Key words
spinal cord - MR imaging - CNS
Introduction
Magnetic resonance imaging (MRI) is considered to be the standard reference modality
for spinal cord imaging. Pathologies such as myelopathy, spinal cord tumors, or demyelinating
lesions can be visualized at a higher sensitivity and specificity compared with other
imaging modalities such as ultrasonography and computed tomography [1]
[2]. New techniques such as spinal functional-MRI (fMRI), diffusion tensor imaging (DTI),
or spectroscopy permit more detailed and specific assessments of the normal anatomy,
physiological changes, and pathologic lesions [3]
[4]
[5].
The spinal cord shows physiological motion within the spinal canal in all three directions,
which has been investigated in several previously published studies [6]
[7]
[8]. It is known that next to the venous plexus, the surrounding cardiac-pulsation-related
cerebrospinal fluid (CSF) is responsible for periodic spinal cord motion within the
spinal canal [9]
[10]
[11]
[12]. This spinal cord motion and pulsation artifacts result in a reduced morphological
assessment as well as limitations in motion-sensitive MR sequences, such as those
used for fMRI or DTI [13]
[14]
[15]. Therefore, several studies attempted to demonstrate the possibility of reducing
such artifacts using techniques such as cardiac gating or respiration triggering [16]
[17].
The literature describes the greatest direction of myelon motion as anterior-posterior
(AP) [8]
[18], while most studies examine cranio-caudal motion [8]
[19]
[20]
[21]. However, none of these studies compared the influence of respiration with physiologic
spinal cord motion induced by heart activity. Until now, this had not been sufficiently
shown by MRI on segmental levels.
Cine-sequence MRI, a common and established cardiac imaging technique [22]
[23] based on a prospective ECG-triggered k-space segmentation, is applicable to demonstrate
cardiac cycle-dependent spinal cord motion [19].
This study was designed to assess the range of physiological spinal cord motion and
the influence of respiration in comparison with the physiological influence of heart
activity on spinal cord motion at each segmental level within the cervico-thoracic-lumbar
spinal canal based on MRI measurements.
Materials and Methods
Volunteers
MRI was performed in 16 healthy volunteers (6 women, 10 men; median age, 25 years;
range, 20 – 33 years; median BMI, 23.8). None of the volunteers suffered from anamnestic
neurological or orthopedic disease, and all claimed to be in a state of good health.
The exclusion criteria were age < 18 years, a history of back pain, a history of spinal
surgery or known spinal disease, and general MRI-related exclusion criteria such as
metallic implants, foreign bodies, or claustrophobia. None of the volunteers who participated
after a public request on the universities notice board had to be excluded due to
one of those exclusion criteria. Therefore, there was no influence for the age or
sex distribution.
The study was conducted according to the Declaration of Helsinki and in accordance
with the ethical standards on human experimentation at our institute. Informed consent
was obtained from all volunteers.
Imaging and Measurements
Data acquisition was performed using a 1.5-T full-body MRI system (Philips Achieva;
Philips Medical Systems, Best, Netherlands). All images were acquired using a spine
coil with volunteers placed in the supine position. Initially, sagittal T1-and T2-weighted
(w) images were performed to exclude pathologic findings, such as degenerative changes
or spinal canal stenosis, followed by gradient echo sequences to assess spinal cord
motion. The latter were performed with three different breathing techniques: first,
during a 30-s breath-hold (to assess only heart activity influences); second, with
normal regular breathing; and finally, during forced breathing. All volunteers were
thoroughly instructed in the commands during an earlier practice run-through outside
the scanner where all procedures were repeated several times.
Normal regular breathing was instructed as a mixture of the abdominal and thoracic
quiet breathing type without using the inspiratory or expiratory reserve volume of
the lungs [24]
[25]. The volume of air breathed in and out at rest equates to the tidal volume and is
approximately 0.5 l in an average, healthy adult [26].
Forced breathing was defined and instructed to the volunteers as a mixture of the
abdominal and thoracic breathing type with use of the inspiratory reserve volume (maximal
inspiration) or expiratory reserve volume (maximal expiration) of the lungs [24]
[25]. This breathing maneuver results in an in/expired air volume of approximately 2.4 – 3.8 l,
equating the inspiratory capacity [26].
