AJR 2005; 184:671-675
© American Roentgen Ray Society
Comparison of MRI and Conventional Radiography for Assessment of Acromial Shape
Marius E. Mayerhoefer1,
Martin J. Breitenseher1,
Andreas Roposch2,
Christina Treitl1 and
Christian Wurnig3
1 Department of Radiology, Medical University of Vienna/Austria, Waehringer
Guertel 18-20, Vienna, 1090 Austria.
2 Present address: Department of Orthopedics, University of Toronto, Toronto,
ON, Canada. At time of study: Medical University of Vienna, Department of
Orthopaedics, Vienna, Austria.
3 Second Orthopaedic Department, Orthopaedic Hospital Vienna-Speising, Vienna,
Austria.
Received May 10, 2004;
accepted after revision July 14, 2004.
Address correspondence to M. E. Mayerhoefer
(marius_mayerhoefer{at}aon.at).
Abstract
OBJECTIVE. Our aim was to determine the value of different MRI
planes independently and in combination for assessment of acromial shape.
MATERIALS AND METHODS. Sixty-one patients with subacromial
impingement syndrome who had undergone acromioplasty after failure to respond
to conservative treatment were included in the study. Parasagittal T2-weighted
MR images and outlet view radiographs of the affected shoulders were acquired
preoperatively. Three-dimensional models of all acromions were constructed
from the MR images, and the Bigliani type of acromion depicted by these models
was determined. Results were compared with the acromial type assessed during
acromioplasty. To provide a reliable reference for further processing and
correlation, we used only those 56 acromions with agreement on acromial shape
between intraoperative findings and 3D models. Then, acromial shape was
determined for three MRI slice positions (S-1, lateral acromial edge; S-2,
just lateral of acromioclavicular joint; and S-3, lateral portion of
acromioclavicular joint), for a combination of S-1 and S-2, and for the
radiographs.
RESULTS. Kappa coefficients were 0.36 (36%) for S-1, 0.41 (41%) for
S-2, and –0.10 (–10%) for S-3. For the outlet view radiographs,
the kappa coefficient was 0.55 (55%), showing better correlation than any
single slice position. Best results, however, were achieved with a combination
of S-1 and S-2, with a kappa coefficient of 0.66 (66%).
CONCLUSION. For determination of acromial shape, outlet view
radiographs are superior to any single MRI slice position, but inferior to a
combination of two MRI slices (S-1 and S-2). If a single MRI slice is being
used, the slice position just lateral to the acromioclavicular joint is
recommended.
Introduction
Subacromial outlet impingement is the result of mechanical irritation of
the rotator cuff. One possible cause for this irritation is an abnormal shape
of the anterior acromial undersurface. On the basis of dried cadaver specimens
and conventional outlet view radiographs, Bigliani et al.
[1] described three different
types of acromion: the flat type 1, the curved type 2, and the hooked type 3.
Whereas type 1 represents the physiologic shape, type 2 and, especially, type
3 acromions are abnormal variants and are regarded as important factors for
the development of outlet impingement.
Because Morrison and Bigliani
[2] originally reported good
correlation of rotator cuff tears with outlet view radiographs, this imaging
method has been considered the gold standard for in vivo assessment of
acromial shape. Subsequent studies, however, have shown that correlation
decreases if the radiographs are obtained by different technologists and that
minor changes of the angle of the central beam may influence considerably the
depiction of acromial shape by these radiographs
[3,
4].
Although in MRI, there is no projection error inherent to conventional
X-ray, determination of acromial shape is considered difficult with this
tomographic method. This is probably due to the observation that the
appearance of the acromial shape is highly dependent on the position of the
MRI plane and therefore may change from slice to slice, as proposed in a
previous study [4]. Although
various different slice positions have been recommended in the literature, to
our knowledge, no direct comparison of these slice positions has been
performed. Because there is no agreement on which slice position to use,
MRI-based determination of acromial shape could be prone to a high
interobserver variability.
