| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Diagnostic Imaging, Brown University School of Medicine, Rhode
Island Hospital, 593 Eddy St., Providence, RI 02903.
2
Department of Orthopaedic Surgery, Brown University School of Medicine, Rhode
Island Hospital, Providence, RI 02903.
Received July 28, 1999;
accepted after revision September 7, 1999.
Address correspondence to G. A. Tung.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. High-field (n = 46) or low-field (n = 21) MR imaging was performed on 41 patients with SLAP tears and 26 patients with normal superior labra. The superior labrum was classified into one of four types on the basis of patterns of intralabral signal intensity. The relative frequency of rotator cuff tears and long head of the biceps tendinopathy was also assessed.
RESULTS. For the diagnosis of SLAP tear, the sensitivity of high-field MR imaging was 90% (95% confidence interval = 74%, 98%), specificity was 63% (35%, 85%), and accuracy was 80% (66%, 91%). The sensitivity of low-field MR imaging was 64% (31%, 89%), specificity was 70% (35%, 93%), and accuracy was 67% (43%, 85%). A branched linear or stellate focus of abnormal intralabral signal intensity was associated with a SLAP tear in 86% of patients. Conversely, two other labral patterns correlated with a normal superior labrum in 71% of patients. Abnormal signal intensity in the biceps tendon was seen in 15% of patients with a SLAP tear. Full-thickness (37%) and partial-thickness (31%) rotator cuff tears were often seen.
CONCLUSION. The performance characteristics of high-field MR imaging are superior to those of low-field MR imaging for the diagnosis of a superior labral tear. Rotator cuff tears can be seen in many patients with superior labral tears, but abnormal signal intensity in the biceps tendon in uncommon.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
At arthroscopy, four main types of SLAP tears and several variants have been described [2, 5, 7]. The relative frequency of the four major types of SLAP tears in 140 patients was reported by Snyder et al. [3]. Marked fraying of the superior labrum without labral detachment or biceps tendon anchor injury is defined as a type I tear (21%). Detachment of the labrum from the glenoid is a type II tear (55%); a type III tear is a vertically displaced bucket-handle SLAP tear (9%). A type IV tear is a bucket-handle tear that extends to a variable extent into the biceps tendon (10%). Any of these four types may occur with more extensive labrocapsular injuries, such as a Bankart lesion or a tear of the middle glenohumeral ligament (5%).
Although arthroscopy is most often used to diagnose SLAP tears, case reports and small case series support the use of nonarthrographic MR imaging and MR arthrography [7,8,9,10,11,12,13]. However, none of these reports present performance characteristics of MR imaging because they do not include a companion cohort of patients with shoulder pain or instability who did not have superior labral injuries at arthroscopy. Other studies from the orthopedic and medical-imaging literature have questioned the clinical efficacy of MR imaging [9, 14].
The purpose of this study was to determine the performance characteristics of high-field and low-field MR imaging for the diagnosis of SLAP tears. We evaluated specific patterns of intralabral signal intensity and the frequency of concomitant injuries to the rotator cuff and long head of the biceps tendons in patients with SLAP tears.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Description of Subjects
Of the 41 patients with SLAP tears at arthroscopy, 27 men and 14 women had
a mean age of 43 years (range, 20-71 years). In this group, the right shoulder
was involved in 26 patients and the left shoulder in 15 patients. Sixteen
males and 10 females in the group of 26 patients had a normal superior labrum
at arthroscopy. The mean age of this group was 45 years (range, 14-70 years;
p = 0.36). The right shoulder was involved in 17 patients, and the
left shoulder was symptomatic in nine patients (p = 0.07).