To avoid intrathoracic pressure changes and arterial as well as venous velocity changes
with the physiological decrease of the heartbeat and during breath-hold, volunteers
were introduced to perform breath-holding in expiration resembling small lung volume
breath-hold [27]
[28].
The scan protocol used in this study included an initial mid-sagittal T1-w turbo spin
echo (TSE) sequence (echo time (TE): 7.40 ms; repetition time (TR): 400 ms; matrix
size: 256 × 256; in-plane resolution 0.97 × 0.97 mm; slice thickness: 5 mm; intersection
gap: 1 mm; flip angle: 90°; field of view (FOV): 250 mm), a sagittal T2-w sequence
(TE: 112 ms; TR: 3160 ms; matrix size: 256 × 256; slice thickness: 4 mm; intersection
gap: 4 mm; field of view (FOV): 240 mm), a sagittal T1-w single-slice gradient echo
(GRE) sequence for respiratory-dependent image acquisition (TE: 2.10 ms; TR: 4.1 ms;
matrix size: 256 × 256; in-plane resolution 0.97 × 0.97 mm; slice thickness: 4 mm;
flip angle: 50°; FOV: 250 mm; and cardiac synchronization with ECG-cardiac triggering
with single heart phases and longest trigger delay). In addition, a sagittal T1-w
GRE sequence was performed for respiratory-dependent image acquisition during breath-holding
(TE: 1.95 ms; TR: 4.0 ms; matrix: 256 × 256; in-plane resolution 0.97 × 0.97 mm; slice
thickness: 4 mm; intersection gap: 2 mm; flip angle: 65°; FOV: 250 mm; and cardiac
synchronization with ECG-cardiac gating with 15 heart phases and shortest gate delay
in breath expiration). Those cine images required the volunteers to hold their breath
for approximately 30 s, which had been tested during the practice run and was feasible
for all participating volunteers without any problems.
A short time interval was taken after each breathing technique of approximately two
minutes to avoid physiological stress to the participants and to give voice prompts
to the volunteers about the beginning of the next scan.
The obtained images permitted cinematic spinal cord motion analysis by lacing the
images together.
The total duration of this examination was < 20 min.
Data Analysis
Measurements were performed by two radiologists on the MR console (Advance Viewer;
Philips Medical Systems, Best, Netherlands) in consensus (SW and CS, blinded for review). The reviewers had 1 and 7 years of experience in spinal radiology.
A horizontal line, perpendicular to the vertebral body long axis in the sagittal T1-w
image and at right angles to the anterior cord surface at each level ([Fig. 1]), was drawn using the electronic caliper tool provided by the MR console software.
The location for this line in the x-axis was defined as the middle of the vertebral
body. This was repeated for all levels, starting with the second cervical and ending
with the second lumbar vertebral body. A one-dimensional histogram was calculated
by the vendor’s application software.
Fig. 1 Sagittal gradient echo magnetic resonance image (MRI) of the cervico-thoracic junction.
Measurement lines were placed on each segmental level.
Abb. 1 Sagittales Gradienten-Echo MRT des zerviko-thorakalen Übergangs. Die Messlinien wurden
auf jeweiliger segmentaler Höhe platziert.
Mathematical oversampling was used with a self-developed spline interpolation algorithm
to increase the achieved low resolution. Therefore, a Canny-algorithm, which has been
frequently used and established in vascular imaging, was applied to permit myelon
definition toward the surrounding CSF and motion detection in a submillimeter range
[29].
Using spline interpolation, the spatial resolution of the one-dimensional histograms
was increased by a factor of 10. At each vertebral level the (temporal) first histogram
served as the standard with a displacement of zero. All (temporal) following histograms
were correlated to that first histogram. For this (one dimensional) voxel-wise displacements
along the axis of the histogram data were tested. The displaced (temporal) following
histograms were then correlated to the (temporal) first histogram. The maximal Pearson’s
correlation coefficient indicated the displacement of the (temporal) following histogram.
Therefore, the displacement was a discrete value with a spatial resolution of 1/10
of the imaging voxel size. This technique allowed for measurements of the spinal cord
motion independent from the physiological movement of the body while breathing.
The offset was calculated to show the highest correlation to the initial profile.
This corresponded to the AP motion, measured in mm.
