Strain
rate
imaging.
Myocardial deformation
imaging by ultrasound / echocardiography
Tissue Doppler
and Speckle tracking
by
Department of Circulation and Medical Imaging,
Faculty of Medicine,
NTNU Norwegian University of Science and Technology
Gateway to strain rate imaging for the novice
researcher and curious clinician. The
gateway. Old town bridge, Trondheim, Norway. It is also called
"the portal of happiness".
Follow
updates
on twitter: @strain_rate
Website updated:
December
2018. Recent
updates:
Link to website
categories,
Link to website
index, Link to list
of tables of normal values from the
HUNT study, link to
guide
to
clinical
use.
What's new:
June 2019:
Finally got some time to get the continuous
revising of the site into the air again
October 2019: I have added the results from the HUNT3 study,
giving the linear
measures of strain in both longitudinal, circumferential and
transmural direction, as well as the inter relation,
indicating fairly little systolic compression of the myocardium. The
concept of strain as a simultaneous deformation in three dimensions,
and the explanation that the three normal strains are coordinates in
three dimensions of this deformation (components of the
3-dimensional strain tensor), has previusly been extended both in
the basic
concepts of motion and deformation section and in the Basic
concepts in myocardial strain and strain rate section.
September 2019: Added a chapter on why
there isn't, and can't be a gold standard for strain, as this
rests on the choice of basic assumptions. As there is no universal algorithm for global strain, of course the concept of global strain as a universal
measure of ventricular function has no exact meaning. It is only a
theoretical concept.The problem goes beyond the inter
vendor differences in speckle tracking. This is also discussed in
the Global
function section. There is no universal global strain and thus
no reference, they are method dependent. This means that strain
values cannot be validated, and different methods cannot be compared
in terms of validity, and finally, normal values do only have
validity within the method used.
January 2019: Just added an example of pacing
induced dyssynchrony to the chapter
on regional
function.
December 2018: For completeness' sake added a new paragraph
to the Doppler
section on relations
between flow velocity and flow, as well as between flow
velocity and pressure. Meant to be theoretical, taking the
Doppler method into the realm of flow physics, and not a full
clinical discourse of all pathological states. Basically the flow
calculation from area and velocity, the continuity equation and the
Bernoully equation.
November 2018: Extended the chapter on phase
and spectral analysis, aliasing
and the Nykvist limit and colour
Doppler in the Basic
Doppler ultrasound section.
Welcome
About the website:
What is this website about?
Mainly giving an overview of the concepts and use of of strain rate
and strain: the fundamental concepts, the necessary technology
and display modes, the relation of deformation to physiology and
pathophysiology, and the clinical use.
The
website is divided into sections to allow
quicker downloading.
Every section can be read separately and are accessible by links, or
the complete website
index.
The text is riddled with links.
Following the links to see the reference, just click on the "back"
button in your browser, and you return to the point in the text
where you were.
The present section is mainly intended as an access portal, with a
full website
index with information about the content of the other
sections. Also, I have added some
general comments to strain and strain rate imaging, a short
introduction explaining very briefly what deformation imaging
(strain rate and strain) is about, and how it differs from the rest
of echocardiography.
In addition, I have given some historical information about
development of strain rate imaging.
The other
sections in this website fall into four main categories:
- Basic
concepts
- Basic concepts of motion and deformation
- Basic concepts in myocardial strain and strain rate
- Basic physiological concepts in strain and strain rate -
relation to load and contractility - what does cardiac imaging
actually measure?
- Technical
aspects
- Basic ultrasound for clinicians
- Basic Doppler ultrasound for clinicians
- How are the basic methods used to obtain strain and
strain rate - Basic strain ultrasound for clinicians
- How is deformation imaging displayed
- Limitations and pitfalls
- Relation
to function
- Relation to global systolic function - Global
systolic functional imaging
- Relation to regional systolic function - Regional
systolic functional imaging
- Relation to diastolic function - Diastolic functional
imaging
- Clinical
use of strain and strain rate imaging - Is
deformation imaging useful?
The
animations and video examples are all in *.gif animation format, so
no special software in the form of various media players will be
necessary for the animation. Also if downloaded (with due credits,
of course), it can be embedded in a power point presentation and
will run in all versions from office 2000 and onwards. It should
then be treated as "picture", not a video, meaning that it is
inserted in the file as a picture, and will then run without the
media player. It also means that it will not be necessary to keep
the loop separate outside the presentation, the animation is fully
embedded in the powerpoint file, just like any other picture, and
will run in a continuous loop when the picture is shown.
The internet is free, so feel
free to use the examples found on this website in
demonstrations and lectures. However, ordinary
ethics dictates that credit should be given to the
author (using the address: from:
http://folk.ntnu.no/stoylen/strainrate).
For publishing, I and the Norwegian University of
Science and Technology retain the copyright to all
material published here. (For written
publication, the acknowledgment should thus be: reproduced with
permission from: http://folk.ntnu.no/stoylen/strainrate).
