J Appl Physiol
90: 329–337, 2001.
innovative techniques
Imaging of skeletal muscle function using 18FDG PET:
force production, activation, and metabolism
GEORGE P. PAPPAS1,2 ERIC W. OLCOTT1,3 AND JOHN E. DRACE1,3
Radiology Service, Veterans Affairs Palo Alto Health Care System, Palo Alto 94304; and
Departments of 3Radiology and 2Mechanical Engineering (Biomechanical Engineering Division),
Stanford University, Stanford, California 94305
1
Received 11 April 2000; accepted in final form 16 August 2000
of skeletal muscle
motion and metabolism would provide information of
significant scientific and clinical benefit. The fundamental understanding of musculoskeletal function
would be greatly advanced by data that elucidate in
vivo muscle contraction, activation, and metabolism.
Such knowledge could also lead to improved treatment
of neuromuscular pathologies such as stroke, spinal
cord injuries, and muscular dystrophies. For example,
the success of muscle-tendon transfer surgery could be
increased by preoperative assessment of potential donor muscles (43); strategies for rehabilitation may also
be optimized by evaluating muscle function and adaptation processes.
Several noninvasive modalities have been used to
acquire in vivo information about skeletal muscle function. Magnetic resonance (MR) imaging techniques
such as cine phase contrast MR can characterize the in
vivo motion and mechanics of contracting skeletal muscle noninvasively (2, 6–8, 27) but cannot measure its
metabolic activity. Modalities currently used for estimating muscle activity and metabolism include electromyography (EMG) and MR techniques such as spectroscopy and T2-weighted imaging (10, 29, 38, 40).
Positron emission tomography (PET) with 18F-labeled 2-fluoro-2-deoxy-D-glucose (FDG) has proven
valuable for assessing organ glucose utilization, primarily in brain, cardiac, and tumor tissue, and is
considered the gold standard for quantifying myocardial viability (4, 25, 32). FDG PET imaging relies on
the principle that active muscle cells exhibit increased
glucose uptake (17, 28). Although it has been employed
in several settings to evaluate muscle tissue metabolism and perfusion, the application of PET to the study
of exercising skeletal muscle remains limited (1, 13, 25,
26, 33, 35, 37). FDG PET provides tomographic spatial
information and allows exercise to be performed conveniently before imaging, the combination of which is
not available with other techniques used for measuring
muscle activity.
Exercising muscle tissue takes up circulating FDG, a
deoxy analog of glucose, which then becomes entrapped
in the intracellular space after phosphorylation. Unlike glucose, FDG does not continue along the usual
glycolytic pathway; rather, it accumulates within exer-
Address for reprint requests and other correspondence: J. E.
Drace, Diagnostic Radiology Center (114), VA Palo Alto Health Care
System, 3801 Miranda Ave., Palo Alto, CA 94304 (E-mail: drace
@stanford.edu).
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
positron emission tomography; fluorodeoxyglucose; intramuscular; tomographic; in vivo
COMPREHENSIVE IN VIVO MEASUREMENTS
http://www.jap.org
329
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
Pappas, George P., Eric W. Olcott, and John E.
Drace. Imaging of skeletal muscle function using 18FDG
PET: force production, activation, and metabolism. J Appl
Physiol 90: 329–337, 2001.—The purpose of this study was to
determine whether [18F]fluorodeoxyglucose (FDG) positron
emission tomography (PET) can be used to evaluate muscle
force production, create anatomic images of muscle activity,
and resolve the distribution of metabolic activity within exercising skeletal muscle. Seventeen subjects performed either elbow flexion, elbow extension, or ankle plantar flexion
after intravenous injection of FDG. PET imaging was performed subsequently, and FDG uptake was measured in
skeletal muscle for each task. A fivefold increase in resistance during elbow flexion increased FDG uptake in the
biceps brachii by a factor of 4.9. Differences in relative FDG
uptake were demonstrated as exercise tasks and loads were
varied, permitting differentiation of active muscles. The intramuscular distribution of FDG within exercising biceps
brachii varied along the transverse and longitudinal axes of
the muscle; coefficients of variation along these axes were
0.39 and 0.23, respectively. These findings suggest FDG PET
is capable of characterizing task-specific muscle activity and
measuring intramuscular variations of glucose metabolism
within exercising skeletal muscle.
330
IMAGING OF SKELETAL MUSCLE FUNCTION USING
cising muscle tissue. This metabolic trapping process
forms the basis of FDG imaging (21).
The purpose of this study was to determine whether
PET is capable of evaluating force production, creating
anatomic images of muscle activity, and measuring the
distribution of metabolic activity within exercising
skeletal muscle. A modality that provides this comprehensive functional information would have considerable scientific and clinical utility. Accordingly, we
tested the following three hypotheses: first, that FDG
uptake as measured by PET reflects the active contractile force and work produced by skeletal muscle; second, that FDG PET can create images that permit
muscle activity to be distinguished during functionally
different exercise tasks; and, third, that FDG PET can
measure and resolve the distribution of glucose metabolism within exercising skeletal muscle.
Subjects. This study included 15 men and 2 women ranging in age from 20 to 78 yr (mean ⫽ 53 yr) referred for
diagnostic FDG PET scans. Scanning was performed for a
variety of reasons, typically to evaluate suspected neoplastic
disease. Patients with diabetes and patients receiving insulin
for cardiac PET imaging were excluded from the study. The
study was performed in accordance with the policies of the
institutional human subjects review board.
