Clinical Elastography & Elasticity Estimation - SD

Introduction

Hi, my name is Brian Garra.

I am a professor of radiology at the University of Vermont College of Medicine and director of ultrasound at Fletcher Allen Healthcare, as well as director of the research programs in the Department of Radiology at the University of Vermont.

I'm going to be speaking to you today about clinical elastic and elasticity imaging and the variations and applications it can be used for.

Today I am going to talk to you about clinical elasticy and elasticity estimation using both ultrasound and some other modalities.

Basis in Tissue Palpation

Elasticy and elasticity estimation have their basis in the ancient technique of tissue palpation, which is very ancient.

It was practiced by the ancient Egyptians as early as 2,600 before the Christian era.

Organs that are commonly palpated include the liver, spleen, thyroid, neck, breast, prostate, aorta, scrotum, and extremities.

The physics of palpation have been recently reviewed, at least briefly by Timothy Hall in 2003.

To be palpable, it's easy to think of it as an optic simply must be harder than the tissue surrounding it.

Otherwise you won't be able to feel it.

Echo Palpation Technique

This is a technique that I use to qualitatively estimate the hardness of tissue.

I call it echo palpation.

And, what you're seeing is I'm, having a gel standoff on the patient's skin.

This is a lump in the patient's groin and I slide the paperclip over the palpable lesion.

You can feel the edges of the palpable lesion through the paperclip, but you can also get an idea of how hard or soft the lesion is.

If the lesion squishes like a sponge, then it's soft, such as this lesion here, or if it moves as a unit, then you would say it's a hard lesion.

So this is a good example of a technique I use both on soft tissue masses and in the breast.

Overview of Elastography

Elastography is the imaging of tissue hardness rather than using standard ultrasound, it's more sophisticated than sono palpation or echo palpation, which I just showed you.

Ultrasound elastography is an outgrowth of tissue motion tracking techniques, which were originally presented, as I remember back in the early 1980s.

It began with m mode tissue motion tracking, and then there were some doppler tissue motion tracking algorithms.

And then even later than that, there were speckle tracking, algorithms.

These methods all sort of coalesced into two types of, elasticity imaging.

One called sono elasticity, which has been primarily developed and promulgated at the University of Rochester.

And static ultrasound elastography, which was, developed in 19 90 19 91 by Jonathan Fer and his associates at the University of Texas.

Other Modalities for Elastography

MRR can also be used to perform elastography and in fact, any method that can image tissue and image changes in tissue position can be used.

This is an example of an MR ELAs Agram.

The image on the left hand side, represents, a displacement image showing, waves of, compression or shear waves traveling through tissue.

And on the right hand side, you can see an image of the reconstructed lesion based on the, waves that were, measured on the, left hand side.

So MR is a very good quantitative method, and you can get a rough estimate of what we call the sheer modulus, which has a direct proportional relationship to the Young's modulus or compression elasticity, parameter for tissue.

Using this technique for Mr.

Other methods that can be used include CT elastography.

I have published along with some other folks at the University of Vermont.

A basic CT elastography technique that was used in the lung optical coherence tomography can be used.

There's some variance of elasticity called po, including poissons ratio imaging, which is basically a measure of the compression or axial ela, strain versus the lateral strain, which we'll be talking about a little later.

Non-imaging quantitative methods are also something we'll be talking about.

We generally create images and try to estimate elasticity from the images, but there are some non quantit non-imaging devices.

Commercial Availability

Elasticity imaging has, grown by leaps and bounds from a purely research technique just a few years ago to now, reaching clinical relevance with at least four major ultrasound manufacturers now, either having the product on the market, or planning to.

You can see on this slide the, in the various companies that have been involved, the largest company being Siemens with two systems, having elastography on the market.

Ultrasonics doesn't have their system on the market yet, but they have, now licensed the technology and are planning to, release it to the market very soon.

Hitachi and Meison are other companies that are already on the market with products.

Organs Amenable to Elastography

Let's talk about what organs can be examined by elastography.

It's usually best for a regular ELAs agram if you can apply pressure, because that's how we create elastin agram.

