IMAGING FRAILTY AND IT’S SKELETAL MUSCLE BIOMARKERS UNDERSTANDING MUSCLE BEFORE IT’S GONE: MULTI-PARAMETRIC CHARACTERIZATION OF SKELETAL MUSCLE BIOMARKERS DERIVED FROM DXA AND MRI IN A FRAIL POPULATION By KONRAD GRALA, BHSc.
A Thesis Submitted to the School of Graduate Studies in Partial Fulfillment of the Requirements for the Degree Master of Applied Science McMaster University © Copyright by
Konrad Grala, September
2024 McMaster University MASTER OF APPLIED SCIENCE (2024) Hamilton, Ontario, Canada
(Biomedical
Engineering) TITLE
: Understanding muscle before it’s gone: Multi-parametric characterization of skeletal muscle biomarkers derived from DXA and MRI in a frail population AUTHOR: Konrad Grala BHSc. (Health, Engineering Science, and Entrepeneur- ship),
McMaster University, Hamilton, Canada SUPERVISOR: Dr.Michael Noseworthy
& Dr.Alexandra Papaioannou
NUMBER OF PAGES
: xvi, 85
ii
Lay
Abstract
A person is diagnosed with sarcopenia when they present symptoms
of low muscle mass, strength, and/or
function. Defining these three criteria with objective measures has been long debated by researchers and clinicians alike. By understanding how different properties, or biomarkers, of skeletal muscle relate to one another and change as a person becomes more frail, we hope to better understand sarcopenia and identify the best measures to classify someone as sarcopenic. Being able to accurately diagnose someone as sarcopenic early allows for more effective treatment of this muscle disease. In this study, non-invasive
magnetic resonance imaging (MRI) and dual energy x-ray absorptiometry (DXA) were used to
measure many different biomarkers
of skeletal muscle at the mid
-thigh. Through characterizing these measures of muscle quality and quantity between different imaging techniques this study aimed to recognize which imaging techniques, and more specifically biomarkers, can best distinguish between a person who is sarcopenic and one who is non-sarcopenic. iii Abstract
Approximately 23% of Canadians over
the
age
of 65
are
considered
frail
, with that number predicted to increase up to 40% for the population over the age of 85. Frailty
is a geriatric syndrome
defined
by the
natural decline in
muscle mass and function
caused by the natural aging process. When developing to an excessive degree, frailty may present as a disease state, which is recognized as
sarcopenia. The
exact
definition of sarcopenia
relies on
the presence of low muscle mass
, strength, and/or function, but quantitative cut-off values are still a topic of debate. Understanding how biomarkers measured via diagnostic imaging
such as magnetic resonance imaging (MRI) and dual-energy x-ray absorptiometry (DXA
) describe skeletal muscle can allow doctors to develop a profile of sarcopenia and define predictors to aide in preventative therapy. 4 male and
9 female (mean age
= 78 ±
6
.5
years
) participants from a frailty study underwent full-body DXA and had their dominant thigh scanned using a 3.0T MRI.
DXA-derived
appendicular
lean mass
(ALM)
and MRI-derived cross-sectional area
(CSA), fat fraction (FF), T2 relaxation (T2),
magnetization transfer ratio (MTR), fractional anisotropy (FA) and mean diffusivity (MD
) from 4 muscle groups at the mid-thigh were defined as muscle biomarkers. Pearson’s correlation was calculated to identify relationships between biomarkers and
a Wilcoxon rank
-sum
test was
per- formed
to
assess
the
agreement
of
low-muscle mass characterization between ALM iv normalized by height (ALMI), ALM normalized by BMI (ALM/BMI), and the gold- standard MRI cross-sectional area. Strong positive correlations between muscle quantity biomarkers such as ALMI and CSA were recognized
within the
quadriceps (
p=0
.0095), adductors (
p=0
.035),
and
sartorius (
p=0
.00065) muscles while muscle quality biomarkers such as FF and T2 showed significant positive correlation within the quadriceps (p=3.58 ∗ 10−5) and the hamstring (p=0.0042) muscles. Finally, ALM/BMI displayed a much stronger agreement in muscle mass quantification with the gold-standard of MRI-CSA over the more commonly researched ALMI from DXA.
The main purpose of this study was to
collate
a
vast array
of
skeletal muscle biomarkers obtained using DXA and MRI on a frail population, and show that signif- icant correlations can be recognized from a single MRI-slice located at the mid-thigh. Additionally, this study recognized the potential of ALM/BMI as the DXA-derived biomarker of choice in muscle mass assessment of frailty. v Acknowledgements To my family and friends, the best support system. ... and free coffee refills at Starbucks. vi Table of Contents Lay Abstract
Abstract Acknowledgements Abbreviations 1 Introduction 2 Background 2.1
Definition of Frailty and Sarcopenia . . . . . . . . . . . . . . . . . . . 2.2 Skeletal Muscle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Muscle Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Literature Review 3.1 Frailty and Sarcopenia . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2
Assessment of Muscle
Strength . . . . . . . . . . . . . . . . . . . . .
3
.3
Assessment of Muscle
Function . . . . . . . . . . . . . . . . . . . . .
3.4
Assessment of Muscle Mass . . . . . . . . . . . . . . . . . . . . . . . vii iii iv vi xiv 1 4 4 6 12 14 14 15 17 19 4 Imaging Techniques in Muscle Assessment 4.1
Magnetic Resonance Imaging (MRI
) . . . . . . . . . . . . . . . . . . 4.2
Dual-energy X-ray Absorptiometry (DXA
) . . . . . . . . . . . . . . .
5 Materials and Methods 5.1 Study Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2
Participant Recruitment . . . . . . . . . . . . . . . . . . . . . . . . .
5.3
Experimental Design . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4
MRI Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 DXA Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 MRI data preprocessing and registration . . . . . . . . . . . . . . . . 6 Results 6.1 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Validating sarcopenia status classification 8 Discussion 9 Limitations and next steps 10 Conclusion 22 22 30 32 32 32 34 35 40 41 44 44 51 54 60 63 viii List of Figures 2.1 Table summarizing definitions of sarcopenia according to each consen- sus group
with cutoff points for low muscle mass, strength, and physical performance
. Content adapted from review by Coletta et al. [1] . . . 6 2.2 Diagram of skeletal muscle hierarchy (
Image credit: download for free at http://cnx.org/contents/14fb4ad7-39a1-4eee-ab6e-3ef2482e3e22@8.119
) 8 2.
3
Anatomy of a
muscle
fiber (
Image credit: download for free at http://cnx.org/contents
/14fb 4.1
39a1-4eee-ab6e-3ef2482e3e22@8.119
) . . . . . . . . . . . . . . . . . . 9 Sample calculation for a single pixel of a T2 map from T2 decay curve 5.1 displayed as a semi-log plot. A sample T2 map of a healthy male adult (top) compared to on of a frail male participant from this study (bottom). 24 Baseline demographics and characteristics of imaging study population. 34 5.2 Localization of the mid-thigh. The length of the femur is measured according to the bony landmarks using a coronal scan. The center slice is placed at the midlength of the femur and is oriented along the axial plane. This becomes the reference for all future acquisitions. . . 36 5.3 Fat fraction image produced by the IDEAL IQ sequence. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). . . . . . . . . 37 ix 5.4 Fractional anisotropy (FA) map produced by the dtifit function from 5.5 FSL suite using DTI data. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). . . . . . . . . . . . . . . . . . . . . . . . . . Mean diffusivity (MD) map produced by the dtifit function from FSL 5.6 suite using DTI data. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnetization transfer ratio (MTR) image produced from saturated 5.7 and unsatureated MT image. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). . . . . . . . . . . . . . . . . . . . . . . . T2 relaxation time map (T2 map) calculated from multi-echo sequence 5.8 images. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample center slice with manually drawn muscle segments. A mask was drawn for each of the four muscle groups; quadriceps (Quad.), hamstrings (Ham.), adductors (Add.) and sartorius (Sart.) on the 6.1 out-phase image (A). The masks can then be applied to the FF images (B) and others to calculate mean biomarker values. . . . . . . . . . . Participant demographics divided by sex and then further by sarcope- nia status as defined by ALMI threshold. . . . . . . . . . . . . . . . . 38 38 39 40 42 45 x 6.2 Appendicular lean mass (ALM) values acquired by DXA and normal- 6.3 ized by height (ALMI) and BMI (ALM/BMI). The participants are further divided by sex and then sarcopenia status according to ALMI. Summary of all MRI results (CSA, FF, MTR, T2, FA and MD) sep- arated by muscle groups (Quad, Ham, Add, Sart, and total muscle cross-section). The data points are further divided by sex and then sarcopenia status according to the ALMI threshold. . . . . . . . . . . 6.4 Pearson correlation table of R-values between all MRI values of the total muscle cross-section and ALMI. . . . . . . . . . . . . . . . . . . 6.5 Pearson correlation table of R-values between all MRI values of the quadricep muscles and ALMI. . . . . . . . . . . . . . . . . . . . . . . 6.6 Pearson correlation table of R-values between all MRI values of the hamstring muscles and ALMI. . . . . . . . . . . . . . . . . . . . . . . 6.7 Pearson correlation table of R-values between all MRI values of the adductor muscles and ALMI. . . . . . . . . . . . . . . . . . . . . . . . 6.8 Pearson correlation table of R-values between all MRI values of the sartorius muscles and ALMI. . . . . . . . . . . . . . . . . . . . . . . . 6.9 Pearson correlation table of R-values calculated between all MRI values and ALMI using all collected data points from all muscles. . . . . . . 6.10 T-test results between groups divided by sex and sarcopenia status. Sarcopenia status was defined by ALMI (left) and ALM/BMI (right). Cells with
p-values < 0.05 are
highlighed
in green
, cells with
p-values <0
.1
are highlighted in
orange, and cells with p-values <0.15 are high- lighted in purple. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 46 47 47 47 48 48 48 49 xi 6.11 Wilcoxon rank test results between groups divided by sex and sar- copenia status. Sarcopenia status was defined by ALMI (left) and ALM/BMI (right). Cells with
p-values < 0.05 are
highlighed
in green
, cells with
p-values <0
.1
are highlighted in
orange, and cells with p- values <0.15 are highlighted in purple. . . . . . . . . . . . . . . . . . 7.1 Summary of MRI data from healthy middle-aged and healthy older population (non-sarcopenic according EWGSOP) from the study by Farrow et al. compared to MRI values collected from this study pop- ulation [2]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Summary of MRI values from non-sarcopenic and sarcopenic (accord- ing to AWGS) populations
from the study by
Fu
et al
. compared
to
this
study population
[3]. . . . . . . . . . . . . . . . . . . . . . . . . . 50 52 53 xii
List of Tables 2.1
Summary
of
medical
imaging techniques
used to asses sarcopenia biomark- ers. Table includes
advantages and disadvantages of each
technique followed by
the specific
sarcopenia biomarkers
that
can be measured. Data collated from review papers by Albano et al. and Chianca et al. [4] [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 xiii Abbreviations EWGSOP AWGS FNIH SDOC MRI DXA ALM ALMI ALM/BMI BMI SPPB TUG BIA
The European working group on sarcopenia in older people The Asian working group for sarcopenia The
Founation
for the National
Institute
of Health
Sarcopenia Definitions and Outcomes Consortium Magnetic Resonance Imaging
Dual Energy X-ray Absorptiometry Appendicular lean mass
Appendicular lean mass
index (ALM
/height2)
Appendicular lean mass
normalized by
BMI Body mass index Short physical performance battery Timed-up-and-go test
Bioelectrical Impedance Analysis xiv IMAT IMCLs CT US SCPT IDEAL-IQ RF TE NMR GRE PDFF FF CSA MTR MTI DTI FA Intermuscular adipose tissue Intramyocellular lipids droplets Computed Tomography Ultrasound Stair climb power test
Iterative decomposition of water and fat with echo asymmetry and least-squares estimation
Radio frequency
echo
time Nuclear
magnetic resonance
Gradient-echo sequence
Proton-density fat-fraction Fat-fraction
Cross-sectional area Magnetization transfer ratio Magnetization transfer imaging
Diffusion tensor imaging Fractional anisotropy
xv DWI
MD
RD AD BMC FM LM Diffusion weighted imaging Mean diffusivity Radial diffusivity Axial diffusivity Bone mineral content fat mass lean mass xvi Chapter 1 Intro duction
1.0
.1
Thesis
structure
The
structure
of this thesis is as follows
. The introduction
chapter
will establish the clinical context of frailty that will be the basis of the research presented in this thesis followed by the primary objective of this research.
The background section will
then
provide an overview of the
problems associated with defining sarcopenia, a foundation in muscle anatomy, and a summary of sarcopenia-related biomarkers that can be assessed with medical imaging techniques.
This is followed by a literature review of the current
methods to assess sarcopenia followed by an explanation of DXA and the MRI techniques employed in this study. The materials and methods chapter will outline the study protocol and explain how the data was acquired and processed. The methods of data analysis and
results of the MRI and
DXA data collected
will be
reported
in
the results
chapter
. A brief comparison of the study population with patient populations from other similar studies will validate this study as elderly and frail. The discussion chapter will go through the results of the biomarker analysis with explanations surrounding the observed biomarker relationships. Finally, the limitations of this study will be identified, followed by future steps for improvement of the study design. The conclusion chapter will comment on the implications of this study as well as final thoughts regarding the use of MRI and DXA in sarcopenia assessment. 1.0.2 Impact of frailty on society With the continuous development of society, life expectancy has experienced a no- ticeable increase mostly in part due
to health and quality of life
improvements.
The proportion of people
in the world
aged 60 years or older is projected to double from
11%
to 22
%
between 2000 and 2050
while
the number of adults
80
years
or
older
are expected to quadruple [6]. Unfortunately, with older age comes the development of health problems and the increase in prevalence of frailty [6].
Approximately 23% of Canadians over
the
age
of 65
are
considered
frail
, with that number predicted to increase up to 40% for the population over the age of 85 [7] [8]. Additionally,
the prevalence of
frailty
is
substantially
higher in women compared to men
[9]. While the term frailty lacks a universal definition, a 2013 consensus statement between six ma- jor international scientific societies agreed upon
the definition of frailty as, “a medical syndrome with multiple causes and contributors that is characterized by diminished strength, endurance, and reduced physiologic function that increases an individual’s vulnerability for developing increased dependency and/or death
” [10]. When exposed to stressors, either internally or externally,
frail older people are highly vulnerable to adverse health outcomes
[6]. With this susceptibility, people with frailty experience more frequent hospitalization
and length of hospital stay
with
a
seven-fold increase over non-frail older people [11]. Frailty
is a geriatric syndrome
defined
by the
natural decline in
muscle mass and function
caused by the natural aging process. Fortunately, this is not irreversible. An individual’s experience with frailty varies, and there is po- tential to reduce one’s level of frailty even in old age [6]. Longitudinal studies have reported up to
17.9% of older adults
improving
their frailty status
, which provides motivation to work towards better understanding the concept of frailty in hopes of de- veloping strategies to reduce frailty severity. The elderly population must be cautious as frailty may present as a disease state, which is recognized as sarcopenia [12]. 1.0.3
Objective The objective of this work is to
characterize skeletal muscle quality and quantity in a frail elderly population. MRI and DXA are each capable of measuring unique skeletal muscle biomarkers to describe both muscle quality and quantity. Measuring these muscle values is an opportunity to understand how sarcopenia presents within the body. Furthermore, understanding which imaging techniques and biomarkers are most sensitive to sarcopenia-related changes would allow for the development of an imaging protocol for sarcopenia diagnosis or early detection. Chapter 2 Background 2.1 Definition of Frailty and Sarcopenia We are in an aging population, the percentage of our population made up of older people is increasing. This introduces a collection of age-related conditions and con- cerns for the population’s health and well-being.