PET imaging. Each subject performed one specific exercise
task beginning shortly after a single intravenous injection of
10–15 mCi of FDG. PET images with a 128 ⫻ 128 pixel
matrix were acquired for 10–15 min beginning ⬃60 min after
the injection, when intravasculular FDG levels had diminished substantially as a result of redistribution into metabolically active tissue and urinary excretion. In view of kinetic
PET data, exercise was performed immediately after injection, when most of the initial FDG dose remained available
within the vasculature, and PET images were acquired at
least 60 min after injection, when ⬍10% of the dose remained
within the vasculature (16).
Scanning was performed with a CTI EXACT scanner (CTI,
Knoxville, TN) with a 16.2-cm axial field of view and a
transaxial resolution of 6 mm (full width half-maximum in
the center field of view without scattering medium). Emission images were corrected for signal attenuation by using a
68
Ga transmission scan acquired for 5–20 min immediately
before or after the emission images. Attenuation-corrected
axial emission images were reconstructed in 47 contiguous
3.4-mm-thick slices using a Hann filter with a 0.7 cycle/pixel
cutoff.
Image analysis. Image processing and region of interest
(ROI) analyses were performed on a Sparcstation 20 microcomputer (Sun Microsystems, Mountain View, CA) using
vendor-supplied software. Glucose uptake within muscles
was assessed by defining ROIs on the attenuation-corrected
axial images. Emission counts per pixel per second were
measured for each ROI. FDG uptake values obtained from
ROI measurements of active muscle were normalized using
values from ROIs defined over inactive muscle tissue. Average uptake within inactive muscle was defined as “basal
uptake.” When exercise was performed unilaterally, the corresponding inactive muscle in the contralateral limb was
employed for determining basal uptake; the inactive deltoid
muscle was employed when exercise was performed bilaterally.
FDG PET
Exercise protocols. Each subject performed a specific, supervised exercise beginning 2 min after a single injection of
FDG, when intravascular levels of FDG were highest. Each
exercise task was explained and demonstrated to the subject
before the FDG injection. Subjects were allowed rest periods
between each exercise set. All subjects were guided through
each task at a constant rate of ⬃2 s per repetition, by the
same investigator, to ensure consistency and correctness of
form. Each subject performed only one type of exercise to
avoid the confounding effects of additive FDG trapping that
would result from the performance of multiple exercise tasks.
The investigation was divided into three studies: a force
production study, a task-specific functional differentiation
study, and an intramuscular spatial distribution study. The
force production study was performed first and provided an
estimate of the amount of exercise required to elicit considerable FDG uptake yet avoid neuromuscular fatigue; information used in the planning of the subsequent studies.
Force production study. Eleven subjects performed specified numbers of repetitions of elbow flexion to determine
whether the FDG PET technique can detect increased metabolic activity in contracting skeletal muscle and to test the
hypothesis that the level of glucose uptake in the biceps
brachii muscle corresponds to the force generated and work
produced by the muscle. The total number of repetitions
ranged from 15 to 600 for the six subjects who used 2-lb.
weights and from 50 to 120 for the five subjects who used
10-lb. weights. Four of six subjects using 2-lb. weights exercised bilaterally, performing unequal numbers of repetitions
with the right and left arms: one subject performed 45 and 15
repetitions, another performed 100 and 200 repetitions, another performed 200 and 400 repetitions, and the remaining
subject performed 300 and 500 repetitions, respectively. The
remaining two subjects using 2-lb. weights performed either
240 or 600 repetitions unilaterally. The five subjects using
10-lb. weights performed 50, 60, 80, 100, or 120 repetitions
unilaterally. Subjects performed all elbow flexion exercises in
a seated, upright position. The wrist was maintained in a
supinated position to ensure recruitment of the biceps brachii
muscle (5). To estimate the average glucose uptake of the
biceps brachii muscle, three ROIs containing the entire muscle cross section were drawn in three axial planes near the
midportion of each subject’s muscle. Regression analysis was
used to determine the relationship between the number of
repetitions of elbow flexion and the average FDG uptake in
the biceps brachii muscle.
Task-specific functional differentiation study. Images of
FDG uptake in exercising skeletal muscle were obtained
during three experiments. The first experiment determined
whether FDG PET could differentiate muscles performing
two dissimilar tasks, elbow flexion and elbow extension. FDG
uptake in the biceps brachii was compared with FDG uptake
in the triceps brachii in two subjects who performed 100
repetitions of either elbow flexion or extension using 10-lb.
weights. The mean value of FDG uptake throughout each
muscle was computed from multiple nonoverlapping ROIs
deposited in a single axial plane located midway along the
upper arm. Ratios of mean biceps activity to mean triceps
activity were computed for each subject. Elbow flexion was
performed in a seated position with the wrist supinated;
elbow extension was performed with the body supine and the
wrist in a neutral position.
The second experiment in this category determined
whether FDG PET could differentiate metabolic activity in
muscles performing two similar tasks, elbow flexion with the
wrist pronated and elbow flexion with the wrist supinated.
