DOI:10.2214/AJR.08.1091
AJR 2008; 191:1406-1411
© American Roentgen Ray Society
CT-Based Patient-Specific Modeling of Glenoid Rim Defects: A Feasibility Study
Gerd Diederichs1,2,
Heiko Seim3,
Henning Meyer1,
Ahi S. Issever1,2,
Thomas M. Link2,
Ralf J. Schröder1 and
Markus Scheibel4
1 Department of Radiology, Charité-Universitätsmedizin Berlin,
Charité Campus Mitte, Chariteplatz 1, Berlin, Germany 10117.
2 Department of Radiology, Musculoskeletal and Quantitative Imaging Research
Group, University of California, San Francisco, San Francisco, CA.
3 Department of Visualization and Data Analysis, Medical Planning Group,
Zuse-Institute Berlin, Berlin, Germany.
4 Center for Musculoskeletal Surgery, Charité-Universitätsmedizin
Berlin, Charité Campus Virchow, Berlin, Germany.
Received April 20, 2008;
accepted after revision June 9, 2008.
Address correspondence to G. Diederichs
(gerd.diederichs{at}charite.de).
There are two versions of the Amira software: a commercial version
developed, maintained, and sold by Visage Imaging, Mercury Computer Systems,
and a scientific version based on the commercial version but developed and
maintained at the Zuse-Institute Berlin, Department of Visualization and Data
Analysis, Medical Planning Group. The scientific version, which was used for
this study, was provided free by the Zuse-Institute Berlin.
Abstract
OBJECTIVE. Reconstruction of glenoid bone defects requires accurate
preoperative planning. The purpose of this study is to present a method for
quantifying the defect size and generating a 3D model of the bone graft for
augmentation by matching the fractured glenoid with the contralateral
side.
MATERIALS AND METHODS. Ten paired shoulders from five cadavers
(subjects: three women and two men; mean age, 85 years) and 60 paired
shoulders in 30 patients (controls: nine women and 21 men; mean age, 21 years)
were examined using CT to determine bilateral comparability by assessment of
the maximum glenoid diameters, surface area, and volume. After creation of a
glenoid rim defect in the study group, repeated CT scans were superimposed
with the data from the contralateral side. The defect size was quantified and
the missing fragment virtually reconstructed. Accuracy was evaluated by
comparing the virtually repaired glenoid with the predefect CT scan.
RESULTS. There were no significant side-to-side differences in
intact shoulders (p < 0.05). After creation of the glenoid
defects, there was a mean decrease of 31% in the anteroposterior diameter, 34%
in surface area, and 19% in volume. The virtually reconstructed glenoids did
not differ significantly from the predefect CT scans. The averaged
predefect-to-postdefect difference was 3% for the anteroposterior diameter
(R2 = 0.71), 6% for the surface area
(R2 = 0.82), and 4% for the volume (R2
= 0.98).
CONCLUSION. A precise 3D model of the glenoid bony defect can be
generated. The computer simulation provides a virtual model of the bone graft,
which may potentially improve arthroscopic bone augmentation.
Keywords: 3D model • CT • defect size • glenoid fracture • glenoid rim • modeling • shoulder instability
Introduction
Traumatic dislocation of the shoulder can be associated with a bony defect
of the glenoid rim and may lead to chronic instability of the glenohumeral
joint. Reports in the literature indicate that an anteroinferior glenoid rim
defect is present in 80–90% of individuals with chronic anterior
instability of the shoulder
[1–3].
Instability increases with the size of the defect
[4], and larger defects have a
poorer clinical outcome after soft-tissue shoulder stabilization procedures
[5,
6]. In cases of large defects,
reconstruction of the glenoid rim can be performed using an autologous iliac
crest bone graft [7,
8]. Whereas initially
autologous glenoid reconstruction was done in an open surgical approach,
recent studies describe arthroscopic techniques as well
[9,
10]. The goal of both
procedures is to fit the bone graft to the defect to restore the original
anatomic configuration of the glenoid concavity. Therefore, accurate
preoperative planning is important if an anatomic reconstruction is to be
achieved. Nevertheless, correct quantification of the bony defect size remains
a challenge.