I do not consider the fact that a signature is
embedded in some pictures sufficient
acknowledgement, without the website address. Even if the
images are taken from other
websites that do acknowledge
the source, I still require
acknowledgment of this website
as the principal source,
the fact that there are publications with
permission/acknowledgement available on the net,
does not alleviate the duty to acknowledge the
original source.
Using the material in papers without acknowledgement,
and
for published material without permission, I consider academic
misconduct, in addition to being copyright violation.
The most extreme example of this kind I have seen so far, is a
paper by the authors Dagianti A, Regna E, Laurito A, Malaj
A, Gossetti B, Fedele F, that I recently came across in
some journal called "Prevention and research". The paper had
used nine figures coming from this website without copyright
permission and with no acknowledgement. Repeated inquiries to
the journal have elicited no response.
List of tables of normal values from
the HUNT study with references.
Left ventricular dimensions
Systolic displacement, velocities, strain and strain rate
Diastolic velocities
Comparison between methods and
reproducibility
Website index:
This section:
Basic concepts
Deals with the basic concepts of motion (displacement and velocity)
and deformation (strain and strain rate), and how these concepts are
inter related. Further, how strain behaves in three dimensional
objects, and how the different strain components are inter related
in an incompressible object. Some of the manthemathical background
have been moved to this section, but the maths is placed at the end
of the section for those interested.
This section extends the basic concept of strain into the specific
geometry of the left ventricle. It is important to understand that
strain is about changes in myocardial dimensions. Thus strains the
components of myocardial volume changes, and only expressions of
geometry. The effects seen by strain rate imaging is thus explained
by geometry, not by material prpoperties of the myocardium, over all
geometry governs the changes and relations between strain
components. This is true of all strains, longitudinal, transmural
and circumferential as well as area strain. Also, the strain
gradient across the wall seen both in transmural and
circumferential strain is due to geometry, not differential fibre
action.
The strain components are simply coordinates of the three
dimensional deformation of the myocardium, and has nothing to do
with material properties of the myocardium, such as anisotropy or
fibre directions in a direct sense. Of course, the total deformation
is a function of fibre shortening, and ultimately, among other
things fibre architecture, but also of load and valve function. And
it is of note, that while systolic deformation deformation continues
to end ejection, myocardial relaxation starts at peak pressure.
- Relations
between tissue velocities and strain rate
- What
are the differences between strain rate and strain?
- Contractility
Basically, strain is the total systolic shortening,
equivalent to the isotonic shortening in experimental models,
and thus very afterload dependent. Peak systolic Strain rate,
on the other hand has been shown to be more closely related to
contractility, but the physiological limits of this
correlation is discussed.
- Stroke
volume It has been shown that strain relates best
to the stroke volume, and thus is both afterload and volume
dependent.
- Timing
Strain rate, being the temporal derivative of strain shows
more detectable shifts, especially in colour M-mode, making it
more useful for timing purposes.
- Normal
left ventricular dimensions Left ventricular
dimensions and geometry is closely related to the geometry of
left ventricular strain. Thus, normal values and the relation to
body size, age and gender are included here. Normal values
provided from the HUNT study.
The data from the HUNT study suggests that
relative wall thickness are both body size and age
dependent, that while wall
thickness increases with age, LV length decreases, invalidating
previously findings of M-mode based LV mass increase with age.
Finally, the ratio
of LV length and external diameter is BSA
independent and (nearly) gender independent, but decreases
with age, being a measure of age related remodeling. Tables of
normal findings from the HUNT study are added.
- Geometry of myocardial strain
- Myocardial
directions - normal strains
- Incompressibility
in the heart
- Longitudinal
strain
- Is
there layer specific longitudinal strain, and can we
measure it?
- Normal
transmural gradient of longitudinal strain The
gradient in longitudinal strain seen by speckle tracking
methods, may be simply a fnding from using a curved ROI of
a certain width. As both midwall and endocardial surfaces
move inwards by wall thickening, there will be an
additional shortening added to the longitudinal wall
shortening, which will be greatest in the endocardium, and
least in the outer contour. Thus, the observed gradient
may just be a geometric effect.
- Even though analysis software will produce layer values
at request, layer strain separation depends on line
density, line width, direction in relation to the wall (in
order to avoid pericardial echoes), and focussing. Number
of lines is again dependent on frame rate. It is difficult
to achieve a sufficient line density as well as narrow
enough lines.
- The
eggshell model
- Circumferential
strain Circumferential strain do NOT
reflect circumferential fibre contraction. There would have
been circumferential shortening even without circumferential
fibres. Circumferential strain is mainly the circumferential
shortening due to inward movement of midwall or endocardial
circumference as the wall thickens. As different vendors use
different definition of circumferential strain (midwall or
endocardial), there is no standard circumferential strain.
- But
circumferential strain is simply a measure of diameter
reduction, the global circumferential strain is the same
as the negative value of fractional shortening.
