AJR 2005; 184:180-184
© American Roentgen Ray Society
Sonography of Full-Thickness Supraspinatus Tears: Comparison of Patient Positioning Technique with Surgical Correlation
Melanie Ferri1,
Karen Finlay2,
Terry Popowich3,
Gary Stamp4,
Peter Schuringa5 and
Lawrence Friedman2
1 Department of Diagnostic Imaging, McMaster University and Hamilton Health
Sciences, 1003-81 Charlton Ave. E, Hamilton, ON, Canada, L8N 1Y7.
2 Department of Diagnostic Imaging, McMaster University and Hamilton Health
Sciences, Henderson General Hospital, 711 Concession St., Hamilton, ON, Canada
L8V 1C3.
3 Department of Diagnostic Imaging, St. Joseph's Hospital, 50 Charlton Ave. E,
Hamilton, ON, Canada L8N 4A6.
4 Department of Orthopaedic Surgery, Guelph General Hospital, 115 Delhi St.,
Guelph, ON, Canada N1E 4J4.
5 Department of Orthopaedic Surgery, St. Mary's Hospital, 911 Queen's Blvd.,
Kitchener, ON, Canada N2M 1B2.
Received October 23, 2003;
accepted after revision June 14, 2004.
Address correspondence to M. Ferri.
Abstract
OBJECTIVE. Sonography has become a popular technique for the
assessment of musculoskeletal disorders. Patient positioning is crucial to a
thorough and accurate assessment of rotator cuff tendons. Two positions, the
Crass and modified Crass, have been routinely used in the research and
clinical settings to examine the supraspinatus tendon. Our study was a
prospective trial to determine whether the Crass or the modified Crass
position affords the most accurate measure of supraspinatus tears when
compared with surgical findings.
SUBJECTS AND METHODS. Twenty-one patients with full-thickness
supraspinatus tears underwent shoulder sonography in both the Crass and the
modified Crass positions. Measurements of supraspinatus tears were performed
in the sagittal and transverse dimensions. Patients subsequently underwent
either arthroscopic or open supraspinatus repair. Intraoperative measurements
were made in two dimensions and were compared with sonographic findings.
RESULTS. Sonography had 100% specificity in detecting full-thickness
supraspinatus tears. No statistically significant difference was seen between
the size of supraspinatus tears in the Crass and modified Crass positions and
surgical findings in the transverse plane (p = 0.55 and 0.61,
respectively). In the sagittal dimension, no statistically significant
difference was seen between surgical findings and the Crass position
(p = 0.14); however, a difference existed when the modified Crass
position was used (p = 0.03).
CONCLUSION. Sonography reliably detects and quantifies supraspinatus
tears. Both the Crass and the modified Crass positions reflected the true size
of supraspinatus tears in the transverse plane. In the sagittal plane, the
Crass position is the more useful to quantify supraspinatus tears because the
modified Crass position overestimates the size of such tears.
Introduction
Sonography of the shoulder joint has become a popular technique for
assessment of musculoskeletal pathology. The role of sonography in the
diagnosis of rotator cuff disease is well documented. A sensitivity of 90-95%,
a specificity of close to 90%, and accuracy in the assessment of both partial-
and full-thickness tears has been reported
[1-4].
These figures compare favorably with those for MRI, with recent MRI studies
reporting a sensitivity of 84%, specificity of 97%, and accuracy of 93% for
combined partial- and full-thickness tears
[5,
6]. Ease of comparison with the
asymptomatic limb, dynamic real-time assessment, patient comfort, easy
accessibility, and low cost are among the advantages of sonography. Rotator
cuff sonography has a long learning curve. Inherent to the detection and
characterization of disorders using sonography is the knowledge of relevant
anatomy and pathology, the experience of the sonographer, the technique used,
and patient positioning.
The purpose of our study was to determine prospectively whether the Crass
or the modified Crass position
[7] affords the most accurate
measure of supraspinatus tendon tears when compared with arthroscopic or open
surgical findings, using intraoperative measurements as the gold standard. To
date, a limited number of trials have examined the correlation of sonography
with surgical findings, none of which correlated measurement of rotator cuff
tears in the Crass and modified Crass positions with surgical measures.
Because technique is a major factor in the accuracy of rotator cuff
sonography, determining the shoulder position that most the accurately
reflects the true size of supraspinatus tears is useful.