We an ELAs agram, we, take an image, then we press slightly and take a second image, and we compare the two images to, get an estimate of displacement.

And for the displacement, we get a strain, value.

The organs that are most amenable to, to compression are superficial organs such as the breast, scrotum, thyroid, and any subcutaneous tissue deep organs, but are, are also accessible.

Those who practice ultrasound know that you can compress the uterus, ovaries, prostate, and kidneys by external pressure.

It just takes a little more pressure.

Organs that are subject to physiological pressure such as the lungs, heart mediastinum can also be imaged because that physiological pressure caused by pulsating arteries or respiration can be used to create an ELAs gram and any organ can be used if you use radiation force methods, which is something we'll be discussing briefly later as well.

This is just an example I have created showing how I can compress the pancreas, which is shown on the left hand slide and the right kidney, which is shown on the right hand part of the slide, showing that the kidney both dis displaces and flattens out slightly in response to the pressure that is more than sufficient to create an ELAs gram or a strain image of the kidney and pancreas.

So the message behind all of this is that any organ can be examined using current ELA osteo methods.

There are different problems that you run into with various organs due to motion out of plane and other causes of noise, but basically the technique can be used anywhere.

Specific Applications Overview

Now, let's move to some specific applications, just a listing of applications here on the next two slides are palpable masses such as cancers, nodules, nodes, things like that.

Diffuse organ diseases also produce changes in stiffness of tissue and are amenable to elasticity imaging.

But the difficulty of that is that ELAs are subjective and they're relative images, and you may not see anything if the whole organ is uniformly increased in stiffness.

There are ways around that that we will discuss.

Muscle contractility is another area that's been studied fairly extensively and muscles can be imaged readily as well as can be the myocardium and blood vessel walls.

Additional applications include evaluation of atheros plaque for soft areas in the plaque that might represent vulnerable plaque that could result in a vessel thrombosis.

Venous thrombosis can be evaluated.

The age of thrombi change, there is stiffness and, can be imaged.

The vessel walls may be imaged as well.

Monitoring of ablation therapy is a very opportune method because ablations always produce much markedly increased stiffness of tissue.

And, since imaging modalities have difficulty imaging the ablated tissue reliably, elasticity imaging may be the solution to this problem.

And finally, monitoring of interstitial fluid movements such as in various types of edema.

So now let's review some of the specific applications.

Breast Mass Evaluation

Breast mass evaluation was the very first application that achieved some popularity using elastic elasticity imaging, superficial their superficial lesions.

So they're easy to compress.

Breast masses are known to be hard when they're cancerous versus the softer, benign lesions, and it's a disease of both medical and medicinal and, political importance.

The initial work was performed by this system.

You see in the picture here, with the patient in the erect position, you can see there's a mammographic machine on the right hand side of the image with a transducer clamped to a compression device mounted to the mammography machine.

Our early ELAs grams were performed in the reposition, so we could do a cranial coddle mammogram and do an exact correlation with the mammogram.

This is an example of after we had moved to supine positioning, we mounted a bar on the, on a table and put the positioning device there, and we did compression against the chest wall, which turns out to be, a much better solution.

The problem with the erect positioning is that you can't get close enough to the chest wall to image many of the lesions, but supine positioning allows all image, all lesions to be imaged.

And we've since moved from using a computerized or computer controlled stepper motor to provide the compression to handheld compression, which also works quite well.

Here's handheld elastography on a chunk of stake and then also on a real patient showing the typical technique where you rest, part of the forearm on the patient and apply a very, very mild or gentle compression to the tissue to perform the ELAs gram.

Typically, the amount of pressure applied is almost imperceptible to the patient.

We only need less than 1% compression of the tissue, so the images actually don't look hardly any different from one another, the pre and post compression images and the patient hardly can feel it, so it's a very, very gentle procedure.

This is an example of a real time ELAs gram performed handheld courtesy of Tim Hall, and you can see the lesion being compressed very, very gently on the sonogram on the left hand side and the ELAs agram resulting from this compression.

