DOI:10.2214/AJR.08.2026
AJR 2009; 193:W33-W37
© American Roentgen Ray Society
Defining Normal Vertebral Angulation at the Thoracolumbar Junction
Michael B. Crawford1,
Andoni P. Toms1 and
Lee Shepstone2
1 Department of Radiology, Norfolk and Norwich University Hospital, NHS
Foundation Trust, Colney Ln., Norwich, Norfolk NR4 7UY, United Kingdom.
2 School of Medicine, Health Policy and Practice, University of East Anglia,
Norwich, Norfolk, United Kingdom.
Received October 29, 2008;
accepted after revision January 2, 2009.
Address correspondence to M. B. Crawford
(Michael.crawford{at}nnuh.nhs.uk).
WEB This is a Web exclusive article.
Abstract
OBJECTIVE. The purpose of this cross-sectional study is to define
the normal range of endplate angulation at T12 and L1 and, by doing so, to
validate the angle measurement tools that are readily available on nearly all
PACS.
MATERIALS AND METHODS. Two hundred consecutive lateral scout CT
scans were examined in patients who were either 25 (n = 100) or 35
(n = 100) years old. The endplate angles for T12 and L1 were measured
using a "Cobb angle" tool on a standard PACS workstation.
Twenty-two cadaveric vertebrae were also imaged, and measurements obtained
from the lateral scout CT image using electronic calipers were compared with
measurements obtained with a goniometer.
RESULTS. The mean endplate angle at T12 measures 4.34° (2 SD,
4.5°) and at L1, 4.48° (4.26°). The normal range of endplate
angulation is therefore -0.16° to 8.84° at T12 and 0.22-8.74° at
L1. No statistically significant difference was seen in the endplate
angulation when men were compared with women or 25- and 35-year-old age groups
were compared. A strong correlation exists between direct and CT-derived
endplate angle measurements.
CONCLUSION. Vertebral endplate angulation can be reliably measured
using widely available PACS workstation tools. The mean endplate angle for T12
and L1 is approximately 4.5°, with an approximate range extending from
0° to 9°. For practical purposes, an endplate angle of 10° or more
can be considered outside the normal range.
Keywords: anatomy • CT • thoracolumbar junction • wedging
Introduction
Anterior angulation or "wedging" of the vertebral bodies
at the thoracolumbar junction is a recognized normal anatomic feature
[1]. The thoracolumbar junction
is also the most common location for osteoporotic and traumatic vertebral
fractures. More than 50% of traumatic fractures occur between T12 and L2
[1,
2]. The lower thoracic spine is
also the most common site of Scheuermann's disease and is a further cause of
vertebral wedging [3,
4]. Therefore, it can sometimes
be difficult to differentiate between grade 1 osteoporotic fractures, mild
traumatic fractures, and normal anatomic wedging. Several techniques have been
described to quantify vertebral deformity. These usually involve measuring and
comparing the relative anterior and posterior heights of vertebral bodies
[1,
5-7].
Although this quantification of vertebral shape provides objective measures
that can assist in the interpretation of thoracolumbar radiographs, the
technique is time-consuming and infrequently used. The purpose of this study
was to define the normal range of angulation of the endplates of the T12 and
L1 vertebral bodies using electronic tools that are widely available on all
diagnostic DICOM workstations. The validity of these tools to obtain
measurements of vertebral angulation was also evaluated.
Material and Methods
An analysis of lateral scout CT scans from 200 consecutive thoracoabdominal
CT examinations (LightSpeed Plus, GE Healthcare) was performed. Two hundred
consecutive patients, 100 each at 25 and 35 years old, were included in the
study beginning January 1, 2002, and ending December 31, 2004. All CT
examinations were harvested from the hospital PACS. Only the first CT scan of
a given patient was included. A lateral tomograph showing the entire
lumbosacral spine was required for inclusion in study. Exclusion criteria were
as follows: major trauma (such as an motor vehicle collision); any disease of
the vertebral column, spinal canal, paravertebral soft tissues, or
retroperitoneum; traumatic vertebral fractures; and known vertebral collapse.
