| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University Hospital Balgrist, Forchstrasse 340,
CH-8008 Zurich, Switzerland.
2 Department of Orthopedic Surgery, University Hospital Balgrist, Zurich,
Switzerland.
3 Biostatistics Unit, University of Zurich, Zurich, Switzerland.
Received August 23, 2007;
accepted after revision January 14, 2008.
Address correspondence to U. Studler.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Cortical irregularities and cysts of the lesser tuberosity were evaluated on radiographs and MR images in 70 patients: 31 with normal subscapularis tendons and 39 with subscapularis tendon tears. These imaging findings were correlated with subscapularis tendon tears diagnosed during arthroscopy and with fatty atrophy of the subscapularis muscle seen on MR images. Correlations between imaging findings, patient age, history of trauma, and the reference standards were assessed using gamma coefficients. Interreader agreement was calculated using kappa values.
RESULTS. For the diagnosis of subscapularis tendon tears, the
sensitivity of cortical irregularities and cysts on radiographs was 44%/51%
(reader 1/reader 2) and 21%/21%, respectively; specificity was 65%/68% and
87%/87%. The sensitivity of cortical irregularities and cysts on MRI was
64%/72% and 36%/39%. Specificity was 48%/45% and 81%/77%. For fatty atrophy of
the subscapularis muscle, the sensitivity of cortical irregularities and cysts
on radiographs was 67%/73% and 17%/27%, respectively; the specificity was
63%/63% and 83%/85%. Significant (p < 0.05) positive correlations
were found between cortical irregularities and surgical grade of subscapularis
tendon tear for reader 2 (radiography, 
= 0.39; MRI, = 0.45)
and between lesser tuberosity abnormalities and patient age ( = 0.11
and 0.43) for both readers. Interobserver agreement of imaging findings varied
from moderate to substantial (
= 0.50–0.76).
CONCLUSION. Abnormalities of the lesser tuberosity are moderately associated with subscapularis tendon tears and patient age. Cysts seen in the lesser tuberosity on shoulder radiographs or MR images are relatively specific for subscapularis tendon tears and muscle atrophy.
Keywords: arthrography • MRI • radiography • shoulder • sports medicine • subscapularis tendon • trauma
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Shoulder radiography is usually part of the initial workup in patients presenting with shoulder pain or impaired function, which also can be caused by subscapularis tendon tears. Using clinical examination findings, it may be difficult to accurately diagnose partial tears of the subscapularis tendon [7, 8]. Therefore, indirect evidence of subscapularis tendon tears seen on radiographs, analogous to radiographic evidence of subacromial impingement, would be helpful. Based on our experience, changes of the lesser tuberosity are a common but rather variable feature on both radiographs and MR images in patients with shoulder pain.
The purpose of this study was to retro-spectively evaluate the association between abnormalities of the lesser tuberosity and subscapularis tendon lesions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One author (first year of musculoskeletal fellowship) searched the hospital database that contained data about all surgical procedures using the key terms "arthroscopy" and "shoulder." From that search, 343 consecutive patients were identified over a period of 12 months (September 2005 through August 2006). Exclusion of 273 patients was performed for the following reasons: patient underwent repeated shoulder surgery (n = 112), patient had a history of proximal humeral fracture (n = 6), patient had a joint infection (n = 3), no specific description of the subscapularis tendon was included in the surgical report (n = 24), or MR images were not available or MRI was performed more than 6 months before surgery (n = 128).
Among the 70 patients (44 men, 26 women; age range, 32–77 years; mean age, 57 years) included in this study, 39 were reported to have a subscapularis tendon tear (27 men, 12 women; age range, 32–77 years; mean age, 58 years) and 31 were reported to have a normal subscapularis tendon (17 men, 14 women; age range, 37–77 years; mean age, 55 years). The indications for arthroscopy in the 70 patients were rotator cuff disorders (n = 67), instability (n = 2), and frozen shoulder (n = 1).
Subscapularis tendon tears were classified according to the description in the surgical report. The classification system used in our hospital is largely based on the craniocaudal extent of the tendon tear. Any abnormality of the tendon, including an articular-side partial-thickness tear, full-thickness tear, or tendon fraying, is referred to as a "tear" in this study. Grade 0 was given when the tendon appeared normal during arthroscopy. Tears confined to the upper one third of the craniocaudal diameter of the tendon were categorized as grade 1 tears. Tendon tears involving more than one third and less than the complete diameter of the tendon were categorized as grade 2 tears, and complete detachments of the subscapularis tendon were categorized as grade 3 tears. The database was also searched to determine whether patients had a history of trauma. Trauma was defined as a history of an injury to the upper extremity that was considered to be responsible for the patient's shoulder symptoms.