Up to 50% of people older than 85 years are estimated to be frail, and these people have a substantially increased risk of falls, disability and lower quality of life
[2]. Sarcopenia is
the
term given to
the
age-related reduction
of
muscle mass and function, known to cause frailty [12]. Rosen- berg first used the term sarcopenia
in 1989 to describe
the
progressive, generalized loss of skeletal muscle mass and
an
accompanying decline in
function with increas- ing age [13]. The name “
sarcopenia” is derived from the Greek “sarcos” referring to flesh and “penia,” a lack of
[14].
The
definition
of
sarcopenia has evolved and been long debated over the years. The generally accepted
criteria for sarcopenia include low muscle mass, low physical function
(based on metrics
such as gait speed
or grip strength)
and low muscle strength
. The decrease
in muscle mass is
often paired with
a reduction in muscle
quality associated with ageing [2]. The specific threshold values that constitute “
low physical function”, and “low muscle strength
”, make
up
a ma- jority of the debate. Various research groups and populations posit their own cutoff points for the physical and biological markers of sarcopenia. The specific biomarkers will be described and elaborated on later in this thesis. These sarcopenia definitions include [1]: The
European working group on sarcopenia in older people (EWGSOP
): The EWGSOP defines
probable sarcopenia by low muscle strength. Diagnosis is confirmed by additional documentation of low muscle quantity or quality
. If
low muscle strength, low muscle quantity/quality, and low physical performance are all
met,
sarcopenia is considered severe
[12].
The Asian working group for sarcopenia (AWGS
): The AWGS
recommends measuring both muscle strength (handgrip strength) and physical performance (usual gait speed) as the screening test
. If a patient falls below the cutoff val- ues for one or both criteria, and they display low muscle mass, they qualify as sarcopenic [15].
The Foundation for the National
Institute
of Health (FNIH): The FNIH
chose
to
avoid the term “sarcopenia” entirely due to confusion and lack of certainty in its definition. Rather, employing the terms “low lean mass” (defined by appen- dicular lean body mass) and “weakness” (defined by grip strength) to describe a patient’s health status [16]. Sarcopenia Definitions
and Outcomes Consortium (SDOC): The
SDOC con- cluded
that both weakness defined by low grip strength and slowness defined by low usual gait speed should be included in the definition of sarcopenia
[17]. A summary of the different definition of sarcopenia including the specific biomarker cutoffs for each definition are outlined in the table below in Figure 2.1 [1]: Figure 2.1: Table summarizing definitions
of sarcopenia according to
each
consensus group
with cutoff points for low muscle mass, strength, and physical performance
. Content adapted from review by Coletta et al. [1] 2.2 Skeletal Muscle A person’s physical capability relies on their range of motion and the amount of power they can exert. Whether actions be as simple as getting out of bed or maintaining one’s posture, or complex physical activity in a social or work setting, the mainte- nance of a person’s health depends on their functional independence [18]. Skeletal muscles are responsible for the body’s mechanical functionality, and as a result
have a
large
impact on quality of life
. Especially
in the
elderly
population
, where muscle quality and quantity is far reduced compared to their younger counterparts, the in- tegrity of their skeletal muscles determine their ability to perform daily chores and reduce their risk of injury and falls [14]. Making up approximately 40% of the hu- man body’s total weight,
muscle is mainly composed of water (75 %), protein (20 %), and other substances including inorganic salts, minerals, fat, and carbohydrates (5
%) [18]. The main cellular components are muscle fibers (myofibers) as well as adipocytes, fibroblasts, satellite cells, endothelial cells,
neurons, and Schwann nerve cells [2
] [19]. Skeletal muscle takes on a hierarchal structure (Figure 2.2). The largest structures are called muscle fascicles, which are made up of
muscle fibers, which are
then
made
up
of myofibrils arranged in parallel
. Myofibrils
are
then
further divided into myofilaments and sarcomeres
which are
arranged in series and
can be
ultimately broken down into structural proteins
[19]. The smallest unit of muscle is the sarcomere, made up thick myofilaments consist- ing
of myosin and thin
myofilaments composed
of actin, troponin, and tropomyosin
[19].
A
myofibril
is
a succession of sarcomeres arranged in series, with the boundaries of a sarcomere defined by two successive Z-lines [19]. The M line is where myosin filaments are anchored together, marking the central most location of the sarcomere. This center region where the M line is located contains only myosin filaments and is called the H-zone [20]. The A band is the larger region of the sarcomere containing the entire length of the myosin fibers which includes regions where actin and myosin Figure 2.2: Diagram of skeletal muscle hierarchy (
Image credit: download for free at http://cnx.org/contents/14fb4ad7-39a1-4eee-ab6e-3ef2482e3e22@8.119
) overlap.
The
I band encompasses
the
ends of two neighbouring sarcomeres and con- tains only actin filaments. During muscle contraction,
the H and I bands
will shorten while
the
A band will remain at a constant length (Figure 2.3) [20]. Myofibrils are the contractile organelles that contain billions of myofilaments in the form of sarcomeres and
each muscle fiber is made of thousands of myofibrils
arranged in parallel.
The size of a
whole
muscle is
predominantly
determined by the
Figure 2.3: Anatomy of a muscle fiber (
Image credit: download for free at http://cnx.org/contents/14fb4ad7-39a1-4eee-ab6e-3ef2482e3e22@8.119
) number and size
of
individual muscle fibers that make it up,
although pathological infiltration by fat and connective tissue may alter this relationship
. The cylindrically shaped skeletal muscle fiber ranges from 2cm up to 50cm in length with a thickness between 10 and 100 µm. The
fiber is surrounded by a
cell
membrane called the sarcolemma
.
Actin and myosin
make up
approximately 70 – 80% of the total protein content of a single
muscle
fiber. Myosin is the main molecular motor
while troponin
and tropomyosin are associated with the actin filament
, playing
very important roles in the activation process that leads to myofilament sliding and force generation
. Not all muscle fibers are created equally. There is heterogeneity throughout human skeletal muscles due to the various
biochemical, mechanical, and metabolic
properties
of individual fibers
[18]. This diversity
is
a very beneficial adaptation that allows for functional specialization within fibers based on
metabolic and mechanical demands
. Even
the capillary supply networks
depend on the fiber type so as to meet these specific demands [21].
Muscle fibers are
classified
into three main groups, type I, type IIa, and type IIb: Type I fibers
, also called slow oxidative fibers, are slow-twitching fibers. The small- est of the fiber types, type I fibers have slow contractile speed, low glycogen con- tent, and also fatigue at a slow rate. This makes fibers in this group well suited for endurance-type contractions, such as maintaining posture and marathon running [22]. Type IIa fibers are fast-twitching fibers also called fast oxidative fibers. They fa- tigue at an intermediate rate and are best suited for activity involving moderate movement over medium durations such as when walking or biking [22]. Type IIb fibers are also fast-twitching fibers but go by the name fast glycolytic fibers. This is because they primarily obtain energy from anaerobic glycolysis. With a fast rate of fatigue, fibers of this type are adapted for brief but intense movements such as sprinting and weight-lifting. These fibers have the largest diameter due to their high density of actin and myosin proteins [22].
The response of muscle fibers to stimuli such as denervation, aging, inactivity, and disease is
unique to each
fiber type. For example, more atrophy is noted in type II fast fibers than in type I slow fibers in conditions associated with muscle wasting such as cancer
[18]. In sarcopenic skeletal muscles,
type II (glycolytic
) fibers
become hypotrophic
resulting in
reduced
fiber
diameter, and become infiltrated with adipose tissue and
potentially fibrotic tissue
at later
disease
stages
[23].
Excess lipids are thought to ‘spill over’ to other
tissues,
especially skeletal muscles
, where they
accumulate as intermuscular adipose tissue (IMAT
), intramuscular
adipose tissue, and intramyocellular lipid droplets (IMCLs
) [24] [25].
Fat infiltration in skeletal muscle is
also called
myosteatosis, and is
an abnormal phenomenon that increases with ageing, negatively correlating with muscle mass, strength, and mobility [
26]. In addition to
myosteatosis, a loss
of muscle mass
due to ageing
can be
seen occurring
in both type I and type II
muscles. More accurately, sarcopenic muscle experience an accelerated rate of atrophy
of type II fibers
, resulting in
the
ratio
of slow
to
fast
twitch
fibers
within the muscle favouring the slow-twitch fibers.
The proportion of total muscle cross-sectional area occupied by type I fibers
increases, resulting
in
a
“switching”
of type II to type I fibers
also referred
to
as a transition from fast to slow muscle [27] [28]. When using medical imaging such as MRI
to assess muscle quantity and quality
in muscle disorders
such as
sarcopenia, the thigh muscles are often the anatomical region of interest. This is because muscles of the thigh appear to be more sensitive to aging-related changes such as intramuscular fat infiltration [29]. Of the thigh muscles, the quadricep muscles are
relatively more affected by aging than other thigh muscle groups
, allowing for the detection of sarcopenia
earlier than whole-body muscle loss
[30] [31].
A study by
Maden-Wilkinson
et al. reported that the
quadriceps
of
older men were 32% smaller than those of young men, and
those of
older women were 28% smaller than younger women
due to age-related atrophy [30]. To measure muscle properties such as muscle volume and intramuscular fat, the best estimates were obtained at 50% femur length [29]. This allows for consistency when trying to scan the same anatomical region between different study participants and different study site locations, even when utilising as few images as a single MRI slice. 2.3 Muscle Biomarkers There are many properties
of skeletal muscles
that
can be assessed with the
help
of
medical imaging modalities. Scientists and medical professionals find value in recording quantitative and qualitative properties of muscles, especially those in the diseased state, to develop diagnostic thresholds, and track disease response to thera- peutic intervention. In sarcopenia research, the most often used imaging techniques are:
magnetic resonance imaging (MRI), dual-energy x-ray absorptiometry (DXA), computed tomography (CT), and
ultrasound (US). Each technique has its own unique capabilities regarding the variety and accuracy of biomarkers that can be detected. A summary of the biomarkers measurable by each modality is presented in Table 2.1. As the focus of this thesis, MRI and DXA will be explored in detail in later chapters. T able 2.1: Summary of medical imaging techniques used to asses sarcopenia biomarkers. T able includes
advantages and disadvantages of each
technique followed by
the specific
sarcopenia biomarkers
that
can be measured. Data collated from review papers by Albano et al. and Chianca et al. [4] [5] Technique Advantages Disadvantages Sarcopenia Biomarkers DXA • Inhomogeneity of results • Affected by hydration status among different brands • Very low radiation exposure • Bidimensional data • Still requires some radiation exposure • • Wide availability • Reproducibility Simplicity • Validated cut-off values • Cheap • •
Appendicular lean mass (ALM) Appendicular lean mass index • (ALMI) Appendicular lean mass
/ BMI (ALM/BMI) 13 MRI • Cross-sectional imaging •
Accurate and reproducible • No
radiation exposure •
High
spatial
resolution
•
No validated cut-off values for sarcopenia • Expensive
• Low availabilty •
Long
acquisition time • Complex
post-processing
• •
Muscle cross-sectional area (CSA
) • Muscle
volume
• Muscle
fat fraction
• T2 relaxation time (experimental) Fractional anisotropy (experimen- • tal) Magnetization transfer ratio (ex- perimental) CT • • Cross-sectional imaging • Accurate and reproducible
Fast acquisition • Widely available • High spatial resolution
• No validated
cut-off values
for sarcopenia • • High radiation exposure Expensive • Complex post-processing • •
Muscle cross-sectional area (CSA
) • Muscle
volume
Muscle
fat fraction
(experimental) US • Portable • Cheap •
No radiation exposure • Real-time imaging
• Fast acquisition • No validated cut-off values
Muscle cross-sectional area (CSA) • • • Muscle volume Muscle
thickness • Low reproducibility
for
sarcopenia • Poor accuracy • Operator skill dependent
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering Chapter 3
Literature Review
3.1
Frailty and Sarcopenia Most definitions of sarcopenia use 3 criteria to identify sarcopenia status:
muscle mass, muscle strength, and muscle function
. Some working
groups
such as the FNIH co-opt alternative terminology such as “weakness” and “slowness”, while others such as the SDOC only recognize 2 of the 3 criteria. Regardless of the disease definition, these groups recognize the wide variety of methods available to assess sarcopenia biomarkers depending on the clinical or research setting present. Researchers and clinicians alike have worked diligently to validate these tests for accuracy, applica- bility, and availability. Diagnostic thresholds or cut-off values to classify a patient as having low muscle strength, mass, or function by each sarcopenia definition are summarized in Figure 2.1 in the Background chapter.
3.2 Assessment of Muscle Strength
Grip
Strength is the most
common tool
used
to
measure muscle strength
[32]. This biomarker of muscle strength requires relatively affordable resources,
does not require a trained specialist
, and has minimal risk on the patient. Hand grip
strength is an
affordable,
reliable, and
well validated
tool
that
can be used in
both
clinical practice and
a
research
setting.
The Jamar dynamometer is recognized as the gold standard
device to measure hand grip strength. To standardize testing conditions, patients are recommended to be seated
in a standard chair with
their
forearms resting flat on the arms
of
the
chair [33]. Ideally, 3 measurements from each arm should be taken and the highest measurement between all 6 trials should be reported, but a recent review paper by Voulgaridou et al. recognizes variation between testing protocols, with most research papers only running 3 trials on the dominant hand [34]. Alterna- tive testing devices are also available for patient populations who cannot properly use the standard dynamometer. A pneumatic dynamometer instead has
patients try to squeeze rubber balls
with the same testing protocol previously mentioned [33].
Low grip strength is correlated with clinical outcomes such as
falls, disability, and impaired health-related quality of life
[35]. Isometric handgrip strength has also shown
good correlation with leg strength, lower extremity power, knee extension torque and calf cross-sectional muscle area
[33].