The former task primarily involves the brachialis and the
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
MATERIALS AND METHODS
18
IMAGING OF SKELETAL MUSCLE FUNCTION USING
FDG PET
331
cising muscle was evaluated. A subject performed 600 repetitions of elbow flexion against a 2-lb. resistance with the
wrist supinated to ensure recruitment of the biceps brachii
(5). FDG uptake was measured along the transverse (mediallateral) and longitudinal (proximal-distal) axes of this muscle. The transverse distribution of FDG uptake was measured using ROIs distributed across an axial image of the
midportion of the biceps (Fig. 5A). The longitudinal distribution of FDG metabolism was measured by positioning circular ROIs over the deep, central portion of the muscle on
contiguous axial images (Fig. 5C); the diameter of each ROI
was chosen to be approximately one-third that of the muscle.
With the use of these ROI values, coefficients of variation for
FDG uptake were computed along the transverse and longitudinal directions for both exercising and inactive biceps
muscles.
RESULTS
Force production study. Two hundred or more repetitions of elbow flexion with 2-lb. weights produced
FDG uptake in the biceps brachii that exceeded basal
levels and was readily visible on projection and axial
images (Fig. 1). Fifty or more repetitions of elbow
flexion with a 10-lb. weight produced substantial increases in FDG levels within the biceps brachii muscle.
Figure 1 shows representative projection and axial
tomographic images of FDG uptake for a subject who
performed repeated 2-lb. elbow flexion with the right
arm. Distinct boundaries between the biceps and brachialis muscles are clearly visible in the anterior
Fig. 1. Projection and tomographic images of [18F]fluorodeoxyglucose (FDG) activity in skeletal muscle were
produced for different exercise tasks. Anatomic maps of FDG uptake were of sufficient resolution to permit
differentiation of adjacent muscles such as the biceps brachii (Bic) and brachialis (Br). A: projection image of a
subject who performed supinated-wrist elbow flexion with the right arm. Substantially increased FDG uptake is
evident within 2 elbow flexor muscles, the biceps brachii and the brachialis. FDG uptake is also seen in parts of
the ipsilateral shoulder. B: examples of corresponding attenuation-corrected serial axial images used for quantitative measurement of FDG uptake. The biceps, brachialis, and triceps (T) all can be distinguished. The brachialis
appears most prominent in the more inferior images (images 70, 75, and 80). The humerus (H) can be visualized
as a signal void due to its lower relative metabolic activity.
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
brachioradialis, with little contribution from the biceps,
whereas the latter additionally recruits the biceps (5). A
subject performed 60 repetitions of elbow flexion with each
arm using 10-lb. weights; one wrist was supinated, whereas
the other was pronated. ROI values were obtained within the
biceps and brachioradialis at multiple levels and averaged
for each muscle. Ratios of mean activity in the right vs. left
biceps and right vs. left brachioradialis were computed.
The third experiment evaluated the ability of PET to
measure differences in FDG uptake that result from varying
the levels of recruitment of given muscles. One subject performed 80 repetitions of ankle plantar flexion, a task that
involves the gastrocnemius and soleus muscles, with 30-lb.
weights on each knee. Each knee was held at either 50° or
110° of flexion to achieve different degrees of muscular recruitment (with the straight leg defined as having zero flexion). Because the gastrocnemius spans both the knee and the
ankle, it is more fully recruited when the knee is less flexed
during this task; the soleus exhibits a compensatory, increased recruitment with increased knee flexion (3, 34). The
average FDG uptake within each of the two muscles was
assessed using eight nonoverlapping ROIs placed on an axial
image at the level of the midcalf. In addition, two other
subjects performed plantar flexion with straight knees and a
resistance applied to the ball of the foot; one subject exercised
bilaterally against resistances of 1 and 4 lb., and the second
subject exercised unilaterally against a resistance of 155 lb.
EMG indicates that gastrocnemius recruitment increases as
plantar flexion is performed under increasing loads (3). Ratios of the mean activity between specific muscles were computed.
Intramuscular spatial distribution study. The intramuscular distribution of glucose metabolism within a single exer-
18
332
IMAGING OF SKELETAL MUSCLE FUNCTION USING
18
FDG PET
Fig. 3. The ability of positron emission tomography to distinguish
muscles performing 2 functionally different exercise tasks is demonstrated by comparing the distribution of FDG uptake during elbow
flexion to FDG uptake during elbow extension. A: axial image of
subject who performed 100 repetitions of elbow extension with the
left arm using a 10-lb. weight. Average FDG uptake measured in the
triceps brachii was 3.52 times that measured in the biceps brachii. B:
axial image of subject who performed 100 repetitions of 10-lb. elbow
flexion with the right arm. Average FDG uptake in the biceps was
1.89 times that measured in the ipsilateral triceps.
Fig. 2. Regression plot of normalized FDG uptake in the biceps
brachii as a function of the number of repetitions of elbow flexion
performed with either 2-lb. (■) or 10-lb. (F) weights. The ratio of the
two regression line slopes is 4.94, which is consistent with the
fivefold ratio in external force and work produced. Only data points
with suprabasal uptake are included.
by the right biceps brachii. The diminished involvement of the right biceps resulted in particularly increased FDG uptake in the right brachioradialis; its
activity measured 1.86 times that of the left brachioradialis. Substantial FDG uptake in these activated elbow flexors was evident in projection images (Fig. 4).
The ankle plantar flexion study, the third functional
differentiation experiment, examined differences in
the relative FDG uptake of the soleus and gastrocnemius muscles with variations in the degree of their
recruitment. In the subject who performed 30-lb. plantar flexion bilaterally with knee angles of 50° and 110°,
the gastrocnemius of the leg held in 50° of knee flexion
utilized 1.31 times the FDG utilized by the contralat-
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
oblique projection image (Fig. 1A) as well as in the
axial images (Fig. 1B).
