Thyroid Elastography - HD

Introduction to Thyroid Elastography

Hi, my name's Manchi Dge.

I'm from the University of Washington,
and my lecture topic today is thyroid elastography.

History of Palpation

Palpation is an important diagnostic tool,
and it was practiced as early as in the 2,600 BC
by the Egyptian physicians.

It said that it's a business of the physician
to know in the first place things that are similar
and dissimilar and that are perceived
by sight, touch, and hearing.

And this was by hypocrites himself.

What he mentioned was that such swellings as a soft
or free from pain and yield to the finger
are less dangerous than the others that are hard
or palpated as hard and large
and indicate danger of speedy death.

Perhaps he was talking about cancers.

Principles of Palpation

What's a principle of palpation?

When you have a focal mass
and you press on it, if the mass is soft,
it gets deformed by pressure.

But if the mass is hard, then it's not deformed by pressure
and is palpable.

Stress is basically higher on fingers when you're
palpating something overlying a superficial
and a hard lesion and lower on the receptors when their
surrounding softer surrounding tissues as well.

Development of Elastography

The tissue motion studies were done in the 1970s
and early 1980s.

No elasticity term was coined in 1988,
and it was basically using doppler ultrasound
to detect motion in response to external vibration.

Ofer in 1990 developed a quantitative method
for imaging elasticity, and that was called as elastography.

It was based on external compression.

There was cross correlation analysis of the pre
and post compression ALINE pairs
to decide if a lesion underwent deformation or not.

It's mainly developed for breast imaging,
but why use it in the thyroid?

Because thyroid is a very superficial organ, mostly,
a few centimeters deep or less than that.

Also, you have these two structures adjacent to the thyroid,
the carotid artery,
which can help in compression on the thyroid itself.

And at some points it does act as a cause for artifacts
and also the thyroid can be compressed against the
trachea, which acts as a heart area.

Thyroid Nodules: Incidence and Challenges

But what's the problem? The problem is
that you have these thyroid nodules
that are increasing in incidents every day, and that's
because of increasing use of imaging where they detected
as incidentalomas.

If you look at this chart,
if you palpate on palpation, you have an incidence
of about 5% of thyroid nodules detected in autopsy.

The number is higher.

About 60% of patients have nodules on ultrasound.

Almost 50% of the population would have some thyroid nodules
detected, but malignancy rate is pretty low.

Out of those detected, only about 10
to 15% will actually have malignancy.

Fine Needle Aspiration Limitations

Now, fining aspiration is a standard procedure.

It is, of course an invasive procedure
and can lead to inadequate samples in about 10 to 20%
of cases, which itself leads to quite a lot of problems,
which is rebiopsy or even surgery at some times.

First Study on Thyroid Elastography

The first study essentially was done by Andre ic.

In this paper described in radiology
where he had 52 nodules on 31 patients,
they did real-time elastography
and did offline analysis of the ELAs grams.

They also did ex vivo biomechanical testing
of these tissue samples
and found that papillary carcinoma was stiffer than
normal or benign lesions.

They calculated something called a strain index value,
which if greater than four, was a strong indicator
of malignancy with a sensitivity of 96%
and specificity of 82%.

Types of Elastography Techniques

Essentially, there are two types
of techniques described strain, and she wave in strain.

Elastography manual compression is used,
which is a most commonly used technique
for strain elastography.

And then the pros of the strain elastography are
that it's widely available, it gives good results.

Cons, of course, are that you need someone who's trained in
doing strain elastography.

The changes in amplitude
and velocity of compression affect the strain.

You may cause pre-train compression itself
where you press on the lesion
before actually applying compression.

And then some other things like carotid artery pulsation can
interfere with the results.

Histology of the nodule can be misleading as well, but,
and that's true for both ShearWave
and strain elastography when cytopathology results may not correlate
with the imaging itself.

Now in she wave elastography, the shewas are generated
by the machine, so there's less variability.

Unfortunately, there are smaller number of studies
with ShearWave elastography,
and if you have a lesion which is located very superficial
like in the isus, ShearWave can cause artifacts.

Large nodule sizes cannot be assessed
because of less penetrability,
especially greater than three centimeters.

And then cystic components are a problem with she
and strain elastography as well.

Strain Elastography Technique

Now in strain elastography, the technique uses detects local defamation
that is strain under some slight pressure, which is usually
by the hand compression is performed by the transducer
or some physiologic movement like carotid pulsation.

What is produced is a color coded ELAs gram superimposed
on the B mode image.

