DOI:10.2214/AJR.04.1629
AJR 2006; 186:795-799
© American Roentgen Ray Society
Axial Rotation of the Lumbar Spinal Motion Segments Correlated with Concordant Pain on Discography: A Preliminary Study
Donna G. Blankenbaker1,
Victor M. Haughton1,
Baxter P. Rogers2,
M. Elizabeth Meyerand2 and
Jason P. Fine3
1 Department of Radiology, University of Wisconsin Medical School, Clinical
Science Center-E3/311, 600 Highland Ave., Madison, WI 53792-3252.
2 Department of Medical Physics, University of Wisconsin, Madison, WI
53706.
3 Department of Statistics and Department of Biostatistics & Informatics,
University of Wisconsin, Madison, WI 53792-3252.
Received October 18, 2004;
accepted after revision February 3, 2005.
Address correspondence to D. G. Blankenbaker.
Abstract
OBJECTIVE. One possible cause of back pain in patients with
intervertebral disk degeneration is decreased stability of the motion segment.
Axial rotations between lumbar spinal vertebrae can be measured noninvasively
with CT. We tested the hypothesis that larger axial rotations are found in
motion segments with disks that test positive for concordant pain, which is
considered by some investigators to be a reasonable, accurate predictor of
spinal instability.
SUBJECTS AND METHODS. Between October 2002 and March 2004, all
patients undergoing discography were evaluated for inclusion in the study,
with the approval of the institutional review board. All patients in whom
concordant pain was detected at discography were enrolled in the study. The
patients were placed supine in the CT scanner on a table that rotated the
pelvis 8° clockwise and then counterclockwise with respect to the thorax.
CT images were obtained with the patient in the two positions of rotation. An
automated program calculated the amount of rotation between each lumbar
vertebra as a result of the table rotations. Rotations were stratified by disk
level and by disk classification (concordant pain, nonconcordant pain, no
significant pain).
RESULTS. We recorded the axial rotations of 94 disks in 16
consecutive patients (10 women, six men; age range, 26–53 years) after
two disks were excluded because of a previous fusion. There were 68 normal
disks by MRI and discography, six disks with nonconcordant pain, and 20 disks
with concordant pain. Rotation averaged 0.6° for the normal disks,
1.4° for disks with nonconcordant pain, and 1.8° for disks with
concordant pain. The differences were significant (analysis of variance,
p < 0.001). Disks at L3–L4 with concordant pain rotated on
average 1.2°, whereas disks classified as normal or nonconcordant pain
rotated on average 0.7° (significant at p = 0.005). Disks at
L4–L5 with concordant pain rotated on average 1.9°, and those
without concordant pain rotated on average 1.4° (significant at p
= 0.05). Disks with concordant pain at L5–S1 had an average rotation of
2.2°, whereas disks without concordant pain had an average rotation of
1.5° (marginally significant difference at p = 0.07).
CONCLUSION. Concordant pain at discography predicts increased axial
rotation at a lumbar disk level.
Keywords: CT • CT technique • discography • spine • vertebra
Introduction
Approximately 46,000 lumbar fusions are performed annually in the United
States in patients with the diagnosis of degenerative spinal instability
[1]. Currently, the clinical
diagnosis of instability is based on clinical findings
[2]. To our knowledge, no
reliable imaging criteria have been described to detect or identify
degenerative changes producing instability. The radiographic evaluation of
instability with lateral radiographs of the spine in flexion and extension has
poor sensitivity and specificity
[3].
Instability implies a functional abnormality of the spine. Although various
clinical characterizations of spinal instability have been suggested
[4,
5], instability has been
defined in the biomechanical literature as loss of stiffness in the motion
segment [2]. According to this
definition, the unstable motion segment responds to a load or torque with an
excessive amount of motion
[5].