The basic problem for accurate reconstruction of the glenoid cavity is that
the initial glenoid morphology before trauma remains unknown. The missing bone
is often fragmented, and therefore its configuration is not obvious because
there is wide interindividual variation in glenoid joint morphology
[11]. Accordingly, shaping of
the bone graft is often done by rough optical estimation, which may be prone
to inaccuracies. A small bone graft can cause recurrent instability, whereas a
large graft may lead to irritation of the rotator cuff, limit the range of
motion of the shoulder, or finally lead to osteoarthritis of the glenohumeral
joint.
Optimal preoperative imaging may be able to provide the surgeon with an
accurate reconstruction of the bone defect, which can be used as a model for
shaping and fitting the bone graft. MRI is primarily precise for the
presurgical imaging of labral and tendon injury in shoulder instability
[12–14].
For assessment of bony defects, MDCT with multiplanar reconstruction (MPR)
enables accurate visualization of glenoid fractures, and a high sensitivity
and specificity in relation to arthroscopy has been reported
[2,
15]. A presurgical approach to
quantify the defect size and display the missing bone fragment is comparison
with the intact contralateral shoulder. Previously published data suggest that
there is only minimal difference between the two glenoids in one individual
[3,
16]. Three-dimensional
reconstruction of the predefect morphology of the glenoids on the basis of the
normal side can be accomplished by acquiring a CT volume data set of both
shoulders, which serves to virtually project the fractured glenoid into the
contralateral glenoid. Using this technique, one can generate images of the
bone fragment, which is theoretically consistent with the ideal bone graft, to
reconstruct the morphology of the glenoid before the trauma.
The aim of this experimental cadaveric study was to develop a method for
quantifying glenoid rim defects and virtually generating a 3D model of the
missing bone fragment to display a standard of reference for the bone
graft.
Materials and Methods
Controls and Study Subjects
The group that served as controls for bilateral glenoid size comparison was
composed of 60 paired shoulders in 30 trauma patients (21 men and nine women)
referred for CT by the emergency department of our hospital. The study
protocol was approved by the institutional review board. The mean age at
scanning was 21 years (age range, 18–45 years). All patients were
examined by a whole-body scanner using a standardized multiple-trauma CT
protocol. Patients with glenoid fractures, osteolysis, and higher-grade
degenerative changes were excluded on the basis of CT morphology. CT scans of
the controls were used only for bilateral glenoid measurements to prove
intersite consistency. The scans were retro spectively evaluated. The reasons
for selecting trauma patients as controls were the consistent ability to
assess the glenoids on the CT scans and the young mean age of the cohort,
providing normal bone morphology.
View larger version (50K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —Three-dimensional surface renderings from CT scans for
analysis of glenoid parameters in 25-year-old man with normal glenoid from
control group. Largest length of glenoid joint surface was measured from top
to bottom (blue line) and in anteroposterior direction (red
line) in en face view of articular surface.
|
|
View larger version (58K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —Three-dimensional surface renderings from CT scans for
analysis of glenoid parameters in 25-year-old man with normal glenoid from
control group. Area for measuring surface area was defined along outer bony
edge of glenoid rim.
|
|
View larger version (53K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —Three-dimensional surface renderings from CT scans for
analysis of glenoid parameters in 25-year-old man with normal glenoid from
control group. For measurement of glenoid volume, articular body was aligned
parallel to joint surface and separated medially at distance of 10 mm (red
line) from joint edge.
|
|
View larger version (52K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —Three-dimensional surface renderings from CT scans for
analysis of glenoid parameters in 25-year-old man with normal glenoid from
control group. Cranially, highest point of edge of joint socket defined plane
for separation from coracoid process.
|
|
The study group for glenoid rim defect creation was composed of 10 paired
shoulders in five cadavers (two men and three women), all of whom had given
written consent donating their bodies to the Institute of Anatomy after death,
in line with local legislative requirements. Four had died from cardiac causes
and one from pneumonia. The mean age at death was 85 years (age range,
77–90 years). Exclusion criteria for the cadaveric studies were the same
as described previously.