Figure 2 shows the results of regression analysis
relating normalized biceps FDG uptake to the number
of repetitions of elbow flexion performed with 2-lb. and
10-lb. weights. Statistically significant positive correlations were found for both the 2-lb. (r ⫽ 0.899, P ⬍
0.01) and 10-lb. (r ⫽ 0.958, P ⬍ 0.05) studies. The
slopes of the 10-lb. and 2-lb. regression lines were 0.029
and 0.006, respectively. The ratio of these slopes was
4.94, nearly equivalent to the fivefold ratio of external
force produced by the elbow flexors under these two
loads (10 lb./2 lb.).
Task-specific functional differentiation study. In the
subjects who performed the first of the functional differentiation experiments, 100 repetitions of either elbow flexion or elbow extension with a 10-lb. weight, the
average activity within the biceps brachii muscle was
markedly different from the average activity within
the ipsilateral triceps brachii muscle. The subject who
performed supinated-wrist elbow flexion showed average FDG uptake in the biceps that was 1.89 times the
corresponding value measured in the triceps muscle of
the exercising arm. The triceps of the subject who
performed elbow extension exhibited 3.52 times the
average activity exhibited by the ipsilateral biceps
muscle (Fig. 3).
Examination of the FDG uptake in the subject who
performed the second functional differentiation experiment, supinated-wrist and pronated-wrist elbow flexion with 10-lb. weights, showed a considerable difference between the left biceps (wrist supinated to
increase biceps recruitment) and right biceps (wrist
pronated to diminish biceps recruitment). The left biceps took up 2.78 times the quantity of FDG taken up
IMAGING OF SKELETAL MUSCLE FUNCTION USING
18
FDG PET
333
eral gastrocnemius with the knee held in the relatively
disadvantaged 110° position. Similarly, the soleus uptake of the leg with the more disadvantaged gastrocnemius (110° of knee flexion) was 1.44 times that of the
contralateral soleus with the knee held at 50°. In the
subject who performed straight-knee plantar flexion
with 1-lb. (right leg) and 4-lb. (left leg) loads, no substantial difference was found between the average gastrocnemius FDG uptake of the two legs. The difference
in soleus activity for these two low loads was also not
substantial. In the subjects who performed plantar
flexion with 1-, 4-, or 155-lb. loads, the ratio of gastrocnemius activity to soleus activity increased with the
amount of weight applied, from 1.24 with a 1-lb. load to
1.41 with a 4-lb. load and 2.36 with a 155-lb. load.
Intramuscular spatial distribution study. FDG uptake varied smoothly and substantially along the
transverse axis of the exercising biceps performing
elbow flexion with 2-lb. resistance, describing a curve
at the level of the muscle’s midsection that was roughly
parabolic in shape but asymmetric (Fig. 5B). The central portion of the exercising biceps exhibited greater
FDG uptake than did the peripheral portions. In the
central region of the muscle belly, FDG uptake along
the transverse axis peaked at over six times the basal
uptake measured in the contralateral biceps. Similarly, the distribution of FDG uptake along the longitudinal axis of the exercising biceps was nonuniform
(Fig. 5D). FDG uptake of almost 4.5 times the average
basal level was found within the proximal muscle belly,
whereas uptake at the extreme proximal and extreme
distal ends of the muscle was ⬃2.5 and 1.5 times the
basal level, respectively. Coefficients of variation for
FDG uptake in exercising biceps muscle were 0.390
along the transverse axis and 0.231 along the longitudinal axis.
In contrast, inactive biceps muscle exhibited less
FDG uptake, and less variability of uptake, than did
the exercising biceps (Fig. 5, B and D). FDG uptake
along the transverse and longitudinal axes of inactive
muscle described curves of markedly different, flatter
shapes than those of the exercising muscle. The coefficient of variation for FDG uptake in inactive muscle
was 0.131 along the transverse axis and 0.156 along
the longitudinal axis.
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
Fig. 4. Projection image of a subject
following 60 repetitions of bilateral elbow flexion using 10-lb. weights. The
left wrist was supinated to recruit the
left biceps brachii muscle, whereas the
right wrist was pronated during exercise. Substantially increased FDG levels are evident in the left biceps
brachii, right brachioradialis, and both
brachialis muscles. Increasingly light
(yellow to white) areas represent increasing FDG uptake. FDG was injected intravenously in the subject’s
right arm. Left anterior oblique view.
R, right; L, left.
