| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Musculoskeletal Imaging |
1 From the Department of Radiology, University of Michigan Medical Center, Taubman Center, Room 2910-G, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0326 (J.A.J.); Advanced Radiology Services, Spectrum Health, Grand Rapids, Mich (S.L.); Department of Radiology, University of Pittsburgh Medical Center, Pa (A.P.); and Departments of Radiology (M.T.v.H., J.G.C.) and Orthopaedic Surgery (P.K.), Henry Ford Hospital, Detroit, Mich. From the 1999 RSNA scientific assembly. Received April 12, 2002; revision requested June 21; final revision received April 25, 2003; accepted May 20. Address correspondence to J.A.J. (e-mail: jjacobsn@umich.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Fifty consecutive ultrasonographic (US) studies of the shoulder in patients who underwent arthroscopic follow-up were retrospectively reviewed by a musculoskeletal radiologist. US images of the supraspinatus tendon were evaluated for tendon nonvisualization, abnormal tendon echogenicity, tendon thinning, greater tuberosity cortical irregularity, cartilage interface sign, joint fluid, and subacromial-subdeltoid bursal fluid. US findings were compared with arthroscopic results. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated for each US sign in the diagnosis of full-thickness tendon tear and again for any type of supraspinatus tendon tear.
RESULTS: Arthroscopy revealed 21 full-thickness tears, five bursal surface partial-thickness tears, 10 articular surface partial-thickness tears, and 14 patients without tear of the supraspinatus tendon. The presence of greater tuberosity cortical irregularity and joint fluid was most important in the diagnosis of full-thickness supraspinatus tendon tear (sensitivity, 60%; specificity, 100%; positive predictive value, 100%; negative predictive value, 78%; accuracy, 84%). For diagnosis of any type of supraspinatus tendon tear (partial or full thickness), tendon nonvisualization, greater tuberosity cortical irregularity, and cartilage interface sign are most important, although a combination of signs did not improve accuracy.
CONCLUSION: Secondary US signs, such as greater tuberosity cortical irregularity and joint fluid, are most valuable in the diagnosis of supraspinatus tendon tear.
© RSNA, 2004
Index terms: Shoulder, injuries, 414.4813 • Shoulder, US, 414.1298 • Tendons, injuries, 414.4813 • Tendons, US, 414.1298
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Relatively less is known about the US appearance of tendinosis. Tendinosis, or tendinopathy, can be defined as tendon degeneration. The term tendinosis is preferred over tendinitis, because eosinophilic, fibrillar, and mucoid degeneration are present, and acute inflammation is typically absent (7). Bachman et al (8) demonstrated that supraspinatus tendon degeneration might appear hypo- or hyperechoic at US. Differentiation from tendon tear may be difficult, as both tendinosis and tendon tear may appear hypoechoic and coexist within the same tendon.
The purpose of our study was to determine which US signs are important in the diagnosis of a surgically identifiable supraspinatus tendon tear. US images of the shoulder were retrospectively reviewed in patients with arthroscopically proved diagnoses of full-thickness tear, partial-thickness tear, and no tear of the supraspinatus tendon.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The original US examinations were performed by one of five musculoskeletal radiologists (J.A.J., M.T.v.H., or J.G.C.) with 3–12 years experience or one dedicated musculoskeletal US technologist with 2 years experience by using 7.5–10.0-MHz linear transducers (HDI 3000, Advanced Technology Laboratories, Bothell, Wash; 5200, Acoustic Imaging, Phoenix, Ariz) and liberal transmission gel in place of a standoff pad. US images included, but were not limited to, transverse and longitudinal planes relative to the supraspinatus, subscapularis, infraspinatus, and long head of the biceps brachii tendons. The US examinations were ordered and completed as part of the routine clinical investigation of the shoulder, as determined by the orthopedic surgeon. Bursoscopy was performed in conjunction with the arthroscopic examination for evaluation of the supraspinatus tendon. Supraspinatus tendon abnormalities were described at arthroscopy according to the Snyder classification (9) as "A" (articular), "B" (bursal), or "C" (full-thickness or complete tear). A designation of A or B was followed by a number, which indicated grade. Grade 0 was classified as normal, a grade 1 lesion had inflamed synovium and minor (< 1 cm) surface fraying, a grade 2 lesion had actual fiber disruption of more than 1 cm but less than 2 cm, a grade 3 lesion had tendon disruption of 2–3 cm, and a grade 4 lesion had complex tendon disruption, such as flap formation and retraction of more than 3 cm (9).