CT Protocols and Image Analysis
All patients of the control group underwent MDCT on the same scanner
(LightSpeed Pro 16, GE Healthcare). They were examined immediately after
admission to the emergency department using the standard helical CT protocol
that is part of the multiple-trauma management at our hospital. The patients
were imaged in the supine position with their arms at their sides and with the
shoulder in neutral rotation. The following scanning parameters were applied:
acquisitions, 16 x 1.25 mm; peak kilovoltage, 120 kVp; tube
current–time product, 140 mAs; and field of view, 358 mm. The initial
examination, performed to exclude fractures or other sequelae of trauma,
extended from the base of the skull to the symphysis.
In the study group, both shoulders were examined with an MDCT scanner
(Aquilion 64, Toshiba Medical Systems) within 48 hours of death using the
following protocol: acquisitions, 64 x 0.5 mm; peak kilovoltage, 120
kVp; tube current–time product, 90 mAs; and field of view, 138 mm. The
scan ranged from the upper margin of the acromioclavicular joint to the
inferior scapular angle.
The CT data sets were analyzed on a standard desktop PC using the software
Amira, version 4 (Visage Imaging, Mercury Computer Systems and Zuse-Institute
Berlin, Department of Visualization and Data Analysis, Medical Planning
Group). After reading in the data sets, surface-shaded MPR images were
generated. For analysis in both groups, all other bones were removed to
generate images that contained only the 3D reconstruction of the glenoid for
analysis. All measurements in the study group, including determination of
accuracy, were done by a single investigator. A second investigator performed
the evaluation of the control group using the procedure described in the next
section.
Morphologic Side-to-Side Comparison
Four measuring parameters were determined using the MPR images in both the
control group and the study group before fracture creation to assess the
side-to-side comparability in healthy subjects (Fig.
1A,
1B,
1C,
1D): the maximum width
(anteroposterior length) and height (superoinferior length) of the bony joint
surface area obtained in the en face view, the surface area of the
glenoid in the en face view, and the volume of the reconstructed
glenoid body. The borders of the glenoid surface area were defined manually
using the outer bone margin for orient ation. For volume determination, the
vertical plane separating the glenoid medially from the scapula was defined
parallel to the glenoid surface at a distance of 10 mm. Superiorly, the
coracoid process per pendicular to the surface area was eliminated using the
uppermost point of the joint area as an anatomic landmark for horizontal
separation.
Quantification of Glenoid Rim Defects
After initial CT with MPR of the intact shoulders and evaluation of the
measuring parameters according to the described protocol, the arms, including
the shoulders, were separated from the trunk. The specimens were stored at
–11°F. An anteroinferior glenoid rim defect was created
arthroscopically in all 10 shoulders via a three-portal technique. The aim was
to resect one fourth to one third of the glenoid at a 90° angle to the
glenoid surface area in the region between the 2- and 6-o'clock positions
based on optical assessment. After the procedure, the fractured glenoids were
again examined by CT with MPR. In the resulting reconstruction images,
residues of the surgically severed fragments were identified and removed
manually. The sizes of the glenoid rim defects were measured by again
determining the four measured parameters. Results were compared with the
parameters obtained on the contralateral side before defect creation.
View larger version (43K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —Virtual glenoid rim reconstruction from CT scans in study
group after fracture creation. Image of right-sided glenoid (A) shows
defect in anterior edge. Contralateral glenoid (B) is intact. First,
intact glenoid is mirrored vertically. Next, joint areas of two glenoid bodies
are aligned and projected into each other (C). In this way former
glenoid rim defect is filled with intact contralateral side (C and
D). To determine precision of reconstruction procedure (E),
reconstructed glenoid (blue and red) was superimposed on
prefracture image (yellow).
|
|
View larger version (46K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B —Virtual glenoid rim reconstruction from CT scans in study
group after fracture creation. Image of right-sided glenoid (A) shows
defect in anterior edge. Contralateral glenoid (B) is intact. First,
intact glenoid is mirrored vertically. Next, joint areas of two glenoid bodies
are aligned and projected into each other (C). In this way former
glenoid rim defect is filled with intact contralateral side (C and
D). To determine precision of reconstruction procedure (E),
reconstructed glenoid (blue and red) was superimposed on
prefracture image (yellow).
|
|
View larger version (45K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C —Virtual glenoid rim reconstruction from CT scans in study
group after fracture creation. Image of right-sided glenoid (A) shows
defect in anterior edge. Contralateral glenoid (B) is intact. First,
intact glenoid is mirrored vertically. Next, joint areas of two glenoid bodies
are aligned and projected into each other (C). In this way former
glenoid rim defect is filled with intact contralateral side (C and
D). To determine precision of reconstruction procedure (E),
reconstructed glenoid (blue and red) was superimposed on
prefracture image (yellow).
