mri of the shoulder

MRI of the Shoulder

Hollis M. Fritts* and Edward V. Craigt

MRI of the shoulder is w ide ly considered the imaging modality of choice in the evaluation of shoulder pain and the clinical impingement syndrome. This is because of its direct evaluation of all of the soft tissue structures of the subacromial space, as well as its abi l i ty to depict the relationship of the overlying osseous and soft tissue structures of the coracoacromial arch. It also provides information regarding the capsulolabral anatomy and, w i th the addit ion of MR arthrography, is becoming recognized as the imaging modality of choice for instability workup. MRI evaluation, when combined with the always important clinical history, physical examination, and radiographs, provides the referring clinician and orthopedic surgeon with the most anatomic and pathological information possible. This, in turn, allows the most informed decision making possible regarding conservative management or surgical treatment. Copyright �9 1994 by W.B. Saunders Company

M RI HAS BECOME the imaging modality of choice, after initial radiographs, for

the evaluation of the painful shoulder. This article relates the experience gained in an outpatient, private practice setting from more than 10,000 shoulder MRI examinations over the course of the past 7 years using 1.5 T and 0.5 T scanners.

Today's cost-conscious environment demands the maximum amount of information possible in the least amount of time. It is the experience of these investigators that two-echo spin-echo imaging with proton-density and T2-weighted technique in three planes provides the most information. After a short axial localizing series using Tl-weighted technique, two-echo spin-echo technique is performed in the oblique coronal, oblique sagittal, and axial planes. Although Tl-weighted images are relatively sensitive for tendon pathology, they do not provide specificity for differentiation of tendinosis, tendinitis, and tears. T2-weighted images are critical for this differentiation. I T2-weighted images are also helpful in detection of other abnormalities such as bursal hypertrophy or bursal fluid accumulation. Proton-density images obtained with two-echo pulse sequences are adequate in providing pseudo-T1 information, including incidental findings of marrow disease.

Axial gradient-echo imaging has been advocated for evaluation of the glenoid labrum in patients with instability. 2-6 Although this technique can be useful in evaluating the labrum, especially on low field-strength scanners, it has not been the investigators' experience that it is more reliable than two-echo spin-echo technique on high-field systems. The inherently greater soft tissue contrast of spin-echo technique is superior to the gradient-echo technique

in evaluating the soft tissues of the shoulder, including the anterior and posterior components of the rotator cuff. Gradient-echo imaging is more susceptible to artifact in the postoperative shoulder than conventional spin-echo imaging.

Intravenous gadolinium-DTPA is rarely used when evaluating soft tissue masses. MR arthrography using a dilute 1:250 mL dilution of gadolinium-DTPA is not used routinely for evaluation of the painful shoulder, but because of its added sensitivity and specificity in the detection of capsulolabral pathology, it is considered for use in young throwing athletes.

The parameters for shoulder MRI examinations when using a 1.5 T scanner are repetition time (TR)/echo time (TE) of 2,000/20, 80 milliseconds, using one-signal excitation with 16-cm field of view. A 3-mm slice thickness with a l-ram gap is used with a 192 x 256 matrix in oblique coronal, oblique sagittal, and axial planes, and the images are filmed with 1.3 to 1.5 magnification. The time of examination from room entry to exit is approximately 35 minutes. Parameters when using a 0.5 T scanner are the same except for the use of a 4-ram slice thickness and a shorter bandwidth of 4.8 MHz.

ROTATOR CUFF DISEASE

Treatment planning for rotator cuff lesions and impingement syndrome is based on mul-

From the *Center for Diagnostic Imaging, Minneapolis, and the ?Department of Orthopaedic Surgery, University of Minne- sota, Minneapolis, MN.

Address reprint requests to Hollis M. Fritts, MD, Center for Diagnostic Imaging, 5775 Wayzata Blvd, Ste 190, Minneapolis, MN 55416.

Copyright �9 1994 by W.B. Saunders Company 0887-2171/94/1505-000255. 00/0

Seminars in Ultrasound, CT, and MRI, Vol 15, No 5 (October), 1994: pp 341-365 341

342 FRI'FI-S AND CRAIG

tiple factors. These include the presence and extent of rotator cuff tear as well as the presence of other impingement lesions including tendon degeneration/tendinitis and bursitis, the expected level of joint function, the severity of pain, and the surgeon's philosophy and skill in choosing various open or arthroscopic methods of treatment. Shoulder MRI is the only imaging modality that provides information regarding all stages of the impingement/cuff degeneration process and rotator cuff abnormalities. It also evaluates the relationship to the overlying coracoacromial arch.

ROTATOR CUFF DEGENERATION

Tears of normal rotator cuff tendon tissue have been described as very uncommon. 7 Tears most commonly occur through weakened degenerative tendon tissue, explaining the predomi- nance of rotator cuff tears in the older age group, s,9 Tears are uncommon in persons youn- ger than 40 years of age. 1~ Factors that have been described as contributing to cuff degeneration include normal aging, hypovascularity of the critical zone, and repeated microtrauma as a result of impingement between the humeral head and the overlying coracoacromial arch. Overuse of the shoulder from athletic or professional activities has also been recognized as promoting rotator cuff injuries. An athletically accelerated aging of the tissue of the rotator cuff has been suggested. 13

Neer popularized the clinical impingement syndrome and proposed a staging classification based on age, duration of disease process, and response to conservative treatment3 ~ Pathologi- cal changes in stage I impingement, including edema and hemorrhage, evolve into fibrosis and tendon degeneration in stage II and eventually to tear of the tendon in stage III.

CORACOACROMIAL ARCH IMPINGEMENT

The majority of rotator cuff disorders have been attributed to the clinical impingement syndrome. Impingement syndrome is a clinical diagnosis that refers to pain associated with the encroachment of the rotator cuff and associated soft tissues including subacromial-subdeltoid bursa and biceps tendon, between the humeral head and the overlying coracoacromial arch. The coracoacromial arch consists of the ante-

rior aspect of the acromion, coracoacromial ligament, coracoid process, and the acromioclavicular joint. Impingement by the coracoacromial arch can be secondary to abnormalities involving one or more of these components, resulting in narrowing of the acromiohumeral distance.

Acromial morphology and its relationship to the subacromial space should be noted on every shoulder MRI examination. Narrowing of acromiohumeral distance can be associated with variations of acromial shape. Bigliani et al described three shapes of the acromion based on lateral-outlet plain-film analysis. 14 A similar plane of visualization is seen on oblique sagittal MRI. Although it must be realized that whereas radiography represents a summation of the entire morphology on one image, MRIs are sequential tomographic slices of that anatomy (Fig 1). The type 1 acromion is relatively fiat in its anteroposterior course. The type 2 acromion curves gently in the anteroposterior dimension, paralleling the superior surface of the humeral head. The type 3 acromion is associated with a congenital or acquired hook or spur of the distal acromion that focally decreases the radius of curvature of the anterior aspect of the acromion. The latter configuration is especially prone to narrowing of the acromiohumeral distance and has been associated with an increased incidence of rotator cuff tears. Subacromial spur formation may also be seen in the oblique coronal plane (Fig 2). 15

Normally the acromion has a nearly horizon- tal orientation, similar to that of the distal clavicle. Lateral downsloping of the anterior acromion, which can be seen with or without type 3 acromion shape, is best appreciated on oblique coronal images and can contribute to narrowing of acromiohumeral distance (Fig 2).16,17

Along with acromial morphology, the presence or absence of an unfused os acromiale should be noted on every shoulder MRI examination. An unfused distal acromial apophysis is commonly mobile and has been associated with an increased incidence of impingement syndrome and rotator cuff tears. 18,19 This may be caused by associated inferior bony hypertrophy that narrows the subacromial space. It is also hypothesized that transient inferior displace-