On the right hand side, this is a benign fibroadenoma.

This is a more recent ELAs agram of a in intraductal invasive intraductal carcinoma.

You can see the irregular sizable dark mass in the left hand image shown right here and the cancer on the sonogram right here.

Now, this cancer is not a real diagnostic issue.

It's pretty obvious from the sonogram that it's a cancer, but there are other situations where the ELAs Graham can provide significant increased specificity Since our original work in our original publication in 1997 where we only had about 40 cases, we have many more now.

And, this shows, a plot of the difference in size between the ELAs gram lesion and the sonogram lesion versus the hardness score on a five point scale.

And what you can see is that there's a fair amount of overlap in lesions, but there is approximately 15% of the lesions that are separable using elastography.

And these are benign lesions that we're talking about.

So when we use elastography, we're often talking about diagnosing benign lesions and avoiding biopsies.

All the lesions on this plot of course, were biopsied because they were considered suspicious by sono sonography, and approximately 15% of them could have been eliminated using elastography according to the plot.

Here, the yellow line represents the line of discrimination between, benign and malignant that you would want to use if you were using elastography to diagnose these benign lesions.

The diamonds are the cancers and the, little pink squares are the benign lesions.

On this slide. A new method to add specificity to the Regular ELAs gram is demonstrated.

This is a method of imaging, the amount of shear strain that it exists at the boundary of a lesion.

Lesions such as a cancer which are anchored the surrounding tissue do not have a free slip boundary and show up as an area of larger shear strain over a broader area.

So if you look at the left hand side of the image, that lesion and that fandom is bound to the adjacent phantom material, and you can see that the areas of shear strain at the boundaries are broad and rather indistinct.

Whereas on the side where there's a slip boundary and the lesions not attached to the surrounding tissue, the areas of shear strain are very thin and sharply demarcated.

We think we can use this feature to add specificity to the diagnosis of breast cancer using elasticity imaging as this feature is quite independent from the overall stiffness of the lesion and the size of the lesion.

So to summarize, breast elastography, will it have a role?

What about lesion detection?

One problem with elasticity imaging is that on any given image, you may see many or several dark areas, which all could be cancers.

So there may be a problem with false positives until recently, real-time display has not been possible, which limited its use as a screening tool.

But perhaps all of these issues can be overcome and it can be used as a detection tool, but not at the current time right now.

Rest elastography primarily is a mass characterization tool.

So the mass is identified using mammography and ultrasound, and then the area is examined using elastography and, increased specificity results hopefully from the gram being added to the sonogram.

The things that are required for characterization, I think you need to use vi video cine loops to look at the entire ELAs gram rather than a few static images.

And real time display is definitely desirable.

Quantification may be helpful in the future to further increase specific specificity.

The problem with breast elastography is that there are many competing modalities.

You have mr you have, infrared imaging.

You have, advances in plain ultrasound and in mammography, tomosynthesis all compet against ultrasound for diagnosing breast cancers.

And it's not clear that elastography will emerge as a dominant player, but since it can be performed at no additional cost as part of the breast ultrasound, which is already a major component, we believe that it will be a useful addition, and will gain a competitive advantage in that way.

Thyroid Nodules

Moving on to thyroid cancer ca, thyroid cancer, at least in my clinic and a lot of others, is the most common reason we do an ultrasound guided ultrasound guided biopsy.

These days, many patients have large numbers of nodules.

And this is a big challenge because the current criteria which calls for biopsying nodules primarily based on their size, has no real firm foundation.

And oftentimes smaller nodules are the cancerous ones and the large ones are benign.

And what do you do with a patient with 20 nodules?

You can't really practically biopsy all of 'em.

You may have to biopsy one or two, and we need a method for which by which we can decide which nodules that are more likely to be cancerous.

Initial elastography results published by, Andre Leic and Radiology in 2005 are promising showing that if you use a strain index greater than four, with a strain index being the strain in the lesion in question versus the strain in adjacent thyroid tissue or muscle, you can achieve reasonable sensitivity and specificity.

And if you combine that with tumor area ratio, you can increase that slightly further.