|
|
View larger version (52K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D —Virtual glenoid rim reconstruction from CT scans in study
group after fracture creation. Image of right-sided glenoid (A) shows
defect in anterior edge. Contralateral glenoid (B) is intact. First,
intact glenoid is mirrored vertically. Next, joint areas of two glenoid bodies
are aligned and projected into each other (C). In this way former
glenoid rim defect is filled with intact contralateral side (C and
D). To determine precision of reconstruction procedure (E),
reconstructed glenoid (blue and red) was superimposed on
prefracture image (yellow).
|
|
View larger version (54K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2E —Virtual glenoid rim reconstruction from CT scans in study
group after fracture creation. Image of right-sided glenoid (A) shows
defect in anterior edge. Contralateral glenoid (B) is intact. First,
intact glenoid is mirrored vertically. Next, joint areas of two glenoid bodies
are aligned and projected into each other (C). In this way former
glenoid rim defect is filled with intact contralateral side (C and
D). To determine precision of reconstruction procedure (E),
reconstructed glenoid (blue and red) was superimposed on
prefracture image (yellow).
|
|
Virtual Glenoid Rim Repair
Computer-based reconstruction of the glenoid defect was performed by
matching the generated surface of the fractured glenoid with that of the
intact and mirrored contralateral glenoid using Amira (Fig.
2A,
2B,
2C,
2D,
2E). For this registration
process, both glenoids were positioned with their surface areas aligned. After
careful adjustment, the defect was "filled" by the intact side,
resulting in a virtually generated bone fragment accurately fitting the
defect. The maximum length, width, and height of the virtual graft were
measured and its volume calculated (Fig.
3A,
3B,
3C,
3D). In a further step, we
tested the hypothesis that glenoid morphology after computer-simulated glenoid
rim recon struction did not differ significantly in dimension from that of the
same intact glenoid before creation of the defect. This was done by
determining the described measured parameters in the virtually reconstructed
glenoids and comparing the results with those obtained in the predefect CT
scans of the same side. The precision of virtual reconstruction of bone
fragments was tested in one individual by performing a series of five virtual
reconstructions of the joint surface area with measurement of the virtually
generated bone grafts.
View larger version (34K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Virtual created model of bony defect (missing bone fragment)
derived from CT scans in study group. After superimposition of fractured with
intact contralateral glenoid, resulting fragment can be separated and
measured. Lateral view (en face) (A), posterior view
(B), anterior view (C), and medial view (D) are
shown.
|
|
View larger version (33K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Virtual created model of bony defect (missing bone fragment)
derived from CT scans in study group. After superimposition of fractured with
intact contralateral glenoid, resulting fragment can be separated and
measured. Lateral view (en face) (A), posterior view
(B), anterior view (C), and medial view (D) are
shown.
|
|
View larger version (37K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —Virtual created model of bony defect (missing bone fragment)
derived from CT scans in study group. After superimposition of fractured with
intact contralateral glenoid, resulting fragment can be separated and
measured. Lateral view (en face) (A), posterior view
(B), anterior view (C), and medial view (D) are
shown.
|
|
View larger version (30K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —Virtual created model of bony defect (missing bone fragment)
derived from CT scans in study group. After superimposition of fractured with
intact contralateral glenoid, resulting fragment can be separated and
measured. Lateral view (en face) (A), posterior view
(B), anterior view (C), and medial view (D) are
shown.
|
|
Statistical Analysis
The means and SDs of measures in unfractured (control group and predefect
study group), fractured, and reconstructed glenoids and re con structed
fragments (only study group) were calculated. All measured values were
normally distributed (Kolmogorov-Smirnov test). A paired Student's t
test was used to test among subgroups in the control group and in the study
group. Because of the smaller sample size, only a Wilcoxon's test was
performed in predefect CT scans of the study group. Values for p <
0.05 were considered statistically significant in all tests. To correlate
bilateral glenoid parameters and virtually reconstructed glenoids versus
ipsilateral predefect CT scans, a linear regression analysis
(R2) was used. The precision of virtual surface area
reconstruction was assessed by calculating coefficients of variance for the
repeat measure ments in the subject in whom the reconstruction was done five
times. The statistical analysis was done with SPSS, version 15.0 (SPSS, Inc.)
and JMP, version 5.1 (SAS Institute).