Shoulder MR Imaging Technique
High-field (1.5-T) MR imaging of the shoulder was performed on 46 patients
with a Magnetom Vision scanner (Siemens, Erlangen, Germany) using a circularly
polarized flexible coil. The routine protocol includes axial and coronal
oblique turbo spin-echo proton density-weighted and coronal oblique and
sagittal oblique turbo spin-echo T2-weighted sequences. The coronal oblique
images were oriented parallel to the longitudinal course of the supraspinatus
tendon, and oblique sagittal images were oriented orthogonal to those of the
coronal oblique plane. The imaging performed had the following parameters:
axial proton density-weighted images (TR/TE, 2000/18; echo train length,
three; field of view, 18 cm; matrix, 252 x 256; acquisition, one; slice
thickness, 3 mm [no gap]; number of slices, 21; scan time, 5.7 min), coronal
oblique proton density-weighted images (2000/18; echo train length, three;
field of view, 16 cm; matrix, 192 x 256; acquisitions, two; slice
thickness, 3 mm [no gap]; number of slices, 21; scan time, 5.4 min), coronal
oblique T2-weighted images with frequency-selective fat saturation (3600/96;
echo train length, seven; field of view, 16 cm; matrix, 252 x 256;
acquisitions, two; slice thickness, 4 mm [no gap]; number of slices, 15; scan
time, 5.7 min), sagittal oblique turbo spin-echo T2-weighted images with
frequency-selective fat saturation (3500/96; echo train length, seven; field
of view, 16 cm; matrix, 252 x 256; acquisitions, one; slice thickness, 4
mm [no gap]; number of slices, 15; scan time, 4.4 min).
Low-field (0.2-T) MR imaging was performed on 21 patients with a Magnetom Open scanner (Siemens) using a multipurpose surface coil. The routine shoulder-imaging protocol includes an axial turbo spin-echo proton density-weighted sequence, both coronal oblique conventional spin-echo T1-weighted and turbo spin-echo sequences, and an oblique sagittal turbo spin-echo T2-weighted sequence. Parameters used were the following: axial proton density-weighted images (2000/24; echo train length, five; field of view, 18 cm; matrix, 250 x 256; acquisitions, two; slice thickness, 5 mm [no gap]; number of slices, 14; scan time, 6.8 min), coronal oblique T1-weighted images (500/26; field of view, 18 cm; matrix, 192 x 256; acquisitions, two; slice thickness, 4 mm [no gap]; number of slices, 11; scan time, 6.5 min), coronal oblique T2-weighted images (3000/102; echo train length, seven; field of view, 18 cm; matrix, 195 x 256; acquisitions, two; slice thickness, 4 mm [no gap]; number of slices, 11; scan time, 5.7 min), and sagittal oblique T2-weighted images (3000/102; echo train length, seven; field of view, 18 cm; matrix, 196 x 256; acquisitions, two; slice thickness, 4 mm [no gap]; number of slices, 11; scan time, 5.7 min).
Evaluation of Superior Glenoid Labrum
All original MR imaging studies were retrospectively reviewed by a
board-certified radiologist with 9 years of clinical experience in
musculoskeletal MR imaging who had no knowledge of either the patient mix or
the arthroscopic diagnoses. MR images of the torn and normal superior labra
were mixed and were evaluated in a random fashion.
Using the nonmagnified coronal oblique turbo spin-echo proton density-weighted (high-field MR imaging) and the nonmagnified conventional spin-echo T1-weighted (low-field MR imaging) images, the reviewer classified the intralabral signal-intensity pattern of the superior labrum into one of four patterns. Homogeneous low signal intensity of the labrum was designated as a type A superior labrum (Fig. 1A,1B). A single linear focus of abnormal increased signal intensity extending from the articular surface of the labrum but neither to nor through the bursal side of the labrum was designated a type B superior labrum (Fig. 2A,2B). The type B pattern was included because it is often seen in the normal superior labrum in the region of the origin of the long head of the biceps tendon [15]. A branched linear focus or a stellate focus of abnormally increased signal intensity involving at least one articular surface of the superior labrum was designated as a type C superior labrum (Figs. 3A,3B,4A,4B,5A,5B). An amorphous globular focus of increased signal intensity within the superior labrum was designated as a type D superior labrum (Fig. 6A,6B).
|
|
|
|
|
|
|
|
|
|
|
|
To determine the sensitivity, specificity, and accuracy of MR imaging, the diagnostic criteria for a normal or normal variant of the superior labrum was labral type either A or B. The diagnostic criteria for a SLAP tear was the presence of labral type either C or D.