Statistical Analysis
Continuous variables were expressed as means ± standard deviations, and categorical
variables as frequencies and percentages. Because the results of the Kolmogorov–Smirnov
test and the visual inspection showed non-normally distributed data, the Wilcoxon
signed-rank test was used to determine the significance between the breathing groups.
Furthermore, the Wilcoxon signed-rank test was used to determine significance between
cervical, thoracic, and lumbar motion and the three breathing types. A P-value < 0.05
was considered statistically significant.
Statistical analysis was performed using the IBM SPSS statistics software (release
20.0; SPSS Inc., Chicago, IL, USA).
Results
The median weight and median body size were validated as normal according to the German
population [30].
[Table 1] demonstrates the mean ± standard deviation and the minimum and maximum results of
spinal cord motion in the AP direction within the spinal canal, measured in mm, with
respect to the acquisitions during breath-holding, normal breathing, and forced breathing.
Table 1
Measured spinal cord motion values (mm) in the AP direction of the vertebral column
for each level with respect to the different breathing modalities used during data
acquisition.
Tab. 1 Ausmaß der ermittelten Rückenmarksbewegungen (mm) in AP-Richtung für die untersuchten
Wirbelkörperhöhen in Abhängigkeit der unterschiedlichen Atemtechniken.
level
|
C2
|
C3
|
C4
|
C5
|
C6
|
C7
|
T1
|
T2
|
T3
|
T4
|
T5
|
T6
|
T7
|
T8
|
T9
|
T10
|
T11
|
T12
|
L1
|
L2
|
breath-hold
|
mean
|
0.19
|
0.18
|
0.21
|
0.19
|
0.19
|
0.21
|
0.29
|
0.28
|
0.20
|
0.17
|
0.17
|
0.19
|
0.14
|
0.15
|
0.16
|
0.15
|
0.14
|
0.18
|
0.19
|
0.17
|
SD
|
0.04
|
0.06
|
0.09
|
0.12
|
0.12
|
0.10
|
0.13
|
0.20
|
0.13
|
0.08
|
0.05
|
0.10
|
0.09
|
0.07
|
0.09
|
0.09
|
0.05
|
0.08
|
0.10
|
0.06
|
minimum
|
0.1
|
0.1
|
0.1
|
0.1
|
0
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0
|
0.1
|
0.1
|
0
|
0.1
|
0.1
|
0.1
|
0.1
|
maximum
|
0.3
|
0.3
|
0.5
|
0.6
|
0.5
|
0.1
|
0.6
|
1.0
|
0.5
|
0.4
|
0.2
|
0.4
|
0.3
|
0.3
|
0.4
|
0.3
|
0.2
|
0.3
|
0.3
|
0.2
|
normal breathing
|
mean
|
0.29
|
0.41
|
0.46
|
0.43
|
0.39
|
0.38
|
0.46
|
0.33
|
0.34
|
0.32
|
0.25
|
0.21
|
0.21
|
0.22
|
0.21
|
0.23
|
0.24
|
0.29
|
0.37
|
0.23
|
SD
|
0.24
|
0.39
|
0.39
|
0.33
|
0.27
|
0.20
|
0.27
|
0.17
|
0.19
|
0.16
|
0.11
|
0.11
|
0.11
|
0.12
|
0.11
|
0.13
|
0.13
|
0.14
|
0.14
|
0.06
|
minimum
|
0.1
|
0.1
|
0.1
|
0.2
|
0.1
|
0.2
|
0.2
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.1
|
0.2
|
0.2
|
maximum
|
1.1
|
1.7
|
1.7
|
1.4
|
1.0
|
0.8
|
1.2
|
0.7
|
0.8
|
0.7
|
0.5
|
0.5
|
0.5
|
0.5
|
0.4
|
0.5
|
0.5
|
0.5
|
0.6
|
0.3
|
forced breathing
|
mean
|
1.16
|
1.59
|
1.98
|
2.10
|
1.82
|
1.49
|
1.46
|
1.36
|
1.41
|
1.36
|
1.04
|
0.85
|
0.76
|
0.94
|
1.02
|
1.06
|
1.19
|
1.44
|
1.49
|
0.80
|
SD
|
0.62
|
0.84
|
1.00
|
1.17
|
1.07
|
0.88
|
0.75
|
0.64
|
0.67
|
0.63
|
0.23
|
0.35
|
0.41
|
0.46
|
0.45
|
0.53
|
0.63
|
0.75
|
0.69
|
0.14
|
minimum
|
0.3
|
0.4
|
0.5
|
0.4
|
0.5
|
0.5
|
0.5
|
0.5
|
0.6
|
0.5
|
0.7
|
0.4
|
0.3
|
0.3
|
0.4
|
0.3
|
0.2
|
0.3
|
0.4
|
0.7
|
maximum
|
2.7
|
3.5
|
4.2
|
4.9
|
4.8
|
3.7
|
2.6
|
2.2
|
2.6
|
3.0
|
1.5
|
1.6
|
1.6
|
1.8
|
1.8
|
2.1
|
2.4
|
2.8
|
2.7
|
0.9
|
The mean maximum AP motion measurements differed significantly between the breath-holding,
normal breathing, and forced breathing tests (P < 0.01 each).