Patients were also excluded if there were fewer or more than five lumbar
vertebrae or segmentation anomalies at the lumbosacral junction. In total, 222
CT studies were examined, with 22 being excluded, leaving a data set of 200.
Eight were excluded because the lateral tomographs were too pixilated to
clearly identify the outline of the vertebrae. Thirteen were excluded because
of lumbosacral ambiguity. One study was excluded because of retroperitoneal
disease.
View larger version (4K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1 —Diagrammatic representation of measurements of variable
endplate morphology on lateral projection. Lines intersecting anterior and
posterior superior corners and anterior and posterior inferior corners were
drawn without regard for shape or orientation of intervening endplate.
|
|
View larger version (46K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Reliability study of CT scout tomography. Lateral photograph
(A) and scout CT scan (B) of part of set of cadaveric lumbar
vertebrae used to assess reliability of CT scout tomograms for reproducing
vertebral geometry.
|
|
View larger version (36K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Reliability study of CT scout tomography. Lateral photograph
(A) and scout CT scan (B) of part of set of cadaveric lumbar
vertebrae used to assess reliability of CT scout tomograms for reproducing
vertebral geometry.
|
|
Measurements
Endplate angles for T12 and L1 were measured using electronic calipers from
the lateral CT tomographs. The superior endplate was defined as a line drawn
between the most anterosuperior and the most posterosuperior corners of the
vertebral body. The inferior endplate angle was similarly defined for the
anteroinferior and posteroinferior corners. The shape of the intervening
endplate was ignored (Fig. 1).
The angle between the two endplates was measured using the "Cobb
angle" tool on the PACS workstation
(Fig. 2).
CT examinations for a total of 200 patients (114 men and 86 women) were
included in the study and were independently examined by two observers. All CT
examinations were obtained with the patient in the supine position. The data
sets were then compared, and when there was a difference of more than 3°,
the case was reviewed by the two reporting observers and remeasured
independently in order to minimize discrepancies between the two observers
that might result from inadvertently measuring the incorrect vertebral level.
The second measurement was then accepted regardless of the difference between
observers. A total of 15 examinations were reviewed for which the second
readings were used in the final data set. Fifty examinations were measured
twice by the same observer (blinded), with an interval of 8 weeks between
measurements.
The rationale for using CT scout tomograms for assessing vertebral
angulation was tested using 22 cadaveric vertebrae. A lateral scout tomogram
was performed on the dry specimens using the standard CT protocol for chest
and abdominal CT (Fig. 3A, and
3B). Thereafter, the vertebral
endplate angles were measured independently by two observers. These
measurements were obtained from the dry bones using a goniometer and from the
CT images using the electronic Cobb angle tool.
Statistical Analysis
For the cadaveric vertebral study, interobserver reliability was tested
using intraclass correlation (ICC); a comparison of the dry bone and CT
measurements was performed using the Bland-Altmann plot and the Pearson's
correlation coefficient test.
For the study of the healthy CT population, the mean endplate angle was
calculated for T12 and L1 with a normal range defined as 2 SD from the mean.
Intra- and interobserver reliability was tested using ICCs. Comparisons
between T12 and L1 were performed using the paired Student's t test,
and between men and women and between patients 25 and 35 years old were
performed using the unpaired Student's t test.
To assess whether the results are applicable to conventional radiography
(in which the thoracolumbar junction is typically projected over the superior
region of the radiographic plate and subjected to magnification and parallax
distortion), a trigonometric model was used to estimate the maximal possible
alteration in endplate angulation. A hypothetic vertebral body measuring 4.5
cm in the anteroposterior dimension, having a 3.5 cm posterior body height,
and with an endplate angle of 9°—the upper limit of the normal
range—was used for the calculations. The hypothetic vertebral body was
considered to be positioned 20 cm from a 42 x 35 radiographic film with
a filmfocus distance of 100 cm and projected over the superior edge of the
film, which is the point of maximal parallax error
(Fig. 4).