Image Evaluation
One musculoskeletal radiologist (J.A.J.) who was unaware of the arthroscopic results evaluated all available hard-copy US images of each patient. The supraspinatus tendon was evaluated for primary abnormalities, such as tendon nonvisualization, abnormal tendon echogenicity (predominantly anechoic or hypoechoic relative to normal hyperechoic tendon), and tendon thinning (flattening or concavity of the normal convex bursal surface). If a focal tendon abnormality was identified, the margins were characterized as either well defined or ill defined, and the extent of the abnormality was characterized as intrasubstance, articular, bursal, or articular surface to bursal surface. The US images were also evaluated for secondary findings, which included cortical irregularity of the greater tuberosity (loss of the normal, smooth hyperechoic surface), the cartilage interface sign (a thin, markedly hyperechoic line at the surface of the hypoechoic, hyaline articular cartilage of the humeral head) (10), joint fluid (any visible anechoic or hypoechoic fluid surrounding the long head of the biceps brachii tendon), and fluid in the subacromial-subdeltoid bursa (greater than 2 mm of anechoic or hypoechoic distention [11], as measured at its point of maximum distention from the acromion to 1–3 cm lateral to the greater tuberosity).
Arthroscopy and Statistical Evaluation
The arthroscopy reports were reviewed by two of the investigators, and results were categorized as full-thickness tear, bursal side partial-thickness tear, articular side partial-thickness tear, and no tendon tear, with respect to the supraspinatus tendon. A grade 1 articular or bursal lesion (minimal fraying and inflamed synovium) was categorized as no tendon tear, similar to what has been previously described in the literature (12). To calculate sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for each US sign, the four categories of possible arthroscopic results were combined to create two columns or categories. In the first calculation, the two categories consisted of full-thickness tendon tear and partial-thickness tendon tear or no tear. In the second calculation, the two categories consisted of full-thickness or partial-thickness tendon tear and no tear. We calculated 95% CIs (Stata; StataCorp, College Station, Tex), except when values were equal to 0% or 100%, in which case we calculated one-sided 97.5% CIs.
Calculation of sensitivity, specificity, positive and negative predictive values, and accuracy was then repeated by using every possible combination of the four secondary US signs (greater tuberosity cortical irregularity, cartilage interface sign, joint fluid, and subacromial-subdeltoid bursal fluid) and the primary US sign (tendon flattening). This was calculated by combining the possible arthroscopic categories into full-thickness supraspinatus tendon tear versus partial-thickness tear and no tear and both partial- and full-thickness tendon tear versus no tendon tear, similar to what has been mentioned previously.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The extent of a focal supraspinatus tendon abnormality was articular surface to bursal surface in 12 (75%) of 16 patients with full-thickness tears, one (20%) of five patients with bursal surface partial-thickness tears, six (60%) of 10 patients with articular surface partial-thickness tears, and three (23%) of 13 patients with tendons without tear. The extent of a focal supraspinatus tendon abnormality was articular in two (13%) of 16 patients with full-thickness tears, one (20%) of five patients with bursal surface partial-thickness tears, two (20%) of 10 patients with articular surface partial-thickness tears, and five (38%) of 13 patients with tendons without tear. The extent of a focal supraspinatus tendon abnormality was bursal in two (13%) of 16 patients with full-thickness tears, two (40%) of five patients with bursal surface partial-thickness tears, none (0%) of 10 patients with articular surface partial-thickness tears, and three (23%) of 13 patients with tendons without tear.
The extent of a focal tendon abnormality was intrasubstance in none (0%) of 16 patients with full-thickness tears, one (20%) of five patients with bursal surface partial-thickness tears, two (20%) of 10 patients with articular surface partial-thickness tears, and two (15%) of 13 patients with tendons without tear.
Supraspinatus tendon thinning was seen in 15 (71%) of 21 patients with full-thickness tears, two (40%) of five patients with bursal surface partial-thickness tears, three (30%) of 10 patients with articular surface partial-thickness tears, and two (14%) of 14 patients with tendons without tear.
The occurrence of the secondary US signs for each arthroscopic result is given in Figure 9. Greater tuberosity cortical irregularity was seen in 18 (86%) of 21 patients with full-thickness tears, two (40%) of five patients with bursal surface partial-thickness tears, five (50%) of 10 patients with articular surface partial-thickness tears, and three (21%) of 14 patients with tendons without tear. Depiction of a cartilage interface was seen in seven (33%) of 21 patients with full-thickness tears, none (0%) of five patients with bursal surface partial-thickness tears, none (0%) of 10 patients with articular surface partial-thickness tears, and none (0%) of 14 patients with tendons without tear. Joint fluid was identified in 13 (65%) of 20 patients with full-thickness tears, none (0%) of five patients with bursal surface partial-thickness tears, two (20%) of 10 patients with articular surface partial-thickness tears, and one (7%) of 14 patients with tendons without tear. There was one patient with a full-thickness supraspinatus tear in whom the images did not include the long head of the biceps brachii tendon; therefore, it was not possible to evaluate for joint fluid. Identification of subacromial-subdeltoid bursal fluid was seen in 13 (68%) of 19 patients with full-thickness tears, one (20%) of five patients with bursal surface partial-thickness tears, two (20%) of 10 patients with articular surface partial-thickness tears, and five (36%) of 14 patients with tendons without tear. There were two patients with a full-thickness supraspinatus tendon tear, in whom the presence or absence of subacromial-subdeltoid fluid could not be assessed.