The Chair
stand
test also called the “Five-Times-Sit-to
-
Stand test” has been
sug- gested as
an alternative measure for muscle strength and can
also
be a substitute for leg muscle strength
. This assessment
measures lower limb strength and balance
and requires strength, coordination, and endurance [33]. In this test, participants are
asked to cross their arms on their chest. The evaluator
provides
the following
stan- dardized
instructions: “I want you to stand up and sit down 5 times as quickly as you can when I say ‘Go’.” Timing begins when the evaluator says “Go” and stops when the participant is fully standing for the fifth time
[35]. Currently, only
the
EWGSOP2 recommends the chair stand test as an assessment tool for muscle strength [12]. While
there is evidence suggesting that the chair stand test is not directly linked to muscle mass, the test has proven effective in gauging overall physical performance and
shows correlation with
a lower prevalence of sarcopenia when compared to grip strength
tests [34]. Knee extensor strength tests can measure lower limb muscle strength, such as that of the quadriceps, using commercial dynamometers [33]. Lower limb muscle strength is proven to be able to predict impairment
in various functional tasks
, in- cluding
walking, rising from a chair, and climbing stairs
[36]. A study by Michal et al documents
that sex-specific isokinetic strength of the knee flexors and extensors at 60◦/s can detect sarcopenia and frailty in individuals
[37].
Knee extensor strength tends to decline more quickly with aging than handgrip strength
which makes
the test valuable for sarcopenia
diagnoses and
cut-off points provide objective markers to identify older adults
’ risk of future mobility limitations [38] [39]. Unfortunately, most sarcopenia definitions don’t set diagnostic thresholds which may be
due to a lack of consistency in
measurements
across studies
[37]. 3.3 Assessment of Muscle Function
Gait speed is the most widely used tool in clinical practice for the assessment of physical performance
/function. It is
employed by almost two-thirds of clinicians that assess physical performance
and has defined cut-off values within almost every defini- tion of sarcopenia
except for the FNIH group which
does
not include muscle function
in their definition [33] [16].
Gait speed is
a
feasible, reliable, and safe
test that requires no specialized equipment [40]. It
is a good predictor of adverse outcomes including disability, cognitive impairment, falls, and mortality
[41] [42]. The gait speed test re- quires a patient to move at their usually walking speed most often over a set distance of 4 or 6m, with the test administrator using a stopwatch to measure gait timing [43] [44]. Some studies have assessed gait speed using distances as short as 2.4m (8-ft) [45]. To make these results comparable to those recorded over longer distances, adjustment equations such as the one published by Guralnik et al. can be used to calculate the gait speed that would have been observed over 4-6m [42] [45]. The Timed-up-and-go (TUG) test
evaluates a well-recognized series of
maneu- vers
used in
a patient’s
daily life and
correlates well with more extensive measures of balance, gait speed and functional abilities
[46]. Patients
are asked to rise from a standard chair, walk to a marker
3m
away, turn around, walk back, and sit down
on the chair
again
[46]. Benefiting from its simplicity and not requiring any specialized equipment, the TUG test has been shows to be a predictor
of sarcopenia in
elderly
hospitalized patients
[47].
The
short physical performance battery (SPPB) is a
3-part
test
used
to
assess
physical performance
in older adults. The components include
a chair stand test
, a
gait speed
test,
and a balance test
where a patient receives a score out of 4 for each component resulting in a total score out of 12 [33]. The battery measures
balance, lower body muscle strength, and mobility
, and a
lower
score predicts
lower quality of life, loss in mobility, disability, and mortality
[48].
A change in
a patient’s
SPPB score by
a single
point
has been
related to a substantial change in mobility
[49]. Overall, the SPPB has shown usefulness in identifying
community-dwelling individuals at risk of
frailty or
functional disability
in a short amount of time without the requirement of expensive equipment or expertise [48]. The stair climb power test (
SCPT) is
another
simple way to measure a
pa- tient’s
leg power
. Patients are
instructed to climb a flight of
8
steps as quickly as possible
without using handrails unless necessary [50]. The test is repeated a second time, and the average between both recorded times is then inputted into equation 3.3.1 below to calculate stair climb power [50]. Although this is a clinically feasible and inexpensive test, it may prove difficult for vulnerable patients in an ambulatory care setting [51]. In response, a shorter 4-step version of the SCPT has been tested and
shows scientific promise as a valid and reliable leg power measurement among community-dwelling older adults
[51]. (Body Weight in kg) ∗ (9.8m/sec2) ∗ (
stair height in meters
) Power (Watts) = (
stair climb time in seconds
) (3.3.1)
The
SARC-F questionnaire is a self-reported questionnaire to screen for sarcope- nia risk recommended by the EWGSOP [12]. It assesses
five components based on features and consequences of sarcopenia: strength, assistance walking, rising from a chair, climbing stairs, and falls
[52]. This pioneer screening tool is an affordable and convenient option
in community healthcare and clinical settings to assess the severity of sarcopenia
[52]. While not necessarily diagnostic, it can help clinicians determine if additional assessments are required. 3.4 Assessment of Muscle Mass
Lean mass (LM) or lean body mass
can be
measured
using bio-electric
impedance analysis (BIA
). BIA
is a
non-invasive
method which estimates the volume of fat and lean body mass
of a patient
based on the relationship between the volume of a
con- ductor (in this case biological tissue with high water content such as muscle) and its electrical resistance [53]. An equation is used to translate the results into
an estimate of
body
fat
volume
and lean body mass
, but
the
accuracy of BIA has been challenged, with reports of it overestimating muscle mass and underestimating fat mass [33].
It is
also
important to note that this
method
is
heavily
dependent on the
testing conditions such as temperature, humidity, and skin condition [54]. For reliable BIA measure- ments hydration status, food intake, and exercise must be controlled [54]. Despite these concerns,
the accuracy of BIA in sarcopenia diagnosis has been validated
, the technology is inexpensive, does not require expertise,
and is relatively easy to use in
a
clinical
or
ambulatory
setting [15].
Appendicular lean mass (ALM) is the
most researched
and
evaluated imaging- based muscle mass biomarker in sarcopenia assessment [15]. Measured
using dual- energy x-ray absorptiometry (DXA
), it
is
the sum of the non-bone and non-fat mass of the four limbs
[33].
To adjust for
patient’s
body size
, ALM is often scaled for height and called appendicular lean muscle index (ALMI = ALM/height2) or it can be scaled for BMI (ALM/BMI) [33]. DXA operates by characterizing body tissue as belonging to one of 3
compartments: lean mass, fat mass
, or
bone mineral density
. While
it is a relatively
quick
and cost effective technique
as far as imaging goes, it is unable
to assess the quality of muscle mass
or
detect adipose tissue within muscles
[34]. DXA results are also recognized to be machine-dependent, with significant variability in results collected from
different manufacturers and models, limiting the comparability of
data collected
from different
research sites [34].
Muscle volume and muscle cross-sectional area (CSA
) are most often col- lected using MRI or CT, both of which are considered
the gold-standard approaches for evaluating body composition
[34]. Both
MRI and CT
are able
to distinguish fat from other soft
tissue within
the body
, quantifying both the
quantity and quality
of
muscle
[34].
Cross-sectional
skeletal muscle
area
at the L4-L5 abdominal region
is highly correlated with total body skeletal muscle volume
, and skeletal muscle index of that same region
has been associated with
better
survival in older adults
[55] [56]. Additionally, muscle cross-sectional area measured at the mid-thigh has been associ- ated
with lower extremity performance in
older
men and women
[57].
The
greatest drawback
of
these technologies is the lack of portability, the high costs associated with these imaging techniques, and high radiation exposure from CT [34]. Anthropometric measures such as calf circumference and midarm muscle circum- ference
have been shown to reflect
older patients’
health, nutritional status
, perfor- mance,
and survival
, as well as correlation with ALM [58]. The greatest benefits from anthropometry is that it is an extremely portable, easy to use, and inexpensive tool suitable for use in a primary care setting [58]. Chapter 4 Imaging Techniques in Muscle Assessment
4.1 Magnetic Resonance Imaging (MRI) 4.1.1
T2 mapping In the context of MR imaging, there are two inherent properties of protons that are observed, T1-relaxation and T2-relaxation. T1 relaxation (
longitudinal relaxation or spin-lattice relaxation
) describes
the
process
of protons
returning
to their
equilib- rium orientation after experiencing a 90 degree RF pulse.
T2 relaxation (transverse relaxation
or
spin-spin relaxation
) describes
the process
by
which the
transverse mag- netization of a system decays as a result of protons dephasing [59]. Different biolog- ical material will have unique T1 and T2 time values depending on their chemical structure. This means that tissues can be visually differentiated in an MR image by techniques that cause T1 or T2 effects to have a greater impact on the signal collected. This is called T1-weighting or T2-weighting. It’s important to note that with any weighting applied to an MRI scan, the resulting MR signal collected still represents a combination of T1, T2, and proton density properties of the tissue. To calculate only
the T2 relaxation time of
protons within
tissue
, a process called T2 mapping is used [59]. T2 mapping begins by using a spin echo sequence to collect multiple acquisitions at different echo times (TEs). For every pixel in the MR image, the signal data from all TEs are fit to the following mono-exponential signal equation [60]: S = S0 exp T2 −T E (4.1.1) By fitting the data, the T2 value of the tissue at each voxel location can be solved for. After processing all of the spin echo data from all echo times, a T2 map is gener- ated whereby every pixel
represents the
average
T2 time of the tissue
present in
that
voxel of space (see Figure 4.1). In biological tissue, especially skeletal muscle, the greatest contributor to T2 relax- ation time is water. As a result, muscle water T2 has strong implications
as an imaging
biomarker
for pathophysiological changes of skeletal muscle tissue
[61].
Muscle
water T2 is considered an indicator for
water mobility in the tissue
, which
is affected by circumstances such as inflammation, myocyte swelling, cell necrosis, denervation
, and even high intensity exercise [61]. Quantitative T2 mapping provides promising results for evaluating inflammation in neuromuscular diseases as well as muscle changes due Figure 4.1: Sample calculation for a single pixel of a T2 map from T2 decay curve displayed as a semi-log plot. A sample T2 map of a healthy male adult (top) compared to on of a frail male participant from this study (bottom). to aging, with increased T2 values identified in frail/prefrail subjects affected by sar- copenia [62]. Aging-associated changes in
water T2
may
be due to changes in
muscle fiber
type, while heterogeneity
in muscle T2 values may
be associated with
muscle tissue
disorganization caused by fibrotic replacement
[62]. 4.1.2 Fat/Water Imaging (IDEAL IQ) The main sources of hydrogen within the body are water and fat, which are respon- sible for a majority of the signal that makes up an MRI image. Hydrogen nuclei in water molecules have a higher resonant nuclear magnetic resonance (NMR) frequency than hydrogen
in fat molecules. Within a homogeneous field, the difference in
reso- nant
frequency between the main peak of fat molecules and that of water molecules is 3.4 ppm
, which translates to a 435 Hz difference between peaks within a 3.0T B0 field. The Dixon technique, invented by Dr. William Thomas Dixon in 1984, exploits
the difference in
NMR
chemical shifts between water and fat to separate
the MR signal
into separate
water-only
and
fat-only images [63]. Because
water and fat protons precess at different rates
according to their resonant frequencies, there will be moments when the spins from both groups will be in-phase and out of phase from one another. By acquiring the MR signal at different TEs in a gradient-echo (GRE) sequence, we can catch these in- and out-phase signals which can then be combined mathematically to separate the fat and water signals [64]. Rather than collect the minimum of 2 TEs as is done in 2-point Dixon method, additional acquisitions may be collected to correct
several confounding factors such as T1 bias, eddy currents, noise
bias,
and T2∗ effects
[65] [66]. A more sophisticated approach is
called Iterative Decomposition of water and fat with Echo Asymmetry and Least squares estimation (IDEAL-IQ
) [65].
In IDEAL-IQ, images are acquired at multiple echo times, and an iterative least-squares decomposition algorithm is
em- ployed
to
simultaneously
solve for a fat fraction map, a water fraction map, and an R2∗ map
[67].
By incorporating an R2∗ map into the algorithm, IDEAL-IQ accounts for T2∗ effects/field inhomogeneity, and yields a proton density fat fraction (PDFF) not confounded by iron overload
[67]. PDFF or fat fraction (FF)
is defined as the ratio of protons from fat to the total
number
of protons from fat and water
within a voxel of tissue, and is the most com- mon biomarker of skeletal muscle quality [68] [69]. FF is a highly accurate biomarker of muscle composition, shown to measure fat infiltration and fatty replacement of muscles in dystrophic patients [70]. Research has also shown correlation of muscle strength with both intermuscular adipose tissue and PDFF of quadriceps muscles in healthy adults [71]. In the context of frailty and sarcopenia, positive correlations between PDFF and age and between intramuscular adipose tissue and frailty have been published [62] [72]. 4.
1.3 Diffusion Tensor Imaging
(DTI)
Diffusion tensor imaging (DTI) is an
extended application
of diffusion weighted imag- ing (DWI
), which relies on the attenuation of MR signal due to Brownian motion. Brownian motion is the term used to describe the seemingly random motion that wa- ter molecules experience. The degree of restriction that water molecules experience depends on the structure of their environment. In an open environment, such as water in a glass, uninhibited motion is present in all directions, this free diffusion is called isotropy [73]. Alternatively, in an environment that restricts or favors
the diffusion of water along a specific
axis such as water in a thin straw, the water molecules are described as being anisotropic. In DWI, a diffusion sensitizing gradient is
applied on either side of the 180
degree refocussing
RF pulse
of
a spin echo sequence
. Movement
of water molecules
along
the
same
direction
as
the diffusion gradient
results in atten- uation of signal proportional to how uninhibited motion is in that direction. DWI is done by repeating the spin-echo sequence with diffusion sensitizing gradients along three orthogonal directions (e.g. x, y, z). DTI, on the other hand, requires at least 6 diffusion sensitizing gradient directions, and is often acquired using 30 or more dif- fusion encoding directions. The DTI approach is far superior to DWI as the former is quantitative and rotationally invariant. The DTI data is subsequently compiled to calculate the direction and degree of anisotropy experienced by water molecules throughout the tissue [74]. Tissue diffusion, at each voxel location, is described best as a rank-2 tensor, which is a 2D matrix of values. From the tensor, diffusion can be characterized using six parameters, including three eigenvectors (E⃗1,E⃗2,E⃗3) to quantify the direction, and three eigenvalues (λ1,λ2,λ3) to quantify the amount along three orthogonal axes [75]. An ellipsoid can be used to represent the probability that a water molecule will move in a given direction within a voxel according to the tensor [75]. The direction of maximal diffusion, called axial diffusivity (AD), is represented by λ1 and is
referred to as the
principal
diffusion direction
, it is
the
largest
of the
3 eigenvalues where λ1 ≥ λ2 ≥ λ3 ≥ 0 [76].
Radial diffusivity
(RD)
is the average of the
two smaller
eigenvalues (RD = (λ2 + λ3)/2
). In the context of muscle tissue, AD describes the diffusivity along the length of muscle fibers while RD refers to the diffusivity perpen- dicular to muscles fibers [76]. The mean diffusivity (MD) describes the overall diffusivity within a voxel regard- less of direction. It is calculated by the following equation 4.1.2 [77]: MD = (λ1 + λ2 + λ3) 3 =λ (4.1.2) Fractional anisotropy (FA), calculated using equation 4.1.3, is the scalar
used to
describe
the degree of anisotropy
within a voxel.