Subjects and Methods
Sonographic signs of full-thickness supraspinatus tears have been reported
by many authors. Small tears are often seen as focal hypoechoic areas within
the supraspinatus tendon. As the tears enlarge, concave hypoechoic defects in
the substance of the tendon become apparent. Compression with the sonographic
transducer can illicit tenderness, flatten the tendon, and cause the deltoid
muscle to drop into the space created by the ill-defined retracted edges of
tendon, also known as "deltoid herniation." In the long axis
(sagittal plane), there is a loss of the upward convex (parrot-beak)
appearance of the supraspinatus tendon insertion on the greater tuberosity
[2,
4,
8-11].
Patients
The study population was selected prospectively from patients with
suspected rotator cuff disorders who were referred for shoulder sonographic
assessment. Twenty-one consecutive patients with full-thickness tears of the
supraspinatus tendon diagnosed on sonography were included in our analysis.
None of these patients had prior rotator cuff surgery. The study population
included seven women and 14 men (age range, 41-83 years). Each patient had
unilateral disease; nine had left-sided disease, and 12 had right. Each
patient underwent complete bilateral sonographic examination using a 700 MR
machine (GE Healthcare) and a 7.5-13-MHz multifrequency linear array probe.
Supraspinatus tears were measured in both the Crass and the modified Crass
positions in the sagittal and transverse planes with respect to the
supraspinatus fibers. The sonographic examinations were performed by one
experienced sonographer and were interpreted by one experienced
musculoskeletal radiologist.
After the sonographic examination, each patient underwent rotator cuff
repair, either arthroscopically or by open technique, by one of two orthopedic
surgeons. Associated procedures, such as acromioplasty and excision of the
distal clavicle, were performed as deemed necessary at the time of surgery.
The average delay between sonographic assessment and surgery was 3 months
(range, 1 month-1 year). Thirteen patients underwent repair within 3 months
from the time of sonography, six tears were repaired between 3 and 5 months
from the time of the sonography, and only one patient each had a delay of 7
months and 1 year. Eleven patients underwent arthroscopic repair, eight of
which repairs were converted to open repairs at the time of surgery. Two
patients proceeded directly to open repair.
Either a sterile ruler (open repair) or the tip of an arthroscopic
instrument of known length (arthroscopic repair) was used to measure the
supraspinatus tears intraoperatively, both in the sagittal and the transverse
planes, relative to the orientation of the tendon fibers (Figs.
1A,
1B,
2A, and
2B). The measurements of the
supraspinatus tears in the Crass and the modified Crass positions in both
planes (Figs. 3A,
3B,
4A, and
4B) were compared with each
other and then with the operative findings using paired t tests.
Technique
All examinations are best performed using a multifrequency 7.5-13-MHz
linear array transducer. It is vital that both the depth and the near field be
continually changed. For example, the anterior shoulder is best assessed using
a near focus and high frequency
[7,
12,
13].
Positioning is a key component of the sonographic examination. In the
neutral position, most rotator cuff is obscured by the overlying acromion.
Crass et al. [7], in 1987,
advocated that the shoulder be positioned in extension, adduction, and
internal rotation to improve visualization of the supraspinatus tendon, thus
augmenting diagnostic accuracy (Figs.
5A,
5B,
5C,
6A,
6B, and
6C). This position has since
been referred to as the Crass position. The patient is seated and the shoulder
is extended, adducted, and internally rotated with the elbow flexed, the palm
facing out, and the fingers pointing toward the contralateral scapula (Figs.
7A and
7B). Internal rotation allows
the supraspinatus to become an anterior structure, and extension draws the
supraspinatus anteriorly from beneath the acromion, allowing the maximal
length of tendon to be visualized. Dynamic scanning is not possible in this
position. Crass et al. [7]
believed that the need for dynamic scanning is eliminated by the additional
information provided by this position. It was also reported that a small
percentage of patients overly internally rotate the shoulder, obscuring the
most anterior portion of the supraspinatus tendon
[7].