- Normal
transmural strain gradients. There is a gradient of both
transmural and circumferential strains towards the
endocardium, but this is trivial, and solely an effect of
geometry.
- Inter
relations of all three major strain components As
the myocardium is nearly incompressible, the three strains must
be interrrelated. Longitudinal shortening will give transmural
thickening, and transmural thickening (inwards) will give inward
motion of endocardial and midwall circumferences. In addition,
there is a small outer circumferential shortening, so all
thickening occurs inwards. This is the cause of the inwards
gradient of transmural and circumferential strain.
- Area strain Area
strain is neither the sum, nor the product of
circumferential strain. A slightly simplified modelling will
give the formula A = L * C + L + C . Thus, area strain is a
function of longitudinal and circumferential strain, not a
"new" parameter.As different vendors use different
definition of area strain (midwall or endocardial), there is
no standard circumferential strain.
- Strains
and fibre direction It is evident from the
discussion of strain components, that they to a very little
degree are related to fibre directions.
The relation between function imaging and physiology -
contractility, load, work and phases of the heart cycle. (Revised
September 2017)
It is important to realise that imaging, always shows the result of
myocardial shortening. This is true of any imaging measure,
whether it is fractional shortening, EF, longitudinal
shortening, strain or strain rate. It is also true,
irrespective of imaging method, MR, CT, MUGA or ultrasound.
Myocardial contraction, on the other hand, generates force
(tension). In an isolated myocyte or unloaded muscle preparation,
this force generates shortening. Unloaded states, however, hardly
exist in the intact heart. The opposite situation, the isometric
contraction, generates force, but with no shortening. This actually
exists during isovolumic contraction. In between, contraction
generates force, but doesn't shorten until the force exceeds the
resistance to shortening (load). Also, active contraction happens
only during pre ejection, where there is no deformation (in fact
about 80% of the systolic work is done during that phase), and first
part of ejection, ejection and systolic deformation (chamber
shortening, length and volume reduction etc), continues well into
myocyte relaxation due to the inertia of the blood as discussed
below.
Thus the notion that deformation indices can be load independent,
is self contradictory, although different parameters may relate
differently to load as discussed here.
And the notion that different imaging methods (like MR) are less
load dependent than others, is simply ridiculous.
However, as segment interaction is part of segmental load, it is in
fact the load dependency (as well as the motion independency) of
strain rate imaging is able to image regional inequalities in
tension, which will result in inequalities in shortening. Here,
however, timing is also of importance.
Technical aspects
This section is the basic ultrasound, B-mode and M-mode part of the
previous "Basic ultrasound, echocardiography and Doppler for
clinicians. Due to the size and number of illustrations, the page
tended to load very slowly. It has now been split into this section
on Basic ultrasound, and another
on Doppler, including tissue Doppler.
is intended as an introduction to basic ultrasound physics and
technology for clinicians without technical or mathematical
background. A basic knowledge of the physical principles underlying
ultrasound, will give a better understanding of the practical
limitations in ultrasound, and the technical solutions used to solve
the problems. This will give a clearer picture of the reasons for
the problems and artifacts. Technical
or mathematical background is not necessary, explanations are
intended to be intuitive and graphic, rather than mathematical. This
section is important for the understanding of the basic principles
described in detail in the section on measurements
of strain rate by ultrasound. Especially in order to understand the fundamental
principles that limits the methods.The principles will also
be useful to gain a basic understanding of echocardiography in
general, and may be read separately, even if deformation imaging is
not interesting.
It is important to realise that the last couple of
years has seen tremendous improvements in both hardware
(allowing a much higher data input to the scanner as well as
processing technology), and software (allowing more data
processing at higher speed). It even allow using input data in
a way that also improves the beamforming characteristics in
processing, as they are used for the generation of a picture.
Thus the simple principles of beamforming outlined here are an
over simplification compared to the most advanced high end
scanners. Thus, present technology is far more complex, and
neither traditional beamforming, focussing nor image
processing conforms to the simple principles described here,
in the most advanced high end scanners, but they will still
serve to give an idea. And
simpler equipment still conform more closely to the
basic principles described here.
The physical principles still apply, although due to the
use of MLA
and similar even more advanced techniques, the
fundamental limitation of the speed of sound has been
dealt with.
This section is the Doppler part of the former "Basic ultrasound,
echocardiography and Doppler for clinicians". Due to the size and
number of illustrations, the page tended to load very slowly. It has
now been split into one section on Basic ultrasound, Doppler, and
this one on Doppler, including tissue Doppler. In addition, the
background paragraphs from the previous mathematics section on the
derivation of the Doppler equation and the phase analysis have
been included here (but can be bypassed, of course, for those not
interested).
Technical or mathematical background is not necessary,
explanations are intended to be intuitive and graphic, rather than
mathematical. This section is important for the
understanding of the basic principles described in detail in the
section on measurements
of strain rate by ultrasound. Especially in order to understand the fundamental
principles that limits the methods.The principles will also
be useful to gain a basic understanding of Doppler echocardiography
in general, and may be read separately, even if deformation imaging
is not interesting.