Different types of arthroscopic SLAP tears are designated by roman numerals, I-IV, whereas the four types of superior labrum based on MR imaging signal-intensity patterns are indicated by capital letters, A-D. Because the number of patients with arthroscopic SLAP tear types III and IV is small compared with the number of types I and II tears in this study, we did not attempt to distinguish among the different arthroscopic types of SLAP tear on MR imaging.
Assessment of the Rotator Cuff and Biceps Tendon
Morphology and signal intensity of the rotator cuff tendons and biceps
tendon were evaluated on oblique coronal or oblique sagittal T2-weighted
images. The primary diagnostic sign for a full-thickness tear was either a
discrete tendon gap or abnormally high signal intensity within the tendon that
was isointense to fluid and that extended through the entire width of the
tendon [16]. The primary
diagnostic sign for a partial-thickness tear was abnormally high signal
intensity (not necessarily as hyperintense as fluid) that did not involve the
entire thickness of the rotator cuff or biceps tendon
[16,17,18]
(Fig.
7A,7B).
A secondary sign of a full-thickness rotator cuff tear was fluid in the
subacromial or subdeltoid bursae or tendon retraction
[16].
|
|
Statistical Analysis
The sensitivity, specificity, and accuracy of both high-field and low-field
MR imaging for the diagnosis of SLAP tears were determined, with arthroscopy
as the gold standard. Sensitivity was defined as the quotient of true-positive
diagnoses of SLAP tears on MR imaging (numerator) and both true-positive and
false-negative diagnoses (denominator). Specificity was defined as the
quotient of true-negative diagnoses of SLAP tears on MR imaging (numerator)
and both true-negative and false-positive diagnoses (denominator). Accuracy
was calculated as the proportion of all patients who were correctly
classified, or the proportion of patients who were either true-positive or
true-negative. Positive predictive value was defined as the quotient of
true-positive diagnoses of SLAP tears on MR imaging (numerator) and all
positive diagnoses (both true-positive and false-positive diagnoses
[denominator]). Negative predictive value was defined as the quotient of
true-negative diagnoses of SLAP tears (numerator) and all negative diagnoses
(both true- and false-negative diagnoses [denominator]).
Because some of the expected frequencies were small (i.e., fewer than five), a two-sided Fisher's exact test was used to test for a statistically significant difference. The null hypothesis was rejected with a p value of less than 0.05.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Performance Characteristics of High-Field MR Imaging
For the diagnosis of SLAP tear, the sensitivity of high-field MR imaging
was 90% (95% confidence interval [CI] = 74%, 98%), specificity was 63% (95% CI
= 35%, 85%), and accuracy was 80% (95% CI = 66%, 91%).
The distribution of labral patterns in the group of 27 true-positive diagnoses included 20 type C and seven type D cases. Of the true-negative diagnoses, the distribution of labral pattern types was three type A cases and seven type B cases. Three false-negative diagnoses were made. On MR imaging, all three had type B labral patterns. At arthroscopy, two missed tears were type II, and one was a type I SLAP tear. Six false-positive diagnoses included two type C and four type D labral patterns.
With a type A or B labral pattern on high-field MR imaging, the negative predictive value for a SLAP tear was 77% (10/13). With a type C or D labral pattern, the positive predictive value for a SLAP tear was 82% (27/33).
Performance Characteristics of Low-Field MR Imaging
For the diagnosis of a SLAP tear, the sensitivity of low-field MR imaging
was 64% (95% CI = 31%, 89%), specificity was 70% (95% CI = 35%, 93%), and
accuracy was 67% (95% CI = 43%, 85%).
The distribution of labral patterns in the group of seven true-positive diagnoses included type C in four patients and type D in three patients. Of the seven true-negative diagnoses, the distribution of labral pattern types was type A in two patients and type B in five patients. Four false-negative diagnoses were found on low-field MR imaging. A type B labral pattern was present in three patients, and the type A labral pattern was seen in the other false-negative patient. At arthroscopy, three of the missed tears were type I SLAP tears, and one was a type II tear. Three false-positive diagnoses included two type C and one type D labral patterns.