Results
Side-to-Side Comparability in Control Group and Predefect Study Group
There were no significant glenoid side-to-side differences in the study
subjects or the control group in the parameters measured: height, width,
surface area, and volume. The mean values, SD, and side-to-side differences
are shown in Table 1. In the
control group, the highest percentage side-to-side difference of 13.1% (range,
4.3–20.4%) was found for glenoid volume. Regarding the other categories,
the differences of 4.8% (0.3–15.2%) for height, 4.3% (0.4–32.8%)
for width, and 3.6% (0.1–7.0%) for surface area were all of comparable
values. Similar results were measured in the older study group before the
fracture was created. Again, glenoid volume showed the highest side-to-side
variation of 7.1% (3.9–15.3%). Values of 2.0% (0.9–4.0%)
difference for height, 4.6% (0.3–13.9%) for width, and 6.8%
(0.7–14.8%) for surface area were not substantially in disagreement
compared with the younger control group.
View this table:
[in this window]
[in a new window]
|
TABLE 1: Glenoid Parameters and Bilateral Comparison in Control Group and Study
Group Before Glenoid Rim Defect Creation
|
|
Quantification of Glenoid Rim Defects
The fractured glenoids differed significantly from the intact contralateral
glenoids regarding width, surface area, and volume
(Table 2). Among the tested
parameters, differences in width of 30.5% (21.0–47.4%) and surface area
of 33.8% (25.4–47.3%) performed better in detecting differences in size.
Even though volume displayed significant differences between the two sides,
averaged percentage values were lower, with a higher range compared with width
and surface area at 18.7% (3.8–38.1%). The only parameter that did not
differ significantly was glenoid height, which was only 3.3% (0–18.1%)
smaller on the side of the defect.
Virtual Glenoidplasty
Comparison with predefect CT scans— Data for glenoid
parameters after virtual reconstruction by matching with the contralateral
side are given in Table 3. The
comparison with the CT scans of the same glenoid before fracture revealed no
significant differences. The highest percentage of difference was 6.4%
(1.9–14.8%) for surface area, showing very similar values compared with
the side-to-side difference in normal glenoids of the study group before
fracture. Also, the difference in width of 3.4% (0.8–9.1%) was in
accordance with the differences found in normal control subjects and normal
glenoids in the study group. Regarding the high side-to-side variation of
volume in normal glenoids, volume after reconstruction showed only a small
pre- to-postdefect difference of 3.7% (0.1–7.9%). In addition to the
previous evaluations, height displayed only a small difference of 2.2%
(0.3–6.5%). The coefficients of variation determined to assess
reproducibility of virtual reconstruction in one case were 2.9% for width,
1.8% for height, 3.3% for surface area, and 4.0% for volume.
Dimensions of reconstructed fragments— The virtually
reconstructed glenoid bony defects of the fractured glenoids were evaluated in
terms of length, width, height, and volume. Mean values and SDs were 12
± 1.8 mm for the largest width, 28 ± 2.4 mm for the largest
height, 14 ± 2.6 mm for the largest thickness, and 1,720 ± 524
mm3 for volume. The repeat measurements in one of the cases yielded
coefficients of variation of 1.7% for width, 2.5% for height, 0.95% for
thickness, and 1.3% for volume.
Discussion
It is generally agreed that a large bony defect of the glenoid should be
treated with bone grafting. The aim of surgically reconstructing the original
glenoid shape is to restore stability of the whole shoulder joint
[8,
17,
18]. Optimal planning before
therapy for glenoid defect repair includes not only correct quantification of
the defect but also a reconstruction of the fragment required to fill the
defect. To the best of our knowledge, this is the first study to present a
technique for computer-based reconstruction of glenoid rim defects. Our
results show that fusion of the fractured glenoid with the intact
contralateral glenoid allows reconstruction of the original morphology in a
virtual model. Virtual imaging-based reconstruction can thus help the surgeon
shape a bone graft that more anatomically fits the defect. Moreover, size
comparison with the intact contralateral glenoid enables quantitative
determination of the amount of bone loss.