Principles and
technology for strain and strain rate imaging by
echocardiography.
the present section takes this further, into
the technical aspects of how deformation is measured using
the basic methods for deformation imaging and measurements.
All ultrasound methods can
be used to assess regional function. And, in fact, regional function
is regional deformation.
Basically, irrespectively of method, the fundamental indices of
motion (velocity and displacement) and of deformation (strain rate
and strain) are the same. Also, the display of the indices can be
used across some of the methods for acquiring them. However, the
application of the methods can vary, so this chapter deals with the
application of the different ultrasound methods to deformation
assessment, similarities and differences.
And the fundamental limitations of ultrasound apply. The poorer the
image quality, the less useful are any of the methods:
Garbage in - garbage out.
Thus all methods have limitations connected to the basic limitations
of ultrasound, and each method have specific limitation as well, and
finally there are specific limitations connected to both numerical
measurement and to the deformation imaging in itself. The
"limitations and pitfalls have been moved to a separate section to
enable quicker uploading.
Finally, there is no universal algorithm for global strain,
of course the concept of global strain as a universal measure of
ventricular function has no exact meaning. It is only a
theoretical concept, but in practice related to each specific
manufacturer as discussed in the section on normalised displacement
and global strain.
As functional imaging is display, the different methods for display
are given here, in order to being able to interpret the different
displays.All methods can give the information as numerical traces,
parametric (colour) images (in 2D or Colour M-mode). 3-/4D
reconstruction has some limitations using segmental strain, however.
Basically, however, irrespectively of method, the fundamental
indices of motion (velocity and displacement) and of deformation
(strain rate and strain) are the same. Also, the display of the
indices can be used irrespectively of the method for acquiring them.
However, some of the methods set limits for how the display can be
made, and this is explained here.
Limitations and artefacts related to
strain rate imaging
Basically, irrespectively of method, the
fundamental indices of motion (velocity and displacement) and of
deformation (strain rate and strain) are the same. Also, the display
of the indices can be used across some of the methods for acquiring
them.
The fundamental limitations of ultrasound apply. The poorer the
image quality, the less useful are any of the methods:
Garbage in - garbage out.
In addition, there are specific problems that arise from the
specific application of the basic methods to deformation
measurement, as well as how the methods are dealt with, which
creates new pitfalls. The present section have been re
edited, so the main headings are now the problems and limitations
themselves, and the effect on each of the different methods are
described under those headings.
Mechanics of
physiology and pathophysiology
An understanding of the concepts of myocardial load and work as well
as the various systolic measurements to each other and to timing in
systole given in the section on what
strain and strain rate actually measure is an advantage.
- The
erroneous comparison between longitudinal strain and
fractional shortening. Increased "radial
function" measured by fractional shortening as
compensation for reduced longitudinal function,
is a conceptual error, due to misunderstanding
of geometry as seen below. Thus, it is doubtful
also that increased "radial function" (meaning wall
thickening) as compensation for decreased longitudinal
function actually exists, it seems theoretically
impossible.
- The
erroneous use of ejection fraction in concentric
geometry Preserved
ejection fraction in heart failure do not reflect
"diastolic heart failure". The whole
point is that ejection fraction (or fractional
shortening) do not measure systolic function in
concentric geometry, All principal systolic strains
can be reduced and still the EF may be preserved. In
eccentric hypertrophy, it is opposite, the EF may be
reduced despite completely normal myocardial function.
All this is due to the faulty use of EF in altered
geometry.
- Wall
measurements - long axis systolic function
- Normalised
displacement Normalised displacement, is MAPSE
/ Wall length, which is relative LV shortening, i.e.
global strain. Thus, this is one method for global strain.
It was thought that global strain would normalise for
heart (and thus body - ) size, but in the hUNT study, the
opposite was the case, BSA dependency increased in
absolute value, although with a negative
correlation.
- Why
is GLS/ normalised MAPSE more BSA dependent than
MAPSE?
Because normalising only for length, do not take into
consideration that bigger ventricles are also wider,
allowing for a bigger stroke volume with the same MAPSE.
- Global
strain As there is no universal algorithm for
global
strain, of course the concept of global strain as a
universal measure of ventricular function has no exact
meaning. It is only a theoretical concept, but in practice
related to each specific manufacturer.
- Why
is GLS/ normalised MAPSE more BSA dependent than
MAPSE?
The hypothesis was that GLS normalises for heart
(and body- ) size.
- Firstly, this was not the case, as the main source
of variability was age, not body size, so there was no
effect of the normalisation. Thus there is even less
reason to believe that GLS has a better accuracy than
MAPSE, as the variability is the same. However, there
are no large, direct comparative studies, so this is
still not decided, as there lacks data on the
separation of the means between healthy and patient
populations for the two measures.