MRI OF THE SHOULDER 343

Fig 1. Acromion configuration in the oblique sagittal plane. (A) Type I acromion has a flat inferior surface (arrows). (B) Type II acromion has a gentle inferior radius of curvature (arrows) that approximately parallels the superior surface of the humeral head. A type III acromion exhibits a focal decrease in the radius of curvature. This may be seen as (C) a smooth-margined congenital hook of the acromion (arrow) or (D) a subacromial spur or enthesophyte (arrow) near the acromial attachment of the r ligament (small arrows).

ment of the unstable os acromiale by the at- tached deltoid muscle with shoulder abduction may lead to a dynamic narrowing of acromiohumeral distance. This variant, which can be seen on axillary plain films, is best detected on axial MRIs (Fig 3). However, its relationship to the subacromial space is best appreciated on coronal and sagittal images, where it can mimic, although can be seen separate from, the normal

acromioclavicular joint. The presence of fluid- like signal intensity in this defect is suggestive of instability, although the absence of this finding does not predict stability at surgery.

Acromioclavicular joint morphology and its relationship to the subacromial space should also be noted. This is best appreciated on oblique coronal and oblique sagittal images where inferior capsular hypertrophy and osteo-

344 FRI3-1"S AND CRAIG

Fig 2. (A) Proton-density and (B) T2-weighted oblique coronal images. Lateral downsloping of the acromion (curved arrow) and thickening of the acromial attachment of the coracoacromial ligament (arrows) contributes to narrowing of acromiohumeral distance. Note the subacromial spur or enthesophyte (arrowhead) beneath the acromial attachment of the coracoacromial ligament and the superficial partial-thickness tear of the distal supraspinatus tendon (open arrow). (C) Oblique sagittal image showing prominent coracoacromial ligament (arrows).

phyte formation can contribute to a narrowing of the subacromial space (Fig 4).

The coracoacromial ligament and its relationship to the subacromial space should be noted. The normal coracoacromial ligament is routinely seen on high-resolution MRI as a low signal intensity structure coursing from the coracoid process to a broad-based attachment to the inferior aspect of the acromion on both oblique coronal and oblique sagittal planes (Fig 5). Thickening at its acromial attachment can contribute to narrowing of the subacromial space and associated impingement of underlying soft tissues (Fig 2).

SOFT TISSUES OF THE SUBACROMIAL SPACE

Subacromial-Subdeltoid Bursa and Peribursal Fat Plane

There has been some confusion in the litera- ture regarding the MRI appearance of the subacromial-subdeltoid bursa and peribursal fat plane. The subacromial-subdeltoid bursa has been erroneously described as a thin band of fat signal intensity and pathological bursal thickening as a wide band of fat on T1- and T2- weighted images. The potential space of the subacromial-subdeltoid bursa is seen as a thin line of low signal intensity on proton-density


ii:i:)

Fig 3. Unfused os acromiale. This 31-year-old man with shoulder pain for the past 10 months is seen to have an unfused os acromiale (oa) that is best appreciated with respect to the acromion (a) and clavicle (cl) on the (A) proton-density axial view. Proton-density (B) oblique coronal and (C) oblique sagittal views show mild inferior displacement with mild narrowing of acromiohumeral distance. {D) Sagittal T2-weighted image shows a small focus of subacromial bursal thickening (arrow). Arthroscopic decompression confirmed mobile os acromiale.

and T2-weighted images, normally separated from the overlying deltoid muscle and underlying rotator cuff by the normal subdeltoid peribursal fat plane (Fig 6). Kaplan et al, using spin- echo T1- and gradient-echo T2*-weighted techniques, found no evidence of detectable fluid in the subacromial-subdeltoid bursa in 30 asymptomatic shoulders. 2~ Neumann et al detected fluid in the subacromial-subdeltoid space in 11 (20%) of 55 shoulders in 32 asymptomatic

volunteers using conventional proton-density and T2-weighted technique. 21 In 6 of these 11, the amount of fluid was minimal (2 to 3 mm). The remaining 5 were 2 mm • 4 mm in dimension or appeared as a thin linear band in the subdeltoid space. Using a fat-suppression technique, Mirowitz described visualization of small amounts of markedly increased signal intensity within the subacromial-subdeltoid bursa in 10 of 15 asymptomatic subjects. 22 He postulated

346 FRI'I-FS AND CRAIG

Fig 4. Large acromioclavicular osteophyte and large chronic rotator cuff tear. T2-weighted image shows a large spur (arrow) arising from the acromioclavicular joint that markedly encroaches on the subacromial space. Large supraspinatus tendon tear with marked tendon retraction to the level of the glenohumeral joint (curved arrow) and severe muscle atrophy (open arrow) reflecting tear chronicity are also seen.

that this fluid-sensitive technique increases detection of a small amount of fluid that is normally present but is below the threshold for delineation with conventional MRI methods. He suggested that one should not regard merely the presence but rather the quantity of bursal

Fig 6. Normal subacromial-subdeltoid bursa. This potential space can be seen on this proton-density oblique coronal image as a line of low signal intensity (large arrows) separated from the overlying deltoid muscle and, at times, from the underlying rotator cuff by peribursal fat (small arrows).

fluid as abnormal. This investigator's experience is that in the normal and asymptomatic shoulder, fluid-like signal accumulation in the subacromial-subdeltoid bursa seen on T2- weighted images is uncommon. Foci larger than 2 mm • 4 mm in thickness may be considered most likely consistent with bursal inflammation

Fig 5. Coracoacromial ligament (arrows), seen in the oblique sagittal plane, courses from the (A) coracoid process (cp) to the (B) undersurface of the acromion (a).


or bursal fluid accumulation. Although this is a common finding in individuals with clinical impingement syndrome, the absence of MRI proof of pathological bursal thickening does not exclude clinically significant bursitis and clinical impingement syndrome.

Fluid-like signal seen in the subacromial- subdeltoid bursa has been described as a secondary sign of rotator cuff tears or other cause such as inflammatory arthritis or infection. 23 Other

investigators have not relied on this sign as an indication of tear. 24,25

Our experience is that pathological bursal thickening and increased signal intensity also can be seen as the earliest and only sign of an overuse or early impingement process, without underlying tendon pathology (Fig 7) and can be a helpful positive finding for impingement process in the MRI examination of the painful shoulder. ~ This is seen as intermediate-low

/

D

Fig 7. Subacromial bursitis without rotator cuff pathology in a 24-year-old man with 4 months of persistent shoulder pain. (A) Coronal proton-density image with a thin layer of intermediate signal intensity that increases in signal on corresponding (B) T2-weighted image (arrows). (C) Oblique sagittal proton-density and (D) T2-weighted images also show bursal thickening and fluid-like signal intensity (large arrows) in the anterior aspect of the subacromial-subdeltoid bursa overlying the rotator cuff interval. Compare this with the signal intensity of fat that decays on T2-weighted images (small arrow).

348 FRITTS AND CRAIG

signal intensity on proton-density-weighted images that increases to high signal intensity on associated T2-weighted images. This high signal intensity on T2-weighted images may be because of frank bursal fluid accumulation or thickened water-laden bursal tissue. Although bursal thickening may be seen immediately beneath the acromion, the immediate subacromial space is often narrowed because of the associated impingement process, providing little room for bursal tissue accumulation. Therefore, bursal hypertrophy tends to accumulate in areas of less confinement anterior and posterior to the acromion, often seen best when comparing oblique sagittal proton-density-weighted and T2-weighted images. Tl-weighted images are not as sensitive for this finding. Bursal hypertrophy or fluid accumulation may also be seen in the oblique coronal and axial planes next to the lateral aspect of the greater tuberosity.