FA is
a value between 0 and 1, where 0 represents complete isotropy (no restriction to diffusion at all) while 1 describes complete anisotropy (diffusion is only possible along a single axis) [77]. FA = √2 √
3 ((λ1 − λ)2 + (λ2 − λ)2 + (λ3 − λ)2) (λ12 + λ22 + λ32
) (4.1.3) √ DTI parameters such as FA and MD can be used to describe the tissue structure in skeletal muscle fibers. It is generally agreed upon
that the principal eigenvector aligns with the local muscle fiber orientation, however
, there is still discussion regard- ing what is described by the second and third eigenvectors [78]. One hypothesis is
that the difference in λ2 and λ3 reflect the cross-sectional shape of the myofibrils
[79]. Studies that combined histology with the tracking
of DTI parameter changes due to physiological or pathological conditions
support this suggested model, as
mean cell diameter was found to be positively correlated with λ2 and λ3. However
, it must be noted that cell swelling in some pathologies was often
accompanied by an increase of extracellular fluid
[80] [81]. There also appears to be conflicting results between research papers studying the effects of ageing on DTI parameters. While some papers did not detect changes in DTI values due to ageing, others noticed
an increase in FA and decrease in MD
with increasing age. These changes could be explained by a reduction in
fiber diameter due to age-induced atrophy
or due to
changes in type 1 and
type 2
fiber composition
[80] [82]. 4.1.4
Magnetization Transfer Imaging (MTI) Magnetization transfer imaging (MTI) is
an MR
technique that
generates contrast through
the
exchange of energy
between
free and bound protons. Conventional MR imaging relies on signal generated by “free” or
mobile protons that have sufficiently long
T2
relaxation times
(greater than 10ms).
Less mobile protons
such as those as- sociated with macromolecules or proteins have T2 values that are too short (i.e. less than 1ms). However, in 1989 Balaban et al. discovered that these “bound” macro- molecular protons are coupled with nearby mobile protons (such as those in liquids) and influence their spin states through energy exchange [83] [84]. Using
an off-resonance radio frequency pulse
, it is possible
to saturate the
bound
protons
, and these saturated spins can be transferred to the pool of free protons. We can then collect an image whereby the MR signal is influenced by
the rate of exchange between the free and bound
proton populations [84]. MTI can function as a quantitative imaging technique by calculating the magne- tization transfer ratio (MTR). The MTR is a percentage value between 0 and 100% that attempts to quantify the amount of energy exchange that occurs based on the two-pool model within each voxel of tissue according to the following equation: MTR = (MToff − MTon) M Tof f (4.1.4) In skeletal muscle MRI,
a lower MTR reflects a decreased quantity and quality of the muscle protein content
[85] [86]. Furthermore, a study by Schwenzer at al. found a decreased MTR
in older
subjects
compared to young men
, indicating
an age-related change in muscle
protein content in healthy adults [85]. 4.2
Dual-energy X-ray Absorptiometry (DXA) Dual energy x-ray absorptiometry (DXA) is
the most
used imaging technique for body composition
analysis [5].
It consists of a whole-body scan
using x-rays of
two different energy
levels (
40 and 70 keV
). DXA relies on
a
three-compartment model, where tissue is classified under one of three categories according to the degree of
X-ray attenuation: bone mineral content (BMC), lipid
(“
fat mass
” or
FM), and lipid-free soft tissue (“lean mass
” or
LM
) [87].
High density tissues (such as bone) attenuate the x-ray
beams
more than low-density tissues (such as soft tissues
) and
based on the
ratio of attenuation between the two different x-rays, lean mass is discriminated from fat mass [87]. Difficulty arises in areas where the x-ray beams pass through bone, roughly
40 – 45% of the total
scan
area
. Because
DXA is
only able to differentiate between bone and
soft tissue (which
encompasses
both FM and LM
), a workaround is
to calculate the exact amount of FM and LM
based on information from pixels immediately adjacent to the bone which don’t contain any bone tissue themselves [88].
LM measurement is an estimation of all non-fat/non-bone tissues
, and the sum of LM from
the upper and lower limbs
of a patient
is defined as appendicular lean mass
(ALM), which is the main DXA biomarker used to measure muscle quantity. Often- times, to correct for height-based differences, ALM
is indexed
by
height and
defined as
appendicular lean mass index (ALMI = ALM
/height2) [5]. Strong correlations have been reported in the literature
between FM and LM
assessed
by DXA
and
the
equivalent
adipose tissue
and skeletal muscle tissue assessed by MRI and CT [88]. Most definitions of sarcopenia have established thresholds to define low muscle mass using DXA-derived ALM or ALMI (refer to Figure 2.1). While a common concern surrounding the use of DXA
is the exposure
of patients
to ionizing radiation, the radiation dose
from a full body DXA scan is roughly
5 µSv, which is lower than the
average amount of
background
radiation a person experiences (6.7 µSv a day) and orders of magnitude lower than even the effective dose of a CT head scan (2 mSv)[89]. One drawback of DXA is that it is unable to quantify intra- muscular adipose tissue which is a key muscle quality parameter in ageing research. Furthermore, DXA results are sensitive to a patient’s condition including body thick- ness, hydration status, and pathology that affects water retention including kidney or liver failure [5].
DXA may overestimate muscle mass in patients with extracellular fluid accumulation
as it is unable to differentiate water from actual LM tissue [5]. Fi- nally, in obese patients, DXA
may overestimate
thigh
muscle mass and underestimate
trunk and thigh
fat mass
[5]. Chapter 5 Materials and Metho ds 5.1 Study Design This imaging study was performed as
part of a larger
community-based
study
mea- suring
the effects of
exercise
on
frailty. The main study is a randomized control trial (RCT) examining if a 4-month
frailty rehabilitation program is an effective community-based intervention to promote healthy aging
. 5.2 Participant Recruitment Participants were recruited from Hamilton and the nearby community by means of doctor referrals, newspaper articles, radio advertisement, and word of mouth. In- formed,
written consent was obtained
for all subjects
prior to participation in the study. Participants were
initially screened over the phone by a researcher to ensure they meet the study’s inclusion/exclusion criteria. Participants must be community- dwelling (not in long-term care); ≥
65 years of age; able to ambulate
25 metres
with or without
a
walking aid; at high risk for mobility disability/functional limitations
as assessed by the John Morley FRAIL scale (score of >2); have medical clearance from referring clinician, or for self-referrals,
medical clearance from
family
physician to
safely
participate in exercise
and take oral nutritional supplements; and
can arrange transportation to the YMCA up to
2 times per
week
. Participants were excluded if they are:
unable to speak or understand English; currently attending a group
exer- cise
program; currently in a drug optimization study/program; currently taking
oral nutritional
supplements daily
; cognitively impaired
where they may have difficulty
fol- lowing
two step commands in group exercise
[90];
receiving palliative/end of life care
; terminally ill; have
unstable angina or
unstable
heart failure
; have another household member enrolled in the study; or are
unable to attend for more than 20% of trial duration
. A randomly selected sample of study participants were invited to partici- pate in this imaging study. Those participants were further screened over the phone to ensure they meet appropriate health and safety criteria for anyone undergoing an MRI and DXA scan. In addition to the previously outlined inclusion criteria, partic- ipants must be able to arrange transportation to St. Joseph’s Healthcare Hamilton hospital; and must have passed both MRI pre-screening and in person MRI safety screening administered by an MR technologist. In addition to the previously out- lined exclusion criteria, participants will not receive an MRI scan if their weight is greater than 250lbs (∼110kg); if they are extremely claustrophobic, or if they have a pacemaker. Participants with hip or knee implants were also excluded for the sake of preserving MRI data quality. All participants of the exercise study underwent a baseline assessment to collect anthropometric measurements including height, weight, age, and sex. A total of 13 participants from the total study received MRI and DXA scans and were analyzed. The demographic data of the analyzed participants used in this study is outlined below in Figure 5.1. Figure 5.1: Baseline demographics and characteristics of imaging study population. 5.3 Experimental Design All MRI scans
were performed
on
a GE Discovery MR750
3.0T MRI (
General Elec- tric, Healthcare, Milwaukee, WI
) scanner using a large GE 16-channel receive-only flex coil. In preparation of their MRI scan,
participants were asked to refrain from
strenuous activity, consuming
alcohol, or
consuming caffeine
24 hours prior to
their scan date [91] [92] [93]. Participants rested on a hospital bed located in a separate room within the imaging research center for 30 minutes prior to being scanned. After resting, participants were transferred to the MRI bed by a researcher with the aid of a wheelchair. The purpose of this preparatory resting is to allow for the normal- ization of muscle compartment size and blood flow which has shown to affect MRI signal in BOLD and DTI techniques [94]. On the MRI scanner bed, the large flex coil was wrapped around the thigh
of the participant’s
dominant leg.
The participant’s
legs
were
immobilized with a strap across both thighs and sandbags on both ankles to prevent involuntary movements that would otherwise result in motion artifacts during scanning. 5.4 MRI Protocol The total MRI scan time was 1hr 15min. The entire scanning protocol consisted of the following scans: 1. Localizer scans: 3-plane fast spin echo (FSE); 10 slices 256 × 256 matrix (in each of 3 orthogonal planes), 5.0mm thickness, 10mm slice gap, TE/TR = 78/1444ms. The localizer scan uses the MRI’s body coil for transmit/receive to check if the entire length of the participant’s femur is within the machine’s FOV. From these images, the mid-
length of the femur
is measured
as the
halfway point
between the greater trochanter and the lateral condyle
of
the
femur.
The
center slice for all MRI sequences in this protocol are placed at the mid-length of the femur (Figure 5.2). Figure 5.2: Localization of the mid-thigh. The length of the femur is measured according to the bony landmarks using a coronal scan. The center slice is placed at the midlength of the femur and is oriented along the axial plane. This becomes the reference for all future acquisitions. 2. Clinical scans and anatomical reference: Routine clinical scans were in- cluded to check for sub-clinical or confounding pathology. Each of these scans had a resolution of 512x512 and
consisted of
25
slices with a slice thickness of
4.0mm. Clinical scans included a coronal T1w with TE/TR = 7.46/646ms, a 2D axial proton density-weighted fat suppressed sequence (FrFSE) with 30.8/2374ms, and an axial
short tau inversion recovery (STIR) sequence with
TE/
TR
= 44.7/6780ms. The 2D proton density-weighted (PD)
image was used as the anatomical reference
during processing and analysis of all MRI data. 3. Fat/Water Imaging:
Fat quantification was performed using iterative
de- composition
of water and fat with echo asymmetry and least-squares
estima- tion (
IDEAL
-IQ). 32 slices, 256 × 256 matrix, 4.0mm slice thickness, TE/TR = 4.944/10.65ms. A separate fat and water image is produced for each participant which is then processed by the MRI machine into a fat fraction image (Figure 5.3). Figure 5.3: Fat fraction image produced by the IDEAL IQ sequence. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). 4.
Diffusion Tensor Imaging (DTI): Spin-echo echo-planar imaging (SE EPI) sequence
with diffusion gradients applied along 60 directions. 33 slices, 64 x 64 matrix, 4.0mm slice thickness, TE/TR = 47/6000ms, b=350 s/mm2.
The diffusion of water
throughout skeletal muscle
is
described
by a set of
eigenvec- tors and eigenvalues summarized by the parameters fractional anisotropy (FA) (Figure 5.4), medial diffusivity (MD) (Figure 5.5), and radial diffusivity (RD). These values may paint a picture of muscle fibre structure and orientation ac- cording to how the fibers influence the Brownian motion of water [95]. Figure 5.4: Fractional anisotropy (FA) map produced by the dtifit function from FSL suite using DTI data. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). Figure 5.5: Mean diffusivity (MD) map produced by the dtifit function from FSL suite using DTI data. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). 5. B0 mapping: 2D double gradient-echo. 32 slices 64 x 64 matrix, 4.0mm thickness, TE/TR = 3.4/400ms. To correct EPI data such as DTI images which are susceptible to B0 inhomogeneities. 6. Magnetization transfer: 2D spoiled gradient recalled echo. 33 slices, 256 x 256 matrix, 10mm thickness, TE/TR = 5.1/50ms, flip angle = 70 degrees. An unsaturated image is collected, followed by a saturated image with a frequency offset of 1.5kHz for the saturation pulse. Figure 5.6: Magnetization transfer ratio (MTR) image produced from saturated and unsatureated MT image. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). 7. T2 map: A T2-weighted (T2w) multi-echo sequence with 8 TEs. 26 slices, 256 x 256 matrix, 4.0mm thickness, TE/TR = 5.65, 11.3, 16.96, 22.61, 28.26, 33.91, 33.56, 45.21 /1200ms. Fitting the signal data
from multiple echo times
to
a mono-exponential decay
function allows for the calculation of T2 within the tissue on a pixel-by-pixel basis (Figure 5.7). T2 is an intrinsic property of tissue which describes the presence of water within muscle as well as oedema and inflammation. Figure 5.7: T2 relaxation time map (T2 map) calculated from multi-echo sequence images. Sample slices taken from a 79 year old man with an ALMI = 9.01 kg/m2 (A), and an 86 year old female with an ALMI = 5.01 kg/m2 (B). 5.5 DXA Protocol Participants underwent a full body DXA (Hologic Discovery A, Hologic, Inc., Marl- borough, MA, USA) at the Hamilton Osteoporosis Diagnostic Services Inc. (HODSI). The same experienced DXA technologist performed the full-body scan for all partici- pants, where they were laid supine on the DXA table. Participants were scheduled to have their DXA scan within a week of their MRI scan. The majority of participants had their DXA and MRI scans on the same day. 5.6 MRI data preprocessing and registration The following steps were followed to preprocess and register the MRI data of each participant in preparation for data analysis: MRI scans were downloaded in
Digital Imaging and Communications in Medicine (DICOM) format and
then promptly converted into
Neuroimaging Informatics Tech- nology Initiative (NIFTI) format
for convenience in data processing. Thigh muscles were manually segmented using FSLeyes from the FSL suite of MRI processing and analysis software [96] [97] [98] [99]. These segments were created
by drawing a region of interest (ROI
) using
the
center slice
of the
iterative decomposition of water and fat with echo
asymmetric
and least-squares estimation (IDEAL IQ
) out-phase image
due to
the dark outline present at the boundary between muscle and surrounding adi- pose tissue. When drawing ROIs, the muscle boundaries were just barely excluded. These ROIs
were used to
calculate
cross-sectional area
(CSA)
for each
muscle group by multiplying
the number of pixels
within an ROI
by the
dimensions
of
a
pixel
. All 4 muscle ROIs were then combined together to create a ’total muscle area’ mask used to calculate all ’total’ average biomarker values. Before applying any of the muscle ROIs to calculate mean biomarker data, the muscle ROIs were eroded by 4 pixels to avoid signal crossover and partial voluming effects from adipose tissue along the muscle boundary (see Figure 5.8).