|
|
|
In the diagnosis of full-thickness tendon tear versus other conditions (partial-thickness tear, tendinosis, or no tear), we found the combination of cortical irregularity of the greater tuberosity and joint fluid had 60% sensitivity, 100% specificity, 100% positive predictive value, 78% negative predictive value, and 84% accuracy. These results are higher than those for any single US sign alone. In the diagnosis of any type of tendon tear (full or partial-thickness) versus no tendon tear, we found the combination of tendon thinning and cortical irregularity of the greater tuberosity had 47% sensitivity, 93% specificity, 94% positive predictive value, 41% negative predictive value, and 60% accuracy. The accuracy of combining these US signs was less than several of the single primary and secondary US signs (Table 2).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
With regard to location of a tendon abnormality, an abnormal hypoechoic or anechoic area extended from the bursal to the articular surface in 12 (75%) of 16 patients with full-thickness tears, and the tendon abnormality was isolated to the bursal surface in two (40%) of five patients with bursal side partial-thickness tears; however, only two (20%) of 10 patients with articular side partial-thickness tears had a tendon abnormality isolated to the articular surface on US images. Surface extension of abnormal tendon echogenicity did not allow differentiation between various tendon tears and no tendon tear. Tendon thinning was seen in 15 (71%) of 21 patients with full-thickness tears, two (40%) of five patients with bursal side partial-thickness tears, three (30%) of 10 patients with articular side partial-thickness tears, and two (14%) of 14 patients without tendon tear. The cause of tendon thinning in these two patients without tendon tear is uncertain. Intrasubstance tears that produce tendon thinning may not be depicted at arthroscopy or bursoscopy. Additionally, apparent thinning of the supraspinatus may be encountered if focal hyperechoic subacromial-subdeltoid bursal thickening is mistaken for tendon fibers or if the adjacent tendon is thickened.
Evaluation of the secondary US signs (greater tuberosity cortical irregularity, cartilage interface sign, joint fluid, and subacromial-subdeltoid bursal fluid) shows that with the exception of bursal fluid, each is present more often in patients with a tendon tear than in patients with no tendon tear. The presence of any secondary US sign in the group of patients without tendon tear at arthroscopy or bursoscopy may be attributed to intrasubstance tendon tears or degeneration not visible at arthroscopy or bursoscopy, as well as shoulder abnormalities not involving the supraspinatus tendon.
The importance of cortical irregularity of the greater tuberosity as an indicator of rotator cuff tears has been debated (3,4,13). We found cortical irregularity, as shown with US, to be present in 18 (86%) of 21 patients with full-thickness supraspinatus tendon tears, two (40%) of five patients with bursal side partial-thickness tears, five (50%) of 10 patients with articular side partial-thickness tears, and three (21%) of 14 patients without tendon tear. We found this sign to be very important, as it had the highest sensitivity and negative predictive value in the diagnosis of any type of tendon tear.
A thin echogenic line at the interface of hypoechoic hyaline articular cartilage of the humeral head and the adjacent rotator cuff (the cartilage interface sign) (10) was seen in seven (33%) of 21 patients with full-thickness tears and was not seen in patients with any other conditions. This sign had 100% specificity and 100% positive predictive value in the diagnosis of full-thickness tendon tear or any tendon tear; however, sensitivities remained low. We recognize that a hyperechoic interface between the hyaline cartilage and the overlying rotator cuff can be depicted normally; therefore, we acknowledge that this sign is subjective. A positive cartilage interface sign should be reserved for patients in whom the interface is markedly hyperechoic. This finding is often more conspicuous when there is hypoechoic or anechoic fluid overlying the articular cartilage.
Similar to the importance of greater tuberosity cortical irregularity, the importance of joint fluid and fluid in the subacromial-subdeltoid bursa as an indicator of rotator cuff tear is debated (5,6,14,15). We found joint fluid in 13 (65%) of 20 patients with full-thickness supraspinatus tendon tears, no (0%) of five patients with bursal side partial-thickness tears, two (2%) of 10 patients with articular side partial-thickness tears, and one (7%) of 14 patients without tendon tear. With regard to bursal fluid, this finding was present in 13 (68%) of 19 patients with full-thickness tendon tears, one (20%) of five patients with bursal side partial-thickness tendon tears, two (20%) of 10 patients with articular side partial-thickness tendon tears, and five (36%) of 14 patients without a tendon tear. The relative high percentage of bursal fluid in these latter patients can be explained by possible isolated bursitis or fluid secondary to shoulder impingement.