One potential pitfall can be encountered when imaging an individual who recently had a therapeutic or diagnostic injection, which can mimic bursitis or bursal fluid accumulation. To avert this problem, appropriate history should be obtained at the time of the examination, and if injection has been performed recently, a diagnosis of bursitis or bursal fluid must be made with caution. It has been suggested to delay MRI after diagnostic or therapeutic injection, although for an unspecified time. 26 MRI examinations were performed recently on four asymptomatic volunteers before and after posterior subacromial injections, two with 10 mL of xylocaine and two with 10 mL of a xylocaine/ cortisone mixture. 27 Imaging was performed within 30 minutes of injection, at 12 hours, and at 24 hours. In all cases, the associated fluid signal intensity was seen primarily posteriorly overlying the distal infraspinatus muscle belly and tendon and was completely gone by 24 hours. Although these results are preliminary, they suggest that subacromial fluid signal intensity seen on MRI examinations performed more than I day after 10 mL of subacromial injection may be considered pathological rather than iatrogenic.

The absence of an intact subacromial-subdeltoid fat plane has been described as an ancillary sign of rotator cuff pathology. However, Miro- witz found that this plane is rarely complete at

the anterior aspect of the supraspinatus tendon insertion in asymptomatic individuals. = Mitchell et al also documented the peribursal fat plane to be most prominent radiographically in inter- nal rotation, where its more prominent posterior aspect is seen tangentially28 MRI evaluation of Kaplan et al's 30 asymptomatic volunteers found discontinuity or obliteration of the subacromial-subdeltoid bursal fat plane in 17 of 30 volunteers, concluding it to be an unreliable sign of rotator cuff pathology. 2~

Normal Rotator Cuff

The normal appearance of the rotator cuff has been described. The normal distal rotator cuff was originally described as uniformly low in signal intensity, perhaps in part because of lower resolution and thicker slice-thickness imaging. However, even the normal distal supraspinatus tendon is not uniformally low signal. In a description of the MRI appearance of the rotator cuff in 15 asymptomatic subjects, Miro- witz described visualization of multiple tendinous slips, usually three, and a gradual transi- tion of muscle to tendon with muscular fibers routinely extending to within approximately 1 cm of the greater tuberosity. 22 These muscle fibers may contribute to increased signal intensity, although a nonfocal configuration and signal intensity identical to that of normal muscle on all pulse sequences suggests that this does not represent tendon pathology.

Using both conventional and fat-suppressed two-echo spin-echo techniques, Mirowitz described foci of relatively increased signal intensity in the substance of the supraspinatus tendon in all 15 normal subjects. 22 These foci were described as being present within the substance of the distal supraspinatus tendon, 8 mm proximal to the greater tuberosity. They were attributed to the critical zone of the rotator cuff, a described site of relative hypovascularity within the supraspinatus tendon that is believed by some to be the initial site of degenerative changes and subsequent tears. 29,3~

Rotator Cuff Tendinitis

Tendinitis has been used in the clinical litera- ture as a relatively nonspecific pathological term including changes associated with tendon


degeneration, inflammation, or edema. The less specific degenerative term, tendinosis, has been proposed as more appropriate. 31 Homogeneous or nonfocal increased signal intensity, especially on Tl-weighted and proton-density-weighted images, with otherwise normal tendon morphology, may merely reflect nonspecific and asymptomatic tendon degeneration. With mild nonspecific degeneration, this increased signal intensity on proton-density-weighted images does not persist on T2-weighted images. However, the presence of tendinous enlargement with an inhomogeneous signal intensity, including mild to moderate increased signal intensity on T2- weighted images, is strongly suggestive of ongoing symptomatic tendinitis (Fig 8). Pathological thickening and increased signal intensity of the subacromial-subdeltoid bursa would support an ongoing and potentially symptomatic process. Rafii et al noted that some similarity and overlap of signal patterns of partial interstitial tears, tendinitis, and tendon degeneration are observed. 25

Although predominantly a plain film diagnosis, calcium hydroxyapatite crystal deposition can occasionally be appreciated as a focus of low signal intensity in the rotator cuff tendons on both proton-density-weighted and T2- weighted images (Fig 9). Because this deposition is of low signal intensity, it may be difficult

Fig 9. Calcific tendinitis of the supraspinatus tendon, without tear, Proton-density oblique coronal image shows a focal collection of decreased signal intensity within the superficial aspect of the distal right supraspinatus tendon (arrow). This represents calcium hydroxyapatJte crystal deposition.

to distinguish from the low signal intensity tendon.

Supraspinatus Rotator Cuff Tears

The value of detection and subsequent repair of full-thickness rotator cuff tears has been documented. Partial-thickness cuff tears are

..... ~o~

Fig 8. Marked intrasubstance degeneration/tendinosis ("tendinitis") of the suprespinatus tendon. Marked fusiform tendon thickening (arrows) associated with ill-defined increased signal intensity is seen on (A) proton-density and (B) T2-weighted axial images.


commonly found in patients with impingement who fail conservative management. Although yet to be proven, it is believed that a partial- thickness tear represents an active evolving process that, with continued impingement, leads to the development of a full-thickness tear. If intervention occurs early, such as with anterior acromioplasty for subacromial decompression, the progressive degenerative process may be arrested or slowed, leading to an improved outcome. This speculation into the evolution and pathogenesis of rotator cuff tears fosters continued attempts of early diagnosis of subacromial space pathologyi including tendinitis and partial-thickness tears, before the development of full-thickness tears.

Rafii et al studied the signal intensity patterns of rotator cuff lesions at MRI in 80 patients who had surgical correlation,'as well as in 13 asymptomatic individuals. 25 The accuracy of MRI in the detection of 31 full-thickness tears was 95%, and of 16 partial-thickness tears was 84%. The most common and most accurate diagnostic findings of partial- and full-thickness tears were regions of increased signal intensity on T2- weighted images, indicating fluid or fluid laden bursal tissue in the cuff defect (Fig 10). This was seen in 22 of 31 fulMhickness tears. High signal intensity on T2-weighted images was the most

accurate, although less common, finding in 7 of 16 partial-thickness tears.

In seven of the full-thickness tears, the diagnosis was made in the absence of high signal on T2-weighted images. The abnormality was seen as low or moderate signal intensity on proton- density and T2-weighted images. This repre- sented a severely degenerated tendon, a low signal intensity of intact bursal and/or synovial surface, or perhaps granulation or scar tissue that replaced the region of torn tendinous fibers and, in some, maintained the continuity of the tendinous margins. The diagnosis was made on the basis of the alteration of normal morphology and contour as well as secondary signs such as subacromial-subdeltoid bursitis, retraction of the musculotendinous junction, and muscle belly atrophy (Fig 4). Rafii et al suggest that the appearance of a tendon tear replaced by connec- tive tissue showing moderate signal intensity may simulate the appearance of tendinitis. 25

Traughber and Goodwin reported a comparison of preoperative MRI findings and careful arthroscopic inspection of both sides of the rotator cuff in 28 patients. 32 MRI correctly identified all five of the full-thickness tears for a sensitivity and specificity of 100%. However, only five of nine partial-thickness tears were diagnosed on MRI. The combined sensitivity

Fig 10. Full-thickness supraspinatus tendon tear. (A) Proton-density coronal image shows abnormal signal intensity (arrows) but poor definition of tear. (B) T2-weighted image shows high signal intensity fluid (arrows) that better defines the size of the full-thickness rotator cuff tear,


and specificity for diagnosis of both partial- and full-thickness tears were 71% and 93%, respectively. It was their impression that short TI inversion recovery (STIR) sequences improved tear conspicuity.