The highest mean values for all levels were found in the forced breathing group (mean
1.32 mm ± 0.38; range, 0.76 – 2.1 mm). The spinal motion values during breath-holding
and normal breathing (mean 0.19 mm ± 0.04, range 0.14 – 0.29 mm; and mean 0.31 mm
± 0.09, range 0.21 – 0.46 mm, respectively) were significantly lower than those during
forced breathing. Finally, the mean values for spinal cord motion during breath-holding
were significantly lower than those during normal breathing measurements (P < 0.01).
Among all three conditions, the AP spinal cord motion values differed significantly
between the grouped test pairs in the cervical, thoracic, and lumbar regions (P < 0.01
each). The corresponding spinal cord motion results with regard to the different vertebral
column regions are shown in [Table 2].
Table 2
Measured spinal cord motion values (mm) in the AP direction of the vertebral column
in different regions with respect to the different breathing modalities used during
data acquisition.
Tab. 2 Ausmaß der ermittelten Rückenmarksbewegungen (mm) in AP-Richtung für die untersuchten
Wirbelsäulenabschnitte in Abhängigkeit der unterschiedlichen Atemtechniken.
region
|
cervical
|
thoracic
|
lumbar
|
breath-hold
|
mean
|
0.20
|
0.18
|
0.19
|
SD
|
0.09
|
0.11
|
0.09
|
minimum
|
0.0
|
0.05
|
0.1
|
maximum
|
0.6
|
1.0
|
0.3
|
normal breathing
|
mean
|
0.39
|
0.27
|
0.34
|
SD
|
0.31
|
0.16
|
0.13
|
minimum
|
0.1
|
0.06
|
0.2
|
maximum
|
1.7
|
1.2
|
0.6
|
moderate forced breathing
|
mean
|
1.59
|
1.11
|
1.40
|
SD
|
1.02
|
0.62
|
0.68
|
minimum
|
0
|
0.05
|
0.4
|
maximum
|
4.9
|
3.0
|
2.7
|
The highest mean AP spinal cord motion values occurred during forced breathing on
levels C4 and C5 in the lower cervical spinal segment (1.98 and 2.1 mm, respectively).
Spinal cord motion during breath-holding was highest at the levels of the first and
second thoracic vertebral bodies (means, 0.29 mm and 0.28 mm, respectively).
The greatest motion during normal breathing was detected on levels C4 and T1 (mean,
0.46 mm each), with a second peak in the lumbar region (level L1; mean, 0.37 mm).
The lowest mean spinal cord motion was detected during breath-holding on all examined
levels compared with data acquired during normal or forced breathing.
[Fig. 2] shows the mean values of AP spinal cord measurements on each vertebral level with
respect to the different breathing types.
Fig. 2 a–c Spinal cord motion values in the AP direction with respect to the different breathing
types at all measured levels. a Breath-holding, b normal breathing and c forced breathing. Peaks are visible from the middle cervical spine up to the upper
thoracic levels with normal and forced breathing. A corresponding peak during breath-holding
is found on the first thoracic levels. The lowest values for all breathing types are
found in the mid-thoracic region.
