A printed version of this article appears in the October 2001 Supplement to Diagnostic Imaging.

Open MR Arthrography of the Shoulder

Daniel Bodor, MD

Introduction

While MRI is the most advanced imaging technique for evaluation of the shoulder, MR arthrography is a further refinement that yields additional detail of intraarticular structures.  When indicated, it offers better detection and assessment of glenoid labral tears, small loose bodies, or cartilage flaps.  It also allows better assessment of the undersurface of the rotator cuff, a frequent site of fibrillation and partial-thickness tears.  The exam can be useful for clarification of inconclusive findings seen on a standard MRI study, or for identification of hidden pathology if a standard MR has not fully explained the patient’s clinical symptoms (Fig 1).  Shoulder MR arthrography can readily be done on open MR units without major limitations, and with a few added advantages.

Imaging

We inject the glenohumeral joint with a mixture of dilute gadolinium contrast (1:200), iodinated contrast and saline (50:50) and perform a brief standard arthrogram prior to transferring the patient to the MRI suite.  Routine MR sequences are performed with external rotation and inferior traction in order to place tension on the biceps superior labral anchor and more optimally assess for SLAP (superior labral anterior-posterior) tears.  As with all of our musculoskeletal MR exams, we try to maximize image resolution by decreasing the FOV as much as is feasible while maintaining acceptable signal to noise.  Our technologists also keep in mind that although some patients may appear large due to abundant subcutaneous fat, their joints are normal in size and the temptation to use a larger FOV must be resisted.

The initial sequences are T1 SE axial and coronal, T2 FSE coronal and oblique sagittal, STIR coronal.  The T1 sequences provide high resolution and high contrast.  The T2 sequences help us to differentiate fat from gadolinium-enhanced joint fluid, allow detection of fluid collections that do not communicate with the joint space, and provide greater contrast between the joint fluid and hyaline cartilage.  The STIR sequence adds greater sensitivity for detection of edema in bones and muscles.  Additional sequences may be performed to address specific needs, e.g. oblique coronal T1 parallel to the intraarticular portion of the biceps long head tendon can be helpful if pathology is suspected at this site.

On all of our shoulder MR arthrograms we perform an additional T1 SE oblique axial sequence in abduction and external rotation (ABER position) perpendicular to the glenoid fossa.  This is easily performed by using a coronal localizer and orienting the scan parallel to the humeral shaft.  The total MR exam time is about 45 min.

Interpretation

The ABER position has been advocated for the assessment of the anteroinferior capsuloligamentous structures, which are a frequent site of pathology and are placed in tension with this position.  Imaging with ABER positioning can usually be performed in any MR unit with small patients, but large patients may be unable to be imaged without the extra space of an open MR system.  The position helps in detecting anterior inferior labral tears (Bankart lesions and variants) and differentiating old healed injuries that have residual signal abnormality from detached and mobile ones that allow influx of fluid into the gap (Fig 2).  The apprehension test position (90^o abduction and full external rotation) is advocated by some as even better than the ABER position in producing tension on the inferior glenohumeral ligament and anterior labrum, thereby increasing sensitivity to lesions at this site [1,2] .  This position can only be achieved in open MR systems.

Since partial-thickness rotator cuff tears are more frequent on the articular than the bursal side, a great added benefit of MR arthrography with shoulder abduction is that it allows a better assessment of the cuff undersurface, an area of limited visibility on standard MR exams.  With the shoulder in the usual adducted imaging position, the cuff is closely apposed to the humeral head, but in abduction it is lifted away, allowing better visualization, particularly when there is fluid distension of the joint.

Pathology secondary to internal impingement (posterosuperior glenoid rim impingement) [3] is much more common than previously thought.  This is supported by the fact that undersurface partial tears are more common than those at the bursal surface.  Pathology from this process is particularly frequent in patients with repetitive overhead arm motions such as baseball players.  When the shoulder is fully abducted and externally rotated, as in the cocking phase of throwing, the undersurface of the cuff comes into contact with the posterior superior glenoid rim.  Repetitive microtrauma results in pathology of the labrum and the cuff, starting with fibrillation and fraying, and progressing to tears.  Chronic bone erosions and cysts are also common.  Typically cuff pathology is posterior, involving the infraspinatus or the posterior portion of the supraspinatus.  Tears begin at the undersurface of the cuff attachment to the humerus and are known as rim-rent tears [4] .  As the tears progress to about half of the cuff, the cuff can delaminate and split longitudinally such that a flap develops.  This flap is typically not detectable on MR unless the arm is placed and imaged in ABER [5] .  In this position the flap of cuff displaces medially and the occurrence becomes apparent (Fig 3).

Glenoid labral cysts are very important to detect on any MR exam of the shoulder since they are nearly always associated with labral tears [6] .  The T2-weighted sequence is important in order to detect these as they will often not fill with gadolinium-enhanced fluid from the joint and will either be entirely missed (Fig 4) or seen as low signal structures on the T1-weighted images potentially confused with soft tissue masses, collagenous structures, or calcification.  Some cysts will be only be partially filled with gadolinium-enhanced fluid, due to presence of septation or contrast dilution with preexisting cyst fluid.  Careful attention must be paid to the labrum, particularly in the vicinity of a cyst.  If no tear is visible, clinical suspicion must still remain high.