Histological differences between the bursal- and joint-side layers of the supraspinatus have been described in a cadaveric study by Na- kajima et al. 33 They further noted different biomechanical properties of these two layers and concluded that the joint-side layer is more vulnerable to a tensile load than the bursal-side layer. The bursal-side layer, which was com- posed primarily of tendon bundles, increased in length to a tensile load but was resistant to rupture. The joint-side layer was a complex of tendon, ligament, and joint capsule that elon- gated more poorly and tore more easily. This may be hypothesized to contribute to the obser- vation by others that joint-side partial-thickness tears are more common than bursal-side tears.

An appreciation of the differences between the joint- and bursal-side layers of the rotator cuff can be seen on MRI of the pathological rotator cuff tendon. First, lamellar orientation of abnormal increased signal intensity between superficial and deep layers of the rotator cuff is frequently observed. Most commonly, this delamination between the superficial (bursal-side) and deep (joint-side) layers is seen adjacent to a partial- or full-thickness rotator cuff tear. A variable degree of moderate increased signal intensity is seen on both proton-density and T2-weighted images, although differentiation from the lower signal intensity of the adjacent layers commonly is best seen on T2-weighted images (Fig 11). This delamination process is documented at surgery as extension into the margins of a partial- or full-thickness rotator cuff tear.

Appreciation of this lamellar arrangement of superficial and deep tendon layers further allows an insight into understanding the MRI appearance of some rotator cuff pathologies. An example is the large, deep surface supraspinatus tendon tear that consists of a distal tear and proximal retraction of the deep layer of the supraspinatus tendon, whereas the superficial layer remains intact. At first impression on MRI, the supraspinatus tendon may look thin but otherwise unremarkable. However, a ridge

of low signal intensity at the deep surface may be appreciated proximal to the greater tuberosity that extends from anterior to posterior, representing the torn and retracted edge of the deep tendon layer (Fig 12). Occasionally, the deep surface of the intact superficial layer may undergo further degeneration that can result in pathological thickening of this superficial layer such that it approximates the thickness of the normal rotator cuff. In this case, the retracted deep layer edge may be the only or most prominent finding heralding a significant deep- surface partial-thickness tear.

Superior (bursal-side) partial-thickness tears can also be difficult to appreciate. An appreciation of the normal relationship of the superficial distal cuff to the level of the lateral aspect of the greater tuberosity can be helpful in the detection of a distal, superficial, partial-thickness tear. The distal superficial aspect of the supraspinatus tendon inserts near the lateral edge of the greater tuberosity (Figs 3 and 6). Visualiza- tion of a step between the level of the greater tuberosity and the superficial suPraspinatus should raise suspicion for a superficial tear or a defect of the distal tendon (Fig 13). Perhaps the most difficult superficial tear to visualize is the shallow abrasion or fraying occurring proximally beneath the overlying acromion. This can be appreciated by loss of definition or interruption of the morphology of the superficial tendon that may be seen best on proton-density or T2- weighted images. Fluid or fluid-laden bursal tissue in the subacromial-subdeltoid bursa can increase the conspicuity of a bursal-side partial- thickness tear. However, without bursal thickening, or with the more common situation of chronic bursal adhesions and scarring that may be intermediate or intermediate-low in signal intensity, these superficial partial tears or abra- sions can be obscured or difficult to appreciate.

Subscapularis Tendon Tears

MRI has led to an increase in the detection of subscapularis tendon degeneration and tears. Subscapularis tendon tears are associated mos t commonly with supraspinatus tendon tears and have been described as occurring rarely in isolation. 34 However, it is our experience that MRI has increased the detection of subscapularis tendon abnormalities and that isolated

352 FRII-FS AND CRAIG

Fig 11. Intrasubstance delamination tear between superficial and deep layers of the infrespinatus tendon. Oblique coronal (A) proton-density and (B) T2-weighted images, as well as oblique sagittal (C) proton-density and (D) T2-weighted images through the infraspinatus tendon show linear increased signal (large arrows) between superficial and deep layers (small arrows).

subscapularis tendon tears are more frequent than originally believed.

Patten reported MRI findings in nine patients with surgically confirmed rotator cuff tears solely or predominantly involving the subscapularis tendon, representing 6% of 149 rotator cuff tears found in 571 patients who had MRI in a period of 3 years. 35 The MRI appearance of subscapularis tendon tears is similar to that previously described for supraspinatus tears, including areas of abnormal and disorganized tendon morphology and abnormal increased

signal intensity on all pulse sequences. As in the supraspinatus, tears are best appreciated when comparing proton-density and T2-weighted axial images (Fig 14), which are preferable to T1- weighted or gradient-echo T2*-weighted images for best tear conspicuity and specificity.

lnfraspinatus and Teres Minor Tendon Tears

Infraspinatus tendon tears are most commonly associated with supraspinatus tendon tears. Isolated, full-thickness, infraspinatus tears are uncommon in the experience of this investi-


Fig 12, Deep-surface partial-thickness tear with retraction of tom edge of deep layer on (A) proton-density and (B) T2-weighted images. A ridge of low signal intensity at the 12 o'clock position (large arrow) represents the torn edge of the deep tendon layer of the supraspinatus. The superficial layer remains intact (open arrows), although there is absence of the distal aspect of the deep layer (small arrows) seen as increased signal intensity on T2-weighted images. Note the large osteophyte arising from the inferior aspect of the acromioclavicular joint (curved arrow).

gator, much less common than isolated subscapularis tendon tears (Fig 15). A full-thickness tear of the entire teres minor tendon is very uncommon, even with massive full-thickness tears involving the rest of the rotator cuff. The investigator has not noted an isolated teres

minor tendon tear, although isolated teres mi- nor muscle belly atrophy without tendon tear can be seen. This has been reported to be associated with impingement and/or denervation of the axillary nerve in the quadrilateral space. 36

,4: %~ ,

Fig 13. Superficial partial-thickness tear of the supraspinatus tendon on oblique coronal (A) proton-density and (B) T2-weighted images, Note abnormally thin supraspinatus tendon caused by the tear of the superficial layer of the tendon (large arrows), replaced by high signal intensity fluid or bursal tissue on the T2-weighted image. The superficial layer that usually extends to the lateral aspect of the greater tuberosity is absent (curved arrow). Note low signal of thickened bursal layer (small arrows) that may mimic superficial tendon fibers.


Fig 14. Isolated full-thickness tear of the subscapularis tendon. Interruption of the low signal intensity of the distal subscapularis tendon (small arrows) by fluid signal (large arrow) on (A) proton-density and (B) T2-weighted axial images.

MRI OF THE BICEPS TENDON

The entire complex course of the biceps tendon is visualized at MRI. Erickson et al described the normal MRI appearance of the biceps tendon and illustrated associated pathological conditions. 37 Biceps tendon medial dislocation and subluxation are commonly associated with subscapularis tendon pathology and tears. MRI is excellent for noninvasive diagnosis of biceps tendon ruptures (Fig 16). Lesser degrees of biceps tendon pathology also can be appreciated on MRI including tendinosis, tendinitis, partial-thickness fraying, and longitudinal splitting.