Imaging Protocols
All patients underwent a two-view radiographic examination. The
anteroposterior view was obtain ed parallel to the glenoid surface with the
humerus in a neutral position. The axillary view was obtained with the humerus
abducted and the beam aligned parallel to the glenoid surface.
MR images were obtained with a 1.5-T system (Symphony or Avanto, Siemens Medical Solutions) using a dedicated shoulder coil. The patient's arm was placed in a neutral position with the thumb pointing upward. All patients underwent MR arthrography after injection of approximately 10 mL of gadopentetate dimeglumine (Magnevist, Bayer Schering Pharma) with a concentration of 2 mmol/L. The routine MR arthrography protocol included the following sequences (matrix size for all sequences, 256 x 512): water-excitation true fast imaging with steady-state free precession (FISP) sequence in the transverse plane (TR/TE, 9.24/3.17; 28° flip angle; 1.7-mm section thickness; no intersection gap; 18-cm field of view [FOV]); T1-weighted spin-echo sequence in the sagittal oblique plane, parallel to the glenohumeral joint (450/12, 4-mm section thickness, 16-cm FOV); fat-suppressed T1-weighted spin-echo sequence (648/12, 3-mm section thickness, 16-cm FOV); T2-weighted fast spin-echo sequence (3,460/79, 4-mm section thickness, 16-cm FOV); and intermediate-weighted fast spin-echo sequence (2,870/13, 4-mm section thickness, 16-cm FOV). The three latter sequences were performed in the coronal oblique plane, perpendicular to the glenohumeral joint space.
|
|
|
The readers assessed shoulder radiographs for cortical irregularities and
cysts. For cortical irregularities, a line was drawn along the contour of the
lesser tuberosity on axial radiographs. Irregularities minimally diverging
from that line were assigned a score of 1 (Figs.
1A,
1B,
1C and
2A,
2B,
2C) and more severe
irregularities (diverging 
2 mm), a score of 2 (Fig.
3A,
3B,
3C). Cysts were defined as
round or oval radiolucent lesions surrounded by a sclerotic margin (Fig.
1A,
1B,
1C). A score of 1 was used when
cysts were small (< 3 mm) and when the number of lesions was fewer than
three. Larger or more numerous cysts were assigned a score of 2. We are aware
that the term "cyst" is not entirely appropriate to describe such
lesions [9]; however,
"cyst" is the common term used to describe these changes, and we
retain it for that reason.
|
|
|
|
|
|
In the next step, readers evaluated MR im ages for cortical irregularities and cysts of the lesser tuberosity. Cortical irregularities in the sagittal oblique plane were given scores using the same definitions as for radiographs. Cystic changes were defined as round or oval lesions with high signal intensity on T2-weighted or true FISP images and were assigned scores in the same fashion as described for the radiographic inter pretations (Figs. 1A, 1B, 1C and 3A, 3B, 3C).
During a separate interpretation session, fatty atrophy of the subscapularis muscle on MR images was evaluated in consensus by both readers according to Goutallier et al. [10]. The Goutallier classification is based on the amount of fat in relation to the amount of muscle. Stage 0 indicates no fatty atrophy; stage 1, some fatty streaks; stage 2, less fat than muscle; stage 3, as much fat as muscle; and stage 4, more fat than muscle. In cases with varying amounts of fatty atrophy in the superior and inferior parts of the subscapularis muscles, the area with the highest stage was used.
Statistical Analysis
Data obtained from each reader were used to calculate the sensitivity and
specificity of radiographic and MRI signs by comparing the reading results
with the arthroscopic findings. In a similar way, sensitivity and specificity
were calculated by comparing the radiographic signs with the stage of fatty
atrophy of the subscapularis muscle on MR images. For these calculations, data
on an ordinal scale (surgical grade, stage of fatty atrophy of subscapularis
muscle, presence and score of cortical irregularities, and presence of cysts)
were dichotomized: normal versus abnormal, except fatty atrophy, where a
cutoff at stage 2 or higher was chosen based on clinical relevance
[11]. The 95% CIs were
calculated using the method proposed by Wilson
[12].