- More surprising was the fact that while gender
differences was only due to body size (as expected),
MAPSE was gender independent and actually less
body size dependent that GLS. There was a weak BSA
dependence of MAPSE, but not enough to give
significant gender difference in 1266 subjects.
Normalising for heart length induced a stronger
negative relation to BSA, as well as significant
gender dependence. This is due to the fact that the
ratio between LV external diameter and length,
remains constant independent of BSA (386).
This again means that the stroke volume increases by
the square of the diameter, without change in MAPSE.
Without change in MAPSE, however, the global strain
decreases inversely by the length. Thus
this is a systematic error due to the fact that
GLS only normalises for one dimension, while
deformation is three dimensional.
January 2018: I have added recent
results from the HUNT study (417),
to the global
systolic function section showing that GLS, which presumably
should reduce biological variability compared to MAPSE, do not do
so.
This is what strain rate imaging is really about. Those of
you who know most echo, but are curious about the physiological
background for regional functional imaging, may well jump directly
here.
- 16, 17 or 18 segments?
- Segmental
interaction (Revised and extended
April 2015) Segment interaction within the AV-plane leads to
the specific patterns of regional dysfunction. This includes
both delayed onset of shortening/intitial stretch, systolic
hypo-/a-/dyskinesia, and post systolic shortening, which all
are part of the same mechanism. This also shows that which has
be shown in studies, mitral annulus motion will not give
information about regional function. Post systolic shortening
is thus not an isolated event, but part of the total pattern
in ischemia, but on the other hand is not limited to ischemia,
being seen both in left bundle banch block and hypertrophy
- Diastolic function
- Strain
and strain rate in the atria There has been a
lot of literature about the so-called "reservoir function" of
the atria, meaning the positive strain of the atria during
ventricular systole. However, the expansion of the atria during
ventricular systole is due to the shortening of the ventricles,
so it is in fact a function of the ventricle Being strain, it is
normalised for atrial size, which means that it decreases with
increasing atrial size, and thus is a composite measure, but do
not add any independent information, as confirmed in the
Copenhagen heart study.
Is deformation imaging
useful?
This section
is intended as a guide for those who want to start using
deformation imaging, but who don't want to go into the basic
theory of deformation, display and physiology/pathophysiology
first. For this reason, it is placed first in the sections,
instead of the conventional last place after all the theory have
been dealt with.
The intention of
this section is to shown how to become friends with the method.
It is usually possible to sniff out extra information not
visible to the eye. In this section I will try to show that the
use of deformation imaging not necessarily needs to be very
complicated or time consuming, as well as suggesting how to
interpret the findings in a quick, and hopefully easy way, and
then giving some more information for those who wish to go
further. The chapter is illustrated illustrated by
examples both of pathology and artefacts. It is an integrated
approach, using both velocity curves and colour SRI, in addition
to strain rate and strain curves.
The section is
intended to be self contained, but presupposes some familiarity
with the concepts and the colour and curve displays. Of course, links are
provided to the relevant more in depth chapters in the other
sections. To understand more about the curves, I
recommend the section on displays.
This
section can also be downloaded as a pdf (without the
animations, of course). The other sections can be used as
reference over time.
References
for all sections
Some general comments
 |
|
| An academic discussion.
Northern fulmars in New Ålesund, Spitsbergen |
Professor and student.
Blue eyed shag and adelie chick- Peterman Island,
Antarctica
|
What is strain
rate imaging about?
Mainly imaging regional
inequalities in myocardial function.
Colour curved anatomical M-mode (CAMM) of strain rate.
Apical infarct to the right, compared to normal on the
left. For explanation see
main chapter as well as specifically
about CAMM below.
Same infarct seen by curves (explained
below) to the left and CAMM to the right.
The
method of strain rate imaging by tissue Doppler was developed
here at the Norwegian
University of Science
and Technology in Trondheim, Norway. It was
the subject of two doctoral theses, one in technology (1)
and one in medicine (2),
and was a result of a successful cooperation between technical
research (in strain and velocity gradients) and medical
research (in long axis function of the left ventricle). One of
the important point of my work with long axis function, was
that this lead to Strain Rate Imaging being applied to longitudinal
velocity gradients, thus making the rough method more robust,
as well as all segments of the ventricle available for
analysis. The method was originally validated in a mechanical
model, in cooperation with the university of Leuven, Belgium (3)
and described in a method article from Trondheim
in 1998 (4)
and 2000 (5).
The basic publications dealt with feasibility (1998) (4),
clinical validation by comparison with echocardiography (6)
and with coronary angiography (7).
Validation of strain measurements (from integrated
strain rate) was done at Rikshospitalet,
Oslo, Norway by comparison with ultrasonomicrometry (8)
and MR, in cooperation with Johns Hopkins
Hospital
(9).
Early work on the feasibility of the method in myocardial
infarction was also done at the university of Linköping
and later at Leuven
(10).
An excellent early review paper was published by the Leuven
group (11).
Why use
strain and strain rate?