Now this is an example of the carotid compressing.

As a carotid increases in size,
it compresses the thyroid in the transverse direction
and increases in the AP dimension.

This is a example of color coded images per produced by elastography.

And you can see that the range, they range from soft to hard
with different color coding used from red to blue,
as in this particular case.

In terms of strain elastography, there are two kinds
of elasticity assessment, either visual scoring of color
that is color-based ELA grams with four
or five scale scoring systems, for example, in this case,
five scoring systems were used.

So score one to score five, right from blue to green,
blue to green with increasing amounts of green
as we go towards the softer lesions
and increasing amount of Blu towards heart lesions.

Examples of Strain Elastography Images

This is an example. So here's a normal thyroid
as seen on the BMO ELA grams.

We see no nodules in the thyroid.

And the color ELA gram shows a similar thing
with uniform green within the thyroid,
the trachea appears is blue because it is stiff.

So that's a normal thyroid.

This is a patient who has a lesion, which has a mixture
of colors in it, red, blue, and green.

It is a solid cystic lesion on ultrasound within,
with vascularity within the solid components.

And this would be something which is stiff a lesion perhaps
in the four or the five.

And this was a papillary carcinoma.

This particular lesion has a mixture of both blue
and green in both these cases,
but this upper lesion turned out
to be a follicular carcinoma, a follicular lesion,
and the lower lesion turned out
to be a multinodular equator.

Quantitative Assessment: Strain Ratio

Now qualitative assessment is very difficult
and people have come out with quantitative assessment using
something called a strain ratio calculation,
where two ROIs are placed, one over the target region,
which is the thyroid nodule in this particular case,
and the other over an adjacent reference region,
which in the thyroid is surrounding normal thyroid
as seen in this example, as the strain ratio increases,
that suggests increased likelihood of malignancy.

Strain can also be induced by carotid pulsations.

In this particular technique,
no free hand compression is applied.

It is less operator dependent
because you just rely on the carotid pulsating
and causing compression on the thyroid.

It is also less operator dependent.

And because it gives uniform pulsations
or compressions on the thyroid
with each systole in the systole flow in the carotid.

Now qualitative and semi-quantitative strain ratio computation can be performed in case
of strain induced carotid pulsation elastography as well.

Our Study: Thyroid Stiffness Index

So this is an example.

This is our study where we did thyroid evaluate,
where we calculated the thyroid stiffness index,
which is the strain near the carotid artery divided
by the strain in the thyroid nodule.

And the higher the TSI, as we called it,
indicate EF a lesion.

So these are examples.

This is a lesion you can see on B mode that has both solid
and cystic areas in it.

The TSI was about 17.4
and on histopathology was a nodular goiter.

This was a lesion which is more uniform, more iso coic
with no solid or with no cystic lesions in it
or calcifications in it.

The TSI is a little higher, 26.5.

This was a nodular goiter with interstitial fibrosis lesion
with microcalcifications and hypoechoic appearance.

Solid TSI of 31.4.

This was papillary carcinoma,
and then this was a hypoechoic nodule,
but with a low TSI of 6.95, which was a colloid nodule.

So when you take all of these results together,
normal thyroid and papillary carcinoma were distinguishable
from other lesions.

And the TSI for papillary carcinoma is higher than the TSI
for other lesions and even the normal thyroid
as can be seen here where papillary carcinoma is much
stiffer compared to these other lesions.

Clinical Case Example

This is a typical example.

This was a patient who was being followed up
for two nodules, one lesion in the upper pole,
one lesion in the lower pole.

Patient was being followed up
for growth in these lesions.

And you can see that on the thyroid elastography,
the TSI in the upper pole lesion was 16.8,
but in the lower pole lesion, those 41.15, with
that high stiffness value,
we suggested finding aspiration.

This came back as papillary carcinoma.

Histopathology confirmed the upper pole lesion
as being nodular GoTo
and the lower pole lesion being solid papillary carcinoma.

Performance Studies and Meta-Analyses

So there these are the studies that have been performed,
some of them in elastography.

And you can see that the sensitivity
and specificity is quite high,
reaching almost a hundred percent specificity range of specificity as such,
and then sensitivity in the high eighties as well.

This is a timeline showing the some
meta-analysis that have been done.

This particular one done by chenga shows gaos,
639 nodules evaluated with strain elastography
and sensitivity of 92% and specificity of 90% in 2012.

Then moon did a retrospective study of 703 nodules
with ELA sensitivity of 65% negative predictive value
of 79%.