Intrinsic Abnormalities of the Biceps Tendon

Oblique sagittal and axial planes are obliga- tory for optimal evaluation of the biceps tendon, with oblique coronal images providing an ancillary role. Although axial images are best for visualization of the biceps tendon at the level of the bicipital groove, proximal biceps tendon degeneration and tendinosis associated with the clinical impingement syndrome can be best appreciated on oblique sagittal images. Here, the intra-articular portion of the tendon is seen in cross-section within the rotator cuff interval and underlying the coracohumeral ligament. In this location, the normal biceps tendon is rela-

Fig 15. Isolated tear of infraspinatus at its musculotendinous junction. Note increased signal intensity at the infraspinatus musculotendinous junction (arrows) on (A) T2-weighted coronal and (B) T2-weighted axial images.


Fig 16. Isolated biceps tendon rupture. The proximally torn biceps tendon is seen as a ser- piginous structure retracted distally to a level below the bicipital groove (curved arrow). The biceps tendon is not seen in the rotator cuff interval (large arrow) between the supraspinatus (ss) and subscapularis (sc) tendons, although the coracohumeral ligament (small arrow) may simulate the biceps tendon in the interval.

tively fiat or slightly ovoid in configuration and normally very low in its signal intensity on proton-density-weighted and T2-weighted images. Biceps tendinosis/tendinitis is seen as abnormal thickening with pronounced ovoid or round cross-section and increased signal intensity on proton-density-weighted images that may decay or persist in signal on associated T2-weighted images, depending on the degree of tendon degeneration (Fig 17). Care must be

taken when evaluating the biceps tendon at the distal rotator cuff interval where it courses through the cuff. In this location, the course of the tendon commonly subtends the "magic angle" of 55 ~ with respect to the direction of the magnetic field of the magnet. 37 Such artifactual increased signal on proton-density-weighted images caused by the magic angle phenomenon decreases in signal on T2-weighted images. Persistence of increased signal intensity on

Fig 17. Marked biceps tendinitis. Marked thickening and increased signal intensity of the biceps tendon (arrows) in the rotator cuff interval on oblique sagittal (A) proton-density and (B) T2-weighted images is consistent with severe intrasubstance tendon degeneration wi thout rupture.


T2-weighted images suggests tendon pathology rather than magic angle effect. When thickening and/or irregularity of the tendon is seen, a diagnosis of biceps tendon pathology can be ensured. Such irregularity, consistent with fraying, as well as longitudinal splitting occasionally can be appreciated (Fig 18). Superior labrum tears or separation can be seen associated with proximal biceps tendon pathology in superior labrum anterior to posterior (SLAP) lesions. 38,39

Positional Abnormalities of the Biceps Tendon

A strong association of subscapularis tendon tears with biceps tendon positional abnormalities has been noted. 35,no Abnormalities range from complete medial dislocation anterior to the glenohumeral joint to incomplete medial subluxation. Cervilla et al described the MRI appearance of six patients with medial dislocation of the biceps tendon that were seen to be of two types.no Five of the dislocations were markedly dislocated to a position anterior to the glenohumeral joint and associated with full- thickness subscapularis tendon tears. One was only mildly dislocated from the groove into degenerated and partially torn subscapularis tendon fibers.

Although complete medial biceps tendon dislocation most commonly is seen associated with full-thickness subscapularis tendon tears (Fig

19), a medially dislocated biceps tendon has been observed with superficial fibers of the subscapularis tendon remaining intact on both MRI and at open surgery (Fig 20). This can be explained with an understanding of the anatomy and relationship of the subscapularis tendon and the transverse humeral ligament. Superfi- cial fibers of the subscapularis tendon contribute to the transverse humeral ligament that attaches to the greater tuberosity at the lateral aspect of the bicipital groove. The medial tendinous wall of the bicipital groove represents intrasubstance and deep-surface fibers of the distal subscapularis that acts as the primary restraint preventing medial displacement of the biceps tendon from the bicipital groove. 33

Incomplete dislocation or subluxation of the biceps tendon may be appreciated. One of the patients in the series of Cervilla et al had incomplete medial dislocation of the tendon into partially disrupted fibers of the subscapularis tendon, n~ This likely reflects an earlier stage of subscapularis tendon degeneration involving its intrasubstance fibers and their contribution to the medial tendinous wall of the bicipital groove. The biceps tendon may then sublux or mildly dislocate into an intrasubstance partial-thickness tear. If entirely intrasubstance, this abnormality may not be appreciated at open or arthroscopic surgery.

IF-

Fig 18. Marked degeneration including longitudinal splitting of the biceps tendon, without complete rupture. One of sequential (A) proton-density and (B) T2-weighted axial images just below the level of the rotator cuff interval showing longitudinal, linear, increased signal intensity (large arrows) within the biceps tendon (small arrows).


Fig 19. Medial dislocation of the biceps tendon from the bicipital groove associated with full-thickness subscapularis tendon tear. (A} Proton-density and (B} T2-weighted axial images show subscapularis tendon tear and retraction (curved arrow} from the greater tuberosity surrounded by fluid signal intensity. The bicipital groove is empty (small arrow} because of medial displa�9 of the biceps tendon (large arrow} to a location anterior to the glenohumeral joint.

MRI OF THE POSTOPERATIVE ROTATOR CUFF

Persistent or recurrent shoulder pain and symptoms after shoulder surgery are common and can be a difficult diagnostic challenge. Differential diagnosis includes recurrent rotator cuff tear, persistent subacromial impingement, deltoid detachment and deficiency, re- sidual or recurrent instability, and nerve injury. 41 Interpretation of MRI of a shoulder after sur-

gery is more difficult than that of a nonoperated shoulder. 42,43 Findings that may be diagnostic of pathology in the unoperated shoulder may be expected after surgery. Knowledge of the type of surgery that has been performed and of what normally is expected as a result of that surgery is essential.

Normal Postoperative Shoulder

Metallic surgical hardware causes a component of signal-void artifact in the immediate surgical vicinity and, if nearby, can obscure the area of interest. Titanium surgical hardware causes minimal artifact because of its nonferro- magnetic nature. Even if no metallic surgical hardware is left behind, it is common to see punctate foci of micrometallic signal-void artifact in the area of surgery that represent micro- particles of metal not visible on plain film radiographs (Fig 21). Because gradient-echo images are more susceptible than spin-echo images to metallic signal-void artifact, an argu- ment can be made to avoid gradient-echo imaging in the postoperative shoulder.

Fig 20. Medial dislocation of the biceps tendon {curved arrow) beneath the subscapularis tendon. T2-weighted axial image shows that the deep layer of the subscapularis tendon is torn from the lesser tuberosity (It}, whereas the superficial layer (small arrows) and its contribution to the transverse humeral ligament remain intact to the greater tuberosity (gt) at the lateral aspect of the bicipital groove (open arrow}.

Anterior Acromioplasty

Neer popularized the clinical impingement syndrome and advocated surgical decompression of the subacromial space with or without rotator cuff repair. 44 In an attempt to decom-

358 FRI'I-I'S AND CRAIG

Fig 21. MicrometaUic artifact and recurrent supraspinatus tendon tear in an individual with a history of open subacromial decompression and rotator cuff repair. (A) Proton-density coronal image shows a large amount of punctate micromatanic artifact involving the subacromial space from anterior acromioplasty (curved arrow). This is seen at the site of the creation o | the trough and suturing of the sulcus (small arrows) of the greater tuberosity (arrowheads), as well as involving the proximally retracted supraspinatus tendon (large arrow), reflecting re-tear of the repair site. (B) Corresponding T2-weighted image better shows the high signal intensity fluid (large arrows) interrupting the supraspinatus tendon and repair. Incidental note is made of a central tendon slip (small arrows) of the overlying deltoid muscle that can falsely mimic an intact superficial layer of the supraspinntus.

press a narrowed subacromial space, the antero- inferior aspect of the anterior acromion is resected in a wedgelike fashion, with the great- est resection anteriorly and tapering posterolat- erally. MRI appearance of postsurgical results can be evaluated best on oblique sagittal images (Fig 22). This resection includes the acromial attachment of the coracoacromial ligament. A common procedure performed in combination with both open and arthroscopic acromioplasty for shoulder decompression is resection of the subacromial bursa. Resection of the peribursal fat layer, as well as fibrosis and scarring, commonly result in the postoperative appearance of obscuration of the subdeltoid fat plane, a normal postoperative finding.