One example:
|
|
Paradoxical motion
of basal basal inferolateral wall. Is this dyskinesia?
|
No, strain rate
imaging shows vigorous systolic shortening of the basal
inferolateral wall. Details of this case can be seen here.
|
- The strain and strain rate
subtracts motion due to the effects of neighboring segments (tethering).
Tethering may both mask pathological deformation and impart
pathological motion to normal segments and deformation imaging
and is necessary to locate and show the true extent of
pathology. And in some situations exclude regional pathology.
This means that motion parameters (displacement and velocity)
reflects global function, and should be applied to the mitral
ring, while deformation imaging (strain and strain rate) shows
regional function within the myocardium.
- However, strain and strain
rate are more susceptible to noise, both by
tissue Doppler and speckle
tracking ( the weaknesses of the last being masked by
smoothing),
and a rough qualitative assessment of regional function can
be done by assessing the offset
between velocity curves, as well as by
colour M-mode.
- Even if velocity curves
indicate abnormal motion, the deformation parameters are
necessary to localise the area of abnormal deformation.
- Deformation parameters are
useful in doing a more comprehensive assessment of local
contraction and relaxation, such as initial stretch,
hypokinesia, post systolic shortening, both in infarcts /
ischemia and in electrical
asynchrony.
- Strain and strain rate are
deformation per length, and thus are normalised for heart
size, also in global deformation, meaning that it reduces
biological variability due to size differences. Clinical
evidence for the advantage of this is emerging, at least in
children where variability in heart size is greatest. The
advantage in adults is still uncertain.
- However, both systolic and
diastolic deformation itself is load dependent, as is the case
with all volume based measures of ventricular function. This,
it appears, cannot be emphasized enough. But as the main value
is in diagnosis of regional
dysfunction, the segment
interaction in combination with the load dependency
enables us to make inferences about uneven contractility, i.e.
regional dysfunction, even if the contractility cannot be
measured directly. Thus, regional deformation imaging shows
regional relative contractility.
- In addition, in different
inotropic states, the changes in contraction will be caused by
changes in contractility, and thus the measures are valid
measures of contractility changes.
Deformation
imaging
has added a lot of knowledge of the regional myocardial properties,
and the fundamental physiological knowledge gained from deformation
imaging is method independent. Thus, the basic principles and the
relation to pathophysiological knowledge comes first, in this
section.
But of course, even with the fundamentals of physiology, an
understanding of the technical aspects of the methods and especially
the limitations and propensity to artefacts is important, thus for
better background the fundamental technology sections on basic
ultrasound, basic
Doppler and the relation
to deformation measurements is important.
Now,
it is important to emphasize that both motion and deformation
imaging are no longer simply tissue Doppler derived
modalities. Both can be derived by tracking the motion of the
myocardium in grey scale pattern, speckle
tracking. The basic concepts are the same; general
principles of motion (velocity and displacement) vs.
deformation (strain rate and strain) apply,
irrespective of method but the
limitations
may differ between the methods.
The
terms:
- Velocity imaging, - Displacement imaging, - Strain
rate imaging, - Strain Imaging, should not be taken
synonymous with tissue Doppler, but should be used
irrespectively of the method employed, and the term
"by tissue Doppler", "by speckle tracking" or
whatever application is used should be added, if
studies are cited.
The fundamental concepts relating to geometry, pumping and
physiology, however, are independent of methods (except where the
methods relate differently to geometry), so even if examples are
taken from one method, the underlying pathophysiology should be the
same with other methods. The sensitivity for specific changes may
vary in different situations according to the strengths of each
method .
In the ideal world, any measurement would give the diagnosis once
correct cut off an normal values are established. But this is not
the ideal world, and image quality is far from perfect in most
cases, which is a fundamental property of ultrasound. Thus, no
single measurement is perfect, an echocardiographic examination
always consists of using all the available, more or less
circumstantial evidence, weighing findings against each other and
arriving at a conclusion. This will usually be fairly certain in the
hands of an experienced clinician, even if single measurements are
not. This is a fundamental property of all echocardiography. With
the limitations
inherent
in basic ultrasound and in the specific methods, clinical
ultrasound will partly be a craft, not pure
science. The careful weighing of the evidence in terms of the
methods limitations is thus an integral part of the examination, and
a knowledge of the methods themselves and the method specific
limitations is essential.
This is also the case with deformation imaging, and this should be
the basic approach, deformation imaging being part of the total
evidence, and can serve as an aid
to diagnosis, as shown in the section on how
to use deformation imaging clinically.
Why is deformation
imaging underutilised?
The number of publications is enormous, but the use in daily clinic
is still limited despite giving additional information, as shown in
"Is deformation imaging useful",
the method, especially tissue Doppler seems to have gone into a
partial eclipse:
Partial solar eclipse, March 20. 2015. Picture taken
from Trondheim
This has some historical reasons:
Tissue Doppler
derived deformation
When tissue Doppler derived deformation (strain rate imaging)
arrived (4),
we concentrated on semi quantitative assessment with colour M-mode (4,
6, 7).