But when they added the grayscale ultrasound features,
the sensitivity increased
and negative predictive value increased as well.

Same year. This particular group suggested
that ultrasound elastography is not superior
to grayscale ultrasound in predicting malignancy.

They had 237 thyroid nodules that reported lower performance
with elastography in comparison with gray scale ultrasound.

However, in 2013, Azizi
did a prospective study using a fourth grade elasticity
score with almost a thousand nodules,
and found that they had,
they had a positive predictive value of 36%
negative predictive value of 97%, which is the,
this positive predictive was value was higher than
for microcalcifications alone or hypo echogenicity alone.

So that's significant. Last Reza, we looked at meta-analysis
of 34 24 studies with 3,531 nodules
and saw that the sensitivity of elasticity score was 82%.

Specificity was 82% as well,
and found that ultrasound elastography increased the
ultrasound accuracy.

So in terms of strain ratio, these are the studies
that have looked at strain ratio with sensitivity ranging from a hundred percent to about 43% or so
and specificity in the high seventies.

Shear Wave Elastography

Moving on to survey wave imaging of the thyroid,
this is a new technique.

ShearWave technique uses these shewas
that are generated by the transducer that displaces tissues,
use an ultrasound ultra fast scanner with a frame rate
of almost 20 kilohertz that receives the information back
and a map of the she wave is cal created.

So this image over here shows the she wave.

This is at two milliseconds, then at three milliseconds.

That is as it is continuing
to displace tissues laterally and spread laterally.

This is at four milliseconds and five milliseconds,
and you can see those two visible parts of the she wave.

Finally of image
of the she wave speed is calculated and created
and displayed for us to see.

So this is just an, again, this showing
how shewas are calculated from volumetric force creation
to ultrafast imaging
and then image acquisition and processing.

This is what is typically shown a color coded map of the lesion with the color,
color images ranging from blue to red,
typically blue indicating soft red indicating stiff.

So it's a new technique.

There are fewer studies performed and,
but it is a quantitative technique
of assessing the thyroid nodules.

Examples of Shear Wave Elastography

This is an example of a case study,
case courtesy from Dr. Mara from Italy.

And what he shows is there's a lesion in the thyroid nodule with a mean stiffness
of 20 kilo pascals 20.8 kilo pascals.

This was a benign nodule
and basically a colloid nodule on ultrasound.

You can see that it's a mixed solid cystic nodule
with no flu in it.

However, this pa particular patient is again,
a 51-year-old woman.

She has two small nodules in her left lobe.

This is the upper pole nodule
and this is the hummus nodule.

And what you can see is when you look at the stiffness in this particular lesion
in the lower pole nodule, the stiffness was lower,
about 31 to 33 for a mean.

But the is para lesion had high stiffness in it
as seen on the color image.

And then the mean stiffness was about 58 to 72 kilo pascals.

FNA of this low paraty lesion was performed,
and that came back as papillary carcinoma.

Studies on Shear Wave Elastography

So these are some studies performed with she wave imaging.

And again, you can see some of these are meta-analysis,
but the number of nodules and
and the sensitivity
and specificity are all over the place.

I think it's because a new technique
and it's still an evolution.

These are some studies. This particular one,
Lynn from China shows that sensitivity of 84%
and specificity of 88% with a good negative predictive value of almost 99%.

This particular study shows, again, similar things,
good sensitivity and specificity.

If you do find that the stiffness is high,
the positive likelihood ratio was about 7.4 according
to them, and they had a good area under the curve as well.

So here's some examples.

This is patient who has a strain elastography performed.

There's high so the lesion here is solid.

You can see the lesion on with a color, color overlay image.

The TSI in this particular case was 20.2.

This turned out to be a nodular goer.

This particular lesion, large lesion, very irregular,
but the TSI
and the color map shows that it's lower stiffness TSI of 9.9
and still turned out to be a papillary carcinoma.

So there's a disconnect over here.

A lesion with a higher TSI turned out to be benign lesion
with a lower TSI turned out to be pap papillary carcinoma.

So there's some these are some pitfalls
of strain elastography.

Similar thing in she wave elastography as well.

Two lesions, one lesion with a high stiffness value
of 76.7 kilo pascals.

Another lesion in the same patient with stiffness value
of 21 kilo pascals,
but still these two lesions both turned out
to be papillary carcinoma.

So here's a lesion, large lesion.

When you look at the color image,
it does have some high stiff high stiffness areas in it.

Maximum kilo pascal value of 1 26 0.9.

But the fine aspirations showed a follicular lesion
with her total cell changes.