View larger version (116K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A. —Axial proton density-weighted fat-saturated MR images of
shoulder in 28-year-old woman. MR images show shoulder in neutral anatomic
position (A), Crass position (B), and modified Crass position
(C). Note supraspinatus tendon (arrow) moves from lateral
position in neutral to anterior in Crass position and anterolateral in
modified Crass position, which affords better sonographic visualization in
Crass and modified Crass positions.
|
|
View larger version (157K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B. —Axial proton density-weighted fat-saturated MR images of
shoulder in 28-year-old woman. MR images show shoulder in neutral anatomic
position (A), Crass position (B), and modified Crass position
(C). Note supraspinatus tendon (arrow) moves from lateral
position in neutral to anterior in Crass position and anterolateral in
modified Crass position, which affords better sonographic visualization in
Crass and modified Crass positions.
|
|
View larger version (156K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C. —Axial proton density-weighted fat-saturated MR images of
shoulder in 28-year-old woman. MR images show shoulder in neutral anatomic
position (A), Crass position (B), and modified Crass position
(C). Note supraspinatus tendon (arrow) moves from lateral
position in neutral to anterior in Crass position and anterolateral in
modified Crass position, which affords better sonographic visualization in
Crass and modified Crass positions.
|
|
View larger version (165K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A. —Sagittal oblique proton density-weighted fat-saturated MR
images of shoulder in 28-year-old woman. MR images show supraspinatus tendon
(arrows) beneath acromion located in neutral position (A),
anteriorly in Crass position (B), and anterolaterally in modified Crass
position (C).
|
|
View larger version (162K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B. —Sagittal oblique proton density-weighted fat-saturated MR
images of shoulder in 28-year-old woman. MR images show supraspinatus tendon
(arrows) beneath acromion located in neutral position (A),
anteriorly in Crass position (B), and anterolaterally in modified Crass
position (C).
|
|
View larger version (169K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6C. —Sagittal oblique proton density-weighted fat-saturated MR
images of shoulder in 28-year-old woman. MR images show supraspinatus tendon
(arrows) beneath acromion located in neutral position (A),
anteriorly in Crass position (B), and anterolaterally in modified Crass
position (C).
|
|
View larger version (90K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7A. —Crass and modified Crass positions. Photographs show Crass
(A) and modified Crass (B) positions. Note that sonographic
probe is moved from anterior to anterolateral position, corresponding to
altered position of supraspinatus tendon.
|
|
View larger version (86K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7B. —Crass and modified Crass positions. Photographs show Crass
(A) and modified Crass (B) positions. Note that sonographic
probe is moved from anterior to anterolateral position, corresponding to
altered position of supraspinatus tendon.
|
|
The anterior portion of the supraspinatus tendon is defined by the position
of the bicipital groove. If the bicipital groove is not seen in the routine
position, Crass et al. [7]
advocated the use of the modified Crass position, wherein the patient places
the hand on the hip with the elbow pulled back. In the modified Crass
position, which is helpful in examination of the anteromedial aspect of the
tendon, the shoulder is extended, the elbow is flexed, and the palm of hand is
against the back pocket [2]
(Figs. 7A and
7B). In 1992, Middleton
[4] preferred the modified
Crass position because it avoids internal rotation and maintains the view of
the intraarticular portion of the biceps tendon and the most anterior portion
of the supraspinatus (Figs. 5A,
5B,
5C,
6A,
6B, and
6C). The appearance of the
supraspinatus tendon changes from the Crass to the modified Crass position in
that it rotates 90° from an anterior to a lateral position (Figs.
5A,
5B,
5C,
6A,
6B,
6C,
7A, and
7B). Localizing the biceps
tendon facilitates appropriate identification of the supraspinatus and
subscapularis tendons [4].
Both sagittal and transverse views of the supraspinatus are crucial to
detection of supraspinatus tears. "Sagittal" and
"transverse" refer to the orientation of the sonographic
transducer to the long axis of the supraspinatus fibers. The rotator cuff and
scapula are oriented at approximately 45° to the body's sagittal and
coronal planes, and sagittal supraspinatus views are obtained in a plane
approximately half way between sagittal and coronal
[4]. Because of the long and
broad nature of the supraspinatus tendon, it is crucial to identify its entire
course to avoid false-negative results.