Several recent studies deal with the quantification of glenoid rim defects
[2,
3,
16,
19–22].
Some of these studies propose methods including only the side of the fractured
glenoid
[19–22].
Gerber and Nyffeler [20]
determined the size of the anteroinferior glenoid rim in relation to the
anteroposterior diameter; this approach was also used by Scheibel et al.
[7] and Warner et al.
[8] and in two clinical
studies. Adequate quantification of the bone loss using only CT scans of the
affected shoulder is difficult because it is not known how much bone was
present before trauma. The study groups of Sugaya et al.
[3] and Griffith et al.
[2,
16] proposed morphologic
comparison of the fractured glenoid with the contralateral intact glenoid.
Assuming that the inferior glenoid portion is nearly circular in the en
face view, Sugaya et al. related this area to the area of the fractured
joint fragment. The authors measured the fragment size directly, which was not
possible in all patients. In cases in which no bone fragment was present, they
made a merely optical comparison with the intact contralateral glenoid but did
not perform measurements. The circle method was then validated by Huijsmans et
al. [23]. Griffith et al.
[2,
16] developed an adequate
method for quantification of bony Bankart lesions using CT of both shoulders
with direct comparison of a variety of measuring parameters. They reported
decreased maximum glenoid width and decreased maximum width-to-length ratio to
be the most useful parameters of bone loss
[16]. In a further study by
the same group, CT measuring parameters were found to have high sensitivity in
quantifying glenoid defects compared with arthroscopy. Here again, the
percentage of difference in maximum glenoid width was used for quantifying the
glenoid rim bony defect
[2].
An approach for imaging bone loss similar to the one used in this study was
used in a recent article by Scalise et al.
[24] focusing on the glenoid
vault in 12 patients with glenohumeral osteoarthritis. Two methods were
evaluated: First, a stereolithography model of a standardized vault shape was
virtually implanted into each glenoid; and second, direct measurements of the
paired glenoid vaults were performed. In contrast to our study the whole
scapula was superimposed to the contralateral side. No control group was
evaluated for side-to-side comparability of the glenoid vault in healthy
patients, which may be useful to quantify the results obtained in abnormal
glenoids. Regarding the bilateral comparison, the authors refer to an anatomic
study performed by von Schroeder et al.
[25], in which no differences
between right and left scapulas were found.
Our results regarding the side-to-side comparison of normal glenoids
correlate well with the results reported by Griffith et al.
[16] in 10 healthy subjects.
Maximum glenoid width and surface area—the two parameters that we
consider highly relevant—showed only small side-to-side variation of
3.7% and 4.1%, respectively. There was likewise no substantial side-to-side
variations in the healthy volunteers examined by Sugaya et al.
[3], although the authors found
slight differences in glenoid width and height but do not provide actual
measurements of the side-to-side comparison. We believe that our data and the
previously discussed reports in the literature confirm that the comparison of
a fractured glenoid with the normal shoulder of the other side is a valid
approach for virtual repair of rim defects in patients with a unilateral
shoulder fracture because side-to-side variation is small.
The largest width, surface area, and volume were found to be well-suited
parameters for quantifying glenoid bone loss. Because the defects were created
in the region between the 3- and 6-o'clock positions, they did not affect the
largest height, which was not an unexpected result. Griffith et al.
[16] analyzed 34 patients with
unilateral dislocations and found the decrease in maximum width to be the best
indicator of glenoid bone loss. They introduced glenoid width-to-length ratio
as a new parameter. No other study group has analyzed glenoid volume. Griffith
et al. consider the cross-sectional area to be a less-useful parameter,
although this parameter differed significantly between normal and dislocated
shoulders. However, the authors investigated small defects in which the
fractured glenoids were on average 7.3% smaller than the intact contralateral
glenoids. In our study, larger defects were created (20–30% of the joint
surface), which may have had a positive impact on the precision of the
different measurement procedures. The focus of our study was on preoperatively
modeling the bone fragment to fill the defect, which is why the fractures had
to be large enough to represent surgical indications in the clinical setting.
Also, the arthroscopic procedure of defect creation led to discontinuous bony
margins. In reality, glenoid bone loss is normally characterized by a
relatively smooth anterior straight edge or smooth anterior concavity to the
glenoid. This has the potential to simplify matters of reconstruction in
practice relative to reconstructions of irregular defects as undertaken in
this study.