334
IMAGING OF SKELETAL MUSCLE FUNCTION USING
18
FDG PET
DISCUSSION
FDG PET demonstrates increased uptake of FDG
within active, exercising skeletal muscles. Quantitative FDG uptake increases as the muscle works against
greater resistance and as the muscle generates more
work, reflecting increased glucose metabolism as the
muscle becomes increasingly active. FDG PET also
demonstrates which individual muscles are active during a given task, by means of both functional images
and the quantitative measurement of FDG uptake. In
addition, FDG PET measures the spatial distribution
of activity within a given muscle, along both the longi-
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
Fig. 5. Intramuscular distribution of FDG activity within the biceps brachii muscle following 600 repetitions of
unilateral, supinated-wrist elbow flexion with a 2-lb. resistance. A: axial image of the midportion of the exercising
biceps, indicating the position of each of the 12 circular regions of interest (ROIs) used to determine the transverse
distribution of FDG uptake in B. FDG activity also is apparent in the brachialis and triceps brachii muscles. B:
distribution of FDG activity within the midportion of the muscle plotted as a function of position along the
transverse axis of the exercising biceps (F) and the contralateral inactive biceps (■). The data points correspond to
the 12 circular ROIs defined in the axial imaging plane within the biceps shown in A. All values are normalized
by the average uptake value of the transversely distributed ROIs in the inactive biceps. The distribution of FDG
activity in the exercising muscle is roughly parabolic in shape, with higher levels of activity evident within the
deep, central portion of the muscle. C: projection image depicting the arrangement of selected circular ROIs along
the biceps brachii muscle used to determine the longitudinal distribution of FDG uptake in D. One ROI was placed
on each of the 44 contiguous axial images encompassing the length of the biceps. Arrowheads indicate the
approximate position of the most proximal (P) and distal (D) ROIs. D: distribution of FDG activity within the deep,
central portion of the muscle is plotted as a function of position along the longitudinal axis of the exercising biceps
(F) and the contralateral inactive biceps (■). All values are normalized by the average uptake value of the
longitudinally distributed ROIs in the inactive biceps. FDG uptake in the exercising biceps is of greater magnitude
and is more variable than is the uptake in the inactive biceps.
IMAGING OF SKELETAL MUSCLE FUNCTION USING
FDG PET
335
correlation between muscle energy consumption and
force production.
Task-specific functional differentiation study. FDG
PET provides images of glucose metabolism and quantitative information that allow active muscles to be
distinguished from each other. This study indicates
that FDG PET can differentiate not only muscle activity during dissimilar tasks such as elbow flexion and
extension but can even differentiate muscle activity
during similar tasks such as elbow flexion with the
wrist pronated vs. supinated. In addition, as the knee
angle or resistance was varied during ankle plantar
flexion, the recruitment of the gastrocnemius changed
and the relative glucose uptake of the gastrocnemius
compared with the soleus changed accordingly.
Intramuscular spatial distribution study. We found
substantial variation in the glucose uptake of exercising biceps brachii muscle and much less spatial variability in inactive muscle tissue. These preliminary
findings of substantial metabolic heterogeneity are
consistent with other human and animal data that
have revealed broad dispersions in regional blood flow
and metabolic rate in even seemingly homogeneous
tissues (1, 23, 37). For example, recent studies with
[15O]H2O PET have demonstrated spatial heterogeneity of blood flow in both inactive and exercising human
skeletal muscle (1, 37).
FDG uptake along both the longitudinal and transverse axes of the exercising biceps brachii was highly
nonuniform. Glucose uptake was greatest in the central regions along both axes and was distributed in a
roughly parabolic fashion (transversely) or a more complex fashion (longitudinally). This suggests that the
central portions of the biceps are the most metabolically active and, therefore, may generate more contractile force and work than do the more peripheral regions
of the muscle. Because muscle tissue requires energy to
maintain tension (15, 24), energy consumption at a
specific point within a muscle should be an indicator of
the active, contractile force at that point. If glucose
uptake accurately reflects energy expenditure, and
metabolic efficiency is constant throughout a given
muscle, then the distribution of FDG activity determined with PET may provide a means of estimating
the distribution of force production within muscle tissue.
Several mechanisms can be considered as possible
explanations for the lower glucose levels measured at
the periphery of the exercising biceps brachii (Fig. 5B).
First, it is conceivable that metabolic efficiency could
differ from one portion of a muscle to another. A theoretically possible mechanism for such spatially nonuniform metabolic efficiency could be muscle fiber-type
regionalization (i.e., spatial variability in the relative
concentration of the more efficient, slow-twitch and the
less efficient, fast-twitch muscle fiber types within a
muscle). However, our results would be consistent with
a greater concentration of more efficient, slow-twitch
muscle fibers being located at the periphery, a finding
contrary to the observation that regionalized muscles,
including the biceps brachii, have greater concentra-
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
tudinal and transverse axes. In contrast to traditional
biomechanical assumptions, which form the basis of
important models of muscle mechanics (18, 44), the
level of intramuscular activity was found to vary considerably along both axes.
Force production study. The uptake of FDG increased as subjects performed either increasing repetitions of elbow flexion against a given load or a given
number of repetitions against increasing loads. The
quantity of work, or, equivalently, energy, required to
accomplish a mechanical task conventionally is defined
as the product of the force exerted and the distance
moved. Therefore, with the assumption that FDG uptake reflects the uptake of ordinary glucose (25, 32),
FDG PET successfully demonstrated increasing glucose consumption with increasing work production.
When elbow flexion was performed with 10-lb. weights,
the slope of the line relating FDG uptake to the number of repetitions was 4.94 times that of the line produced when 2-lb. weights were used; this result is
consistent with the fivefold relationship between the
quantities of work produced at these two levels of
resistance (10 lb./2 lb.). Therefore, within the range of
motion, level of resistance, and total number of repetitions performed by the subjects, FDG uptake as measured by PET appears to vary linearly with the quantity of work performed. Although we have employed
linear regression to fit lines to the data points relating
FDG uptake to the number of repetitions performed, it
is possible that the true relationship is more complex.
Further experience involving a greater number of subjects will be required to address this issue.
