would shift over,
do the same thing with the next group.
This will result in pulses and
therefore scan lines going outta many directions from many
different, origin locations and a lot of overlapping.

So each location in the anatomy will be interrogated
by more than one pulse.
The result at each pixel location will be simply the average
of the echoes that have come back from that location,
from the multiple pulses that interrogated it,
and the result is shown here in the lower right.

Two things will happen, at least with this.
One is that speckle noise will be reduced
because it's a random process,
and so averaging reduces that,
but brings up the echo information from the structures,
relative to the noise.
Secondly,
and this is an example of the reduction
of the speckle noise.
When we do the compounding on the lower right,
also specular reflectors are smooth things
that are only imaged well when the beam is parallel to them.
Would in the conventional approach only be imaged?
Well, if they were horizontal,
but since pulses go out in different directions in this mode
of operation, smooth surfaces that aren't necessarily
horizontal will have improved imaging.

Again, we have the conventional approach in this example
and then the compounded
image as shown.

Volume Imaging

Moving on to volume imaging
or 3D imaging
or as it is sometimes called
four D imaging when it is live or real time.

Fundamentally, it's not different from 2D imaging in the
sense that in fact, we do acquire the information
in 2D, the 3D volume of information that is acquired
and stored in the memory
of the instrument is acquired in 2D form.
What we do is simply get several 2D images
that back to back are parallel to each other,
and they will form a 3D volume
of information just like the slices
form a loaf of bread.
So our 3D volume is actually acquired with many two Ds.
So fundamentally in the acquisition process,
it's not different than what we've been doing for years.

However, the way that we can now present that 3D volume
of information on the display
is much different than the 2D presentations
that we have been limited to up to now.
We can have surface renderings,
we can have multiplanar renderings,
we can have volume renderings,
and we can have traditional 2D presentations,
but we can choose whatever orientation of the 2D plane in
that 3D volume that we want, many
of which would've never been accessible
to conventional 2D imaging.
For example, the coronal plane.

So examples of surface rendering.
They're the ones that are most popular
and that we see most often
and are, very dramatic, especially in obstetric imaging.
But here's an example with gallstones
and an example with, a fetus.
And another one, and we can
do this in real time just like we can do it,
2D in real time.
And this is an example with the heart.
And this is an example with, fetus
going for a swim.

In addition to the surface presentation,
we can have the Multiplanar 2D presentation
where we're seeing two dimensions,
but orthogonal planes simultaneously.

Volume rendering would include everything that is in the,
in the volume that has been acquired
and stored somewhat like a chest film in x-ray.
We can also choose to show just certain echo strengths.
In this case, the strongest echoes, which
of course come from the bones
and we have what's called, here in x-ray mode kind of,
kind of looks like a radiograph in fact,
but that is echo information, the strongest ones,
and we can do that in 3D.
Also Present it in 3D,
and we get the feeling of the 3D by rotating

advantages of volume imaging.
We can, as we've already seen, simultaneously view
several planes.
We can get image orientations in 2D
that we would've not been able
to get in conventional 2D imaging.
We can, orient the 2D plane to have continuity
of curve structures and vessels.
We can rotate to enhance our perception
of complicated anatomy.
And, the 3D presentation of it.
We can improve volume measurements
and interventional guidance.

Progression of Ultrasound Imaging

Well over the many years that,
I've been involved in diagnostic ultrasound,
we have progressed from 1D imaging, which
is a mode
and m mode, one dimension in that,
although we always use two dimensional displays in both
of these cases, the vertical dimension is depth
as it always is, but the horizontal dimension
and a mode is the amplitude of the echo.
And in m mode it is time.
And both of these modes use a stationary beam.

Then we went to two dimensional anatomic imaging
that was black and white
because we didn't have the ability in our memories
to store various echo strengths and
therefore to be able to show gray levels.
And these were static images.
We could only generate an image over a few seconds
because it required
mechanically moving the transducer over the surface
of the patient to acquire all the echo information.

Then we moved to static gray scale when
memories came along that had the ability to store
many values of echo strength at each location rather than
just two, namely black or white.
And then we went to real time note that,
the original A mode and m mode were real time,
but they were limited to one spatial dimension.
Now we have two spatial dimensions, a 2D image,
and we can do it in real time.

And then we had the ability to do 3D imaging,
but it was static, and then we were able
to do it in real time, and that was
and is live 3D or four D.

So we went from a mode
and m mode to 2D
by stable or black
and white to static gray
scale imaging of triplets in this example
to grayscale real time
to static 3D
and then to real time 3D.
It's been a wonderful progression over many decades.