The colour WMS seemed to equivalent with B-mode WMS both in
feasibility and accuracy (6,
7), but no better. This is hardly be surprising, as the expert
WMS is a fairly precise method, an colour WMS is still semi
quantitative. However, with increased understanding, it beacme
evident that the colour M-mode offered additional information, both
on the time
course of the deformation (like initial akinesia and post
systolic shortening), as well as being very robust against noise.
noise
However, the possibility to go from semi-quantitative assessment of
regional function to quantitative measurement, seemed interesting.
And with objective measurement, the experience dependency might be
bypassed. So objective measurement became the goal. And principally
peak vaues. But as strain rate is the result of a derivation, it is
much more noisy
than velocity curves. Even if smoothing and the integration to
strain took care of much of the random noise, non random noise
(especially clutter)
remained a problem.
Thus, curves and especially peak values seemed to be difficult to
interpret, as it was difficult to see whether the curves were due to
pathology or artefacts. In addition the fact that tissue Doppler was
angle dependent, and that strain rate derived from a set of
velocities was even
more angle dependent, lead to a lot of erroneous speculation
that it could for instance not be used in the apical segment, which
of course is
nonsense. And in this setting, it became evident that using
tissue Doppler with reliance was just as experience dependent as
B-mode.
Thus, inherent
weaknesses in the tissue Doppler method, especially angle
dependency and vulnerability
to noise, especially to clutter,
made the method little approachable, and this led to the
method being little used.
And in all this, colour tissue Doppler was not reinstated, despite
it's ability to:
- Give information as wall motion score similar to B-mode, and
- Give additional information on timing beyond the B-mode
information (also due to higher temporal resolution), and
- Visualise clutter artefacts in a way that makes timing
information still available, and hence,
- Might guide the placement of the ROIs to extract reliable
curves.
Thus, interpretation of artefacts, as well as timing is easy with
colour M-mode. The curved M-mode will then give the possibility of
assessing both if regional shortening is normal, and to look at the
timing. Actually, the curved M-mode has better
spatial resolution that strain rate or strain curves.
Also, in the user interface provided, all difficulties were
eminently visible, but instead of being considered an advantage, it
disgusted a lot.
The problem being both the experience dependency, initially poor
user friendliness, and variability of results. The limitations of
the methods may scare potential users as well.
One of the main points about assessing regional function by
deformation parameters, however, is that peak values is not the most
important thing at all. As seen below, the colour M-mode and the
shape of the curves gives almost all information qualitatively in a
quick and easy way:
- Reduced systolic peak strain and strain rate can be seen
simply by comparing the curves or colour display from the
affected segments with curves from non-affected segments.
- The pathological time course of ischemia with delayed onset of
shortening and post systolic shortening can be seen directly in
the same way.
Both initial delay in shortening of a segment, reduced peak values
and post systolic shortening are all phenomenons that tells about
the relative reduction in rate of tension development (strain rate)
and total tension, all measures of relative reduced contractility of
a segment or region, as well as delay in relaxation, another
assessment of both ischemia and contractility.
In my opinion, however, tissue Doppler seems to be able to fly,
although it is a rather heavy bird:
|
|
| Wandering albatross
is heavy in taking off from water. Drake Passage. |
However, this bird
is fully able to take off and fly. |
Speckle
tracking derived deformation
When speckle tracking was launched, it seemed to solve some of the
problems from tissue Doppler.
The challenger. Musk oxen, Grønnedal, Greenland.
- It came with an alluringly user friendly interface which
actually had nothing to do with speckle tracking per se).
Especially for deriving peak values, which is attractively easy,
but poor use of data, as it discards additional information
about timing.
- In addition, the low frame rate makes it less suited for
timing, especially in the spline smoothed versions.
- Also, it was maintained that speckle tracking was angle
independent, and able to track in all directions,
- Which is a half truth at best, as
- the axial resolution of B-mode is so much better than the
lateral, this means that tracking
is better in the axial direction.
- The line density decreases with depth, and so, of course
do the lateral
resolution and
- With
increasing frame rate of B-mode, the line density is
reduced, and especially in 3D speckle tracking, where line
density is severely reduced.
- There was an apparent reduction in noise, (which again did not
have anything to do with speckle tracking per se, but with the
applications using a generous amount of smoothing), and the
smoothing made the effects of clutter less visible, and the
method apparently more robust.
- It is true that speckle tracking is done in harmonic mode,
while tissue Doppler is done in fundamental, removing some of
the clutter, but
- Most of the clutter is smeared out due to the spline
smoothing, which may lead to absurd examples.
- The only method free of this that I know of, is the
segmental tracking application developed by NTNU,
described elsewhere,
which can track segmental deformation by speckle
tracking alone, or in
combination with tissue Doppler.