Factors contributing to persistent narrowing and associated persistent impingement include inadequate acromial resection and failure to address inferior hypertrophy of the acromioclavicular joint (Fig 23).

Rotator Cuff Repair

Rotator cuff repair is most commonly performed as open surgery, although, if small, an

Fig 22. Adequate subacromial decompression including anterior acromioplasty. Proton-density oblique sagittal image shows residua of anterior acromioplasty with resultant Bigli- ani type 1 acromion (a) morphology (long arrows) and partial resection of the adjacent portion (*) of the coracoacromial ligament (short arrows). Note small low signal intensity loci of micrometanir artifact (arrowheads), as well as obscuration of the normal subdeltoid fat plane.


Fig 23. Massive recurrent rotator cuff tear with marked inferior acromioclavicular joint hypertrophy that may have contributed to the demise of rotator cuff repair; T2-weighted oblique coronal image shows a large recurrent supraspinatus tear (small arrows) and retracted supraspinatus tendon (large arrow), as well as marked inferior osteophyte and capsular hypertrophy of the acromioclavicular joint (curved arrow) that markedly narrows the subacromial space.

arthroscopically guided mini-open-incision op- eration may be performed. A subacromial- subdeltoid bursectomy is performed, followed by anterior acromioplasty, including resection of a portion of the coracoacromial ligament. If hypertrophic, the distal clavicle and acromioclavicular joint may be resected. Small tears may be sutured in a side-to-side fashion, although larger tears generally require the creation of a bone trough in the humeral sulcus, or more proximal aspect of the head, down to bleeding bone followed by suture fixation of the torn tendon edge into the trough.

The normal postoperative MRI appearance after rotator cuff tear repair includes findings seen after anterior acromioplasty, discussed previously. A shallow defect or trough often is seen in the humeral sulcus between the greater tuberosity and the articular surface of the humeral head (Fig 21). Increased signal intensity is commonly seen on proton-density images of postoperative cuff repair, usually decaying to low or mildly increased signal intensity on T2-weighted images.

Diagnosis of recurrent rotator cuff tear is

more difficult because of the increased signal intensity that can be seen in a normal repair. Owen reported an overall accuracy of 90% in the diagnosis of recurrent full-thickness rotator cuff tears. 4z Full-thickness high signal intensity on T2-weighted images or nonvisualization of a portion of the cuff is diagnostic of full-thickness recurrent tears. However, small foci of high signal intensity may be seen normally, without frank recurrent tear. This makes differentiation of normal postoperative changes from tendinitis or small partial-thickness recurrent tears more difficult. Another potential pitfall can occur in a patient with a very large or massive chronic cuff tear, in which attempted mobilization of the cuff may be inadequate to provide enough length to bring the tendon to the humeral sulcus without undue tension on the repair. In this instance, a trough may be created in a more proximal portion of the humeral head rather than in the humeral sulcus, thereby allowing function but minimizing stress on the repair site. At postoperative MRI, this can result in the absence of visualization of cuff tissue more distally at the greater tuberosity and suleus, thereby falsely simulating a full-thickness rotator cuff tear.

Fig 24. Normal glenoid labrum. Proton-density axial image shows uniformly low signal intensity of the anterior and posterior glenoid labrum (long arrows), Note the linear increased signal intensity on proton-density images at the base of the labrum that represents the normal hyaline articular cartilage of the glenoid to which the fibrous labrum attaches (short arrows). This is not to be mistaken for a labrum tear,


Fig 25. Acute posterior labrum tear and avulsion of the posterior capsule off the humerus. This 18-year-old man had a shoulder injury 3 days previously. Proton-density axial image shows posterior labrum tear (large arrows), as well as posterior capsule detachment from the humerus (curved arrow) and abnormal undulation (small arrows).

Fluidlike signal intensity in the subacromial- subdeltoid bursa on T2-weighted images is even less specific than in the unoperated shoulder. Not only can this be present as bursal inflamma-

tion with an intact cuff, but cuff repair may not produce a watertight seal of the cuff. This could result in bursal fluid collection that may be difficult to differentiate from a pathological recurrent subacromial-subdeltoid bursitis or he- matoma/seroma.

MRI can detect and delineate the extent of deltoid deficiency, a clinical condition consisting of deltoid detachment from the acromion that can occasionally result from a postoperative injury or inadequate reattachment of the deltoid to the acromion after open surgery. Deltoid deficiency is a severe complication that can lead to a flail shoulder and chronic pain that may require glenohumeral arthrodesis. 41

CAPSULOLABRAL STRUCTURES AND PATHOLOGY

Glenoid Labrum Tears

MRI has also proved to be popular for visualization of the soft tissues of the capsulolabral complex including labral anatomy and labrum tears (Fig 24). The sensitivity for detection of labrum tears ranges from 44% 2 to 95%. 3 How- ever, pitfalls in interpretation of MRIs outnum- ber those of C-F arthrography, resulting in greater interobserver and intraobserver variability.

Fig 26. A 22-year-old man with a recent shoulder injury and MRI findings consistent with residua of anterior shoulder dislocation injury. (A) T2-weighted axial image shows small and shallow Hill-Sachs' deformity with underlying marrow edema/hemorrhage (arrows). (B) T2-weighted axial image through the level of the inferior glenoid shows a Bankart lesion with displaced anterior labrum tear (curved arrow), associated with injury and thickening of the inferior glenohumeral ligament (small arrows), and mild medial capsular stripping with dissection of extracapsular fluid between the subscapularis and bony scapula (arrowheads). The middle glenohumeral ligament is seen as a focal thickening of the anterior capsule (long arrow).


Fig 27. A 25,year-old throwing athlete with persistent shoulder pain especially with throwing motion and a clinical picture consistent with impingement syndrome. (A) Proton-density oblique coronal image shows linear increased signal intensity (arrows) involving the superior glenoid labrum. The superior la- bruin tear is seen to better advantage on axial (B and D) proton,dens'~y and correspond, ing {C and E) T2-waighted ira- ages as linear loci of increased signal intensity (arrows) through the normally low signal intensity superior labrum. A SLAP lesion was documented at arthroscapic surgery,

MR arthrography has been described as useful in the detection of labral abnormalities. Flannigan et al correctly diagnosed nine labral tears in 23 patients, whereas only three of the nine were diagnosed by conventional MRI. 45

i 'ii~!i ,i ~i~ ~!~i ~ ~

Chadnani et al compared conventional MRI, MR arthrography, and CT arthrography with surgical findings in 28 patients with a mean age of 27 years. They found that MR arthrography is more sensitive than CT arthrography or conven-

362 FRITI-S AND CRAIG

Fig 28. Superior glenoid labrum tear consistent with SLAP lesion, Coronal Tl-weighted fat saturation image after arthrography using dilute gadolinium solution shows high signal contrast in the axillary recess (curved arrow} as well as extending into a superior labrum tear (straight arrow).

tional MRI in the detection of abnormalities of the labrum. 46 Labrum tears were found at surgery in 26 of 28 patients. Conventional MRI, MR arthrography, and CT arthrography found 93%, 96%, and 73% of these tears, respectively. A surgically proven detached labral fragment was found in 46%, 96%, and 52%, respectively. Labral degeneration consisting of flaying and/or attenuation was found in 11%, 56%, and 24%, respectively. Capsular injuries can occasionally be shown, especially in acute injuries (Fig 25).