Results
First, all full-thickness tears detected on sonography were confirmed in
the operating room, resulting in 100% specificity in the detection of
full-thickness supraspinatus tears for the study population. None of the
full-thickness tears diagnosed on sonography were found to be
partial-thickness tears at surgery. The mean transverse tear size in the Crass
and the modified Crass positions were 1.7 and 1.9 cm, respectively (p
= 0.09). In the sagittal plane, the tears measured 1.7 cm in the Crass and 1.9
cm in the modified Crass positions (p = 0.13). Comparing the
measurements in the Crass position with operative findings, the mean
transverse measurements were 1.7 and 1.8 cm during sonography and surgery,
respectively (p = 0.55). Similar results were found in the sagittal
plane, with a mean length of 1.7 cm in the Crass position and 1.5 cm in the
operating room (p = 0.14).
In the modified Crass position, the mean transverse length was 1.7 cm,
which was not statistically significant when compared with the operative mean
of 1.8 cm (p = 0.61). However, a significant difference was seen
between the modified Crass sagittal measurement (1.9 cm) and the corresponding
operative finding of 1.5 cm (p = 0.03).
Discussion
Hodler et al. [14] in 1988
correlated sonographic and surgical findings in 51 patients using the Crass
position and dynamic imaging. A complete tear was defined as a concave
supraspinatus margin, absence of the outer tendon margin, or a region of
hypoechogenicity spanning the width of the tendon. Prospective sensitivity was
100%, specificity was 75%, and accuracy was 92%. Retrospectively, sonography
correlated well with intraoperative findings for small and moderate lesions;
however, large lesions often were underestimated
[14].
In a large trial by Wiener and Seitz
[3] in 1992, preoperative
sonography in the modified Crass position was performed on 225 patients and
findings were correlated with arthroscopic inspection or open surgery.
Partial-thickness tears were diagnosed when any of the following criteria were
met: a focal hypoechoic zone, multiple small hypoechoic discontinuities, or
large dominant linear echogenic foci within the substance of the cuff.
Full-thickness tears were defined as the presence of a hypoechoic zone
extending through the entire substance of the cuff or loss of substance with
visualized tear margins. Massive tears were those in which the rotator cuff
was not visualized and the deltoid muscle approximated the humeral head. In
206 (92%) of 225 patients, the sonographic findings were confirmed. Sonography
underestimated the extent of cuff injury in 11 patients (5%) and overestimated
it in eight patients (4%)
[3].
Farin et al. [15], in 1996,
examined the use of sonography, double-contrast arthrography, and
contrast-enhanced CT arthrography in assessing the site and size of rotator
cuff tears. Patient positioning was not described. Sonography detected 80% of
partial- and 90% of full-thickness tears with one false-positive finding. The
size of the tear was accurately identified on sonography in 70% of cases;
tears were overestimated in 2% and underestimated in 9% of cases
[15].
We reviewed the literature regarding the sensitivity and specificity of
sonography of the rotator cuff compared with surgical findings. Most authors
advocated the use of sonography to both detect and quantify supraspinatus
tears. Unfortunately, most authors did not provide reproducible descriptions
of patient positioning during sonographic examination, and therefore
conclusions regarding the accuracy of tear measurement as related to shoulder
positioning cannot be made on the basis of the current literature.
Previous trials have examined the use of sonography in detecting and
accurately measuring the size of supraspinatus tears. However, either the
position of the shoulder joint was not specified, or sonography was performed
with the shoulder held in one position: the neutral, the Crass, or the
modified Crass position. Because patient positioning is crucial to appropriate
sonographic investigation of the shoulder, we believed it relevant to assess
the ability of the Crass and modified Crass positions to quantify accurately
the size of supraspinatus tears as compared with operative findings. Because
the size of the supraspinatus tear can affect the operative approach, it is
important to quantify accurately the size of full-thickness supraspinatus
tears.
A limitation of this trial was our small sample size of 21 patients.
Furthermore, an arthroscopic instrument of known length was used to measure
the tears in the small arthroscopic operative field; it is unknown whether
measurement accuracy was compromised. During the open supraspinatus repair,
intraoperative measurements were made along the contour of the greater
tuberosity because the malleable ruler can assume the shape of the tuberosity.
During sonography, measurement of supraspinatus tears was made using straight
lines between calipers, which may have affected our data adversely. Finally,
the sonographer and radiologists involved in our trial are experienced in
musculoskeletal radiology, which likely contributed to the accuracy of the
sonographic and operative findings. Therefore, our findings may not be
reproducible among radiologists who are not experienced in musculoskeletal
sonography.