Gamma () coefficients of Goodman and Kruskal
[13], which are measures of
ordinal association, were used to measure correlations between imaging
findings and the reference standards defined by either arthroscopy findings or
the stage of fatty atrophy of the subscapularis muscle on MR images. A gamma
coefficient was also used to assess the correlation of imaging findings with
patient age and a history of trauma.
Interreader agreement on the dichotomized scales was quantified by kappa statistics. A kappa value of 0–0.20 indicates slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81– 1.00, almost perfect agreement according to Landis and Koch [14].
A p value of < 0.05 was considered to indicate a statistically significant difference. All analyses were performed with SPSS software (release 10.1.3, SPSS) for Windows (Microsoft).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Osseous Abnormalities of the Lesser Tuberosity
For both readers, sensitivity, specificity, and gamma coefficients of
cortical irregularities seen on radiographs are listed in
Table 2 (when compared with
arthroscopy) and Table 3 (when
compared with fatty atrophy on MRI), respectively. Cortical irregularities
detected on radiographs by reader 2 were significantly correlated with the
surgical grade of subscapularis tendon tears ( = 0.45, p =
0.007) and with fatty atrophy of the subscapularis muscle on MR images
( = 0.49, p = 0.02). An increase in the frequency of cortical
irregularities on radiographs was observed between surgical grades 1 and 2
(Fig. 4A).
|
|
|
Cysts seen on radiographs showed the highest specificity of all imaging features when compared with findings at arthroscopy (87%) or with stage of fatty atrophy on MR images (83–85%). There was a clear difference between the frequency of radiographic cysts in patients with surgical grade 1 and 2 tendon tears (Fig. 4B) similar to our findings for cortical irregularities. Both readers found no cysts on radiographs in patients with grade 3 surgical tears. Gamma coefficients revealed no significant correlation between cysts on radiographs and the reference standards.
|
= 0.39, p = 0.02). Cysts seen on MR images were, as
those seen on radiographs, relatively specific for subscapularis tendon tears
at arthroscopy (Table 2).
Nine of the 10 calculated gamma coefficients showed a positive correlation
between imaging findings and patient age (range, 0.11–0.43). The
following five correlations were significant: cortical irregularity ( =
0.43, p 
0.00) and cysts ( = 0.41, p = 0.001)
seen on MR images by reader 1; and cortical irregularity ( = 0.35,
p = 0.003) and cysts ( = 0.27, p = 0.04) seen on MR
images and cortical irregularity ( = 0.26, p = 0.04) seen on
radiographs by reader 2. We also performed logistic regressions with a
dichotomized version of the image findings as the response variable and with
age and a dichotomized version of arthroscopy as explanatory variables. The
influence of age remained positive for nine of the 10 imaging findings and was
significant for four of them, whereas the relationship of patient age with
arthroscopy findings was not significant.
Gamma coefficients for the correlation between imaging findings and a positive history of trauma ranged between –0.37 and 0.20. No significant correlation could be found and hence no relationship between imaging findings and trauma could be established.
Kappa values () ranged from moder ate (cysts and cortical
irregularities on radiographs, = 0.50) to substantial (cortical ir
regular ities on MR images, = 0.76)
(Table 2).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We found that all radiographic features were insensitive ( 51%) when
compared with the arthroscopic reference of subscapularis tendon tears. On the
other hand, lesser tuberosity cysts seen on radiographs were relatively
specific for subscapularis tendon tears (87%). Because subtle osseous
changes—cysts, in particular—may not be detected on radiographs,
we included patients with both shoulder MR images and radiographs so that
lesser tuberosity changes visible only on MR images could be incorporated into
our analysis.
As we expected, the sensitivities of cortical irregularities and of cysts seen on MR images were higher for both readers when compared with the radiographic results. However, the specificities of these features detected on MR images were lower than the specificities obtained from radiographs by about the same amount. Therefore, osseous abnormalities seen on MR images may overestimate subscapularis tendon tears.
Osseous changes probably occur with a certain delay after subscapularis tendon tearing. Thus, osseous abnormalities may reflect long-standing subscapularis tendon tears. In patients with rotator cuff tears, the degree of fatty atrophy of rotator cuff muscles also increases over time [10]. Moreover, fatty atrophy of the subscapularis muscle has been reported to be highly specific for subscapularis tendon tears [20]. We therefore compared radiographic findings with the fatty atrophy of the subscapularis muscle as depicted by MR images. Similar to our comparison with arthroscopic tendon tears, cysts were shown to be specific for fatty atrophy of the subscapularis muscle (83–85%).