A functional imaging study that accurately distinguishes stable and
unstable segments would be helpful in selecting patients who might benefit
from fusion for the treatment of degenerative spinal instability. One strategy
is to measure spinal motion segments when a specific load or torque is
applied. Disks with a radial tear undergo significantly more rotation than
disks lacking such a tear, especially when an axial rotatory torque is applied
[6]. Normal motion segments of
the lumbar spine rotate approximately 1–2° when subjected to the
axial torque, and degenerated disks undergo rotation of 2° or more
[7]. Discography, despite its
invasiveness, is sometimes performed to select the spinal level for fusion in
patients with suspected degenerative spinal instability
[1]. During discography, the
patient is asked to report if the injection of contrast medium in the disk
produces typical pain (concordant pain) or pain of a different character
(nonconcordant pain). McCormick
[1], on the basis of a
meta-analysis of the discography literature, suggested that the presence of
concordant pain at discography predicts outcome from spinal fusion with a fair
degree of accuracy, although discography is a controversial procedure
[8,
9]. Hypothetically, if
concordant pain on discography suggests a positive result from spinal fusion,
then increased axial rotation would correlate with the presence of concordant
pain. Therefore, we initiated this prospective study to correlate axial
rotation of the lumbar spinal motion segments with pain on discography.
View larger version (53K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1 —Image shows table insert that provides clockwise and
counterclockwise rotation at lumbar spine. Inset is placed on CT gantry.
Patient is positioned on insert with head and thorax on longer segment and
hips on shorter segment. Each segment is on rollers that permit them to rotate
8° clockwise and counterclockwise with the axis of rotation 10 cm above
segment so that spine is at isocenter of rotation.
|
|
View larger version (105K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —Images illustrating CT data acquired for calculation of vertebral
rotation. First image displays frontal digital scout image that shows patient
rotated 8° in one direction. Parts of table top and roller devices that
permit rotation also are seen.
|
|
View larger version (83K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C —Images illustrating CT data acquired for calculation of vertebral
rotation. show two images, which have been cropped, obtained with left and
right rotation. Software program rotates image D with respect to
C to obtain best superimposition and then reports amount of rotation
required to achieve that superimposition.
|
|
View larger version (83K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D —Images illustrating CT data acquired for calculation of vertebral
rotation. show two images, which have been cropped, obtained with left and
right rotation. Software program rotates image D with respect to
C to obtain best superimposition and then reports amount of rotation
required to achieve that superimposition.
|
|
Subjects and Methods
Patient Selection
Institutional review board approval was obtained. Between October 2002 and
March 2004, an experienced staff musculoskeletal radiologist screened all
patients scheduled for discography and enrolled all patients who had
concordant pain detected at discography in the study. Excluded from the study
were patients who had undergone previous multilevel spinal surgery; were
younger than 18 years; weighed more than 200 pounds (91 kg); or had a
neurologic deficit, systemic illness, or spine fracture. MR images were
reviewed and the intervertebral disks were classified as normal or abnormal.
Each patient was informed of the purpose of this study and signed a written
consent statement.
Discography Technique
We performed discography using standard techniques
[10,
11]. Each patient's pain
history was reviewed in detail before each procedure. The patient's pain
severity before needle placement, after needle placement, and during injection
of contrast medium was rated by the patient on a 10-point scale. At each
level, the disk was classified as normal if the disk was not injected or if
the injection produced no pain or only mild pain. The disk was classified as
nonconcordant if it produced pain that was not typical of the patient's usual
pain pattern and as concordant if the injection produced pain typical in
severity and character of the patient's usual pain. The staff radiologist
performing discography classified the patient's pain response. The discography
images were saved.
View larger version (93K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —MR images of spine showing technique for obtaining rotation
measurements. Images were acquired with clockwise and counterclockwise
rotations, respectively, collimated to a 30-pixel radius.
|
|
View larger version (97K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —MR images of spine showing technique for obtaining rotation
measurements. Images were acquired with clockwise and counterclockwise
rotations, respectively, collimated to a 30-pixel radius.
|
|
View larger version (78K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —MR images of spine showing technique for obtaining rotation
measurements. One image in C is then rotated with respect to other by
means of a pixel-shifting method to minimize misalignment and amount of
rotation required is recorded.
|
|
For the CT study, the patient was placed on a specially built table that
provided separate supports for the torso and for the hips and legs
(Fig. 1). By rotating the hip
support, isocentric rotation around the spinal axis was achieved
[12]. With the patient
positioned comfortably on the table and the hip support rotated 8°
clockwise, a localizer image was acquired and a series of axial CT sections of
1.25 mm in thickness from T12 to S1 was obtained (Figs.