- Thus, the segmental values obtained by this kind of speckle
tracking are not true segmental values, results
of a spline (or similar) function taking into account not
only the tracking of speckles in the actual segments, but the
over all tracking as well as the global deformation. the
segmental values are then partly splines of the global
function. Given the obvious curvature
dependency in addition, the method has a serious
sensitivity issue for reduced segmental function. Thus the
apparent lower vulnerability and better reproducibility of
speckle tracking over tissue Doppler is due to more smoothing.
When the same smoothing is applied to tissue Doppler. results become
near identical.
- It also seems that measurements are software dependent, even
software updates from the same vendor results in different
values, and also that reproducibility, even within one software
frame is lower than previously reported. Not unsurprising for
those of us who have tried this in practice. (377).
|
|
Same small apical infarct analysed by speckle tracking
(left) and tissue Doppler (right), showing fairly low sesitivity
of speckle tracking in this particular instance, as described in
more detail here,
and technical reasons here.
Thus, for regional function, the flight capability of speckle
tracking is low:
|
 |
Molting gentoo penguin, Port
Lockroy, Antarctica.
|
Jumping penguins. Paradise
Harbour, Antarctica. |
A direct comparison between speckle tracking and tissue Doppler in
clinical examples can be seen here.
It is thus no surprise that the publications on speckle tracking
lately, has been about global
strain, which is a global function parameter. (Thus, a method
that started out as a method for regional function, has devolved to
being a global measure, due to the inherent weaknesses in the
speckle tracking application). Here, it may prove useful, especially
when compared with the most used and poorest method of them all; ejection
fraction. It has been instrumental in shifting emphasis from
ejection fraction (which in fact doesn't
work in small ventricles), to long axis function. However, so
far it has not been proven that global longitudinal strain adds
information to that of simple longitudinal shortening . However, the
value of long axis function was shown early in the 90ies (31,
32,
33,
34,
35,
64,
65,
66)
by mitral annular plane displacement(MAPSE).
It has not been definitely proven that normalising for LV length
(which is global strain), is advantageous in adults, although of
value in children (159,
214,
288).
Also, global strain is not load independent, and do not
measure contractility.
The different advantages and disadvantages, which is extensively
discussed in a separate section on measurements
of
strain and strain rate by ultrasound. It may also be a
surprise to some, but spectral Doppler has a fairly low temporal
resolution. Even if sampling rate can be as high as 1000, the spectral
analysis requires a long sequence of frames, thus reducing the
effective frame rate, the temporal resolution of both pulsed Doppler
flow and tissue
Doppler is usually around, and may even be below 100 FPS. This
is described in slightly more detail in the basic
Doppler section.
Modern ultrasound technology utilising the ability of new probes
being able to transmit more data, as well as increased computing
speed and capacity for more advanced post processing, has
increased both line density and frame rate of B-mode, and this
development will continue. This means that some of the
reservations towards speckle tracking in terms of frame rate and
lateral resolution will become less important in time. This,
however, do not lend more validity to earlier results. And so
far, it does not apply to 3D
speckle tracking
Concerning nomenclature:
The terms: - Velocity
imaging, - Displacement imaging, - Strain rate
imaging, - Strain Imaging, should not be taken
synonymous with tissue Doppler, but should be used
irrespectively of the method employed, and the term
"by tissue Doppler", "by speckle tracking" or
whatever application is used should be added, if
studies are cited.
There is still a need for standardisation of nomenclature in the
field. Especially for newer measures relating to deformation
imaging, which are not covered by The ASE standards (146).
The systolic mitral annular excursion is a useful measure of global
systolic function. It has had various names, the term AVPD
(atrioventricular plane descent) is unfortunate, as it don't
separate beteen mitral and tricuspid excursions, event though they
are different, the term MAE has been used from the beginning, but as
TAPSE has been firmly established for the right ventricle, I think
the term MAPSE (mitral annular plane systolic excursion) should be
used in order to harmonise. The text and figures in this website
have been adjusted accordingly.
The nomenclature about positive and negative deformation has been a
mess.The original definition of strain makes shortening and
shortening velocity negative values. For the longitudinal and
circumferential functional measures, this means that the more the
contraction, the lower the negative values. Then "increased" strain
and strain rate would mean "less contraction", which is absolutely
counter intuitive. Also, the literature has through all the time
strain and strain rate has been evident, talked about "peak values"
strain and strain rate. Consistent with the usage of negative
values, that actually should have been "trough values". For
transmural strain, the case is opposite, wall thickening is positive
strain and peak values are really peak values. The new definitions
paper (287)
recommends the use of absolute numbers, which will make the
discussions more intuitive, and I wholeheartedly concur.
However, as we have two reference systems for strain and strain
rate: Lagrangian
and Eulerian, I'm no fan of the term "natural strain" for
Eulerian strain, I can't see why one reference system is more
"natural" than another. Also using mathemathician's names should be
symmetrical.
Back
to website index
Editor:
Asbjørn Støylen Contact address:
asbjorn.stoylen@ntnu.no