Hill-Sachs' deformity represents a compres- sion fracture of the posterosuperior aspect of the humeral head caused by impaction of the head into the inferior glenoid at the time of dislocation. Small lesions can be detected with greater sensitivity and specificity on axial CT and MRI (Fig 26). 47 Medial capsular stripping can be appreciated on MRI after acute or subacute anterior dislocations. Chronic thickening of the medial capsule can be seen as residua of a more chronic anterior dislocation injury.

Superior Labrurn Tears (SLAP Lesions)

In 1990, Snyder et al described the SLAP lesion, an acronym that stands for superior labrum, anterior to posterior, including the origin of the biceps tendon. 38 This injury may occur after acute trauma sustained during a fall

on an outstretched arm, traction of the biceps tendon by a sudden traumatic pull, or with repetitive stress in throwing athletes. Hurley and Anderson noted an association between superior labrum tears and the impingement syndrome in a population of athletes and raised the possibility of secondary impingement caused by instability. 48

Although superior labrum tears can be detected at CT arthrography, 49 the ability of MRI to obtain direct coronal images, which are optimal for evaluation of the superior labrum and biceps tendon anchor, allows better evaluation of these lesions (Fig 27). The addition of intra-articular saline or dilute gadolinium solution adds to the accuracy of MRI by providing an intra-articular contrast material that, similar to CT arthrography, is absorbed into the abnormal labrum and labrum tears (Fig 28). MR arthrography has now been established as the imaging modality of choice for detection of these lesions. 46,5~

OSSEOUS CHANGES

Burk et al, 53 in an MRI study of 10 symptomatic professional baseball players and one asymp-

Fig 29. Nondisplaced radiographicany occult greater tuberosity fracture. This 32-year-old woman fell onto an abducted shoulder when descending a hill while cross-country skiing 3 weeks previously. Initial emergency department radiographs were negative. Persistence of pain and weakness prompted MRI examination to rule out rotator cuff tear. A proton-density coronal image shows a nondisplaced fracture line (large arrows} and surrounding marrow edema/hemorrhage (small arrows}, as well as an intact and normal rotator cuff, Pain and weakness gradually resolved over the course of 2 months.

MR1 OF THE SHOULDER 363

Fig 30. Large posterosuperior ganglion cyst extending with impingement on the infraspinatus branch of the suprascapular nerve in the spinoglenoid notch associated with denervation signal change in the infraspinatus muscle. (A) T2-weighted coronal image shows ganglion cyst (large arrows) in the spinoglenoid notch. (B) Proton.density and (C) T2.weighted axial images show associated slender neck (small arrow) extending through the posterior component of superior labrum tear (curved arrow). (D) T2-weighted sagittal image shows generalized increased signal intensity of the infraspinatus [is) muscle belly (arrows) relative to supraspinatus (ss) and teres minor (tm) muscles reflecting denervation signal alteration.

tomatic player, described cortical irregularity and/or subchondral cyst formation at the posterior aspect of the greater tuberosity near the insertion of the infraspinatus tendon in five of seven players with rotator cuff tears. These findings were believed to represent chronic avulsion changes resulting from the decelera- tion stresses of the follow-through motion. How- ever, similar findings were noted in the asymptomatic volunteer and one of the three players

without cuff tear. It is our experience that such findings of signal alteration of subcortical marrow at the posterior aspect of the humeral sulcus are common, increasing in frequency with age. This finding is not necessarily associated with cuff pathology nor is it confined to throwing athletes. This is the location of the anatomic humeral bare area, an intra-articular portion of the humerus that is not covered by articular cartilage. Lack of protective articular

364 FRI3-1"S AND CRAIG

cartilage is expected to lead to increased incidence of erosions and subchondral cyst formation in this region. These erosions have been documented at arthroscopy without being inti- mately related to the infraspinatus insertion.

Occult osseous injury of the greater tuberosity, including bone contusion and radiographi- caUy occult fracture of the greater tuberosity, is a cause of persistent shoulder pain and weakness after trauma to the shoulder (Fig 29). This may occur from a fall directly onto the shoulder or on an outstretched arm. This finding is most commonly seen as an incidental finding on MRI examination that was performed to rule out rotator cuff tear. MRI rules out rotator cuff tear and establishes the diagnosis that may be expected to resolve without sequelae with further conservative management.

Overt greater tuberosity fractures also can be evaluated preoperatively to determine the degree of fracture displacement or elevation, as well as to detect any associated rotator cuff tears.

GANGLION CYSTS

MRI is the modality of choice for evaluation of clinical suspicion for suprascapular neuropathy. 54,55 Differential diagnosis includes posterosuperior ganglion cysts that, if large enough, can encroach on branches of the suprascapular nerve. If extending into the supraspinatus fossa and/or suprascapular notch, encroachment on the suprascapular nerve can cause innervation of both the supraspinatus and the infraspinatus. Extension into the spinoglenoid notch can encroach on the branch to the infraspinatus muscle resulting in denervation with isolated infraspinatus atrophy and weakness (Fig 30). In a recent study, 5614 patients with MRI findings of posterosuperior ganglion cysts were identified. All except one were found to have a labrum tear, usually a SLAP lesion, at arthoscopy. That one patient had a primary excision of the cyst. Arthroscopy performed 2 years later for recurrent pain and ganglion cyst showed a very small SLAP lesion that, retrospectively, is believed to have been missed at initial arthroscopy.

REFERENCES 1. Fritts H: MRI and shoulder imaging. Lippincott's

reviews. Radiology 1:373-405, 1992 2. Garneau RA, Renfrew DL, Moore TE, et al: Glenoid

labrum: Evaluation with MR imaging. Radiology 179:519- 522, 1991

3. Legan JM, Burkhard TK, Goff WB, et al: Tears of the glenoid labrum: MR imaging of 88 arthroscopically confirmed cases. Radiology 179:241-246, 1991

4. Legan JM, Burhard JK, GoffWB II, et ai: Tears of the glenoid labrum: MR imaging of 88 arthroscopicaily confirmed cases. Radiology 179:241-246, 1991

5. Habibian A, Stauffer A, Resnick D, et al: Comparison of conventional and computed arthrotomography with MR imaging of the shoulder. J Comput Axial Tomogr 13:968- 975, 1989

6. Iannotti JP, Zlatkin MB, Esterhai JL, et al: Magnetic resonance imaging of the shoulder. Sensitivity, specificity and predictive value. J Bone Joint Surg 73A:17-29, 1991

7. McMasters PE: Tendon and muscle ruptures: Clinical and experimental studies on the causes and location of subcutaneous ruptures. J Bone Joint Surg 15A:705-722, 1933

8. Codman EA: The Shoulder. Boston, MA, Thomas Todd, 1934

9. Meyer AW: The minute anatomy of attritional lesions. J Bone Joint Surg 13A:341-360, 1931

10. Neer CS II: Impingement lesions. Clin Orthop 173:70- 77, 1983

11. Cofield RH: Current concepts review rotator cuff disease of the shoulder. J Bone Joint Surg 67A:674-679, 1985

12. Nevaiser JS: Ruptures of the rotator cuff. Clin Or- thop 3:92-98, 1954

13. Craven WM: Tranmatie avulsion tears of the rotator cuff, in Andrews JR, Wilk KE (eds): The Athlete's Shoul- der. New York, N-Y, Churchill Livingston, 1994, pp 129-140