2A,
2B,
2C, and
2D). The hip support was then
rotated 8° counterclockwise and the CT images were reacquired. These
studies were performed on a 16-MDCT scanner (LightSpeed, GE Healthcare) with
the parameters set at 140 kV and 300 mAs.
The image data set was sent to a workstation where the rotations were
measured with specially written software
[12]. An investigator blinded
to the results of discography performed the measurements. For each vertebra
from T12 to S1, a pair of axial slices was selected at each level: one from
the clockwise rotation and one from the counterclockwise rotation. The slice
that was chosen was the one that best displayed the transverse and superior
articular process or alae and sacral promontory. The selected images were
smoothed with a 21-neighbor median filter. One image from the pair was
arbitrarily chosen to serve as a reference image and the other, as the
floating image. In each image, a pivot point was selected in the midline near
the posterior margin of the vertebral body. Each image was then cropped to
include all pixels in a 60-pixel radius of the pivot point. The reference and
floating images were coregistered by maximizing a normalized cross-correlation
measure as a function of rotation around the reference image's pivot point
(Figs. 3A,
3B,
3C, and
3D). The angle through which
the floating image was rotated to align with the reference image (i.e.,
registration angle) was recorded for each disk level. The relative rotations
between adjacent vertebrae were then calculated as the difference in their
respective rotations.
Statistical Analysis
The measurements were tabulated in a spreadsheet together with disk
appearance as normal or abnormal on MRI and presence or absence of concordant
pain.
The analysis of variance was used to test for the effect of disk level and
disk classification on the measured rotation. Differences in the rotations
between the patient groups were tested post hoc for significance with the
Student's t test. Significance was set at a p value of less
than 0.05.
Results
Sixteen consecutive patients were enrolled in the study (10 women, six men;
age range, 26–53 years; mean age, 39 years). All patients had back pain;
the presence or absence of radiculopathy was not always recorded in the
referral. All patients who were enrolled in this study completed the study. No
patients were excluded after enrollment. CT data and discography results were
considered technically adequate in each case. Two levels in one patient were
excluded because of a fusion at one level and a possible junctional disk at
the adjacent level.
Sixty-eight intervertebral disks were classified as normal by MRI and
discography, six as nonconcordant pain, and 20 as concordant pain. At
T12–L1 and L1–L2, all disks were classified as normal. At
L2–L3, 15 disks were classified as normal and one as concordant pain. At
L3–L4, 11 were classified as normal, one as nonconcordant, and four as
concordant. At L4–L5, five disks were classified as normal on
discography, three as nonconcordant, and seven as concordant. At L5–S1,
five disks were classified as normal at discography, two as nonconcordant, and
eight as concordant.
The mean rotation measurement (in degrees) and range are listed in
Table 1.
Table 2 lists the rotation
values for normal disks in patients with back pain. For T12–L1,
L1–L2, and L2–L3, 47 of the 48 disks were classified as normal on
discography. Rotations averaged 0.3–0.7° at these levels. At
L3–L4, four disks classified as concordant pain had, on average,
1.2° of rotation, whereas 12 disks classified as normal or nonconcordant
pain had average rotations of 0.7° or 1.1°, respectively. At
L4–L5, seven disks with concordant pain had an average rotation of
1.9°, and eight classified as normal or nonconcordant had rotation on
average of 0.7° or 1.8°, respectively. At L5–S1, two disks were
classified as nonconcordant pain. One of these had a rotation of 3.4° at a
level with no pain recorded on injection, an extruded disk, and spondylolysis
defects in the L5 pars. The average for the two disks classified as
nonconcordant pain was 2.6°. Disks classified as producing concordant pain
had an average rotation of 2.2°, and those classified as normal had an
average rotation of 1.0°.