As with any other quantitative approach, it is likely that measuring is
more difficult and inaccurate when the method is applied to small glenoid rim
defects. The results show that there is slight side-to-side variation in
glenoid size in healthy individuals. In the older subjects of the study group,
the difference in surface area was an even 7%, although the difference was not
significant. These data suggest that small defects of about 5% of the maximum
glenoid width may be difficult to distinguish from normal side-to-side
variation. Griffith et al.
[16] therefore emphasize that
the diagnosis of a glenoid rim defect requires not only a reduced width
compared with the normal contralateral shoulder but also the presence of a
straight anterior line in the en face view of the reconstructed
images.
As with published techniques for glenoid defect quantification, the
approach presented here is limited in that image analysis is
examiner-dependent because the method is not yet fully automated. Another
limiting factor is that image analysis is time-consuming. Virtual glenoid rim
reconstruction takes about 10–15 minutes, which is acceptable in the
routine clinical setting in selected cases only. Accuracy and time
requirements can be optimized by installing the required algorithms on
conventional viewing workstations. Of note, comparison with the contralateral
normal side requires that CT of both shoulders be performed in all cases.
Moreover, as already noted by Griffith et al.
[16], the technique cannot be
used in patients with bilateral shoulder instability. Finally, radiation
exposure is a relevant problem, particularly because most patients are young.
However, because the bones of the shoulder girdle are high-contrast objects,
CT might be performed with an acceptable image quality at a markedly reduced
radiation dose.
In this study, we presented a new method for computer-based reconstruction
of glenoid rim defects by matching with the contralateral normal side. With
this new approach, a precise 3D model of the lost bone fragment can be
generated, which provides important information for the preoperative planning
of surgical bone augmentation. The model represents the bone graft that needs
to be modeled to optimally fill the defect in glenoid rim augmentation.
Practical use during surgery can be using either the images or a printed 3D
model as a guide for shaping of the iliac bone graft. Future studies may
investigate how closely the surgical repair corresponds to the
computer-generated model.
References
- Edwards TB, Boulahia A, Walch G. Radiographic analysis of bone
defects in chronic anterior shoulder instability.
Arthroscopy 2003;19
: 732–739[Medline]
- Griffith JF, Yung PS, Antonio GE, Tsang PH, Ahuja AT, Chan KM. CT
compared with arthroscopy in quantifying glenoid bone loss.
AJR 2007; 189:1490
–1493[Abstract/Free Full Text]
- Sugaya H, Moriishi J, Dohi M, Kon Y, Tsuchiya A. Glenoid rim
morphology in recurrent anterior glenohumeral instability. J Bone
Joint Surg Am 2003; 85:878
–884[Abstract/Free Full Text]
- Malicky DM, Soslowsky LJ, Blasier RB, Shyr Y. Anterior glenohumeral
stabilization factors: progressive effects in a biomechanical model.
J Orthop Res 1996;14
: 282–288[CrossRef][Medline]
- Boileau P, Villalba M, Hery JY, Balg F, Ahrens P, Neyton L. Risk
factors for recurrence of shoulder instability after arthroscopic Bankart
repair. J Bone Joint Surg Am 2006;88
:1755
–1763[Abstract/Free Full Text]
- Mologne TS, Provencher MT, Menzel KA, Vachon TA, Dewing CB.
Arthroscopic stabilization in patients with an inverted pear glenoid: results
in patients with bone loss of the anterior glenoid. Am J Sports
Med 2007; 35:1276
–1283[Abstract/Free Full Text]
- Scheibel M, Nikulka C, Dick A, Schroeder RJ, Gerber Popp A, Haas
NP. Autogenous bone grafting for chronic anteroinferior glenoid defects via a
complete subscapularis tenotomy approach. Arch Orthop Trauma
Surg [Epub ahead of print] 2008 Jan15
- Warner JJ, Gill TJ, O'Hollerhan JD, Pathare N, Millett PJ. Anatomic
glenoid reconstruction for recurrent anterior glenohumeral instability with
glenoid deficiency using an autogenous tricortical iliac crest bone graft.