14. Bigliani LU, Morrison D, April EW: The morphology of the aeromion and its relationship to rotator cuff tears. Orthop Trans 10:228, 1986 (abstr)

15. Epstein RE, Schweitzer ME, Frieman BG, et al: Hooked acromion: Prevalence on MR images of painful shoulders. Radiology 187:479-481, 1993

16. Seeger LL, Gold RH, Bassett LW, et al: Magnetic resonance imaging of glenohumeral joint disease. Invest Radio123:650-661, 1988

17. Crues JV, Fareed DO: Magnetic resonance imaging of shoulder impingement. Top Magn Reson Imag 3:39-54, 1991

18. Bigliani LU, Norris, Fischer J, et al: The relationship between the unfused acromial epiphysis and subacromial impingement lesions. Orthop Trans 7:138, 1983 (abstr)

19. Mudge MK, Wood VE, Frykman GK: Rotator cuff tears associated with os acromiale. J Bone Joint Surg 66A:427-429,1984

20. Kaplan PA, Bryans KC, Davick JP, et al: MR imaging of the normal shoulder: Variants and pitfalls. Radiology 184:519-524, 1992

21. Neumann CH, Holt RG, Steinbach LS: MR imaging of the shoulder: Appearance of the supraspinatus tendon in asymptomatic volunteers. AIR Am J Rocntgenol 158:1281- 1287, 1992

22. Mirowitz SA: Normal rotator cuff: MR imaging with


conventional and fat-suppression techniques. Radiology 180:735-740, 1991

23. Zlatkin MB, Ianotti JP, Roberts MC, et al: Rotator cuff tears: Diagnostic performance of MR imaging. Radiol- ogy 172:223-229, 1989

24. Burk KL, Karasick D, Kurtz AB, et al: Rotator cuff tears: Prospective comparison of MR imaging with arthrography, sonography, and surgery. AJR Am J Roentgenol 153:87-92, 1989

25. Rafii M, Firooznia H, Sherman O, et al: Rotator cuff lesions: Signal patterns at MR imaging. Radiology t77:817- 823, 1990

26. Seeger LL: Magnetic resonance imaging of the shoulder. Clin Orthop 244:48-62, 1989

27. Wright RW, Fritts HM, Tierney GS, et al: Impinge- ment test: The effect on the timing and results of shoulder MRI. Submitted to American Shoulder and Elbow Sur- geons for the l l th Annual Open Meeting, Orlando, FL, February 1995

28. Mitchell MJ, Causey G, Berthoty DP: Peribursal fat plane of the shoulder: Anatomic study and clinical experience. Radiology 168:699-704, 1988

29. Rathbun JB, Macnab I: The microvascular pattern of the rotator cuff. J Bone Joint Surg 52B:540-553, 1970

30. Lohr JF, Uhthoff HK: The microvascular pattern of the supraspinatus tendon. Clin Orthop 254:29-34, 1990

31. Kjellin I, Ho CP, Cervilla V: Alterations in the supraspinatus tendon at MR imaging: Correlation with histopathologic findings in cadavers. Radiology 181:837- 841, 1991

32. Traughber PD, Goodwin TE: Shoulder MRI: Arthro- scopic correlation with emphasis on partial tears, J Comput Assist Tomogr 16:129-133, 1992

33. Nakajima T, Rokuuma N, Hamada K, et al: Histo- logic and biomechanical characteristics of the supraspinatus tendon: Reference to rotator cuff tearing. J Shoulder Elbow Soc 3:79-87, 1994

34. McAuliffe TB, Dowd GS: Avulsion of the subscapularis tendon: A case report. J Bone Joint Surg 69A:1454- 1455, 1987

35. Patten RM: Tears of the anterior portion of the rotator cuff (the subscapularis tendon): MR imaging findings. AJR Am J Roentgenol 162:351-354, 1994

36. Linker CS, Helms CA, Fritz RC: Quadrilateral space syndrome: Findings at MR imaging. Radiology 188:675-676, 1993

37. Erickson S J, Fitzgerald SW, Quinn SF, et al: Long bicipital tendon of the shoulder: Normal anatomy and pathologic findings on MR imaging. AJR Am J Roentgenol 158:1091-1096, 1992

38. Snyder S J, Karzel RP, Del Pizzo, et al: SLAP lesions of the shoulder. Arthroscopy 6:274-279, 1990

39. Cartland JP, Crues JV, Stauffer A: MR imaging in the evaluation of SLAP injuries of the shoulder: Findings in 10 patients. AJR Am J Roentgenol 159:787-792, 1992

40. Cervilla V, Schweitzer ME, Ho C: Medial dislocation of the biceps brachii tendon: Appearance at MR imaging. Radiology 180:523-526, 1991

41. Post M: Complications of rotator cuff surgery. Clin Orthop 254:97-104, 1990

42. Owen RS, Iannotti JP, Kneeland JB, et al: Shoulder after surgery: MR imaging with surgical validation. Radiol- ogy 186:443-447, 1993

43. Haygood TM, Oxner KG, Kneeland JB, et al: Mag- netic resonance imaging of the postoperative shoulder, Magn Res Imag Clin North Am 1:143-155, 1993

44. Neer CS II: Anterior acromioplasty for the chronic impingement syndrome in the shoulder: A preliminary report. J Bone Joint Surg 54A:41-50, 1972

45. Flannigan B, Kursunoglu-Brahme S, Snyder S, et al: MR arthrography of the shoulder: Comparison with conventional MR imaging. AIR Am J Roentgenol 155:829-832, 1990

46. Chadnani VP, Yeager TD, DeBerardino T, et al: Glenoid labrum tears: Prospective evaluation with MR imaging. A JR Am J Roentgenol 161:1229-1235, 1993

47. Workman TL, Burkhard TK, Resnick D, et al: Hill-Sachs lesion: Comparison of detection with MR imaging, radiography, and arthroscopy. Radiology 185:847-852, 1992

48. Hurley JA, Anderson TE: Shoulder arthroscopy: Its role in evaluating shoulder disorders in the athlete. Am J Sports Med 18:480-483, 1990

49. Hunter JC, Blatz DJ, Escobedo EM: SLAP lesions of the glenoid labrum: CT arthrographic and arthroscopic correlation. Radiology 184:513-518, 1992

50. Karzel RP, Snyder SJ: Magnetic resonance arthrography of the shoulder: A new technique of shoulder imaging. Clin Sports Med 12:123o136, 1993

51. Palmer WE, Brown JH, Rosenthal DI: Labral- ligamentous complex of the shoulder: Evaluation with MR arthrography. Radiology 190:645-651, 1994

52. Tirman PF, Applegate GR, Flannigan BD, et al: Magnetic resonance arthrography of the shoulder. Magn Res lmag Clin North Am 1:125-142, 1993

53. Burk DL, Torres JL, Marone P J, et al: MR imaging of shoulder injuries in professional baseball players. J Magn Reson Imaging 1:385-389, 1991

54. Zeiss J, Woldenberg LS, Saddemi SR, et al: Case report: MRI of suprascapular neuropathy in a weight lifter. J Comput Assist Tomogr 17:303-308, 1993

55. Fritz RC, Helms CA, Steinbach LS, et al: Suprascapu- lar nerve entrapment: Evaluation with MR imaging. Radiol- ogy 182:437-444, 1992

56. Moore TP, Fritts HM, Buss DD: Suprascapular nerve entrapment secondary to supraglenoid cyst compres- sion. Presented at the American Academy of Orthopaedic Surgery Annual Meeting, San Francisco, CA, February 1993. J Bone Joint Surg (submitted)

mri of the shoulder

Documents