View larger version (7K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4 —Diagrammatic representation of magnification and parallax
distortion of thoracolumbar vertebra projected onto superior edge of 42
x 35 cm radiographic plate. With film-focus distance of 100 cm (A + B),
beam at edge of film subtends angle of 11.9° to horizontal
[tan-1 (21 / 100)]. A vertebral body with endplate angle of 9°
(2 1), measuring 3.5 cm in posterior body height
(x1) and 4.5 cm in anteroposterior width
(y1), positioned 20 cm (A) from film will be magnified and
distorted because of parallax errors. From this the following can be estimated
using simple trigonometry. Position of vertebra from central beam (C) equals
16.8 cm, anterior body height (z1) equals 3.15 cm, and
superior and inferior endplate angles (1) measure 4.5°.
Posterior vertebral body height of projected image increases to 4.4 cm,
anterior body height increases to 3.5 cm, superior and inferior endplate
angles (2) decrease to 4.35° and 4.39°,
respectively. Dimensions x2, y2, and
z2 refer to posterior vertebral body. β = angle from
center to edge of x-ray beam.
|
|
Results
The mean endplate angle for T12 was 4.34°
(Table 1). Therefore, the
calculated normal statistical range is from -0.16° to 8.84°. The mean
endplate angle for L1 was 4.48°, with a normal statistical range of
0.22-8.74° (Fig. 5). No
statistical difference was seen between the two levels (p = 0.40)
(Table 2).
In women, the mean angle at T12 was 4.12° and at L1, 4.49°. In men,
the mean endplate angle at T12 was 4.5° and at L1, 4.7° degrees (Figs.
6 and
7). No significant difference
was seen between men and women at either T12 (p = 0.23) or L1
(p = 0.94) (Table
2).
In the 25-year-old patients, the mean vertebral endplate angle at T12 was
4.6° and at L1, 4.44° (Table
1). In the 35-year-old patients, the mean endplate angulation at
T12 measured 4.12° and at L1, 4.5° (Figs.
8 and
9). Again, no significant
difference was seen in the two age groups at either T12 (p = 0.16) or
L1 (p = 0.77) (Table
2).
The ICC for intraobserver reliability, which was based on the first 50 CT
examinations, was 0.800 (95% CI, 0.674-0.962) for T12 and 0.794 (0.665-0.923)
for L1. The interobserver reliability for all 200 measurements was 0.858
(0.794-0.923) for T12 and 0.808 (0.723-0.893) for L1. The mean difference for
the two observers was 1.2° at both T12 and L1 (2 SD, 1.8°).
The interobserver reliability for the direct (ICC, 0.972; 95% CI,
0.932-0.988) and CT-derived (0.963; 0.91-0.984) measures of the dry cadaveric
vertebra were very good. The correlation between the mean observed measures
showed a strong correlation between direct and CT-derived endplate angle
measurements (r = 0.961, p < 0.01). The Bland-Altman plot
shows that the mean CT measurement is 0.8° smaller than the dry bone
measurement (95% CI, -3.2° to 1.6°).
Trigonometric modeling of the alteration in endplate angulation on
conventional radiography yielded the following results. The angle of the
superior endplate is reduced by a maximum of 0.15°, and the inferior
endplate angle is reduced by a maximum of 0.11°
(Table 3).
View this table:
[in this window]
[in a new window]
|
TABLE 3 : Trigonometric Estimate of Effect of Parallax Error During Conventional
Radiography on Endplate Angulation of a Vertebra Projected Onto Edge of 42
x 35 cm Radiographic Plate
|
|
Discussion
The purpose of this study was to define the normal anatomic vertebral
angulation that occurs at the thoracolumbar junction. CT lateral scout
tomographs were used in this study for two reasons. The first is that the
lateral scout CT images do not have any geometric distortion. Although a fan
beam is used to obtain the image, the reconstruction algorithm corrects any
anteroposterior magnification. The validity of this was confirmed using the
dry bone phantom. The second reason is that for most patients in this age
group, thoracolumbar CT was incidental to the primary reason for examining the
patient. In contrast, radiographs and MR images of the spine almost always are
of symptomatic patients.
One may argue that the sample examined is not reflective of a normal
population because all patients included had an indication for a CT
examination. However, by selecting patients between the ages of 25 and 35
years, it is unlikely that any would by affected by osteoporosis. We believe
that in the normal population of women, vertebral morphology changes little
until the menopause [1]. The
lower age limit of 25 was defined to ensure that bone growth was complete, and
the incidence of osteoporosis is rare in the 35-year age group. Careful
selection of the criteria for CT and exclusion of abnormal findings should
also have excluded those with abnormal vertebral morphology due to
disease.