Here's an example of a thyroid ELAs gram courtesy of Dr.

Leic showing a benign nodule outlined on the left hand image here and the, and the benign nodule on the ELAs gram here showing that it's only minimally more stiff than the adjacent thyroid tissue right here.

The cancer on the other hand, the outlined on the left hand slide here shows up as quite a bit darker or harder than the adjacent thyroid tissue.

And this may be a useful way to determine which nodules to biopsy like thyroid nodules.

Lymph Nodes

Lymph nodes are also a problem.

One of the most common sites of tumor involvement include lymph nodes.

They're present in all area of the bodies and including superficial areas.

And a lot of these superficial nodes are biopsied.

Again, if there's multiple nodes, it'd be nice to know which nodes are most suspicious so we know which ones to biopsy.

An example of benign nodes is shown here.

The lymph nodes on the sonogram are shown on the left hand slide.

And on the right hand slide you can see that the lymph nodes are relatively soft, showing a large strain value, which displays as bright on the image, suggesting that they represent benign nodes.

Their shape is also suggestive of benign lymph nodes.

These nodes, however, are more rounded, which might be a suggestive of malignancy, although that's not a reliable criterion by itself On the ELAs gram, however, you can see there's an area of, of increased stiffness in those in the region of those lymph nodes making them suspicious for cancer.

So the potential for lymph node elastography is quite great.

It's a common technique to biopsy, biopsy lymph nodes and knowing which ones to biopsy would be very useful.

It could possibly can be combined with PET CT and regular B mode ultrasound to create a powerful contra comprehensive tool for metastatic malignancy diagnosis.

Prostate Cancer

Prostate cancer is another area that can be examined, that elastography may be of help in prostate cancer was actually the first type of lesion that was imaged using elastography back in 19 90, 91.

And the current methods of diagnosing prostate cancer remain an elevated prostate specific antigen or PSA plus random biopsies of the prostate gland.

With the biopsies being guided, by ultrasound.

This random technique works pretty well, but is quite invasive, and we don't yet have a reliable imaging tool yet for prostate cancer.

Though high field MR appears to be a promising tool, that hasn't been fully validated yet.

Prostate nodules are hard to palpation.

And here on this sono elasticity imaging image provided courtesy of the University of Rochester, we can see the normal prostate gland showing up as green and the hard prostate nodule showing up as darker.

This is an example of one of our early prostate ELAs grams showing normal anatomy.

There's no cancer here.

This is the em and that you typically see this cross pattern occurring in the prostate cland.

This dark area was not cancerous and suggests that like all ELAs grams, it will have to be viewed in conjunction with another modality to increase its specificity.

Here's another example of a prostate cancer courtesy of Rees Suan from France, showing the lesion on the sonogram and the hardness of the lesion on the ELAs gram.

And yet another lesion more recently imaged a courtesy of Kaiser holla at Riverside Research Institute showing a small prostate cancer showing up as a small dark hard area on the ELA gram and not really showing up on the sonogram.

What is the potential for prostate elasticity imaging?

There are image quality challenges due to the fact that prostate imaging probes are very tightly curved, making it hard to apply uniform pressure to the prostate cland.

But these problems may be solvable.

Prostate cancers may vary in stiffness.

Some cancers are harder than others, and that's a problem that, means that probably elasticity imaging will not be able to reliably image all cancers.

But continuing work at Riverside and elsewhere suggests that if you combine elastography with other features, perhaps even high field Mr.

The identification of cancers may be quite reliable.

And of course, it's a great, method for ablation monitoring if you decide to treat the prostate cancer with ablation rather than surgery or radiation.

Diffuse Organ Disease

Diffuse organ disease is another area where elasticity imaging probably plays a great role.

There are many diffuse organ diseases, and I've listed a couple of the processes in what they might be expected to do to the tissue edema and necrosis or decay of the tissue would produce softer tissue, whereas inflammation and fibrosis will tend to produce harder tissues.

And a couple of examples include rejection of a transplant, which is harder than a normal transplant, and hepatic cirrhosis, which also produces increased stiffness in the organ, the liver.