View this table:
[in this window]
[in a new window]
|
TABLE 1: Average Rotations for the Disks Classified as Normal, Nonconcordant
Pain, or Concordant Pain at Discography
|
|
Of the disks classified as normal on discography because injection produced
no significant pain, five appeared abnormal on MRI. These included at
L4–L5, one with a radial tear in the posterior annulus fibrosus and two
others with a focally abnormal disk margin classified as a protrusion; at
L3–L4, one disk with a bulging annulus fibrosus; and at L5–S1, one
disk with a disk fragment in the spinal canal classified as an extrusion of
the disk.
The analysis of variance test showed significant (p 
0.001)
differences among the three disk classifications. At the lower three lumbar
disk levels, sufficient concordant pain disks were present to permit
statistical testing. At L3–L4, the difference between concordant disks
and the other categories was significant at a p value of 0.004. At
L4–L5, the difference between concordant and other disks was not
significant (p = 0.14). If the disks with abnormal morphology on MRI
were excluded from the disks classified as normal, the significance of the
difference was 0.05. At L5–S1, the difference between concordant disks
and the other two categories was not significant (p = 0.2). If the
disk with an extrusion was excluded from the disks classified as normal on the
basis of discography, the difference between the concordant pain group and the
others had marginal significance (p = 0.07).
Rotations greater than 1° were found in 17 (25%) of 68 normal disks and
20 (74%) of 27 nonconcordant or concordant disks. Rotations less than 1°
were found in 51 (75%) of 68 normal disks and seven (26%) of 27 disks with
nonconcordant or concordant pain.
Discussion
The largest rotation measurements were found at lumbar disk levels at which
concordant pain was found at discography. Statistical significance was
achieved in two of the three levels at which concordant pain was most
frequently encountered.
The patients with nonconcordant pain or absence of pain on injection were
lumped together on the assumption that both groups tested negative for
instability on the basis of discography. However, in this group were some
disks that had signs of extensive degeneration such as a disk herniation or
protrusion—despite the absence of concordant pain on discography. If
these disks were excluded from the normal group, increased statistical
significance was found.
Both spinal level and disk classification significantly correlated with
rotation measurements. Therefore, the disk level must be considered a
confounding factor in the measurement of vertebral rotation. To eliminate the
confounding effect of level, we analyzed results level by level, with
relatively small numbers of cases at each level.
This study is a preliminary one to estimate the utility of rotation
measurements in the lumbar spine to select patients for spinal fusion. Our
patient group was highly selected. We evaluated only patients who were
referred to us by spine surgeons for discography. Discography is not a gold
standard for determining the need for spinal fusion; however, the presence of
concordant pain at the level in question has been considered, at least by some
investigators, as an indication that spinal fusion at that level would have a
good outcome [1]. Because the
sample size was small, statistical inference must be made cautiously. Larger
sample sizes may not be feasible because referrals for discography have been
declining due to decreasing confidence in its capability for selecting spinal
fusion patients. The classification of the patient's pain experience during
discography is not always clear-cut. Some disks may have been incorrectly
classified as concordant or nonconcordant in this study. The effect of other
variables, such as a pars defect at one level, cannot be evaluated.
Conventional anatomic MR or CT images do not provide a reliable means with
which to detect the presence of degenerative spinal instability. Some studies
have suggested that MRI findings correlate with instability
[13], but these studies have
not in general been confirmed. Investigators found that decreased signal
intensity in the vertebral endplates on T1-weighted images correlates with
excessive translation at that level in flexion–extension radiographs
[14]. In another study,
annular tears were not accurate predictors of abnormal translation in
flexion–extension radiographs
[15]. No definite radiographic
or imaging findings are relied on in selecting patients for spinal fusion. No
MRI findings have sufficient predictive value to distinguish unstable motion
spinal segments.
The amount of rotation found in our study concords with previously
published studies. As in the study of Friberg and Hirsch
[16], we found disk
degeneration and instability were most frequent at L4–L5. As other
investigators have shown, we found that rotation varies with disk level
[17–20].
The relative amount of rotation for each level in our study does not agree
with previous reports, probably because of the small number of disks in our
study.
The preliminary data suggest that CT axial rotation measurements may be
useful in identifying a spinal level at which spinal fusion will have positive
results. Further study is needed to evaluate the predictive value of the CT
rotation study.