The fat fraction image was
automatically
calculated by the
MRI
scanner
from the IDEAL IQ sequence [65]. To mark pixels that should be removed from the analysis, Figure 5.8: Sample center slice with manually drawn muscle segments. A mask was drawn for each of the four muscle groups; quadriceps (Quad.), hamstrings (Ham.), adductors (Add.) and sartorius (Sart.) on the out-phase image (A). The masks can then be applied to the FF images (B) and others to calculate mean biomarker values. two masks were created. A mask was drawn by hand to remove the femur bone as well as the skin around the thigh. A second mask was generated to ignore voxel values above 100% fat fraction and below 0% fat fraction. Furthermore, a ”fat mask” was applied to all images whereby any voxel associated with an FF ≥ 50 in the FF data, was given a ’NaN’ value. This way voxels that should most like be considered as fat tissue will be excluded from all biomarker calculations except for CSA [68] [100]. The T2 map was generated using custom MATLAB code to fit the T2w signal to a monoexponential decay function to calculate the T2 value for each voxel of tissue (Figure 4.1). The T2w images were pre-processed with a thigh mask applied to all 8 echo time acquisitions. Removing all background signal reduces
the number of data- points
processed
in the
MATLAB script
to
only those associated with tissue from the thigh. This reduces the time required to produce a T2 map.
The magnetization transfer ratio
(MTR) image
is calculated from the
magnetization transfer image with pulse saturation (MTon) and the unsaturated image (MToff). The two images were co-registered to each other before calculating the MTR as a percentage according to equation 4.1.4 as previously described [84]. An additional mask was generated to identity voxel values outside the threshold of 0 – 100% MTR so that they may be ignored in downstream analysis. The 60 direction DTI data was processed using FSL’s diffusion toolbox software suite to generate FA, MD, RD, and ADC maps. The FSL function “dtifit” produces an image for each of the 3 principle eigenvalues and eigenvectors. These images are used to calculate FA, MD, RD, and ADC according to the previously described equations 4.1.3, 4.1.2 [101]. All generated images/maps (ie. T2 map, MTR, FA and MD) were registered to the fat fraction image so that the muscle ROIs which were drawn according to the FF resolution could be applied. A custom MATLAB script applied the muscle ROIs to each co-registered image, including any additional masks previously described, to cal- culate the average biomarker values and standard deviation for each muscle. Results were saved to a text file and excel file for statistical analysis. Chapter 6 Results 6.1 Statistical Analysis A total of 13 participants received a thigh MRI and full-body DXA scan within the same week, with most participants having both scans performed on the same day. The participants
were classified as either
‘sarcopenic’
or ‘non
-sarcopenic’
based
solely
on the
fulfilment of the low muscle mass criteria stated in most sarcopenia definitions. According to the sex-based threshold for appendicular lean mass index (ALMI) proposed by Coin et al., female participants were deemed ‘sarcopenic’ with an ALMI < 5.47 kg/m2 and male participants were deemed ‘sarcopenic’ with an ALMI < 7.59 kg/m2 [102]. Participant demographics
for each
group
are summarized in table
6.
1. Figure
6.1: Participant demographics divided by sex and then further by sarcopenia status as defined by ALMI threshold. The principal biomarker collected from the full-body DXA was appendicular lean mass (ALM) which was normalized
by height-squared to calculate appendicular lean mass index (ALMI) (kg/m2) and
normalized by BMI to calculate ALM/BMI. The mean DXA values separated by sex and sarcopenic status are summarized in table 6.2 below: Figure 6.2: Appendicular lean mass (ALM) values acquired by DXA and normalized by height (ALMI) and BMI (ALM/BMI). The participants are further divided by sex and then sarcopenia status according to ALMI. MRI data was pre-processed, and thigh muscles were segmented into four mus- cle groups using the center slice; quadriceps (Quad), hamstrings (Ham), adductors (Add), and sartorius (Sart) as described in Chapter 5, Materials and Methods. A muscle mask containing all 4 muscle groups was also created to calculate the average MRI values across across the entire muscle cross-section (total). The MRI biomark- ers collected are cross-sectional area (CSA), fat fraction (FF), magnetization transfer ratio (MTR),
T2-relaxation time (T2), fractional anisotropy (FA), and mean
diffu- sivity (
MD
). Post-processing
and
statistics
were calculated using
built-in
MATLAB
functions (
The MathWorks
Inc.,
Natick, USA). The
resulting mean values and stan- dard deviation are summarized in Figure 6.3, with participants separate by sex and sarcopenia status according to ALMI. Figure 6.3: Summary of all MRI results (CSA, FF, MTR, T2, FA and MD) separated by muscle groups (Quad, Ham, Add, Sart, and total muscle cross-section). The data points are further divided by sex and then sarcopenia status according to the ALMI threshold. Pearson’s correlation, R and p-values, were calculated between all MRI biomarkers and ALMI for each muscle group as well as for the total muscle area, tables 6.4 to 6.8. A final set of correlation values was calculated from the combination of data points across all segmented muscle groups (Quad, Ham, Add, and Sart), presented in table 6.9: Figure 6.4: Pearson correlation table of R-values between all MRI values of the total muscle cross-section and ALMI. Figure 6.5: Pearson correlation table of R-values between all MRI values of the quadricep muscles and ALMI. Figure 6.6: Pearson correlation table of R-values between all MRI values of the hamstring muscles and ALMI. Figure 6.7: Pearson correlation table of R-values between all MRI values of the adductor muscles and ALMI. Figure 6.8: Pearson correlation table of R-values between all MRI values of the sartorius muscles and ALMI. Figure 6.9: Pearson correlation table of R-values calculated between all MRI values and ALMI using all collected data points from all muscles. Finally, a 2-sample T-test and
Wilcoxon rank sum test (also known as
a
Mann- Whitney U test) was
performed
to
calculate
the significance of between
group differ- ences. It is known that ALMI was used as the threshold to characterize the popula- tions as ‘sarcopenic’ and ‘non-sarcopenic’ so these tests are intended to evaluate the degree by which all other MRI biomarkers agree with this grouping. After perform- ing all the data analysis, the research population was re-characterized as ‘sarcopenic’ and ‘non-sarcopenic’ using the biomarker ALMI/BMI
to define low lean mass. The
threshold
for low lean mass
was an ALM/BMI < 0.512 for females and an ALM/BMI < 0.789 for males as defined by the FNIH [16]. The T-test and Wilcoxon rank sum
results are shown in Figure 6.10
and
Figure 6.11. Figure 6
.10:
T
-test
results
between groups divided by sex and sarcopenia status. Sarcopenia status was defined by ALMI (left) and ALM/BMI (right). Cells with
p-values < 0.05 are
highlighed
in green
, cells with
p-values <0
.1
are highlighted in
orange, and cells with p-values <0.15 are highlighted in purple. Figure 6.11: Wilcoxon rank test results between groups divided by sex and sarcopenia status. Sarcopenia status was defined by ALMI (left) and ALM/BMI (right). Cells with
p-values < 0.05 are
highlighed
in green
, cells with
p-values <0
.1
are highlighted in
orange, and cells with p-values <0.15 are highlighted in purple. Chapter 7 Validating sarcop enia status classification Performance metrics to quantify muscle strength or muscle function, and frailty scores were unavailable for these study participants. To informally validate the claim that these participants fall within the frailty and sarcopenia populations, demographic and MRI-biomarker data from other MRI studies assessing skeletal muscle were collated. General trends in DXA and MRI muscle biomarkers within groups of this study pop- ulation are compared to those from studies with healthy and sarcopenic populations characterized by official measures to provide context as to the degree by which this study population represents the general frailty population. Farrow et al. investigated quantitative MRI biomarkers and their relationships be- tween skeletal muscles of healthy adults from 3 different age groups (18 - 30 year, 31 - 68 years, and >69 years) [2]. The study population for this thesis has an average age between the mean age of Farrow’s middle-aged and older groups, leaning heav- ily towards the older group as shown in Figure 7.1. The mean fat fraction for both hamstring muscles and quadriceps of this study’s population also falls between the middle-aged and older FF values recorded by Farrow. The mean T2 values of ham- string and quadricep muscles from this study’s population don’t necessarily match a specific age group from Farrow’s study, but they still fall within standard deviation of recorded mean T2s. Figure 7.1: Summary of MRI data from healthy middle-aged and healthy older population (non-sarcopenic according EWGSOP) from the study by Farrow et al. compared to MRI values collected from this study population [2]. A study by Fu et al.
aimed to
explore
the association between
muscle CSA
and
FF
in a
sarcopenic
population
[3]. Their study population
was divided into
’non-
sarcopenic’ and ’sarcopenic’ based on the
AWGS definition of sarcopenia. The trend of mean thigh muscle CSA in Fu’s population whereby CSA of the non-sarcopenic group < CSA of the sarcopenic group matches the total thigh muscle CSA values of this study’s population. When comparing mean FFs though, thigh muscle FF in Fu’s sarcopenic group is greater than the non-sarcopenic group, but that trend is only seen in the male population of this study according to table 6.3. While this study’s total population has a larger mean ALMI than Fu’s group, the sarcopenic group’s ALMI is less than the non-sarcopenic group’s, same as for this study (
Figure 7.2). Figure 7.2
: Summary
of
MRI
values
from non-sarcopenic and sarcopenic (according to AWGS) populations
from the study by
Fu
et al
. compared
to
this
study population
[3]. Chapter 8 Discussion
The purpose of this study was
two-fold;
to
characterize
the
relationships
between
MRI-biomarkers
of
skeletal muscle
quality and
quantity with DXA biomarkers in an elderly frail population, and to identify the between group differences in MRI- biomarker values. These comparisons were assessed on an individual muscle-group basis as well as from the total muscle cross-section. According to the Pearson correlation results in Figures 6.4 to 6.8, ALMI correlated strongly with total CSA (p=0.0047) as well as CSA of most thigh muscles including the quadriceps (p=0.0095), adductors (p=0.035), and sartorius (p=0.00065). This was expected as ALMI is meant to function as a DXA measure of muscle mass just like MRI-derived CSA. This study found no significant correlations between FF and ALMI (p-value of total, quadriceps, hamstrings, adductors, sartorius, and combined = 0.86, 0.9, 0.23, 0.94, 0.15, 0.81 respectively), which is interesting as many re- ports in the literature recognize a significant correlation between FF of the thigh and ALMI (P=0.011) [3]. A very strong positive correlation between FF and T2 was shown within the quadriceps (
R=0.9, p
=3.5*
10
−5)
and
the hamstrings (
R=0
.83,
p=0
.00042) as expected of skeletal muscle tissue. Fat fraction and T2 relaxation are both
properties of
muscle
tissue
that
are dependent on the
amount
of
water present within
the
imaging voxel,
an increase in
fat fraction
corresponds to a decrease in the
amount of water present within the tissue, which results in longer T2 relaxation times [103]. For this reason, T2 relaxation is an MRI biomarker with applications in the detection of microstructural tissue
changes such as edema
or
inflammation
in neurological
and
muscular diseases [103]. The literature also reports significant correlation between these two biomarkers and MTR, but this pattern was only no- ticed within the hamstring muscles of this study (p=0.02) [104]. A strong negative correlation between FF and CSA was expected according to the literature but was only seen significantly within the sartorius muscle (
R= -0
.51,
p=0
.078)
as well as
when all
muscle
data values were pooled together (R= -0.56, p=1.5 ∗ 10−5) [105]. A significant correlation (p-values<0.05) between both diffusion metrics FA and MD was found in all muscle groups except for the sartorius. FA and FF significantly correlated positively in the hamstrings (p=0.00028), the adductors (
p=0
.011),
and total
cross-section (
p=0
.015). Interestingly FA and MTR correlated positively
in the
quadriceps (
R=0
.62,
p=0
.024)
and
sartorius (
R=0
.55,
p=0
.051)
but
negatively
in the
hamstrings (R= - 0.58, p=0.038). This may be a result of some sort of error such as partial voluming effects from the femoral artery and vein that lies right next to the sartorius muscle as Li et al. has also calculated a significant positive correlation between FA and MTR [106]. Alternatively, this relationship may be describing a difference in muscle fiber type distribution between the sartorius and other muscle groups. Finally, MD and FF show
positive correlation
within
the
quadriceps (
R=0
.6,
p=0
.032)
and
adductors (
R=0
.7,
p=0
.0083) which goes against other reports. Far- row et al. explains that increased fatty infiltration has been demonstrated to restrict diffusion, which should present as a decrease in MD [107] [103]. The hamstring muscle data has produced many strong correlations between many of the observed biomarkers as shown in Figure 6.6. With such a small sample size, these results may be too good to be true and rather a result of error such as par- tial voluming effects and difficulty segmenting the hamstring ROI accurately. Of all the muscle groups, the hamstring muscles have the highest tendency to separate into smaller muscle subsections due to IMAT fat infiltration which increases the chances of accidental inclusion of fat tissue within the muscle group ROI. Alternatively, a possible explanation may be that hamstring muscle are more susceptible to aging and frailty-related
changes in skeletal muscle compared to the
other
muscle
groups. Farrow et al. shows a markedly stronger significance in FA values of the hamstrings (p=0.2) between different age-groups over the quadriceps (p=0.7), and Yoon et al. recognizes high correlation between age and mean FF and mean FA of the posterior compartment (hamstrings) of the thigh [2] [108]. Comparisons can be made between muscles according to mean biomarker values in Figure 6.3. The standout muscle was the sartorius, responsible for the largest FF, T2 and FA values as well as the lowest MTR
compared to the other
muscle
groups. The
quadriceps were
the
muscle
group
with
the
lowest mean FF. Besides the sartorius, T2 values were very similar between muscle groups, with the trends for FF and T2 matching
results from a study by
Morrow
et al
. [104]
The
largest MTR was seen in the adductor muscles. Finally, MD values were very similar between all muscle groups with the quadriceps and adductors having a slightly larger mean MD (MD = 0.0017) compared to the hamstrings and sartorius muscles (MD = 0.0016). Figures 6.10 and 6.11 show differences in mean biomarker values between groups divided by sex and sarcopenia status. Sarcopenic
low muscle mass was
first
defined according to
ALMI
and
then redefined according to ALM/BMI.