It was the aim of our study to compare the diagnostic value of three
different MRI slice positions, independently and in combination, with the
value of conventional outlet view radiographs. In addition, we tested the
hypothesis that acromial shape frequently changes between sections and appears
more curved or hooked on medial sections.
Materials and Methods
Sixty-one patients with clinically confirmed impingement syndrome were
included in this retrospective study. Written informed consent was obtained
from each patient. Impingement syndrome was diagnosed if both Neer's
impingement sign (pain at passive elevation and internal rotation) and Neer's
impingement test (pain ceases after injection of local anesthetics into the
subacromial space) were positive. All patients had undergone arthroscopic
acromioplasty after failure to respond to conservative treatment for at least
6 months. For preoperative evaluation of the acromion, conventional outlet
view radiographs, obtained by different technologists, and parasagittal
T2-weighted MR images, perpendicular to the supraspinatus tendon, were used.
After planning on an axial localizer, we obtained MR images with a 1.0-T MRI
scanner (Philips) equipped with a dedicated shoulder surface coil, using a TR
of 2,000 msec, a TE of 140 msec, a 256 x 256 matrix, and a slice
thickness of 4 mm, which are often used for examinations of the shoulder. All
MR images were stored in DICOM format to allow optimal image processing.
Three-dimensional models of the acromions were constructed from the
parasagittal MR images with the 3D Slicer software package (Massachusetts
Institute of Technology) by manual outlining of the acromion on each slice.
Then, the Bigliani type of acromion depicted by the 3D models was determined
in consensus by a radiologist and an orthopedic surgeon on the basis of work
by Epstein et al. [5], who
categorized an acromion with down-sloping in its anterior third as Bigliani
type 3 and an acromion with down-sloping in the middle section as Bigliani
type 2. The results were compared with the acromial shape assessed during
acromioplasty. To provide a reliable reference for evaluation of the different
MRI slices and outlet view radiographs, we used only those 56 acromions
(90.3%) with agreement on acromial shape between intraoperative findings and
the 3D model for further processing and assessment of correlation because the
combination of the 3D model and intraoperative findings appeared to us to be
the best in vivo approximation of direct evaluation of cadaver specimens.
From each of the remaining 56 MRI data sets, three different MRI slices
were extracted. The first slice position (S-1), which was used by Peh et al.
[4] for assessment of acromial
shape, was located 4 mm from the lateral margin of the acromion. The second
slice position (S-2), described by Epstein et al.
[5], was located just lateral
to the acromioclavicular joint. Finally, the third slice position (S-3),
located just medial to S-2, depicted the most lateral portion of the
acromioclavicular joint, where the lateral part of the joint capsule and the
acromioclavicular ligament, but usually not yet the clavicula, are visible
(Fig. 1).
View larger version (68K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —Three-dimensional reconstruction of acromion and clavicula
with manually schematized acromioclavicular joint capsule (yellow),
illustrating the three MRI slice positions: S-1, S-2, and S-3. S-3 already
depicts most lateral part of joint capsule. Note that because of individual
differences in acromial size, S-1 and S-2 are not always located directly next
to each other.
|
|
Assessment of acromial shape in parasagittal MR images was performed
mathematically using the Osiris software package (Fig.
2A,
2B). For every slice position
(S-1, S-2, and S-3), a line connecting the most caudal margins of the acromial
undersurface was drawn and its length measured. Then, the line was divided by
three and, with the help of two orthogonal lines, the acromial undersurface
was divided into three segments of equal length. Then, the angle between the
anterior third and the posterior two thirds of the acromion was measured.
Acromions with an angle of 10° or less were considered type 1 acromions,
and those with an angle between 11° and 20°, type 2 acromions. If an
angle of more than 20° was found, the angle between the posterior third
and the anterior two thirds was also measured. If this angle was 10° or
less, the acromion was classified as type 3; otherwise, it was classified as
an extreme type 2. In addition, the arithmetic mean of the angles measured in
S-1 and S-2 was calculated, and the results were classified as described.