In the 14 patients without tendon tear in this study, 10 (71%) had abnormal hypoechoic areas, and three (21%) had abnormal anechoic areas. While the cause of this abnormal tendon echogenicity is unknown, it is possible that these findings represent tendinosis; however, this cannot be confirmed, as arthroscopy and bursoscopy are insensitive in the diagnosis of tendinosis and intrasubstance tendon tears. The relative low percentage of tendon thinning (14%) and cortical irregularity (21%) compared with patients with tendon tears suggests the majority of these patients have tendinosis. No patients in this study had abnormal increased areas of tendon echogenicity, which suggests that this may not be a prevalent finding in tendinosis.
In the diagnosis of full-thickness supraspinatus tendon tear versus other conditions (partial-thickness tendon tear or no tendon tear), the following conclusions can be made. Cortical irregularity of the greater tuberosity is the most sensitive US sign (86%). Tendon nonvisualization and cartilage interface sign are the most specific signs (100%). Tendon nonvisualization and the cartilage interface sign have the highest positive predictive values (100%). Greater tuberosity cortical irregularity has the highest negative predictive value (86%), while overall, the presence of joint effusion is most accurate (80%). In other words, no one US sign is very helpful. We found that a combination of secondary US signs (greater tuberosity cortical irregularity and joint fluid) is, in fact, most helpful (sensitivity, 60%; specificity, 100%; positive predictive value, 100%; negative predictive value, 78%; accuracy, 84%).
In the diagnosis of partial- or full-thickness supraspinatus tendon tear versus no tendon tear, greater tuberosity cortical irregularity is most sensitive (69%), tendon nonvisualization and cartilage interface sign are most specific (100%) and have the highest positive predictive values (100%), and greater tuberosity irregularity has the highest negative predictive value (50%) and highest accuracy (72%). We found that a combination of US signs did not improve accuracy in the detection of full-thickness tear or any other tear when compared with any single primary or secondary US sign.
The limitations of this study include the retrospective review of US images, which relies on the technique of the individual performing the US examination and the diligence of that individual to document adequately the US findings on static images. Additionally, it was not possible to categorize the primary and secondary US signs as independent observations, as many of the findings were present on the same image. Our use of distention of the long head of the biceps brachii tendon sheath as an indicator of joint fluid is another limitation, as it is possible that not all cases of joint fluid will distend this tendon sheath, and true tenosynovitis may simulate joint fluid on static US images. Selection bias may be present, as patients with abnormal US findings may be more likely to undergo arthroscopy. The use of arthroscopy and bursoscopy can be viewed as a limitation as well, as intrasubstance tendon abnormalities cannot be diagnosed.
In conclusion, in the diagnosis of full-thickness supraspinatus tear, the secondary US signs of greater tuberosity cortical irregularity and joint effusion are most helpful. In the diagnosis of partial-thickness supraspinatus tear, tendon nonvisualization, greater tuberosity cortical irregularity, and cartilage interface sign are important, although a combination of US signs did not improve accuracy.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  U. Studler, C. W. A. Pfirrmann, B. Jost, V. Rousson, J. Hodler, and M. Zanetti Abnormalities of the Lesser Tuberosity on Radiography and MRI: Association with Subscapularis Tendon Lesions Am. J. Roentgenol., July 1, 2008; 191(1): 100 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. B. Fritz, H. A. Ouellette, T. A. O'Hanley, A. Kassarjian, and W. E. Palmer Cystic Changes at Supraspinatus and Infraspinatus Tendon Insertion Sites: Association with Age and Rotator Cuff Disorders in 238 Patients Radiology, July 1, 2007; 244(1): 239 - 248. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Herold, M. Bachthaler, O. W. Hamer, R. Hente, S. Feuerbach, C. Fellner, M. Strotzer, M. Lenhart, and C. Paetzel Indirect MR Arthrography of the Shoulder: Use of Abduction and External Rotation to Detect Full- and Partial-Thickness Tears of the Supraspinatus Tendon Radiology, July 1, 2006; 240(1): 152 - 160. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
|
|
|
|
A. Papatheodorou, P. Ellinas, F. Takis, A. Tsanis, I. Maris, and N. Batakis US of the Shoulder: Rotator Cuff and Non-Rotator Cuff Disorders RadioGraphics, January 1, 2006; 26(1): e23 - e23. [Abstract] [Full Text]
|
|
|
|
|
|
J. B. Moosikasuwan, T. T. Miller, and B. J. Burke Rotator Cuff Tears: Clinical, Radiographic, and US Findings RadioGraphics, November 1, 2005; 25(6): 1591 - 1607. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