Certain physiological and biomechanical properties
of muscle can affect the relationship between force
production and energy consumption and were considered when designing the study. First, all tasks were
constructed to require substantially submaximal effort
and included rest periods between consecutive sets of
exercise, to maintain muscular metabolism within the
aerobic range (12). By allowing our subjects to avoid
fatigue and the possible concomitant entry into anaerobic pathways, we sought to maintain a fixed relationship between glucose utilization and contractile force
production.
Second, the energy consumption of a muscle depends
on whether it is shortening or lengthening and also on
the rate at which it shortens or lengthens (9, 24). In
addition, a muscle’s ability to produce force depends on
its relative length and contraction speed, as described
by its “force-length curve” and “force-velocity curve,”
respectively (44). To mitigate the influence of these
effects on our force production data, all subjects performed the same task (elbow flexion) with the same
speed and range of motion; therefore, each biceps
brachii should have operated within a similar range of
contraction speed and relative muscle length. Under
the controlled conditions of the study, biceps glucose
metabolism was found to be proportional to the amount
of resistance and work performed. Further study is
required to define the range of physiological conditions
over which PET can provide an accurate quantitative
18
336
IMAGING OF SKELETAL MUSCLE FUNCTION USING
FDG PET
yielding a combination of neuromuscular activation
and metabolism data that are not available with any
other modality. In addition, PET could be used to
ascertain the contribution of individual muscles to
force production, information that is important for simulating musculoskeletal motion and optimizing rehabilitation regimens (43, 44).
A current constraint of the FDG PET technique is
limited temporal resolution. FDG PET images represent a summation of all of the muscle metabolic activity that occurs during the time between FDG injection
and PET imaging. As a result, PET may be insensitive
to more short-lived features of metabolic activity and
therefore may fail to resolve temporal variations in
neuromuscular recruitment. In view of this, exercise
protocols were designed to include low-resistance exercise, sets with relatively few repetitions, and lengthy
rest periods between sets to minimize the onset of
fatigue and ensure that motor unit recruitment remained invariant for the duration of each task (12, 14).
The amount of intramuscular glycogen utilized during exercise, and the rate at which it is replenished by
circulating glucose, may influence FDG uptake. Because glycogen utilization depends on exercise intensity and duration (17, 31), the short-duration, low-intensity exercise protocols used in this study mitigated the
effect of this variable. Further work is required to
characterize fully the impact of exercise intensity and
fuel source on quantitative FDG PET.
Radiation exposure does occur with PET; however, it
may be possible to obtain images of exercising muscle
with substantially lower doses of FDG than used in
conventional clinical and diagnostic studies (35).
In summary, FDG PET is capable of characterizing
task-specific muscle function and measuring intramuscular variations of glucose metabolism within skeletal
muscle tissue. Images of muscle activity with sufficient
resolution for biomechanical and physiological analyses of muscle contraction can be generated. We envision an expanded role for PET as a scientific and
clinical tool. Alone or in conjunction with other modalities, PET may provide a powerful means for studying
the basic science of biomechanics, advancing the design
of functional restoration surgeries, allowing study of
muscle pathology and adaptation, and allowing for
improved neuromuscular rehabilitation.
We thank Drs. George M. Segall and Marguerite T. Hays. We also
thank Dr. Marie Carlisle, Jennifer Lawry, and Julie Loero for clinical
support and Dr. Felix E. Zajac for manuscript review.
This work was supported by the Department of Veterans Affairs,
National Institutes of Child Health and Human Development Grant
HD-31493, and a Stanford-NIH Biotechnology Training Grant.
REFERENCES
1. Ament W, Lubbers J, Rakhorst G, Vaalburg W, Verkerke
GJ, Paans AMJ, and Willemsen ATM. Skeletal muscle perfusion measured by positron emission tomography during exercise.
Pflügers Arch 436: 653–658, 1998.
2. Axel L and Dougherty L. MR imaging of motion with spatial
modulation of magnetization. Radiology 171: 841–845, 1989.
3. Ballantyne BT, Kukulka CG, and Soderberg GL. Motor unit
recruitment in human medial gastrocnemius muscle during combined knee flexion and plantarflexion isometric contractions. Exp
Brain Res 93: 492–498, 1993.
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
tions of slow-twitch fibers located in their deep, rather
than their superficial, regions (20, 22, 41). Therefore,
fiber-type regionalization is not likely to explain our
results. Other mechanisms that could affect the decreased glucose uptake observed in the periphery of the
biceps include preferential recruitment of motor units
within the deep, central portion of the muscle and
nonuniform contraction mechanics (e.g., muscle fiber
length or contraction speed that vary throughout the
muscle). Further investigation will be required to elucidate the mechanisms responsible for the spatial variation in muscular glucose uptake.
Potential applications. Because PET appears able to
provide three-dimensional information regarding skeletal muscle metabolism throughout the body, it may be
useful for evaluating aspects of skeletal muscle physiology that are not readily assessable with existing
modalities. For example, EMG provides real-time data
about neuromuscular activity but only limited information about the spatial distribution of that activity.