The 2-sample T-test and Wilcoxon rank sum test were
applied
to
both variations of sarcopenia classifica- tion. In Figures 6.10 and 6.11, results with cells containing
p-values less than 0.05 are highlighted in
green,
p-values less than 0.10 are highlighted in orange, and p-values less than 0
.15
are highlighted in
purple. It’s important to be aware that T-tests as- sume a normalized distribution, which isn’t realistic in such a small sample size. For this study it is used as a preliminary assessment and
the Wilcoxon rank-sum
, being
a non-parametric
statistical test, is
the
more relevant assessment. Comparing both DXA biomarker methods for sarcopenia classification, the great- est observation is that ALM/BMI has a stronger association with MRI-CSA than ALMI. In the female population, total CSA, quadriceps CSA, hamstring CSA and Sartorius CSA result in T-test p-values of 0.123, 0.006, 0.106, 0.054 respectively. In the male population, only the mean quadricep CSA shows somewhat
significant differences between the sarcopenic and non-sarcopenic populations
with a p-value of 0.181, but
this may be explained by the
very
small
sample size
of n
=4 for the entire male population in this study. Furthermore, the Wilcoxon rank test results also recognize the
significant difference in
muscle CSA
between the
female sarcopenic
and
female
non
-sarcopenic
groups
. It cannot be understated that these results show that ALM/BMI is strongly associated
with the gold standard of muscle
mass quan- tification while ALMI shows marginal associations at best. This contrast between different methods of normalizing ALM has actually been reported once by Bani Has- san et al. using CT as the standard for reliable muscle mass quantification [109]. Interestingly, in the ALM/BMI defined female sarcopenic and non-sarcopenic popu- lations, FA shows some significant between-group differences when calculated from total muscle, the quadriceps, and the hamstrings with Wilcoxon
p-values of 0
.07,
0
.07,
and 0
.143
respectively
. The NIH defines obesity as a BMI > 30 kg/m2, which means that 6 (46%) out of the 13 study participants are considered obese according to that definition. It is appar- ent
that this study
population
has a larger
than average BMI
compared to
the other study populations referenced from Farrow et al. and Fu et al. [2] [3]. This may also be a contributing factor to the improved results from adjusting ALM for BMI rather than participant height as delineation of ALM is difficult with obese patients [110]. The stronger association between ALM/BMI and muscle mass over ALMI is scarcely reported in the literature, with most MRI-related associations limited to comparisons of functional tests rather than direct muscle mass values [109]. Further investigation is required, as the small samples size may still be an alternative explanation in the context of this study. A possible explanation for the improvement in Wilcoxon re- sults is that using ALM/BMI over ALMI increases the ratio of ‘non-sarcopenic’ to ‘sarcopenic’ women from 7:2 to 3:6. This increase in the smallest group size from 2 to 3 strengthens the sample size. To support this, no noticeable improvements in Wilcoxon values for the male ‘non-sarcopenic’ vs. ‘sarcopenic’ comparison were seen. Despite switching from ALMI to ALM/BMI threshold, the resulting participant dis- tribution remained unchanged at 2:2 because the total number of men participants were so low to begin with. Chapter 9 Limitations and next steps This study is presented as a preliminary assessment of DXA and MRI based biomark- ers using a subset of participant data from a larger study on frailty. As a result, limitations are recognized. The greatest study
limitation is the small sample size
. With a total
of
13
study
participants that are then divided into subgroups based on sex and ‘sarcopenia sta- tus’, multiple groups are left with as few as 2 participants. Such a limited number of data points reduces statistical power and the applicability of study results to the general population. Regarding the processing of MRI data, there is no standardized protocol or pro- gram for the registration of thigh muscle images. Acquiring data from different MRI techniques for each participant results in images of varying resolutions with the possi- bility of slight patient movement between sequences. Despite best efforts at registering images to adjust for resolution differences and patient movement through trial and error, there is a possibility that voxels from different acquisitions don’t correspond to the exact same anatomical location. Furthermore, thigh muscles were manually segmented on a single MRI slice. Manual muscle segmentation by a single observer introduces human error into the process of discerning tissue as muscle or fat and fur- ther categorizing the muscle group it belongs to. It is possible that pixels containing no muscle tissue may have been included within the muscle ROIs, although this error was reduced with the inclusion of a fat mask. All MRI data was collected
from a single slice
centered
at
the
mid-thigh
. Quality of data can be improved by including data from 2 or 3 slices above and below the center slice, except for muscle CSA values which must only be taken from a single slice. Finding significant results from limited data holds its own merit, as minimizing scan time through optimized protocols would increase the usefulness of MRI in sarcopenia assessment and diagnosis and reduce preprocessing requirements [105]. Research has already recognized strong correlation between muscle CSA at the midthigh and mus- cle volume for all thigh muscle groups to validate it as a measure of muscle mass [111]. To progress this research further, it would benefit from a larger sample size. In- creasing the number of data points would allow for sex and BMI stratification as well as improve the significance of any true correlations between muscle groups and biomarkers. The study analysis could also be repeated using ALM/BMI rather than ALMI to further explore the correlations with FF and sarcopenia status according to alternate assessment methods. Additionally, non-imaging metrics of frailty and physical function can be included in the overall analysis. Using the full definition of sarcopenia and categorizing participants
according to the
most accepted
criteria
of
low muscle mass, low muscle function, and low
muscle strength, the MRI-biomarker results will gain more diagnostic applicability. Chapter 10 Conclusion
Sarcopenia is a
multi-faceted
skeletal muscle disorder
resulting
in skeletal muscle atrophy and muscle
fat infiltration beyond what is expected from the natural aging process. This
disorder characterized by a
combination
of
reduced
muscle mass, muscle
function, and muscle
strength
still seeks a unifying definition. Quantitative measures from imaging and non-imaging assessment tools alike may assist in a more objective diagnosis of sarcopenia. DXA is the current most researched and validated imag- ing modality in sarcopenia assessment to measure muscle mass. The DXA-derived biomarker ALM which is often normalized by height to produce ALMI
may not be the best
metric
of muscle mass
. Compared to ALMI, ALM normalized by BMI displays stronger agreement with MRI-derived muscle CSA which is the current gold standard in muscle volume quantification. The results from this study promote ALM/BMI as a more reliable DXA biomarker of muscle mass in sarcopenia assessment, especially in populations with larger BMIs and higher obesity rates. To my knowledge no other research studies to date have directly reported on correlations between ALM/BMI and MRI-derived thigh muscle CSA. MRI is an extremely versatile imaging modality that provides avenues
to assess both
skeletal
muscle quantity and quality
through
a
variety of muscle biomarkers such as CSA, FF, MTR, T2, FA, and MD as described in this study. By further correlating MRI biomarkers with currently accepted tools for sarcopenia assessment such as frailty scores and DXA,
MRI has the potential to be
recognized
as a
valid
and
reliable
tool
for
the assessment of
most, if not all 3, criteria that define sarcopenia. So long as the benefits can outweigh the high cost and expertise that limits accessibility to MRI, an all-encompassing MRI protocol may be designed and optimized to objectively diagnose sarcopenia in the future. Bibliography [1]
G. Coletta and S. M. Phillips, “An elusive consensus definition of sarcopenia impedes research and clinical treatment: A narrative review,” Ageing
Research Reviews,
vol. 86, p. 101883
, 4 2023. [2]
M. Farrow, J. Biglands, S. F. Tanner, A. Clegg, L. Brown, E. M. Hensor, P. O’Connor, P. Emery
, and
A. L. Tan, “The effect of ageing on skeletal muscle as assessed by quantitative MR imaging: an association with frailty and muscle strength,” Aging Clinical and Experimental Research
, vol. 33, pp. 291–301, 2 2021. [3] C. Fu, D. Yan, L. Wang, F. Duan, D. Gu, N. Yao, M. Sun, D. Wang, X. Lin, Y. Wu, X. Wang, X. Cheng, and D. Zhang, “
High prevalence of sarcopenia and myosteatosis in patients undergoing hemodialysis
,” Frontiers in Endocrinology, vol. 14, p. 1117438, 2023. [4]
D. Albano, C. Messina, J. Vitale
, . Luca, and
M. Sconfienza, “Imaging of sarcopenia: old evidence and new insights
Bone mineral content CSA Cross- sectional area CT Computed tomography DTI Diffusion tensor imaging DXA Dual-energy X-ray absorptiometry EWGSOP European Working Group on
Sar- copenia
in Older People FA Fractional anisotropy
,” [5]
V. Chianca, D. Albano, C. Messina, S. Gitto, G. Ruffo, S. Guarino, F. Del Grande, and L. M. Sconfienza, “Sarcopenia: imaging assessment and clinical application,” Abdominal Radiology
2021 47:9, vol. 47, pp. 3205–3216, 10 2021. [6]
G. Kojima, A. E. Liljas, and S. Iliffe, “Frailty syndrome: implications and challenges for health care policy,” Risk Management and Healthcare Policy, vol. 12
, p.
23
, 2019. [7]
C. C. Kennedy, G. Ioannidis, K. Rockwood, L. Thabane, J. D. Adachi, S
. Kirk- land,
L. E. Pickard, and A. Papaioannou
, “
A Frailty Index predicts 10-year fracture risk in adults age 25 years and older: results from the Canadian
Mul- ticentre
Osteoporosis Study (CaMos),” Osteoporosis international : a journal established as result of cooperation between the European Foundation for
Osteo- porosis
and the National Osteoporosis Foundation of the USA
, vol. 25, p. 2825, 12 2014. [8]
K. Rockwood, X. Song
, and
A. Mitnitski, “Changes in relative fitness and frailty across the adult lifespan: evidence from the Canadian National
Popula- tion
Health Survey,” CMAJ : Canadian Medical Association Journal
, vol. 183, p. E487, 5 2011. [9]
R. M. Collard, H. Boter, R. A. Schoevers, and R. C. Oude Voshaar, “Prevalence of frailty in community-dwelling older persons: A systematic review,” Journal of the American
Geriatrics
Society, vol. 60, pp. 1487–1492
, 8 2012. [10]
J. E. Morley, B. Vellas, G. Abellan van Kan, S. D. Anker, J. M. Bauer, R
. Bern- abei,
M. Cesari, W. C. Chumlea, W. Doehner, J. Evans, L. P. Fried, J. M
. Gu- ralnik,
P. R. Katz, T. K. Malmstrom, R. J. McCarter, L. M. Gutierrez Robledo, K. Rockwood, S. von Haehling, M. F. Vandewoude, and J. Walston, “Frailty consensus: A call to action,” Journal of the American Medical Directors
Asso- ciation,
vol. 14, no. 6, pp. 392–397, 2013
. [11]
L. Han, A. Clegg, T. Doran
, and
L. Fraser, “The impact of frailty on
health- care
resource use: a longitudinal analysis using the Clinical Practice Research Datalink in England,” Age and Ageing
, vol. 48, pp. 665–671, 9 2019. [12]
A. J. Cruz-Jentoft, G. Bahat, J. Bauer, Y. Boirie, O. Bruyère, T. Cederholm, C. Cooper, F. Landi, Y. Rolland, A. A. Sayer, S. M. Schneider, C. C. Sieber, E. Topinkova, M. Vandewoude, M. Visser, M. Zamboni, I. Bautmans, J. P. Baeyens, M. Cesari, A. Cherubini, J. Kanis, M. Maggio, F. Martin, J. P. Michel, K. Pitkala, J. Y. Reginster, R. Rizzoli, D. Sánchez-Rodríguez, and J. Schols, “Sarcopenia: revised European consensus on definition and diagnosis,” Age and Ageing, vol. 48, pp. 16–31
, 1 2019. [13]
I. H. Rosenberg, “Sarcopenia: origins and clinical relevance,” The Journal of nutrition, vol. 127, no. 5
Suppl,
1997
. [14]
W. K. Mitchell, J. Williams, P. Atherton, M. Larvin, J. Lund, and M. Narici, “Sarcopenia, dynapenia, and the impact of advancing age on human skeletal muscle size and strength; a quantitative review,” Frontiers in Physiology, vol. 3
JUL,
p
. 28458, 7 2012. [15]
L. K. Chen, L. K. Liu, J. Woo, P. Assantachai, T. W. Auyeung, K. S. Bahyah, M. Y. Chou, L. Y. Chen, P. S. Hsu, O. Krairit, J. S. Lee, W. J. Lee, Y. Lee, C. K. Liang, P. Limpawattana, C. S. Lin, L. N. Peng, S. Satake, T. Suzuki, C. W. Won, C. H. Wu, S. N. Wu, T. Zhang, P. Zeng, M. Akishita
, and
H. Arai
, “Sar- copenia
in Asia: Consensus Report of the Asian Working Group for
Sarcope- nia,”
Journal of the American Medical Directors Association
, vol. 15, pp. 95– 101, 2 2014. [16] S.
A. Studenski, K. W. Peters, D. E. Alley, P. M. Cawthon, R. R. McLean, T. B. Harris, L. Ferrucci, J. M. Guralnik, M. S. Fragala, A. M. Kenny, D. P. Kiel, S. B. Kritchevsky, M. D. Shardell, T. T. L. Dam
, and
M. T. Vassileva, “The FNIH Sarcopenia Project: Rationale, Study Description, Conference
Rec- ommendations,
and Final Estimates,” The Journals of Gerontology: Series A
, vol. 69, pp. 547–558, 5 2014. [17] S. Bhasin,
T. G. Travison, T. M. Manini, S. Patel, K. M. Pencina, R. A
. Field- ing,
J. M. Magaziner, A. B. Newman, D. P. Kiel, C. Cooper, J. M. Guralnik, J. A. Cauley, H. Arai, B. C. Clark, F. Landi, L. A. Schaap, S. L. Pereira, D. Rooks, J. Woo, L. J. Woodhouse, E. Binder, T. Brown, M. Shardell, Q. L. Xue, R. B. DAgostino, D. Orwig, G. Gorsicki, R. Correa-De-Araujo, and P. M. Cawthon, “Sarcopenia Definition: The Position Statements of the Sarcopenia Definition and Outcomes Consortium
,”
Journal of the American Geriatrics
So- ciety,
vol. 68, pp. 1410–1418
, 7 2020. [18]
W. R. Frontera and J. Ochala, “Skeletal Muscle: A Brief Review of Structure and Function,” Behavior Genetics, vol. 45, pp. 183–195
, 3 2015. [19]
D. Cretoiu, L. Pavelescu, F. Duica, M. Radu, N. Suciu
, and S.
M. Cretoiu, “Myofibers
,”
Advances in Experimental Medicine and Biology, vol
. 1088,
pp
. 23– 46, 2018. [20]
J. M. Squire, “Architecture and function in the muscle sarcomere,” Current Opinion in Structural Biology, vol. 7, pp. 247–257
, 4 1997. [21]
G. Liu, F. Mac Gabhann, and A. S. Popel, “Effects of Fiber Type and Size on the Heterogeneity of Oxygen Distribution in Exercising Skeletal Muscle,” PLoS ONE, vol. 7, p
. 44375, 9 2012. [22]
C. McCuller, R. Jessu
, and
A. L. Callahan, “Physiology, Skeletal Muscle,” StatPearls
, 7 2023. [23]
L. B. Verdijk, T. Snijders, M. Drost, T. Delhaas, F. Kadi, and L. J. Van Loon, “Satellite cells in human skeletal muscle; from birth to old age,” Age, vol. 36, no. 2
, p.
545
, 2014. [24]
J
. Pie Ń Kowska,
B. Brzeska, M. Kaszubowski, O. Kozak, A. Jankowska
, and
E. Szurowska, “MRI assessment of ectopic fat accumulation in pancreas, liver and skeletal muscle in patients with obesity, overweight and normal BMI in correlation with the presence of central obesity and metabolic syndrome
,” Di- abetes,
Metabolic Syndrome and Obesity, vol. 12, pp
. 623–636,
2019
. [25]
C. w. Li, K. Yu, N. Shyh-Chang, Z. Jiang, T. Liu, S. Ma, L. Luo, L. Guang, K. Liang, W. Ma, H. Miao, W. Cao, R. Liu, L. j. Jiang, S. l. Yu, C. Li, H. j. Liu, L. y. Xu, R. j. Liu, X. y. Zhang, and G. s. Liu, “Pathogenesis of sarcopenia and the relationship with fat mass: descriptive review
,” Journal of
Cachexia, Sarcopenia
and
Muscle
, vol. 13, p. 781, 4 2022. [26]
M. Nachit, M. De Rudder, J. P. Thissen, O. Schakman, C. Bouzin, Y
. Hors- mans,
G. Vande Velde
, and
I. A. Leclercq, “Myosteatosis rather than sarcopenia associates with non-alcoholic steatohepatitis in non-alcoholic fatty liver
dis- ease
preclinical models
,” Journal of Cachexia,
Sarcopenia
and
Muscle
, vol. 12, pp. 144–158, 2 2021. [27]
T. Hosoi, M. Yakabe, H. Sasakawa, T. Sasako, K. Ueki, S. Kato, S. M. Tokuoka, Y. Oda, M. Abe, T. Matsumoto, M. Akishita
, and S.