There's a non-imaging device that has been, extensively studied for evaluation of liver cirrhosis.

That is a fibro scan produced by a French company called Echos.

This, transducer is placed up against the ch abdominal wall between ribs and a pulse of ultrasound energy is sent into the, liver to image the change in liver caused by the vibration of the transducer.

So the vi the transducer does a mechanical vibration that transmits vibration waves into the liver, which produces shear waves.

The ultrasound beam then follows and images the shear waves and from the images or analysis, it detects the velocity of the shearer waves and then calculates the stiffness of the liver or the sheer modulus denoted in the equation as e.

This device has been extensively studied, mostly in Europe and elsewhere outside of the United States, and is showing very good performance with an area under the ROC curve of 0.94 for cirrhosis and a high sensitivity and specificity, but it doesn't seem to work quite as well for lesser degrees of, of liver fibrosis.

R can also be used to evaluate for liver cirrhosis.

The MR also uses sheer waves, which travel quite slowly through tissue.

A vibrator is placed against the liver and then a pulse sequence that is sensitive to phase is used to help image the shear waves as they traverse the liver.

The speed at which the shear waves travels is directly proportional or proportional to the, sheer modulus within that tissue.

And you can see the difference between normal and cirrhotic or fibrotic liver shown on this slide.

Diffuse renal disease and elasticity imaging should be a fruitful area for further study.

Pilon nephritis, interstitial nephritis and glomerulonephritis are all Entities that are hard to image and where elastography could be helpful.

Renal transplants also develop processes that are amenable to elastogenic imaging, including rejection, acute tubular necrosis and fibrosis caused by chronic rejection.

This is an example from a paper that was recently published about how to whether you could see the compress compression of a kidney by external compression.

And in this case it was used to evaluate normal versus abnormal.

The average strain in one patient with chronic rejection was approximately one third of that in a patient with a normal transplant.

So even accrued method like this where they're simply measuring the shape of the kidney and whether it flattened during compression has proven useful for diagnosis and suggests that elastography in a more advanced form would work to help diagnose rejection in the uterus.

Several things can be studied.

Endometrial tumors can be evaluated primarily to distinguish between cancers, polyps and fibroids.

Within the myometrium, fibroids and adeno myomas could be monitored.

Perhaps it could be used to grade the severity of adenomyosis.

It's not clear whether adenomyosis will produce increased or decreased stiffness of the myometrium.

This, entity remains to be studied and there's been some recent work on uterine changes during pregnancy suggesting that softening of the cervix, which is an indi indicator that the woman's ready to deliver may be Studied using graphic imaging.

So the potential in summary for diffuse organ disease is, that it's got great potential.

The adv, the advent of quantitative methods for strain estimation such as the sheer modulus estimation, means that, you could possibly use ELA graft evaluation to replace a number of the liver biopsies.

Also, imaging can help with grading of severity and also identifying areas that are more involved than others.

Many diffuse diseases do not involve an organ uniformly, but involve it asymmetrically or in patches.

And this can be useful information in, helping to determine what therapy is appropriate for a patient.

Cardiovascular Applications

In the cardiovascular system, strain imaging has been performed for a very long period of time.

They use somewhat different techniques 'cause these organs auto produce strain as a normal course of their process of contraction.

For strain imaging of the myocardium or heart evaluation of areas of ischemia and infarction can be evaluated using strain imaging showing less strain in the areas of infarction and ischemia.

Arterial walls can be evaluated, but in more particularly atheros plaque can be evaluated to look for areas of vulnerable plaque which, are softer than normal plaque and should appear.

This is an example of a vulnerable plaque or soft plaque that was imaged using strain imaging.

The area pointed out in the arrow in yellow is an area of increased softness on the strain image and it correlates with an area of vulnerable plaque on the histology.

So the potential for cardiovascular strain imaging is pretty high.

However, there are a number of methods that compete with strain imaging for evaluation of infarct and other abnormalities.

In addition, all the real high quality images of vessels have occurred using intravascular imaging, which is expensive and very invasive.