On the basis of our findings, both the Crass and the modified Crass
positions reflect the true size of full-thickness supraspinatus tears in the
transverse plane when compared with operative findings. Conversely, sagittal
measures in the modified Crass position significantly overestimate the size of
full-thickness supraspinatus tears, yet the Crass position reflects the true
size of tears in the sagittal plane. It is conceivable that the two positions
could create a difference in tension across full-thickness supraspinatus
tears, thus affecting the resultant measurement in the sagittal plane.
External rotation, a component of the modified Crass position, may increase
tension along the length of the tendon fibers, resulting in overestimation of
tear size in the longitudinal plane. In contradistinction to the modified
Crass position, the Crass position more closely mimics the adducted position
of the shoulder at the time of surgery, thus potentially contributing to the
improved accuracy of tear measurement in this position.
In conclusion, we recommend that the radiologist performing and
interpreting shoulder sonography recognize that the sagittal measurement of
full-thickness supraspinatus tears is more accurately performed with the
shoulder in the Crass position.
Acknowledgments
We thank Bruce Weaver of McMaster University.
References
- Mack LA, Masten FA, Kilcoyne RF, Davies PK, Sickler ME. US
evaluation of the rotator cuff. Radiology1985; 157:205
-209[Abstract/Free Full Text]
- Crass JR, Craig EV, Feinberg SB. Ultrasonography of rotator cuff
tears: a review of 500 diagnostic studies. J Clin
Ultrasound 1988;16:313
-327[Medline]
- Wiener SN, Seitz WH. Sonography of the shoulder in patients with
tears of the rotator cuff: accuracy and value for selecting surgical options.
AJR 1993;160:103
-107[Abstract/Free Full Text]
- Middleton WD. Ultrasonography of the shoulder. Radiol
Clin North Am 1992;30:927
-940[Medline]
- Quinn SF, Sheley RC, Demlow TA, Szumowski J. Rotator cuff tendon
tears: evaluation with fat-supressed MR imaging with arthroscopic correlation
in 100 patients. Radiology1989; 172:223
-229[Abstract/Free Full Text]
- Steinbach G. MRI of the rotator cuff. Semin
Roentgenol 2000;35:217
-230[Medline]
- Crass JR, Craig EV, Feinberg SB. The hyperextended internal
rotation view in rotator cuff ultrasonography. J Clin
Ultrasound 1987;15:416
-420[Medline]
- Bouffard JA, Lee SM, Dhanju J. Ultrasonography of the shoulder.
Semin Ultrasound CT MR 2000;21
: 164-191[Medline]
- Friedman L, Finlay K, Popowich T, Jurriaans E. Ultrasonography of
the shoulder: pitfalls and variants. J Can Assoc
Radiol 2002;53:22
-31
- Ahovuo J, Paavolainen P, Bjorkenheim JM. Ultrasonography in lesions
of the rotator cuff and biceps tendon. Acta Radiol1989; 30:253
-255[Medline]
- Paavolainen P, Ahovuo J. Ultrasonography and arthrography in the
diagnosis of tears of the rotator cuff. J Bone Joint Surg
Am 1994;76A:335
-340
- Seibold CJ, Mallisee TA, Erickson SJ, Boynton MD, Raasch WG, Timins
ME. Rotator cuff: evaulation with US and MR imaging.
RadioGraphics1999; 19:685
-705[Abstract/Free Full Text]
- Thain LMF, Adler RS. Sonography of the rotator cuff and biceps
tendon: technique, normal anatomy, and pathology. J Clin
Ultrasound 1999;27:446
-458[Medline]
- Hodler J, Fretz CJ, Terrier F, Gerber C. Rotator cuff tears:
correlation of sonographic and surgical findings.
Radiology1988; 169:791
-794[Abstract/Free Full Text]
- Farin PU, Kaukanen E, Jaroma H, Vaatainen U, Miettinen H,
Soimakallio S. Site and size of rotator-cuff tears: findings at ultrasound,
double-contrast arthrography and computed tomography with surgical
correlation. Invest Radiol1996; 31:387
-394[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
M. J. C. M. Rutten, G. J. Jager, and J. G. Blickman
From the RSNA Refresher Courses: US of the Rotator Cuff: Pitfalls, Limitations, and Artifacts.
RadioGraphics,
March 1, 2006;
26(2):
589 - 604.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