With a type A or B labral pattern on low-field MR imaging, the negative predictive value for a SLAP tear was 64% (7/11). With a type C or D labral pattern, the positive predictive value for a SLAP tear was 70% (7/10).
Diagnosis of SLAP Tear Based on Specific Labral Patterns
A normal superior labrum at arthroscopy was often associated with glenoid
labral patterns A or B on high-field and low-field MR imaging
(Table 2). The ratios of a
true-negative to false-negative diagnosis of SLAP tear for labral patterns A
and B were 5:1 and 2:1, respectively. The negative predictive values for
labral type A and for type B were 83% (5/6) and 67% (12/18), respectively.
With either type A or B, the negative predictive value of MR imaging was 71%
(17/24). Conversely, a tear of the superior labrum was often predicted by
labral patterns C or D on high-field and low-field MR imaging
(Table 3). The ratios of a
true-positive to a false-positive diagnosis of a SLAP tear for labral patterns
C and D were 6:1 and 2:1 respectively. The positive predictive values for
labral type C and type D were 86% (24/28) and 67% (10/15), respectively. Given
labral type either C or D, the positive predictive value of MR imaging was 79%
(34/43) for the diagnosis of a SLAP tear.
|
|
Rotator Cuff and Long Head of the Biceps Tendons
Among the 41 patients with SLAP tears at arthroscopy, 15 had full-thickness
rotator cuff tears, 13 had partial-thickness tears, and 13 had
normal-appearing rotator cuff tendons on MR imaging. In the 26 patients with a
normal superior labrum at arthroscopy, nine had full-thickness cuff tears, six
had partial tears, and 11 had normal rotator cuff tendons on MR imaging
(p = 0.38). All fullthickness tears and articular-side partial
rotator cuff tears were confirmed at arthroscopy.
Abnormally high signal intensity in the biceps tendon was noted in six patients (15%) with arthroscopically proven SLAP tears. Abnormal signal intensity was found in only one (4%) patient who had no evidence of a SLAP lesion at arthroscopy (p = 0.01). On both high-field and low-field MR imaging, abnormal signal intensity in the biceps tendon was most often shown on sagittal oblique images of the shoulder. On these images, it was often difficult to distinguish the origin of the biceps tendon from the anterior margin of the rotator cuff. However, we found that the coracohumeral ligament provided an important landmark for the origin of the long head of the biceps tendon on these images (Fig. 7A,7B).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In other studies of high-field MR imaging, the torn superior labrum has been correlated with several different patterns of abnormal intralabral signal intensity. The more common patterns that have been previously described include linear signal intensity that transgresses both the bursal and articular surfaces of the labrum, complex linear patterns of increased intralabral signal intensity, and a large globular focus of increased signal intensity [7,8,9, 11, 12, 14]. Our study confirms the value of specific labral patterns for predicting the presence or absence of a SLAP tear on MR imaging. The positive predictive value for labral types either C or D is 79%, and in particular, the type C pattern is associated with a SLAP tear in 86% of 28 patients. Conversely, the negative predictive value for labral types either A or B is 71%. False-negative and false-positive diagnoses can occur, particularly with labral types B and D, respectively. The type B labrum was designated as normal because the sublabral recess and the origin of the biceps tendon may appear as a linear focus of increased signal intensity in the anterior and middle part of the superior labrum [4, 11, 13, 14, 20, 21]. However, in one third of patients in this study, the type B pattern was associated with a SLAP tear at arthroscopy. In most of these false-negative findings, the linear focus of abnormal signal intensity was located posterior to the long head of the biceps tendon. Therefore, as has been mentioned by others [15, 21], the type B labral pattern should be considered a SLAP tear if it is seen posterior to the biceps tendon. Conversely, a false-positive diagnosis of SLAP tear was made in one third of patients in whom a globular focus of abnormal signal intensity was seen in the superior labrum (type D pattern). In these false-positive findings, we postulate that abnormal labral signal intensity may be attributed to fibrovascular tissue and mucoid or eosinophilic degeneration of the labrum [21, 22]. Finally, three of the four missed SLAP tears on low-field MR imaging were arthroscopic type I SLAP tears. The type I SLAP tear—marked fraying of the superior labrum without labral detachment—is the most difficult to diagnose on MR imaging and MR arthrography [7, 9, 23]. Some controversy exists as to whether this lesion is truly an injury or a degenerative lesion [3, 4].