Selecting 2 SD is a conventional statistical description of a normal range
in a normal Gaussian distribution of data. However, it has been previously
highlighted that this is an artificial definition, which will mean that a few
healthy patients will fall outside this range
[7].
A potential selection bias may have occurred because patients were excluded
for poor quality of the lateral tomographs. This process would tend to select
out patients with larger body mass indexes but whose vertebral morphology may
be considered entirely normal.
Despite these limitations, this sample is likely to be as close to normal
as can be achieved in an analysis using clinical data. To perform normal
population sample studies—that is, drawn directly from the community
without any symptoms or known conditions—with radiography or MRI would
not be financially or ethically viable.
The range of values for angulation of the T12 and L1 vertebrae were derived
from the mean of the two observers' measurements of 200 cases. The very good
intraclass coefficient for the intra- and interobserver variability suggests
that this is a reliable and reproducible set of measurements. The
interobserver correlation may be slightly higher than in normal clinical
practice because 15 observations were repeated, and because there was a
greater than 3° difference between the observers. This repetition was
performed to minimize any measurement errors attributable to mistakes in
counting vertebral levels and transcribing measurements. These are part of
variance between observers, but the repeat measurements were performed so that
most of the interobserver variability could be attributed to measurement
errors only.
The results of this study differ from previous published data in several
areas. Previous studies have mostly included women because of their greater
propensity for osteoporosis [1,
8-10],
have included fewer patients
[1,
8-12],
and have used different methods to evaluate vertebral body morphology
[1,
8,
9,
11,
12]. Most previous methods
involved measuring the relative heights of the anterior, middle, and posterior
aspects of vertebral bodies [1,
8,
9,
11,
12]. Panjabi et al.
[13,
14] showed similar results for
vertebral angulation at T12 (4.0° ± 1.11°) but a slightly
higher degree of angulation at L1 (6.7° ± 1.61°). However,
their results were derived from the dry bones of only 12 individuals (mean
age, 46.3 years), some of whom had significant comorbidity at death (six
patients had cancer). We would argue that our sample is a better and more
normal population sample. The advantages of our approach are that we have
included a larger number of patients, have included both men and woman, and by
using CT we have reduced as much as possible the influence of geometric
distortion. Angulation of endplates is not dependent on size of the vertebra,
whereas absolute height measurements may vary with sex and height of
individual.
The mean angle was slightly greater in men than women at T12 and the
converse was true at L1. However, neither of these is statistically
significant. With mean differences at T12 of 0.4° and at L1 of 0.02°,
the normal ranges for both levels and sexes are therefore very similar. These
findings differ from previously published data in which men appear to
consistently have a greater degree of vertebral wedging than women
[15]. This discrepancy may be
explained by the relatively young age of our sample group. Although wedging in
the women is constant before menopause, postmenopausal women have an increased
incidence relative to men of a similar age. The similarity in the measurements
of vertebral wedging in our two age groups supports a recently published
cadaveric study, which suggested that vertebral morphology is constant with
increasing age [16].
One question that arises after taking data from a source with limited
geometric distortion is whether the values for normal range of angulation can
be translated into conventional radiographic images of the lumbar spine. The
trigonometric modeling of the parallax effect on endplate angulation suggests
that the maximum alteration is an apparent reduction in endplate angle of
0.26°. This calculation suggests that, for practical purposes, the
measures derived from the lateral CT radiograph are applicable to conventional
radiography of the thoracolumbar junction. This finding is supported by a
previous study that concluded that cephalograms and CT scanograms are
comparable for depicting angular relations of structures
[17].