Ogawa
, “Sar- copenia
phenotype and impaired muscle function in male mice with fast-twitch muscle-specific knockout of the androgen receptor,” Proceedings of the National Academy of Sciences of the United States of America
, vol. 120, p. 2218032120, 1 2023. [28]
K. Ohlendieck, “Proteomic profiling of fast-to-slow muscle transitions during aging,” Frontiers in Physiology, vol. 2
DEC,
p
. 18432, 12 2011. [29]
J. Y. Hogrel, Y. Barnouin, N. Azzabou, G. Butler-Browne, T. Voit, A. Moraux, G. Leroux, A. Behin, J. S. McPhee, and P. G. Carlier, “NMR imaging estimates of muscle volume and intramuscular fat infiltration in the thigh: variations with muscle, gender, and age,” Age, vol. 37, pp. 1–11
, 6 2015. [30] T.
Maden-Wilkinson, H. Degens, D. A. Jones, and J. S. McPhee, “Comparison of MRI and DXA to measure muscle size and age-related atrophy in thigh muscles
.,” 9 2013. [31]
T. Abe, R. S. Thiebaud, J. P. Loenneke, M. Loftin, and T. Fukunaga
, “Preva- lence
of site-specific thigh sarcopenia in Japanese men and women,” Age, vol. 36, no. 1
, p.
417
, 2014. [32]
R. W. Bohannon, “Grip Strength: An Indispensable Biomarker For Older Adults,” Clinical Interventions in Aging, vol. 14, pp. 1681–1691
, 10 2019. [33]
C. Beaudart, E. McCloskey, O. Bruyère, M. Cesari, Y. Rolland, R. Rizzoli, I. Araujo de Carvalho, J. Amuthavalli Thiyagarajan, I. Bautmans, M. C. Bertière, M. L. Brandi, N. M. Al-Daghri, N. Burlet, E. Cavalier, F. Cerreta, A. Cherubini, R. Fielding, E. Gielen, F. Landi, J. Petermans, J. Y. Reginster, M. Visser, J. Kanis
, and
C. Cooper, “Sarcopenia in daily practice: assessment and management,” BMC
Geriatrics, vol.
16
, pp. 1–10, 10 2016. [34]
G. Voulgaridou, S. Tyrovolas, P. Detopoulou
, D. Tsoumana, M. Drakaki, T. Apostolou, I.
P. Chatziprodromidou
, D.
Papandreou
, C.
Giaginis, and
S.
K. Papadopoulou, “Diagnostic Criteria and Measurement Techniques of
Sarcope- nia:
A Critical Evaluation of the Up-to-Date Evidence,” Nutrients
, vol. 16, 2 2024. [35]
M. K. Beauchamp, Q. Hao, A. Kuspinar, C. D’Amore, G. Scime, J. Ma, A
. May- hew,
C. Bassim, C. Wolfson, S. Kirkland, L. Griffith
, and
P. Raina
, “Reliabil- ity
and Minimal Detectable Change Values for Performance-Based Measures of Physical Functioning in the Canadian Longitudinal Study on Aging
,”
The Journals of Gerontology: Series A, vol. 76, pp
. 2030–2038, 10 2021. [36]
L. L. Ploutz-Snyder, T. Manini, R. J. Ploutz-Snyder, and D. A. Wolf
, “Func- tionally
Relevant Thresholds of Quadriceps Femoris Strength
,” The
Journals of Gerontology
: Series
A, vol. 57
, pp. B144–B152,
4
2002. [37] M. Steffl
and P. Stastny, “Isokinetic testing of muscle strength of older
individ- uals
with sarcopenia or frailty: A systematic review,” Isokinetics and Exercise Science
, vol. 28, pp. 291–301, 1 2020. [38]
D. Samuel, K. Wilson, H. J. Martin, R. Allen, A. A. Sayer
, and
M. Stokes, “Age-associated changes in hand grip and quadriceps muscle strength ratios in healthy adults,” Aging clinical and experimental research
,
vol. 24, no. 3, pp. 245–250, 2012
. [39]
T. M. Manini, M. Visser, S. Won-Park, K. V. Patel, E. S. Strotmeyer, H. Chen, B. Goodpaster, N
. De
Rekeneire, A. B. Newman, E. M. Simonsick, S. B. Kritchevsky, K. Ryder, A. V. Schwartz, and T. B. Harris, “Knee Extension Strength
Cutpoints
for Maintaining Mobility,” Journal of the American
Geri- atrics
Society, vol. 55, pp. 451–457
, 3 2007. [40]
O. Bruyère, C. Beaudart, J. Y. Reginster, F. Buckinx, D. Schoene, V. Hirani, C. Cooper, J. A. Kanis, R. Rizzoli, E. McCloskey, T. Cederholm, A. Cruz- Jentoft
, and
E. Freiberger, “Assessment of muscle mass, muscle strength and physical performance in clinical practice: An international survey,” European Geriatric Medicine
, vol.
7
, pp. 243–246, 6 2016. [41]
G. Abellan Van Kan, Y. Rolland, S. Andrieu, J. Bauer, O. Beauchet, M. Bonnefoy, M. Cesari, L
. M.
Donini, S. Gillette-Guyonnet, M. Inzitari, F. Nourhashemi, G. Onder, P. Ritz, A. Salva, M. Visser, and B. Vellas, “Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force
,” The
Journal of nutrition, health and aging, vol. 13, pp. 881–889
, 12 2009. [42]
J. M. Guralnik, L. Ferrucci, C. F. Pieper, S. G. Leveille, K. S. Markides, G. V. Ostir, S. Studenski, L. F. Berkman, and R. B. Wallace, “Lower Extremity
Func- tion
and Subsequent Disability: Consistency Across Studies, Predictive Models, and Value of Gait Speed Alone Compared With the Short Physical Performance Battery,” The Journals of Gerontology: Series A, vol. 55
, pp. M221–M231,
4
2000. [43]
M. Maggio, G. P. Ceda, A. Ticinesi, F. De Vita, G. Gelmini, C. Costantino, T. Meschi, R. W. Kressig, M. Cesari, M. Fabi, and F. Lauretani, “Instrumental and Non-Instrumental Evaluation of 4-Meter Walking Speed in Older
Individ- uals,”
PLOS ONE, vol. 11
, p. e0153583,
4
2016. [44]
E
. Rydwik,
A
.
Bergland, L. Forsén, and K. Frändin, “Investigation into the reliability and validity of the measurement of elderly people’s clinical
walk- ing
speed: A systematic review,” Physiotherapy Theory and Practice
, vol. 28, pp. 238–256, 4 2012. [45]
P. M. Cawthon, S. M. Patel, S. B. Kritchevsky, A. B. Newman, A
. San- tanasto,
D. P. Kiel, T. G. Travison, N. Lane, S. R. Cummings, E. S. Orwoll, K. A. Duchowny, T. Kwok, V. Hirani, J. Schousboe, M. K. Karlsson, D
. Mell- ström,
C. Ohlsson, Ljunggren, Q. L. Xue, M
. Shardell, J.
M. Jordan
, K.
M. Pencina
, R.
A. Fielding
, J.
Magaziner
, R.
Correa-De-Araujo
, S.
Bhasin
, and T.
M. Manini
, “
What Cut-Point in Gait Speed Best Discriminates Community- Dwelling Older Adults With Mobility Complaints From Those Without? A Pooled Analysis From the Sarcopenia Definitions and Outcomes Consortium,” The Journals of Gerontology: Series A
, vol. 76, pp. e321–e327, 9 2021. [46] J. D. Podsiadlo, S. Bscpt, and M. D. J.
Richardson, “The Timed “Up & Go”: A Test of Basic Functional Mobility for Frail Elderly Persons,” Journal of the American Geriatrics Society, vol. 39, pp. 142–148
, 2 1991. [47]
B. P
. Martinez,
I. B. Gomes, C. S. de Oliveira, I. R. Ramos, M. D. M. Rocha, L. A. Forgiarini, F. W. R. Camelier, and A. A. Camelier, “Accuracy of the Timed Up and Go test for predicting sarcopenia in elderly hospitalized
pa- tients,”
Clinics, vol. 70, no. 5
, p.
369
, 2015. [48] S.
Y. Lee, P. L. Choo, B. W. J. Pang, L. K. Lau, K. A. Jabbar, W. T. Seah, K. K. Chen, T. P. Ng, and S. L. Wee, “SPPB reference
values
and performance in assessing sarcopenia in community-dwelling Singaporeans – Yishun study,” BMC Geriatrics, vol. 21
, 12 2021. [49] S.
Perera, S. H. Mody, R. C. Woodman, and S. A. Studenski, “Meaningful Change and Responsiveness in Common Physical Performance Measures in Older Adults,” Journal of the American Geriatrics Society, vol. 54, pp. 743–749
, 5 2006. [50]
C. Riwniak, J. E. Simon, N. P. Wages, L. A. Clark, T. M. Manini, D. W. Russ
, and B.
C. Clark, “Comparison of a Multi-Component Physical Function Battery to Usual Walking Speed for Assessing Lower Extremity Function and Mobility Limitation in Older Adults,” The Journal of nutrition, health
and
aging
, vol. 24, pp. 906–913, 8 2020. [51]
M. Ni, L. G. Brown, D. Lawler, and J. F. Bean, “Reliability, Validity, and Minimal Detectable Change of Four-Step Stair Climb }Power Test in Community-Dwelling Older Adults,” Physical Therapy, vol. 97
, p.
767
, 7
2017
. [52]
T. K. Malmstrom, D. K. Miller, E. M. Simonsick, L. Ferrucci, and J. E. Morley, “SARC-F: a symptom score to predict persons with sarcopenia at risk for poor functional outcomes,” Journal of Cachexia, Sarcopenia and Muscle, vol. 7
, p.
28
, 3
2016
. [53]
M. D. Ballesteros-Pomar, E. González-Arnáiz, B
. Pintor-de-la
Maza, D. Barajas-Galindo, D. Ariadel-Cobo, L. González-Roza
, and
I. Cano- Rodríguez, “Bioelectrical impedance analysis as an alternative to dual-energy x-ray absorptiometry in the assessment of fat mass and appendicular lean mass in patients with obesity,” Nutrition
, vol. 93, p. 111442, 1 2022. [54]
I. Janssen, S. B. Heymsfield, R. N. Baumgartner, and R. Ross, “Estimation of skeletal muscle mass by bioelectrical impedance analysis,” Journal of Applied Physiology, vol. 89, no. 2, pp. 465–471, 2000
. [55]
W. Shen, M. Punyanitya, Z
. M.
Wang, D. Gallagher, M. P. St.-Onge, J. Albu, S. B. Heymsfield, and S. Heshka, “Total body skeletal muscle and adipose tissue volumes: Estimation from a single abdominal cross-sectional image,” Journal of Applied Physiology, vol. 97
, pp.
2333–2338
, 12 2004. [56]
L. Lenchik, K. M. Lenoir, J. Tan, R. D. Boutin, K. E. Callahan, S. B. Kritchevsky
, and
B. J. Wells, “Opportunistic Measurement of Skeletal Muscle Size and Muscle Attenuation on Computed Tomography Predicts 1-Year
Mor- tality
in Medicare Patients
,”
The Journals of Gerontology: Series A, vol. 74, pp
. 1063–1069, 6 2019. [57]
M. Visser, S. B. Kritchevsky, B. H. Goodpaster, A. B. Newman, M. Nevitt, E. Stamm, and T. B. Harris, “Leg muscle mass and composition in relation to lower extremity performance in men and women aged 70 to 79: The Health
, Ag- ing
and Body Composition Study,” Journal of the American Geriatrics Society, vol. 50, no. 5, pp. 897–904, 2002
. [58]
A. M. Martone, E. Marzetti, R. Calvani, A. Picca, M. Tosato, R. Bernabei
, and
F. Landi, “Assessment of sarcopenia: From clinical practice to research
,”
JOURNAL OF
GERONTOLOGY
AND
GERIATRICS,
vol
. 67,
no. 1, pp
. 39– 45,
2020
. [59]
D. W. McRobbie, E. A. Moore, M. J. Graves, and M. R. Prince, “MRI from Picture to Proton
,” 4 2017. [60]
Y. Fatemi, H. Danyali, M. S. Helfroush
, and
H. Amiri, “Fast T2 mapping using multi-echo spin-echo MRI: A linear order approach
,”
Magnetic Resonance in Medicine, vol. 84, pp. 2815–2830
, 11 2020. [61] S.
Schlaeger, D. Weidlich, E. Klupp, F. Montagnese, M. Deschauer, B. Schoser, S. Bublitz, S. Ruschke, C. Zimmer, E. J. Rummeny, J. S. Kirschke
, and
D. C. Karampinos, “Water T2 Mapping in Fatty Infiltrated Thigh Muscles of Patients With Neuromuscular Diseases Using a T2-Prepared 3D Turbo Spin Echo With SPAIR,” Journal of Magnetic Resonance Imaging
, vol. 51, pp. 1727–1736, 6 2020. [62]
N. Azzabou, J. Y. Hogrel, and P. G. Carlier, “NMR based biomarkers to study age-related changes in the human quadriceps,” Experimental Gerontology, vol. 70, pp. 54–60
, 10 2015. [63]
W. T. Dixon, “Simple proton spectroscopic imaging,” Radiology, vol. 153, no. 1, pp. 189–194, 1984
. [64]
C. F. Lins, C. E. G. Salmon, and M. H. Nogueira-Barbosa, “Applications of the Dixon technique in the evaluation of the musculoskeletal system
,” Radiologia
Brasileira, vol. 54, no. 1
, p. 33, 2021. [65] S.
B. Reeder, A. R. Pineda, Z. Wen, A. Shimakawa, H. Yu, J. H. Brittain, G. E. Gold, C. H. Beaulieu, and N
. T.
Pelc, “Iterative decomposition of water and fat with echo
asymmetry
and least-squares estimation (IDEAL): Application with fast spin-echo imaging,” Magnetic Resonance in Medicine, vol. 54, pp. 636–644
, 9 2005. [66] S.
B. Reeder, C. A. McKenzie, A. R. Pineda, H. Yu, A. Shimakawa, A. C. Brau, B. A. Hargreaves, G. E. Gold, and J. H. Brittain, “Water–fat separation with IDEAL gradient
-
echo imaging,” Journal of Magnetic Resonance Imaging, vol. 25, pp. 644–652
, 3 2007. [67] S.