View larger version (68K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —Mathematic determination of acromial shape. Line connecting
most caudal points of acromial undersurface is drawn on parasagittal
T2-weighted MR image, and with help of two orthogonal lines, acromion is
divided into three segments of equal length.
|
|
View larger version (66K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B. —Mathematic determination of acromial shape. Then on
parasagittal T2-weighted MR image, angle between anterior third and posterior
two thirds is measured. In this case, anterior angle of deflection is 180°
– 168° = 12°, indicating slightly curved (type 2) acromion in
S-2 position.
|
|
The type of acromion depicted on the outlet view radiographs was also
determined in consensus by a radiologist and an orthopedic surgeon, again on
the basis of the recommendations of Epstein et al.
[5].
To determine the correlation of acromial shape from the different slice
positions and the outlet view radiographs with our reference (agreement
between the 3D model and the intraoperative record), kappa coefficients and
95% confidence intervals (CI) were calculated. Following Fleiss's
[6] interpretation scheme, we
regarded kappa coefficients as poor, if below 0.40; as fair, if between 0.41
and 0.59; as good, if between 0.60 and 0.74; and as excellent if 0.75 or
higher.
To provide a more detailed insight into the strengths and weaknesses of
each method and slice position, sensitivity and specificity were calculated
for the three Bigliani types of acromion.
Results
Of the 56 acromions that showed agreement between the 3D model and the
intraoperative findings, 10.7% (n = 6) were classified as type 1;
82.1% (n = 46), as type 2; and 7.1% (n = 4) as type 3.
Distributions of acromial shape assessed at different slice positions and in
outlet view radiographs are listed in Table
1. From the lateral edge of the acromion to the acromioclavicular
joint, an increase in hooked acromions (7.1% in S-1, 16.1% in S-2, and 37.5%
in S-3) and a decrease in flat acromions (30.4% in S-1, 21.4% in S-2, and
19.6% in S-3) were noted on MR images.
Kappa coefficients were 0.36 (poor) (95% CI, 0.14–0.59) for S-1 and
0.41 (fair) (95% CI, 0.18–0.64) for S-2. As expected, S-3 proved to be
the least valuable slice position, with a kappa coefficient of –0.1 (95%
CI, –0.16 to 0.13), which indicated correlation slightly worse than
chance. The kappa coefficient of the outlet view radiographs was 0.55 (fair)
(95% CI, 0.31–0.80), thus showing better correlation than any single MRI
slice position. However, the arithmetic mean of S-1 and S-2 scored even better
with a kappa coefficient of 0.66 (good) (95% CI, 0.44–0.89).
Sensitivity and specificity were calculated as listed in
Table 2. Although for the
detection of the highly abnormal type 3 acromions, S-2 was more sensitive than
radiographs, specificity was substantially higher in radiographs for this type
of acromion. Again, the arithmetic mean of S-1 and S-2 generally provided even
better sensitivity and specificity than radiographs.
Discussion
We attempted to compare the value of different MRI slice positions for
determination of acromial shape. It was of particular interest to test whether
the different MRI slice positions, either independently or in combination,
would provide better results than outlet view radiographs, which are most
commonly used for evaluation of the acromion.
The clinical significance of acromial shape was first described by Bigliani
et al. [1], who emphasized the
importance of anterior acromial down-sloping for the development of rotator
cuff tears. According to these authors, acromions can be divided into a
nonpathologic flat acromion (type 1), pathologic curved (type 2), and hooked
(type 3) acromions (Figs. 3A,
3B,
3C,
4A,
4B,
4C,
5A,
5B,
5C). This rather subjective
definition was later refined by Epstein et al.
[5], who proposed that curved
acromions are characterized by down-sloping in the middle third of the
acromion, whereas in hooked acromions, down-sloping occurs in the anterior
third.
View larger version (174K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A. —41-year-old man with impingement syndrome of the shoulder.
Both outlet view radiograph (A) and MR image (B) in position S-2
show flat type 1 acromion, which was also confirmed by 3D model (C) and
intraoperatively.
|
|
View larger version (127K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —41-year-old man with impingement syndrome of the shoulder.