Future Developments in Ultrasound

So we've looked at some of our contemporary technology
that is relatively new,
has come along in just the last few years.
Pulse coating, harmonic panoramic compound
and volume imaging.

Where are we going in the future?
Well, we've already been down this road a long way,
but we're continuing down the road
that computers have gone down,
and that is that computers have gotten smaller and faster.
That's generally true of electronics, so it's certainly true
of our sonographic instruments.

Also, as computers went from being very large
and requiring huge rooms to becoming something
that could be put on a desk
or carried around like a briefcase
or even put in a pocket.
So has happened to ultrasound.
Diagnostic ultrasound systems haven't required large rooms,
but they have been large devices
that weigh 300 pounds or so.
But we have gone from that to those that can be carried
by hand, which have been commercially available
for a few years now, having the advantage of,
of course, portability
and price smaller generally is cheaper,
and although the quality of the smaller units
is normally not as good as the larger units, the quality
of them compares
to the larger units from just a few years ago.
So it just tends to lag the ability
of the larger units by a few years.

And this is an example of one
that is in fact a laptop computer with the
electronics unique to sonography being
placed in this small box.
And that's an image from that.
And then we have smaller units.
There is an image from that one, including power doppler.
And then we have the pocket units that are available
just in the last, couple of years now commercially.

So where are we going in the future?
Well, miniaturization will continue
and that will make a ultrasound available to more
and more, physicians and more
and more medical environments.
Point of care ultrasound is what it's being dubbed now,
but other things, are in process
and will be happening in the future as listed here.

High intensity focused ultrasound is a therapeutic
application, and therapy
actually predates ultrasound imaging,
but it's had a revival in the last few years with a lot
of research going on where ultrasound,
high intensity ultrasound is being used
to treat disease.
And these are examples that are available today.

Microbubble therapy, the ability to deliver
drugs and to do gene therapy by tagging things to bubbles
and then by ultrasound locally
breaking those bubbles at the site where we want them.
The tagged agents to
act

tissue characterization has been discussed
and worked on for decades, continues
to be investigated.
That is the ability to do quantitative things, for example,
here to measure the attenuation in tissue
and determine that one tissue is different from another
or maybe normal versus abnormal based on some
measurable acoustic quantity.
In this case attenuation.
But there are other things that can, be measured such
as sound speed and scattering strength and so on.

Elastography

Elastography is an emerging area now
that is commercially available.
It, corresponds to the physical exam, the ability
to ultrasonically image the hardness
of tissues and to show that a particular region
of tissue is harder than that, that is around it
corresponding to the physical exam.
The gram is shown here in the center.
The actual anatomy is shown on the right
and the conventional sonogram shown on the left.

Elastography requires a compression of the tissue.
It can be simply a physical compression by pushing
with the transducer on the surface, the skin,
or it can be generated in other ways.
It could be generated with a low frequency driving pressure
or maybe a high, intensity, short pulse
of sound,
but somehow the tissue has to be compressed.
And then there is a correlation process
with the uncompressed
and compressed image to show tissue motion.
And in fact, relative tissue motion where softer tissue
compresses and moves relative to harder tissue
that's not compressing well.

The harder tissue is shown in blue on the right various ways
that this can be presented in color.
And this, shows a low frequency
driving, compression
and how the elastography
image is built up.
Blue being hard

here, we have a lesion in the breast,
could be a cyst.
It is, hypo coic.
In fact, there's a little bit of enhancement beyond it,
which would be consistent with a cyst,
but we know that,
carcinomas sometimes can look like this, in the breast.
And the ELAs agram indicates that in fact, this is hard,
which would not be consistent with it being a cyst.

And here a prostate tumor, is it a tumor?
Well, it's hypoechoic in the conventional ultrasound image,
but here in the gram it's clear that it is hard.

Contrast Agents

Contrast agents in the case of ultrasound are bubbles.
Here we have a liver mass
that, is imaged.
It's not imaged real well, but it's least suggested here.
Contrast agents are injected ven.
They are bubbles that will produce echoes.
And in this image in the lower right,
we can see the improvement in the imaging of that mass
because the bubbles have penetrated the normal liver
better than the mass.

This is approved in the United States, for cardiac
applications, but not for other applications as compared to,
other countries in the world
where it's being used in many ways, investigational,
in the US except
for the cardiac applications which are approved

as, instruments get faster in processing
and in acquiring, the echo information,
we are progressing to automation.
There are some things that are automatically done,
optimization of the instrument of the,
image on some instruments, for example.

Well, these are the things that we have,
discussed
that have come along in the last few years
and also, some look into the future.
And our guess is about where we are going.

Ultrasound has been an exciting field for a long time.
It's been a dynamic field. It has not become static.
It continues to progress and will, so in the future.