Eskreis-Winkler
, G.
Corrias
, S.
Monti
, J.
Zheng
, M.
Capanu
, S.
Krebs
, M.
Fung
, S.
Reeder, and
L.
Mannelli, “IDEAL-IQ in an oncologic population: Meeting the challenge of concomitant liver fat and liver iron,” Cancer Imaging
, vol. 18, pp. 1–7, 12 2018. [68]
K. Engelke, O. Chaudry, L. Gast
,
M. A
. Eldib,
L
. Wang,
J. D
. Laredo,
G
. Schett, and
A. M
. Nagel, “
Magnetic resonance imaging techniques for the quantitative analysis of skeletal muscle: State of the art
,” Journal of Orthopaedic Transla- tion, vol. 42, p. 57, 9 2023. [69]
M. Codari, M. Zanardo, M
. Eugenia
di Sabato, E. Nocerino, C. Messina, L
. Maria
Sconfienza
, and
F. Sardanelli, “MRI-Derived Biomarkers Related to Sarcopenia: A Systematic Review
,” 2019. [70]
A. Grimm, H. Meyer, M. D. Nickel, M. Nittka, E. Raithel, O. Chaudry, A
. Fried- berger,
M. Uder, W. Kemmler, K. Engelke
, and
H. H. Quick, “A Comparison Between 6-Point Dixon Mri and Mr Spectroscopy to Quantify Muscle fat in the Thigh of Subjects with Sarcopenia,” Journal of Frailty and Aging
, vol. 8, pp. 21–26, 1 2019. [71]
T. Baum, S. Inhuber, M. Dieckmeyer, C. Cordes, S. Ruschke, E. Klupp, P. M. Jungmann, R. Farlock, H. Eggers, H. Kooijman, E. J. Rummeny, A. Schwirtz, J. S. Kirschke, and D. C. Karampinos, “Association of Quadriceps Muscle Fat With Isometric Strength Measurements in Healthy Males Using Chemical Shift Encoding-Based Water-Fat Magnetic Resonance Imaging
,”
Journal of
Com- puter
Assisted Tomography, vol. 40
, no.
3, p. 447
, 2016. [72] O.
Addison, M. J. Drummond, P. C. Lastayo, L. E. Dibble, A. R. Wende, D. A. McClain, and R. L. Marcus, “Intramuscular fat and inflammation differ in older adults: The impact of frailty and inactivity,” The Journal of nutrition, health
and
aging
, vol. 18, pp. 532–538, 5 2014. [73]
D. M. Koh and D. J. Collins, “Diffusion-weighted MRI in the body
: Applica- tions
and challenges in oncology,” American Journal of Roentgenology, vol. 188, pp. 1622–1635
, 6 2007. [74] L. R. Ranzenberger, J. M. Das, and T. Snyder, “Diffusion Tensor Imaging,”
Surgery of Spinal Cord Tumors Based on Anatomy: An Approach Based on Anatomic Compartmentalization
, pp. 147–151, 11 2023. [75]
R. E. Roberts, E. J. Anderson, and M. Husain, “White matter microstructure and cognitive function,” Neuroscientist, vol. 19, pp. 8–15
, 2 2013. [76]
L. J. O’Donnell and C. F. Westin, “An Introduction to Diffusion Tensor Image Analysis,” Neurosurgery Clinics of North America, vol. 22, pp. 185–196
, 4 2011. [77]
J. Oudeman, A. J. Nederveen, G. J. Strijkers, M. Maas, P. R. Luijten, and M. Froeling, “Techniques and applications of skeletal muscle diffusion tensor imaging: A review,” Journal of Magnetic Resonance Imaging, vol. 43, pp. 773– 788
, 4 2016. [78]
C. C
. V.
DONKELAAR, L. J. G. KRETZERS, P. H. M. BOVENDEERD, L. M. A. LATASTER, K. NICOLAY, J. D. JANSSEN, and M. R. DROST, “Diffusion tensor imaging in biomechanical studies of skeletal muscle function,” Journal of Anatomy, vol. 194, pp. 79–88
, 1 1999. [79]
D. C. Karampinos, K. F. King, B. P. Sutton, and J. G. Georgiadis, “Myofiber
el- lipticity
as an explanation for transverse asymmetry of skeletal muscle diffusion MRI in vivo signal,” Annals of Biomedical Engineering, vol. 37, pp. 2532–2546
, 12 2009. [80]
A. Esposito, L. Campana, A. Palmisano, F. De Cobelli, T. Canu, F. Santarella, C. Colantoni, A. Monno, M. Vezzoli, G. Pezzetti, A. A. Manfredi, P. Rovere- Querini
, and
A
. De
Maschio, “Magnetic Resonance Imaging at 7T Reveals Common Events in Age-Related Sarcopenia and in the Homeostatic Response to Muscle Sterile Injury,” PLOS ONE
, vol. 8, p. e59308, 3 2013. [81]
M. Scheel, P. von Roth, T. Winkler, A. Arampatzis, T. Prokscha, B. Hamm, and G. Diederichs, “Fiber type characterization in skeletal muscle by diffusion tensor imaging,” NMR in Biomedicine, vol. 26, pp. 1220–1224, 10
2013. [82]
O. Yanagisawa, D. Shimao, K. Maruyama, M. Nielsen, T. Irie, and M. Niitsu, “Diffusion-weighted magnetic resonance imaging of human skeletal muscles: gender-, age- and muscle-related differences in apparent diffusion coefficient,” Magnetic Resonance Imaging, vol. 27, pp. 69–78
, 1 2009. [83] S.
D. Wolff and R. S. Balaban, “Magnetization transfer contrast (MTC) and
tis- sue
water proton relaxation in vivo,” Magnetic Resonance in Medicine, vol. 10, no. 1, pp. 135–144, 1989
. [84]
R
. M.
Henkelman, G
. J.
Stanisz, and S
. J.
Graham, “Magnetization transfer in MRI: A review,” NMR in Biomedicine, vol. 14, no. 2, pp. 57–64, 2001
. [85]
N. F
. Schwenzer,
P. Martirosian, J. Machann, C. Schraml, G. Steidle, C. D. Claussen, and F. Schick, “Aging effects on human calf muscle properties assessed by MRI at 3 Tesla,” Journal of Magnetic Resonance Imaging
, vol. 29, pp. 1346– 1354, 6 2009. [86]
G. A. Power, M. D. Allen, W. J. Booth, R. T. Thompson, G. D. Marsh, and C. L. Rice, “The influence on sarcopenia of muscle quality and quantity derived from magnetic resonance imaging and neuromuscular properties,” Age, vol. 36
, no. 3,
pp. 1377–1388, 2014
. [87] R.
J
.
Toombs, G. Ducher, J. A. Shepherd, and M. J. De Souza, “The Impact of Recent Technological Advances on the Trueness and Precision of DXA to Assess Body Composition,” Obesity, vol. 20, pp. 30–39
, 1 2012. [88]
G. Guglielmi, F. Ponti, M. Agostini, M. Amadori, G. Battista, and A
. Baz- zocchi, “
The role of DXA in sarcopenia,” Aging Clinical and Experimental
Re- search, vol.
28
, pp. 1047–1060, 12 2016. [89]
C. Messina, D. Albano, S. Gitto, L. Tofanelli, A. Bazzocchi, F. M. Ulivieri, G. Guglielmi
, and
L. M. Sconfienza, “Body composition with dual energy X-ray absorptiometry: from basics to new tools,” Quantitative Imaging in Medicine and Surgery
, vol. 10, p. 1687, 8 2020. [90]
M. F. Folstein, S. E. Folstein, and P. R. McHugh, ““Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician,” Journal of Psychiatric Research, vol. 12, pp. 189–198
, 11 1975. [91]
P. Vermathen, P. Saillen, A. Boss, M. Zehnder, and C. Boesch, “Skeletal muscle 1H MRSI before and after prolonged exercise. I. muscle specific depletion of intramyocellular lipids,” Magnetic Resonance in Medicine, vol. 68, pp. 1357– 1367
, 11 2012. [92]
F. Esposito, G. Pignataro, G
. Di
Renzo, A. Spinali, A. Paccone, G. Tedeschi, and L
. Annunziato, “
Alcohol increases spontaneous BOLD signal fluctuations in the visual network,” NeuroImage, vol. 53, pp. 534
–543, 11 2010. [93] H. C. S.
Yang, Z. Liang, J. F. Yao, X. Shen, B. d
. B.
Frederick, and Y. Tong, “Vascular effects of caffeine found in BOLD fMRI,” Journal of neuroscience research, vol. 97, p. 456
, 4
2019
. [94]
A. H. Elzibak and M. D. Noseworthy, “Assessment of diffusion tensor imaging indices in calf muscles following postural change from standing to supine
posi- tion,”
Magnetic Resonance Materials in Physics, Biology and Medicine, vol. 27, pp. 387–395
, 10 2014. [95]
A. M. Heemskerk and B. M. Damon, “Diffusion Tensor MRI Assessment of Skeletal Muscle Architecture,” Current medical imaging reviews, vol. 3
, p.
152
, 7
2007
. [96]
M. Jenkinson, C. F. Beckmann, T. E. Behrens, M. W. Woolrich, and S. M. Smith, “FSL,” NeuroImage, vol. 62, pp. 782–790
, 8 2012. [97] S.
M. Smith, M. Jenkinson, M. W. Woolrich, C. F. Beckmann, T. E. Behrens, H. Johansen-Berg, P. R. Bannister, M
. De
Luca, I
. Drobnjak,
D. E. Flitney, R. K. Niazy, J. Saunders, J. Vickers, Y. Zhang, N
. De
Stefano, J. M. Brady, and P. M. Matthews, “Advances in functional and structural MR image analysis and implementation as FSL
,” NeuroImage,
vol. 23, pp. S208–S219
, 1 2004. [98]
M. W. Woolrich, S. Jbabdi, B. Patenaude, M. Chappell, S. Makni, T. Behrens, C. Beckmann, M. Jenkinson, and S. M. Smith, “Bayesian analysis of
neuroimag- ing
data in FSL,” NeuroImage, vol. 45, pp. S173–S186
, 3 2009. [99]
M. D. Noseworthy, A. D. Davis, and A. H. Elzibak, “Advanced MR imaging techniques for skeletal muscle evaluation
,” Seminars in musculoskeletal
radiol
- ogy,
vol. 14, no. 2, pp. 257–268, 2010
. [100] J. Linge,
M
. Petersson,
M
. F. Forsgren,
A
.
J
. Sanyal, .
Olof
, and
D
. Leinhard, “
Adverse muscle composition predicts all-cause mortality in the UK Biobank imaging study,” 2021
. [101]
A. L. Alexander, J. E. Lee, M. Lazar, and A. S. Field, “Diffusion tensor imaging of the brain,” Neurotherapeutics, vol. 4
, p.
316
, 7
2007
. [102]
A. Coin, S. Sarti, E. Ruggiero, S. Giannini, M. Pedrazzoni, S. Minisola, M. Rossini, A. Del Puente
, E. M. Inelmen,
E. Manzato
, and
G. Sergi
, “Preva- lence
of Sarcopenia Based on Different Diagnostic Criteria Using DEXA and Appendicular Skeletal Muscle Mass Reference Values in an Italian Population Aged 20 to 80
,”
Journal of the American Medical Directors Association, vol. 14, pp. 507–512
, 7 2013. [103]
L. A. Otto, W. L. van der Pol, L. Schlaffke, C. A. Wijngaarde, M. Stam, R. I. Wadman, I. Cuppen, R. P. van Eijk, F. L. Asselman, B. Bartels, D. van der Woude, J. Hendrikse
, and
M. Froeling, “Quantitative MRI of skeletal muscle in a cross-sectional cohort of patients with spinal muscular atrophy types 2 and 3,” NMR
in Biomedicine, vol.
33, 10 2020
. [104]
J. M. Morrow, C. D. Sinclair, A. Fischmann, M. M. Reilly, M. G. Hanna, T. A. Yousry, and J. S. Thornton, “Reproducibility, and age, body-weight and gender dependency of candidate skeletal muscle MRI outcome measures in healthy volunteers,” European Radiology
, vol. 24, no. 7, p. 1610, 2014. [105]
Y. X. Yang, M. S. Chong, W. S. Lim, L. Tay, S. Yew, A. Yeo, and C. H. Tan, “Validity of estimating muscle and fat volume from a single MRI section in older adults with sarcopenia and sarcopenic obesity,” Clinical Radiology
, vol. 72, pp. 9–427, 5 2017. [106] K. Li, R.
D. Dortch, E
. B.
Welch
, N.
D. Bryant
, A.
K. Buck
, T.
F. Towse
, D.
F. Gochberg
, M.
D. Does
, B.
M. Damon, and
J.
H. Park, “Multi-parametric MRI characterization of healthy human thigh muscles at 3.0 T - relaxation
, magneti- zation
transfer, fat/water, and diffusion tensor imaging,” NMR in Biomedicine
, vol. 27, no. 9, pp. 1070–1084, 2014. [107]
M. Farrow, J. D. Biglands, A. J. Grainger, P. O’Connor, E. M
. Hensor,
A. Ladas, S. F. Tanner, P. Emery
, and
A. L. Tan, “Quantitative MRI in myositis patients: comparison with healthy volunteers and radiological visual assessment,” Clinical Radiology
, vol. 76, pp. 1–81, 1 2021. [108]
M. A. Yoon, S. J. Hong, M. C. Ku, C. H. Kang, K. S. Ahn
, and
B. H. Kim, “Multiparametric MR imaging of age-related changes in healthy thigh muscles,” Radiology
, vol. 287, pp. 235–246, 4 2018. [109]
E. Bani Hassan, S. Phu, S. Vogrin, G. Escobedo Terrones, X. Pérez, I. Rodriguez-Sanchez
, and
G. Duque, “Diagnostic Value of Mid-Thigh and Mid- Calf Bone, Muscle, and Fat Mass in Osteosarcopenia: A Pilot Study
,” Calcified
Tissue
International, vol.
105
, pp. 392–402, 10 2019. [110]
E. M. Evans, M. M. Misic, and D. M. Mallard, “A technique to assess body composition and sarcopenia using DXA: application for an obese population,” European Journal of Clinical Nutrition
2010 64:2, vol. 64, pp. 218–220, 11 2009. [111] S.
Cotofana, M. Hudelmaier, W. Wirth, M. Himmer, S. Ring-Dimitriou, A. M. Sänger, and F. Eckstein, “Correlation between single-slice muscle anatomical cross-sectional area and muscle volume in thigh extensors, flexors and adductors of perimenopausal women,” European Journal of Applied Physiology
, vol. 110, pp. 91–97, 9 2010.
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering M.A.Sc. Thesis
– K. Grala;
McMaster University
– Biomedical
Engineering
M.A.Sc. Thesis – K
. Grala;
McMaster University
– Biomedical
Engineering
1 2 3 4 5 6 7 8 9 10 11 12 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