Both outlet view radiograph (A) and MR image (B) in position S-2
show flat type 1 acromion, which was also confirmed by 3D model (C) and
intraoperatively.
|
|
View larger version (33K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C. —41-year-old man with impingement syndrome of the shoulder.
Both outlet view radiograph (A) and MR image (B) in position S-2
show flat type 1 acromion, which was also confirmed by 3D model (C) and
intraoperatively.
|
|
View larger version (164K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —47-year-old woman with impingement syndrome of shoulder. Both
outlet view radiograph (A) and MR image (B) in position S-2 show
curved type 2 acromion, which is also confirmed by 3D-model (C) and
intraoperatively.
|
|
View larger version (151K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —47-year-old woman with impingement syndrome of shoulder. Both
outlet view radiograph (A) and MR image (B) in position S-2 show
curved type 2 acromion, which is also confirmed by 3D-model (C) and
intraoperatively.
|
|
View larger version (36K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C. —47-year-old woman with impingement syndrome of shoulder. Both
outlet view radiograph (A) and MR image (B) in position S-2 show
curved type 2 acromion, which is also confirmed by 3D-model (C) and
intraoperatively.
|
|
View larger version (137K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A. —63-year-old man with impingement syndrome of shoulder.
Whereas outlet view radiograph (A) depicts curved acromion (type 2), MR
image (B) in position S-1 reveals blunt acromial hook (type 3), which
is also confirmed by 3D model (C) and intraoperatively.
|
|
View larger version (81K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B. —63-year-old man with impingement syndrome of shoulder.
Whereas outlet view radiograph (A) depicts curved acromion (type 2), MR
image (B) in position S-1 reveals blunt acromial hook (type 3), which
is also confirmed by 3D model (C) and intraoperatively.
|
|
View larger version (42K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C. —63-year-old man with impingement syndrome of shoulder.
Whereas outlet view radiograph (A) depicts curved acromion (type 2), MR
image (B) in position S-1 reveals blunt acromial hook (type 3), which
is also confirmed by 3D model (C) and intraoperatively.
|
|
In their original study from 1986, Morrison and Bigliani
[2] describe good correlation
of rotator cuff tears and acromial shape as depicted by outlet view
radiographs, all of which were obtained by a single technologist. A subsequent
study by the same authors revealed that correlation decreases significantly if
the radiographs are obtained by different technologists
[3]. In 1993, Peh et al.
[4] confirmed that acromial
morphology in outlet view and Y radiographs is sensitive to minor variations
in radiographic technique, especially in angulation of the central beam, as
might occur between different technologists. Despite these results, outlet
view radiographs remained the standard for in vivo assessment of acromial
shape.
When MRI was introduced into clinical routine, it was expected that as a
tomographic method, it would be superior to radiographs in identifying
acromial shape and would provide more consistent results. However, a
comparison of the diagnostic value of MRI with that of radiographs, with
regard to the determination of acromial morphology, resulted in a higher
correlation of rotator cuff abnormalities with radiographic rather than with
MRI findings [4]. In addition,
interobserver variability of acromial shape was significantly higher for MRI
than for radiographs [7]. These
surprising results were generally attributed to the observation that in MRI,
acromial shape may change with the slice position chosen for review
[4]. Difficulty in comparing
results from the previously mentioned studies is due to the fact that
different MRI sections were used in each study. Whereas Epstein et al.
[5] chose the section just
lateral to the acromial joint, Peh et al.
[4] preferred a section 4 mm
from the lateral edge of the acromion. Haygood et al.
[7] allowed reviewers to choose
the shape most likely associated with rotator cuff abnormalities from the
entire set of MR images.
On the basis of these studies, we compared the diagnostic value of
different slice positions. S-1 was considered equivalent to the slice position
used by Peh et al. [4], and S-2
was considered similar to that used by Epstein et al.
[5]. Slice position S-3,
located in the lateral portion of the acromioclavicular joint, where according
to our experience, the acromion most frequently appears hooked, seemed
comparable to the slices examined by Haygood et al.