The results of this study indicate that measuring vertebral endplate
angulation at T12 and L1 using electronic calipers is a reliable and
reproducible technique. In routine practice, this would be quick and easy to
do if there was any concern about the amount of angulation of the
thoracolumbar vertebra after visual inspection. The sum of the normal
statistical range of endplate angulation at T12 and L1 (2 SD = 8.75°),
estimated geometric distortion (0.26°), and the mean interobserver
differences (1.2°) is 10.2°. We would suggest that 10° might be a
useful rule of thumb for the maximal statistically "normal"
endplate angulation measurement from a lateral radiograph, and that
measurements outside the normal range may indicate a fracture or collapse even
if cortical or trabecular disruption is not visible. The converse, however, is
not true; an endplate angle of less than 10° does not exclude a
fracture.
In conclusion, measurement of vertebral body endplate angulation using the
method described in this article is a reliable and reproducible technique.
Assuming that our study sample is a normal population, the range of normal
angulation is approximately 0-9° for T12 and L1. For practical purposes, a
vertebral angulation of 10° or more could be considered to be outside the
normal range.
References
- Gallagher JC, Hedlund LR, Stoner S, Meeger C. Vertebral
morphometry: normative data. Bone Miner1988; 4:189
-196[Medline]
- Holmes JF, Miller PQ, Panacek EA, Lin S, Home NS, Mower WR.
Epidemiology of thoracolumbar spine injury in blunt trauma. Acad
Emerg Med 2001; 8:866
-872[CrossRef][Medline]
- Ali RM, Green DW, Patel TC. Scheuermann's kyphosis. Curr
Opin Pediatr 1999; 11:70
-75[CrossRef][Medline]
- Tribus CB. Scheuermann's kyphosis in adolescents and adults:
diagnosis and management. J Am Acad Orthop Surg1998; 6:36
-43[Abstract]
- Jensen KK, Tougaard L. A simple x-ray method for monitoring
progress of osteoporosis. Lancet 1981;2
: 19-20[CrossRef][Medline]
- Black DM, Cummings SR, Stone K, Hudes E, Palermo L, Steiger P. A
new approach to defining normal vertebral dimensions. J Bone Miner
Res 1991; 6:883
-892[Medline]
- Leidig G, Minne HW, Sauer P, et al. A study of complaints and their
relation to vertebral destruction in patients with osteoporosis.
Bone Miner 1990;8
: 217-229[CrossRef][Medline]
- Davies KM, Recker RR, Heaney RP. Normal vertebral dimensions and
normal variation in serial measurements of vertebrae. J Bone Miner
Res 1989; 4:341
-349[Medline]
- McCloskey EV, Spector TD, Eyres KS, et al. The assessment of
vertebral deformity: a method for use in population studies and clinical
trials. Osteoporos Int 1993;3
: 138-147[CrossRef][Medline]
- Ryan PJ, Fogelman I. Osteoporotic vertebral fractures: diagnosis
with radiography and bone scintigraphy. Radiology1994; 190:669
-672[Abstract/Free Full Text]
- Dai L. Radiographic measurements of the heights of vertebral bodies
in thoracic and lumbar spine. Chin Med Sci J1996; 11:117
-119[Medline]
- Hurxthal LM. Measurement of anterior vertebral compressions and
biconcave vertebrae. Am J Roentgenol Radium Ther Nucl
Med 1968; 103:635
-644[Medline]
- Panjabi MM, Goel V, Oxland T, et al. Human lumbar vertebrae.
quantitative three-dimensional anatomy. Spine1992; 17:299
-306[CrossRef][Medline]
- Panjabi MM, Takata K, Goel V, et al. Thoracic human vertebrae.
quantitative three-dimensional anatomy. Spine1991; 16:888
-901[Medline]
- Lauridsen KN, De Carvalho A, Andersen AH. Degree of vertebral
wedging of the dorso-lumbar spine. Acta Radiol Diagn
(Stockh) 1984; 25:29
-32[Medline]
- Masharawi Y, Salame K, Mirovsky Y, et al. Vertebral body shape
variation in the thoracic and lumbar spine: characterization of its asymmetry
and wedging. Clin Anat 2008;21
: 46-54[CrossRef][Medline]
- Chidiac JJ, Shofer FS, Al-Kutoubi A, Laster LL, Ghafari J.
Comparison of CT scanograms and cephalometric radiographs in craniofacial
imaging. Orthod Craniofac Res 2002;5
: 104-113[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