In summary, concordant pain at discography correlates with increased axial
rotation of the segment measured by a technique that uses CT and a program to
measure vertebral rotations. Further study is warranted to determine whether
this functional imaging test may be useful in selecting patients who might
benefit from spinal fusion.
References
- McCormick PC. Selection criteria for degenerative lumbar spine
instability. Clin Neurosurg 1997;44
: 29-39[Medline]
- Pope MH, Panjabi MM. Biomechanical definitions of spinal
instability. Spine 1985;10
: 255-260[Medline]
- Hayes MA, Howard TC, Gruel CR, Kopta JA. Roentgenographic
evaluation of lumbar spine flexion–extension in asymptomatic
individuals. Spine 1989;14
: 327-331[Medline]
- Pope MH, Frymoyer JW, Krag MH. Diagnosing instability.
Clin Orthop 1992;279
: 60-67
- Kirkaldy-Willis WH, Farfan HF. Instability of the lumbar spine.
Clin Orthop 1982;165
: 110-123
- Schmidt TA, An HS, Lim TH, Nowicki BH, Haughton VM. The stiffness
of lumbar spinal motion segments with a high-intensity zone in the annulus
fibrosis. Spine 1998;23
: 2167-2173[CrossRef][Medline]
- Nowicki BH, Haughton VM, Schmidt T, et al. Occult lumbar lateral
spinal stenosis in neural foramina subjected to physiologic loading.
Am J Neuroradiol 1996;17
: 1605-1614[Abstract]
- Carragee EJ, Tanner CM, Yang B, Brito JL, Truong T. False-positive
findings on lumbar discography. Spine1999; 24:2542
-2547[CrossRef][Medline]
- Carragee EJ, Tanner CM, Khurana S, et al. The rates of
false-positive lumbar discography in select patients without low back
symptoms. Spine 2000;25
: 1373-1381[CrossRef][Medline]
- Anderson MA. Lumbar discography: an update. Semin
Roentgenol 2004; 39:52
-67[Medline]
- Fenton DS, Czervionke LF. Discography. In: Fenton DA, ed.
Image-guided spine intervention. Philadelphia, PA:
Saunders, 2003: 227-255
- Rogers BP, Haughton VM, Arfanakis K, Meyerand ME. Application of
image registration to measurement of intervertebral rotation in the lumbar
spine. Magn Reson Med 2002;48
: 1072-1075[CrossRef][Medline]
- Toyone T, Takahashi K, Kitahara H, Yamagata M, Murakami M, Moriya
H. Vertebral bone-marrow changes in degenerative lumbar disc disease: an MRI
study of 74 patients with low back pain. J Bone Joint Surg
Br 1994; 76:757
-764
- Weishaupt D, Zanetti M, Hodler J, Boos N. MR imaging of the lumbar
spine: prevalence of intervertebral disk extrusion and sequestration, nerve
root compression, end plate abnormalities, and osteoarthritis of the facet
joints in asymptomatic volunteers. Radiology1998; 209:661
-666[Abstract/Free Full Text]
- Bram J, Zanetti M, Min K, Hodler J. MR abnormalities of the
intervertebral disks and adjacent bone marrow as predictors of segmental
instability of the lumbar spine. Acta Radiol1988; 39:18
-23
- Friberg S, Hirsch C. Anatomical and clinical studies on lumbar disc
degeneration. Acta Orthop Scand 1949;19
: 222-242[Medline]
- Panjabi M, Yamamoto I, Oxland T, Crisco J. How does posture affect
coupling of the lumbar spine? Spine 1989;14
: 1002-1011[Medline]
- Pearcy MJ. Stereo radiography of lumbar spine motion.
Acta Orthop Scand 1985;212
[suppl]: 1-45
- Tibrewal SB, Pearcy MJ, Portek I, Spivey J. A prospective study of
lumbar spinal movements before and after discectomy using biplanar
radiography: correlation of clinical and radiographic findings.
Spine 1985; 10:455
-460[Medline]
- Haughton VM, Rogers B, Meyerand ME, Resnick DK. Measuring the axial
rotation of lumbar vertebrae in vivo with MR imaging. Am J
Neuroradiol 2002; 23:1110
-1116[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