So a method for performing strain imaging from external to the vessel without having to insert a catheter will be required for this to really become popular Ablation monitoring.

Ablation Monitoring

I've already mentioned there are many ablation therapies available including radiofrequency, cryoablation, chemo laser, and high intensity focused ultrasound.

All of these produce increased tissue stiffness, which should be imageable using electrographic imaging, and inexpensive means of monitoring therapy is not currently available.

Ultrasound can see the lesion during the process of ablation, but cannot identify the lesion after the ablation has been completed.

And Mark can identify but is very expensive.

So what an inexpensive means would be very helpful.

The zone of coagulation being much harder than normal tissue is readily imaged as shown in this slide where a chunk of meat has been ablated locally.

And you can see the area of ablation as a slight area of increased echogenicity, but it shows up quite clearly on the elastic gram is an area of increased hardness high.

This is an example in a prostate of an ablation caused by high intensity focused ultrasound.

The outlined area is the area that was ablated showing markedly increased stiffness compared to the remainder of the prostate gland.

Poroelasticity and Fluid Movement

Poor elasticity is the monitoring of fluid movement using graphic imaging.

Tissue is not perfectly elastic, but an ideal tissue is perfectly elastic.

In other words, when you press down 10%, the tissue moves 5% in one direction and 5% in the other direction.

And thus you can, if you take the ratio of strain in the vertical direction versus the horizontal direction, you get a hoons ratio of 0.5 in tissue that contains fluid.

However, what happens is the pressure pushes the tissue out initially, but the tissue gradually moves back as fluid within the tissue dribbles away out of the field of view.

This is, known as pitting edema.

On a clinical examination on an ELAs agram, it appears as a decrease in the poissons ratio.

Thus the poissons ratio will drop from 0.5 down to some lower value in the 0.2 to 0.3 range.

The rate of change of the poissons ratio can be used to estimate how fast the fluid is moving away from the compartment in question.

We hope that this parameter can be used to help evaluate the, and quantify the amount of fibrosis that is present in tissues with edema versus the amount of actual fluid.

And mark can be used to evaluate fluid directly, but again, it's very expensive.

This real time sequence shows the shows a poor gram as we generate them courtesy of, the folks at the University of Texas who are my collaborators.

Initially, you notice the image started out as nearly solid red and solid red on one of these images represents a poisson's ratio of 0.5.

Remember I said that the poissons ratio will start at 0.5 and gradually drift downward as fluid leaves the area, being compressed by the transducer.

As you can see, the it image is gradually turning bluish, showing that the poissons ratio was falling down towards point, one and 0.2.

Thus the fluid is leaving, but it happens over a very slow time period.

This is a, a several minute examination showing that the fluid takes some time to leave, so it'll never be a really fast method of imaging.

The rate at which the fluid leaves, we hope is an indication of the viscosity of the fluid and of the material through which the fluid is escaping.

Thus, it may give us an indication of the amount of fibrosis, or other impediments to flow that might be present in the tissue.

Tissue that is purely fibrotic, will have a, will be read and will have a POIs signs ratio of 0.5 as it doesn't contain appreciable amounts of fluid.

This is an example in a patient with lymphedema of the change type of changes we expect to see.

Here you can see in the normal arm the three images on the top of the slide.

If you compress a normal arm, there's very little free fluid in the tissue.

And what you get is a poissons ratio of 0.5 shown in red.

Here, the arm with edema or free fluid inside the tissues starts out with a lower poisson ratio, and as time progresses, it gets even lower as some of the fluid drifts away from the area of interest causing relaxation of the tissues towards the center line and a decrease in the lateral strain.

Conclusion

So in conclusion, thanks to enthusiastic creative researchers, the number of potential applications for elasticity assessment technology has dramatically increased and, it's moving into the clinical realm much faster than anyone expected.

Even myself, quantitative elasticity applications have, been developing rapidly, and now I believe they can take their place alongside the traditional qualitative elastogenic imaging as a potentially very important tool for assessment of both tumors and for diffuse organ disease.

Thank you very much.