Abb. 2 a–c Ausmaß der Rückenmarksbewegung in AP-Richtung auf den gemessenen Wirbelkörperhöhen
in Abhängigkeit der unterschiedlichen Atmungstechniken. a Atemanhalten, b normale Atmung und kräftige Atmung, c die höchsten Werte bei normaler und kräftiger Atmung sind von der mittleren Halswirbelsäule
bis zur oberen Brustwirbelsäule zu erkennen. Die höchsten Werte bei angehaltener Atmung
sind auf Höhe der ersten Brustwirbelkörper zu finden. Die geringsten Werte wurden
bei allen Atmungstechniken auf Höhe der mittleren BWS ermittelt.
Discussion
In this study, we investigated the influence of respiration on spinal cord motion
compared with physiological motion induced by cardiac cycle caused by heart activity.
Furthermore, a table of normal spinal cord motion values in the AP direction for each
level of the cervico-thoracic-lumbar vertebral column was compiled.
Respiration was shown to have a significantly higher impact on spinal cord motion
within the spinal canal when compared to the influence of heart pulsations as measured
during breath-holding. Next to the physiologic CSF, which circulates along the spinal
canal, this respiratory-based spinal cord motion seems to be the causative factor
behind the motion artifacts described in several MRI studies [13]
[14]
[15]. In particular, new imaging techniques such as spinal fMRI, spinal DWI or DTI could
be considerably influenced by high respiration rates and high breathing depth [4]
[15], since the measured structures are in a sub-millimeter range [5] as well as by the limited signal-to-noise ratio and spatial resolution [31].
Based on the results of our study, further studies are required to evaluate the impact
of improved respiratory gating on the amount of artifacts, since most of the recent
studies focus on cardiac triggering [32].
According to our results, the highest spinal cord motion values were found in the
region of the cervico-thoracic junction, as well as in the region of the thoraco-lumbar
junction. This finding agrees with former studies of spinal cord motion [8]
[18] and is likely based on regional differences in CSF pulsatility that originate from
the cerebral ventricles [33] and anatomical circumstances of the intumescentia cervicalis and lumbosacralis,
respectively.
Previous studies have shown only minor motion of the spinal cord in left-right or
right-left direction [19]. Therefore, we did not investigate this aspect of spinal cord motion in this study.
Knowledge about the considerable influence of respiration, compared to heart activity,
might contribute to a reduction in motion artifacts in MRI. Attempts such as reducing
the acquisition time, increasing the phase-gradient amplitude, or using interleaved
multishot echo planar sequences have been undertaken [3]. In addition to the general advice that patients should not move during MRI examinations,
the avoidance of deep breathing at a high respiration rate could also be recommended.
Therefore, respiration-gated techniques seem to be more promising than cardiac-triggering
techniques.
Limitations
First, a relatively small number of volunteers were included. A higher number would
have increased the statistical power. Second, the data were acquired from a normal
healthy population, which may be different from the usual clinical population with
frequent cases of age-related degenerative spinal disease. However, we tried to assess
normal values in healthy subjects without any degenerative spinal cord changes. Finally,
our study did not investigate inter-reader or intra-reader observer variability associated
with consensus readings. However, the main spinal cord measurements were performed
with the software algorithm, which permits automated measurements in a submillimeter
range and should therefore not be a major source of error. Nevertheless, repeated
independent readings by different readers, as well as a higher number of volunteers
would be desired to proof the reliability of the applied technique in future studies.
Conclusion
Normal physiological spinal cord motion values for each level of the cervico-thoraco-lumbar
vertebral column were measured by MRI. Furthermore, respiration was demonstrated to
be a considerable cause of motion, with a significant impact, compared to heart-related
motion. This knowledge could help to reduce motion artifacts in spinal cord MRI and
could suggest the use of respiration-triggering methods such as navigator echo as
well-established techniques in cardiac MRI.
Clinical Relevance of the Study
-
In high-resolution or functional imaging of the spinal cord, artifact-free images
are desired.
-
Physiological spinal cord motion is one of the sources of motion-dependent artifacts
in spine imaging.
-
Respiration was demonstrated to be a considerable cause of motion, with a significant
impact, compared to heart-related motion.
-
This might help to reduce motion artifacts in spinal cord MRI and could suggest the
use of respiration-triggering methods such as navigator echo as well-established techniques
in cardiac MRI.
-
In addition, a table with normal values of spinal cord motion for cervical, thoracic
and lumbar levels was compiled.