In additional, EMG is not practical for evaluating large
groups of muscles or simultaneous measurement of
whole body muscular activity (35). 31Phosphorous-MR
spectroscopy has been used to study the metabolic
effects of muscular fatigue and various metabolic and
vascular pathologies (38–40); however, creating spectroscopic images of metabolic activity would require
prohibitively long acquisition times and concurrent
exercise within an MR scanner. T2-weighted MR imaging has also been applied to measure exercise intensity in skeletal muscle (10, 11, 29, 38, 42). Increased T2
relaxation times are putatively caused by exerciseinduced edema and increased tissue temperature; however, because these phenomena are governed by diffusion and conduction, respectively, the ability of this
technique to measure spatial variations in intramuscular activity is questionable. Recent PET studies with
[15O]H2O have assessed perfusion and estimated blood
flow heterogeneity in skeletal muscle (1, 37); however,
blood flow within skeletal muscle is not strongly linked
to regional uptake of glucose (19, 30, 36) and, accordingly, is not likely to be closely associated with energy
production. Overall, FDG PET may provide an attractively more direct and accurate means to measure
muscular metabolism.
FDG PET conceivably could play a valuable role in
the management of neurological or muscular disease
and in the advancement of biomechanical and physiological science. For example, images that characterize
the distribution and severity of functional deficits and
adaptation could be useful for patients with stroke or
neuromuscular pathology. FDG PET could provide a
noninvasive means to monitor the progress of rehabilitation and reinnervation in patients with stroke or
spinal cord injury. Tomographic information provided
by PET imaging could also play a role in improving
experimental studies of muscle biomechanics and exercise physiology, which currently rely solely on EMG
(3, 5, 34). For example, a combined PET-EMG examination could use EMG to measure muscle coordination
patterns in real time and FDG PET to provide detailed
spatial information about muscle energy consumption,
18
IMAGING OF SKELETAL MUSCLE FUNCTION USING
FDG PET
337
29. Ploutz-Snyder LL, Nyren S, Cooper TG, Potchen EJ, and
Meyer RA. Different effects of exercise and edema on T2 relaxation in skeletal muscle. Magn Reson Med 37: 676–682, 1997.
30. Raitakari M, Nuutila P, Ruotsalainen U, Laine H, Teras M,
Iida H, Makimattila S, Utriainen T, Oikonen V, Sipila H,
Haaparanta M, Solin O, Wegelius U, Knuuti J, and YkiJarvinen H. Evidence for dissociation of insulin stimulation of
blood flow and glucose uptake in human skeletal muscle: studies
using [15O]H2O, [18F]fluoro-2-deoxy-D-glucose, and positron
emission tomography. Diabetes 45: 1471–1477, 1996.
31. Romijn JA, Coyle EF, Sidossis LS, Gastaldelli A, Horowitz
JF, Endert E, and Wolfe RR. Regulation of endogenous fat and
carbohydrate metabolism in relation to exercise intensity and
duration. Am J Physiol Endocrinol Metab 265: E380–E391,
1993.
32. Smith GT. Diagnosis of skeletal muscle ischemia by radionuclide imaging and PET. In: Clinical Positron Emission Tomography, edited by Hubner KF, Collman J, Buonocore E, and
Kabalka GW. St. Louis, MO: Mosby, 1992, p. 78–90.
33. Smith GT, Wilson TS, Hunter K, Besozzi MC, Hubner KF,
Reath DB, Goldman MH, and Buonocore E. Assessment of
skeletal muscle viability by PET. J Nucl Med 36: 1408–1414,
1995.
34. Tamaki H, Kitada K, Akamine T, Sakou T, and Kurata H.
Electromyogram patterns during plantarflexions at various angular velocities and knee angles in human triceps surae muscles.
Eur J Appl Physiol 75: 1–6, 1997.
35. Tashiro M, Fujimoto T, Itoh M, Kubota K, Fujiwara T,
Miyake M, Watanuki S, Horikawa E, Sasaki H, and Ido T.
18F-FDG PET imaging of muscle activity in runners. J Nucl Med
40: 70–76, 1999.
36. Utriainen T, Nuutila P, Takala T, Vicini P, Ruotsalainen
U, Ronnemaa T, Tolvanen T, Raitakari M, Haaparanta M,
Kirvela O, Cobelli C, and Yki-Jarvinen H. Intact insulin
stimulation of skeletal muscle blood flow, its heterogeneity and
redistribution, but not of glucose uptake in non-insulin-dependent diabetes mellitus. J Clin Invest 100: 777–785, 1997.
37. Vicini P, Bonadonna RC, Utriainen T, Nuutila P,
Raitakari M, Yki-Jarvinen H, and Cobelli C. Estimation of
blood flow heterogeneity distribution in human skeletal muscle
from positron emission tomography data. Ann Biomed Eng 25:
906–910, 1997.
38. Weidman ER, Charles HC, Negro-Vilar R, Sullivan MJ,
and MacFall JR. Muscle activity localization with 31P spectroscopy and calculated T2- weighted 1H images. Invest Radiol 26:
309–316, 1991.
39. Wiener DH, Fink LI, Maris J, Jones RA, Chance B, and
Wilson JR. Abnormal skeletal muscle bioenergetics during exercise in patients with heart failure: role of reduced muscle blood
flow. Circulation 73: 1127–1136, 1986.
40. Wilson JR, McCully KK, Mancini DM, Boden B, and
Chance B. Relationship of muscular fatigue to pH and diprotonated Pi in humans: a 31P-NMR study. J Appl Physiol 64:
2333–2339, 1988.
41. Yamaguchi GT, Sawa AGU, Moran DW, Fessler MJ, and
Winters JM. A survey of human musculotendon actuator parameters. In: Multiple Muscle Systems: Biomechanics and Movement Organization, edited by Winters JM and Woo SL-Y. New
York: Springer-Verlag, 1990, p. 717–759.