Am J Sports Med 2006;34
: 205–212[Abstract/Free Full Text]
- Scheibel M, Kraus N, Diederichs G, Haas NP. Arthroscopic
reconstruction of chronic anteroinferior glenoid defect using an autologous
tricortical iliac crest bone grafting technique. Arch Orthop Trauma
Surg [Epub ahead of print] 2007 Nov22
- Mochizuki Y, Hachisuka H, Kashiwagi K, Oomae H, Yokoya S, Ochi M.
Arthroscopic autologous bone graft with arthroscopic Bankart repair for a
large bony defect lesion caused by recurrent shoulder dislocation.
Arthroscopy 2007;23
:677.e1
–677.e4
- De Wilde LF, Berghs BM, Audenaert E, Sys G, Van Maele GO, Barbaix
E. About the variability of the shape of the glenoid cavity. Surg
Radiol Anat 2004; 26:54
–59[CrossRef][Medline]
- Probyn LJ, White LM, Salonen DC, Tomlinson G, Boynton EL. Recurrent
symptoms after shoulder instability repair: direct MR arthrographic
assessment—correlation with second-look surgical evaluation.
Radiology 2007;245
: 814–823[Abstract/Free Full Text]
- Shankman S, Bencardino J, Beltran J. Glenohumeral instability:
evaluation using MR arthrography of the shoulder. Skeletal
Radiol 1999; 28:365
–382[CrossRef][Medline]
- Waldt S, Burkart A, Imhoff AB, Bruegel M, Rummeny EJ, Woertler K.
Anterior shoulder instability: accuracy of MR arthrography in the
classification of anteroinferior labroligamentous injuries.
Radiology 2005;237
: 578–583[Abstract/Free Full Text]
- Stevens KJ, Preston BJ, Wallace WA, Kerslake RW. CT and 3D
reconstructions of shoulders with anterior glenohumeral instability.
Clin Anat 1999;12
: 326–336[CrossRef][Medline]
- Griffith JF, Antonio GE, Tong CW, Ming CK. Anterior shoulder
dislocation: quantification of glenoid bone loss with CT.
AJR 2003; 180:1423
–1430[Abstract/Free Full Text]
- Chen AL, Hunt SA, Hawkins RJ, Zuckerman JD. Management of bone loss
associated with recurrent anterior glenohumeral instability. Am J
Sports Med 2005; 33:912
–925[Abstract/Free Full Text]
- Churchill RS, Brems JJ, Kotschi H. Glenoid size, inclination, and
version: an anatomic study. J Shoulder Elbow Surg2001; 10:327
–332[CrossRef][Medline]
- Itoi E, Lee SB, Berglund LJ, Berge LL, An KN. The effect of a
glenoid defect on anteroinferior stability of the shoulder after Bankart
repair: a cadaveric study. J Bone Joint Surg Am2000; 82:35
–46[Abstract/Free Full Text]
- Gerber C, Nyffeler RW. Classification of glenohumeral joint
instability. Clin Orthop Relat Res 2002;400
: 65–76[CrossRef][Medline]
- Bigliani LU, Newton PM, Steinmann SP, Connor PM, McLlveen SJ.
Glenoid rim lesions associated with recurrent anterior dislocation of the
shoulder. Am J Sports Med 1998;26
: 41–45[Abstract/Free Full Text]
- Saito H, Itoi E, Sugaya H, Minagawa H, Yamamoto N, Tuoheti Y.
Location of the glenoid defect in shoulders with recurrent anterior
dislocation. Am J Sports Med 2005;33
: 889–893[Abstract/Free Full Text]
- Huijsmans PE, Haen PS, Kidd M, Dhert WJ, van der Hulst VP, Willems
WJ. Quantification of a glenoid defect with 3D CT and MRI: a cadaveric study.
J Shoulder Elbow Surg 2007;16
: 803–809[CrossRef][Medline]
- Scalise JJ, Bryan J, Polster J, Brems JJ, Iannotti JP. Quantitative
analysis of glenoid bone loss in osteoarthritis using 3D CT scans.
J Shoulder Elbow Surg 2008;17
: 328–335[CrossRef][Medline]
- von Schroeder HP, Kuiper SD, Botte MJ. Osseous anatomy of the
scapula. Clin Orthop Relat Res 2001;383
: 131–139[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