In several studies, investigators have examined the relationship between rotator cuff tears and osseous changes around the shoulder using different imaging techniques [4, 21–26]. Because those studies mainly dealt with supraspinatus tendon tears, the comparison of osseous abnormalities with rotator cuff disorders focused on changes of the greater tuberosity. Only two of those studies also investigated the lesser tuberosity for the presence of cysts and subscapularis tendon tears in small patient subgroups [22, 26]. The results from those two studies are conflicting as to whether cysts are associated with tendon tears. Furthermore, those investigations are limited by a lack of surgical confirmation of tendon tears. Although we found that the correlation of cysts with subscapularis tendon tears was not significant in our study, we noted a trend toward a higher prevalence of cysts in more advanced surgical grades of subscapularis tendon tears.
Histologic investigations of cysts located in the proximal humerus have shown cysts to be filled with fibrous or vascular connective tissue [6, 27] and cystic spaces to be covered by a lining of mesothelial cells and, in a few cases, by a synovial lining that communicates with the joint space. In an article, Resnick [28] reviewed two fundamental pathogenic theories explaining the formation of osseous cysts. The first theory is that repetitive mechanical impact may result in areas of bone necrosis and subsequently in cyst formation [29]. This condition may be encountered in the shoulder. Gerber and Sebesta [18] proposed a concept of anterosuperior impingement within the glenohumeral joint. They suggested that elevation and internal rotation of the arm lead to a mechanical conflict between the anterosuperior glenoid rim and the articular side of the subscapularis tendon.
The second theory that has been advocated is that intact articular cartilage prevents the subchondral bone from being intruded by synovial fluid, a condition that may induce trabecular resorption and cyst formation [30]. Analogous to the covering function of cartilage, intact attachment of tendons protect the underlying cortical bone from being exposed to synovial fluid at the tendon insertion site. Fritz et al. [21] observed that cysts of the greater tuberosity were more common in partial-thickness articular-sided tears than in bursa-sided supraspinatus tears, lending support to the theory of synovial fluid intrusion.
Although our study was not designed to address the cause of cysts, we assume that either of these mechanisms—impingement of the lesser tuberosity or fluid intrusion through cortical infractions—substantially contributed to the formation of cysts in our patients with subscapularis tendon tears. However, our observation of lesser tuberosity cysts on radiographs (13%) and MR images (23%), respectively, in arthroscopically intact tendons indicates that other factors might be causing cystic changes. These findings are consistent with results from previous studies in which investigators reported cysts occurring in the greater tuberosity of asymptomatic subjects. Needell et al. [25] found humeral head cysts in 13% and Huang et al. [23] in 8–16% of shoulders with an intact supraspinatus tendon on MRI. Moreover, in our study, we noted a significant association of patient age with cortical irregularities and cysts, respectively. These osseous abnormalities therefore may in part reflect a degenerative process associated with aging.
There are some limitations to our study. First, the control group did not consist of a symptomatic age-matched healthy volunteers. All patients underwent shoulder arthroscopy because of a variety of shoulder problems, which means that some patients with a normal subscapularis tendon at arthroscopy may have experienced direct anterior trauma to the shoulder. Second, some subscapularis tendon tears may not have been identified at arthroscopy. Subscapularis tendon tears can be missed when tendon degeneration has occurred within the tendon substance (intratendinous tear) and the surface of the tendon appears normal during arthroscopy [31]. Theoretically, these limitations could have contributed to an underestimation of the diagnostic accuracy of osseous abnormalities. This study was further limited by the lack of histologic confirmation of cysts. In addition, retrospective data acquisition in our study caused a significant patient selection bias. However, to our knowledge, ours is the first series comparing osseous abnormalities in the lesser tuberosity with arthroscopically confirmed subscapularis tendon tears.
In summary, abnormalities of the lesser tuberosity are associated with subscapularis tendon tears and patient age. Lesser tuberosity cysts seen on radiographs and MR images are relatively specific for subscapularis tendon tears and fatty atrophy of the subscapularis muscle. Cortical irregularities are less useful for the identification of subscapularis tendon tears. Cysts seen in the lesser tuberosity on radiographs should be reported by the radiologist to prompt clinicians to focus on subscapularis tendon tears. In the presence of ambiguous undersurface subscapularis tendon tears on MR images, cysts indicate supportive evidence of tendon tears and, therefore, increase diagnostic certainty.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