They, we, the cytopathology said,
could not exclude neoplasia.

And so the patient was operated on a hemithyroidectomy done, which showed
that this was a nodular hyperplasia, multinodular goiter
with a background of lymphocytic thyroiditis.

Pitfalls and Artifacts in Elastography

So now to talk about some of the artifacts in Elastography,
some of the artifacts are related
to the nodule characteristics itself
and some to the elastography technique.

In terms of elastography technique,
when you have deeper lesions, you can have lack
of penetration and absence of information.

If you set the color scale incorrectly,
you can always have a neo field artifact in the closest
region, and then the same area,
you'll have a compression artifact as well.

In addition, you can have artifact compression artifact from
carotid artery pulsation.

Some of the nodule characteristic, which can cause artifacts
and that need to be recognized are when you have a cystic
and solid lesion, when you have rim calcified nodules,
which may not transmit the compression.

And then you have a large nodule when the compression may
not be transmitted,
transmitted all the way deep into the lesion.

Examples of Artifacts

So this is an example of neo field artifact
where you have the reddish areas,
or and transmission of all the high stiffness color
through the lesion right up to the trachea.

And this was an ISTE lesion.

You can see that it's very close to the transducer.

Now, if we use a standoff pad
and then took the measurements, again, you can see
that we avoided this neo field artifact.

Sometimes you have compression artifact
and these are seen as finger-like projections, and that's
because of the ma transducer being click placed
too hard on the neck,
causing a compression artifact on the neck.

It's on the thyroid nodule in the neck.

And another way to recognize this is that it extends
through the image suggesting
that it is a compression artifact.

This is an example again showing the effect of compression.

This particular image was acquired without compression.

This particular image was acquired with compression,
and you can see that the strain ratio, the strain changed.

You are the maximum kilo pascal value in the surrounding thyroid lesion is 9.2.

With the le in the lesion about 16 kilo pascals
with compression, it changed to maximum
of 14.5 in the thyroid, normal thyroid
and a maximum of 18.8 in the nodule itself.

So you can have variability in kilo pascal values,
even on shareway with compression when you
use a strain ratio.

Now, the strain ratio did not change
as much without compression 0.5 with compression 0.7,
so it may be better to use a strain ratio even on
ShearWave elastography.

Then absolute strain values.

This is an example of color gain being set incorrectly when
you have over gain and you can see these reddish areas
distributed through the lesion.

Those are artifacts because of over gain
and the color color being set too high.

Similarly on under gain, you'll have areas
that do not show any color,
and in this case, increasing the color gain
on elastography would be helpful.

This is an example of lack of adequate contact
where you can see even on b mode image, there's lack
of information along the lateral aspect of the nodule.

So in this particular area, everything's black
because there's no information with ultrasound.

Similarly, there's no information with elastography as well.

This can be avoided by making sure
that the sheer wave box is within the central part
of the nodule and node off to the side and with good contact
and a good contact is maintained
with the nodule and the surface.

This is an example showing artifact due
to cystic areas in the nodule and b more gram.

You can see that there cystic areas here superficially
with solid areas deeper.

This shows high stiffness values
posterior to the cystic area.

So you can see reddish color beyond the fluid.

And that is due to the sheer wave speed being affected
by the fluid portions of the nodule compared
to the solid portions itself.

This is an example showing lack of penetration
in deeper nodules
where the lesion is located more than three centimeters
and there's lack of information on elastography
because of the deeper location of the nodule.

Calcified nodules can not transmit the ultrasound waves.

And so you can see lack of information here
where you can see the calcifications rim calcifications on
BM mode image causing lack of ShearWave information in
that same region.

Similarly, a calcified rim calcified lesion shows lack
of information on the she wave ELAs gram images as well.

Conclusion: Role of Elastography in Clinical Practice

So in reality,
no surgery will be performed without fine needle aspiration.

If malignant features are seen on ultrasound,
fine needle aspiration would be performed for those lesions.

And it may be more prudent
to exclude benign lesions
with elastography rather than trying to diagnose malignancy.

And the question remains what happens with indeterminate
or non-diagnostic samples,
in which case you might need elastography
and fine aspiration to give complimentary information.

So can elastography help? Yes, it can.

Of course most studies are needed,
more prospective studies are needed.

It is also prudent to use carefully
with common sense in terms of how the images are acquired
and recognizing artifacts as well.

So robust techniques are needed with better guidelines
and more experience,
and that can help elasticy help us
with finding aspiration in the thyroid nodules.

Thank you.