Rotator cuff tears and long head of the biceps tendinopathy may accompany a superior labral tear. In our series, rotator cuff tears on MR imaging were seen in 68% of patients with SLAP tears, compared with other studies that have reported cuff tears in as many as 55% of patients with SLAP lesions [2, 3, 5, 6]. Rotator cuff tears are more common in older patients with SLAP lesions, and in these patients, the superior labral lesion may be degenerative, not traumatic [4]. Both chronic overuse tears of the posterosuperior labrum and articular-side partial tears of the supraspinatus and infraspinatus tendons may result from internal impingement in the overhead position [24, 25]. Abnormal signal intensity in the biceps tendon was identified in 15% of patients with SLAP tears; it occurred significantly more often than in those with a normal superior labrum. Contiguous T2-weighted images in the oblique sagittal plane show the origin and proximal part of the long head of the biceps tendon in cross-section. The most medial and proximal part of the long head of the biceps tendon can be identified just inferior to the coracohumeral ligament at the point where this ligament merges with the supraspinatus tendon [26] (Fig. 7A,7B). This part of the long head of the biceps tendon is most often injured in patients with SLAP tears [1,2,3].
Our study had several limitations. First, it was a retrospective study that included a large number of SLAP tear patients in the cohort. Therefore, bias was introduced in favor of diagnosing SLAP tears on MR imaging even though the reviewer was unaware of the final arthroscopic diagnosis. Second, this study was subject to verification or workup bias: when a study is restricted to patients who require an invasive test to verify disease, then the study population may be biased toward patients with more severe disease. In this study, the reference standard was shoulder arthroscopy. Patients with negative findings on MR imaging would be less likely to undergo arthroscopy, and this tendency may result in a report of relatively higher sensitivity and lower specificity for MR imaging than is true because the number of negative tests (both false-negative and true-negative) is decreased. Third, although we made an attempt to match groups of patients with and without SLAP tears, the number of patients in those groups was not equal. In addition, the number of patients who had low-field MR imaging was small compared with the number who had high-field studies. Finally, only one radiologist evaluated the MR imaging studies; therefore, variability among reviewers for the classification of labral patterns was not assessed. Furthermore, because only one interpretation was evaluated, its reproducibility could not be addressed in this study.
In conclusion, the performance characteristics of high-field MR imaging were superior to those of low-field MR imaging for the diagnosis of SLAP tears. We defined patterns of intralabral signal intensity that correspond to the normal and to the torn superior labrum. In particular, a branched linear pattern of abnormal intralabral signal intensity is most consistent with a SLAP tear. T2-weighted oblique sagittal images should be evaluated for abnormal signal intensity in the long head of the biceps tendon, which was seen in 15% of patients with SLAP tears. Finally, partial-thickness and full-thickness rotator cuff tears are often associated with superior labral tears.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Gobezie, D. Zurakowski, K. Lavery, P. J. Millett, B. J. Cole, and J. J. P. Warner Analysis of Interobserver and Intraobserver Variability in the Diagnosis and Treatment of SLAP Tears Using the Snyder Classification Am. J. Sports Med., July 1, 2008; 36(7): 1373 - 1379. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Magee, M. Shapiro, and D. Williams Comparison of High-Field-Strength Versus Low-Field-Strength MRI of the Shoulder Am. J. Roentgenol., November 1, 2003; 181(5): 1211 - 1215. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Dubrulle, J. Delomez, A. Kiaei, P. Berger, C. Vincent, F.-M. M. Vaneecloo, and L. Lemaitre Mass Screening for Retrocochlear Disorders: Low-Field-Strength (0.2-T) versus High-Field-Strength (1.5-T) MR Imaging AJNR Am. J. Neuroradiol., June 1, 2002; 23(6): 918 - 923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Erratum Am. J. Roentgenol., August 1, 2000; 175(2): 548 - 548. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