[7]. To make full use of the
multiplanar imaging capabilities of MRI, we also combined the measurements of
slice positions S-1 and S-2. Whereas a previously described mathematic scheme
for classification of acromial type relies on measuring the angle between a
line through the acromial midsubstance and a second line through the point of
deflection in the acromion [8],
we focused on the acromial undersurface, mathematically dividing it into three
segments of equal length. This was done because we believe that the previously
mentioned point of deflection can hardly be localized accurately by an
observer, particularly in curved (type 2) acromions. In contrast, the angles
measured with our own method are precisely defined. Mathematic determination
of acromial morphology, however, was only used for MR images, but not for
outlet view radiographs, because as a consequence of the radiographic
summation technique, the acromial undersurface was, in many cases, not clearly
visible in its entirety. A factor that may account for the high percentage of
type 2 (curved) acromions (82%) assessed in our study was the adoption of the
interpretation of Epstein et al.
[5] of the Bigliani
categorization, which also influenced our mathematic classification scheme for
MR images. However, imaging of most of the patients who undergo subacromial
decompression depicts a type 2 acromion
[9]. In addition, a similar
percentage of type 2 acromions (77%) in symptomatic patients was also assessed
by Peh et al. using radiographs.
Conclusions that may be drawn from our results can be summarized as
follows: Apparent acromial morphology is highly sensitive to changes in the
imaging plane, with an increase in abnormal shapes from the lateral to the
medial sections. For determination of acromial shape, outlet view radiographs
are superior to any single MRI slice position. If, for any reason, a single
slice is being used for examination of acromial morphology, we recommend the
slice position just lateral to the acromioclavicular joint (S-2), where
neither the joint capsule nor the acromioclavicular ligament is yet visible.
If this slice position depicts a type 3 acromion, the slice located 4 mm from
the lateral edge of the acromion (S-1) may be helpful for evaluation because
it has a high specificity for this type of acromion. In our study, the best
results were achieved through a combination of measurements in slice positions
S-1 and S-2 (calculation of arithmetic mean of angle measurements). With a
kappa coefficient of 0.66, this two-slice combination appears to be superior
to outlet view radiographs and may, therefore, be recommended, if time
permits. We caution against using a slice position inside the
acromioclavicular joint for determination of acromial shape because results
assessed in our S-3 position indicate only chance correlation.
References
- Bigliani LU, Morrison DS, April EW. The morphology of the acromion
and its relationship to rotator cuff tears. (abstr) Orthop
Trans 1986;10:228
- Morrison DS, Bigliani LU. Roentgenographic analysis of acromial
morphology and its relationship to rotator cuff tears. Orthop
Trans 1987;11:439
- Morrison DS, Bigliani LU. The clinical significance of variations
in acromial morphology. Orthop Trans1987; 11:234
- Peh WC, Farmer TH, Totty WG. Acromial arch shape: assessment with
MR imaging. Radiology1995; 195:501
–505[Abstract/Free Full Text]
- Epstein RE, Schweitzer ME, Frieman BG, Fenlin JM Jr, Mitchell DG.
Hooked acromion: prevalence on MR images of painful shoulders.
Radiology1993; 187:479
–781[Abstract/Free Full Text]
- Fleiss JL. portions, Statistical methods for rates and
proportions 2nd ed. New York, NY: Wiley, 1981:218
- Haygood TM, Langlotz CP, Kneeland JB, Iannotti JP, Williams GR Jr,
Dalinka MK. Categorization of acromial shape: interobserver variability with
MR imaging and conventional radiography. AJR1994; 162:1377
–1382[Abstract/Free Full Text]
- MacGillivray JD, Fealy S, Potter HG, O'Brien SJ. Multiplanar
analysis of acromion morphology. Am J Sports Med1998; 26:836
–840[Abstract/Free Full Text]
- Park JY, Levine WN, Marra G, Pollock RG, Flatow EL, Bigliani LU.
Portal-extension approach for the repair of small and medium rotator cuff
tears. Am J Sports Med2000; 28:312
–316[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