42. Yue G, Alexander AL, Laidlaw DH, Gmitro AF, Unger EC,
and Enoka RM. Sensitivity of muscle proton spin-spin relaxation time as an index of muscle activation. J Appl Physiol 77:
84–92, 1994.
43. Zajac FE. How musculotendon architecture and joint geometry
affect the capacity of muscles to move and exert force on objects:
a review with application to arm and forearm tendon transfer
design. J Hand Surg [Am] 17: 799–804, 1992.
44. Zajac FE. Muscle and tendon: properties, models, scaling, and
application to biomechanics and motor control. Crit Rev Biomed
Eng 17: 359–411, 1989.
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on October 2, 2016
4. Bergmann SR. Quantification of myocardial perfusion with
positron emission tomography. In: Positron Emission Tomography of the Heart, edited by Bergman SR and Sobel BE. New York:
Futura, 1992, p. 97–127.
5. Buchanan TS, Rovai GP, and Rymer WZ. Strategies for
muscle activation during isometric torque generation at the
human elbow. J Neurophysiol 62: 1201–1212, 1989.
6. Drace JE and Pelc NJ. Measurement of skeletal muscle motion in vivo with phase-contrast MR imaging. J Magn Reson
Imaging 4: 157–163, 1994.
7. Drace JE and Pelc NJ. Skeletal muscle contraction: analysis
with use of velocity distributions from phase-contrast MR imaging. Radiology 193: 423–429, 1994.
8. Drace JE and Pelc NJ. Tracking the motion of skeletal muscle
with velocity-encoded MR imaging. J Magn Reson Imaging 4:
773–778, 1994.
9. Fenn WO. The relation between the work performed and the
energy liberated in muscular contraction. J Physiol (Lond) 85:
277–297, 1924.
10. Fisher MJ, Meyer RA, Adams GR, Foley JM, and Potchen
EJ. Direct relationship between proton T2 and exercise intensity
in skeletal muscle MR images. Invest Radiol 25: 480–485, 1990.
11. Fleckenstein JL, Canby RC, Parkey RW, and Peshock RM.
Acute effects of exercise on MR imaging of skeletal muscle in
normal volunteers. Am J Roentgenol Radium Ther 151: 231–
237, 1988.
12. Fox EL, Bowers RW, and Foss ML. Skeletal muscle: structure
and function. In: The Physiological Basis for Exercise and Sport
(5th ed.). Madison, WI: Brown & Benchmark, 1993, p. 94–135.
13. Fujimoto T, Itoh M, Kumano H, Tashiro M, and Ido T.
Whole-body metabolic map with positron emission tomography
of a man after running. Lancet 348: 266, 1996.
14. Garland SJ, Enoka RM, Serrano LP, and Robinson GA.
Behavior of motor units in human biceps brachii during a submaximal fatiguing contraction. J Appl Physiol 76: 2411–2419,
1994.
15. Hartree W and Hill AV. The regulation of the supply of energy
in muscular contraction. J Physiol (Lond) 55: 133–158, 1921.
16. Hays MT and Segall GM. A mathematical model for the distribution of fluorodeoxyglucose in humans. J Nucl Med 40: 1358–
1366, 1999.
17. Holloszy JO, Kohrt WM, and Hansen PA. The regulation of
carbohydrate and fat metabolism during and after exercise.
Front Biosci 3: D1011–D1027, 1998.
18. Huijing PA. Muscle, the motor of movement: properties in
function, experiment and modelling. J Electromyogr Kinesiol 8:
61–77, 1998.
19. Iversen PO and Nicolaysen G. Local blood flow and glucose
uptake within resting and exercising rabbit skeletal muscle.
Am J Physiol Heart Circ Physiol 260: H1795–H1801, 1991.
20. Johnson MA, Polgar J, Weightman D, and Appleton D.
Data on the distribution of fibre types in thirty-six human
muscles. An autopsy study. J Neurol Sci 18: 111–129, 1973.
21. Kearfott KJ, Elmaleh DR, Goodman M, Correia JA, Alpert
NM, Ackerman RH, Brownell GL, and Strauss WH. Comparison of 2- and 3-18F-fluoro-deoxy-D-glucose for studies of tissue metabolism. Int J Nucl Med Biol 11: 15–22, 1984.
22. Kernell D. Muscle regionalization. Can J Appl Physiol 23: 1–22,
1998.
23. Kuikka JT. Effect of tissue heterogeneity on quantification in
positron emission tomography. Eur J Nucl Med 22: 1457, 1995.
24. McMahon TA. Muscles, Reflexes, and Locomotion. Princeton,
NJ: Princeton Univ. Press, 1984.
25. Ng CK, Soufer R, and McNulty PH. Effect of hyperinsulinemia on myocardial fluorine-18-FDG uptake. J Nucl Med 39:
379–383, 1998.
26. Nuutila P. Applications of PET in diabetes research. Horm
Metab Res 29: 337–339, 1997.
27. Pipe JG, Boes JL, and Chenevert TL. Method for measuring
three-dimensional motion with tagged MR imaging. Radiology
181: 591–595, 1991.
28. Ploug T, Galbo H, and Richter EA. Increased muscle glucose
uptake during contractions: no need for insulin. Am J Physiol
Endocrinol Metab 247: E726–E731, 1984.
18