Special Considerations

Because of the complicated and variable anatomy, imaging evaluation of the capsular-labral-ligamentous complex of the shoulder remains one of the most difficult areas of MRI interpretation for radiologists no matter which MR system is used.  To optimally assess this area, one must have a thorough knowledge of the anatomy and its variants, as well as its patterns of injury.  These have been described in several well-written articles [7-11] .  For each exam, the interpreting physician must also be able to reconstruct a three dimensional mental picture of this complex area from the two dimensional images.  Special considerations for interpretation of studies on open MR units include altered tissue signal intensities, lack of frequency-selective fat suppression, and altered locations of magic angle artifacts.

In low field systems, T1 relaxation is faster and contributes to some changes in tissue signal intensities, including higher contrast.  Underlying physics aside, tissue signal alterations that we have encountered include somewhat higher T1 signal intensities of hyaline cartilage and lower T1 signal in fibrocartilage and unenhanced simple fluid.  These factors make correlation of corresponding T1- and T2-weighted images even more important when interpreting MR studies on low field units.  While T1-weighted images have higher signal-to-noise and higher resolution with excellent contrast between the fibrocartilaginous labrum and the bright intraarticular fluid, the contrast difference is more subtle for hyaline cartilage.  Correlation with T2-weighted images offers higher contrast between bright gadolinium-enhanced fluid and dark hyaline cartilage with increased conspicuity of some lesions such as flaps or hyaline cartilage loose bodies.

Since we cannot perform frequency-selective fat suppression with low field MRI units as we do in our high field units when we image MR arthrograms, we must be more attentive at the time of image interpretation.  Fat suppression has generally been advocated for MR arthrography in order to be able to detect high T1 signal gadolinium-enhanced fluid in the subacromial-subdeltoid bursa that has escaped through a full-thickness rotator cuff tear.  We have not found this to be a significant problem, particularly since the detection or exclusion of a full thickness tear is already made at the time of the standard arthrogram.  The true benefit of MR arthrography over regular MR is not so much in the detection of full thickness rotator cuff tears, but rather in the assessment of intraarticular pathology such as undersurface partial thickness rotator cuff tears and fibrillation, as well as pathology of the glenoid labral-ligamentous complex, that are best assessed with joint distension.  In fact, if a full thickness tear of the cuff is present, this can potentially limit the advantages of MR arthrography over standard MR for detection of intraarticular pathology since optimal joint distention may not be achieved.

It appears that an additional advantage of imaging in an open MR system is the lack of magic angle artifact in the distal rotator cuff that is nearly always present on low TE sequences (T1-, proton density-, or T2*-weighted) when patients are imaged in standard MR units.  With a longitudinal main magnetic field, artifactually increased signal occurs where the cuff collagen is oriented 55o relative to the craniocaudal axis.  Since the magnetic field is vertically oriented in open MR units, magic angle effects occur at different anatomic locations.  They occur at 55o relative to the anterior-posterior direction rather than relative to craniocaudal.  Because there is no portion of the cuff with this collagen fiber orientation, images obtained in open MR units are spared of this potential interpretation pitfall.  However, magic angle artifact remains a potential cause of diffusely increased signal in the fibrocartilage of the glenoid labrum with either field orientation, similar to that found when imaging knee menisci.  It must also be recognized when assessing articular hyaline cartilage, as we see on knee exams where differing signal intensities are present at the medial versus lateral facet cartilages of the patella.

Patient Selection

MR arthrography is not necessary for all patients.  Most patients are referred for MRI evaluation of suspected rotator cuff tears; for this, standard MRI is excellent.  It allows accurate assessment of the size and location of tears, degree of tendon retraction, as well as detection of muscle edema or atrophy.  Other unsuspected causes of shoulder pain are readily detected and occasionally identified, such as occult fractures of the proximal humerus or distal clavicle, avascular necrosis, or calcific tendinitis and bursitis.

MR arthrography is indicated if labral pathology is suspected such as in a patient with shoulder instability, clicking, or catching.  It can also be beneficial when there is an inconclusive finding on standard MRI that needs further clarification.  Clinical symptoms and signs not accounted for by findings on standard MR can be further investigated by MR arthrograms to detect otherwise hidden abnormalities such as small labral or chondral flaps, loose bodies, and undersurface partial rotator cuff tears.

Conclusion

MR arthrography is a valuable technique for optimizing assessment of internal shoulder conditions, particularly those involving the glenoid labroligamentous complex and the undersurface of the rotator cuff, and can be performed well in open MR systems.

References

1.  Wintzell G, Haglund-Akerlind Y, Larsson H, Zyto K, Larsson S.  Open MR imaging of the unstable shoulder in the apprehension test position: description and evaluation of an alternative MR examination position.  Eur Radiol. 1999;9(9):1789-95.

2.  Wintzell G, Larsson H, Larsson S.  Indirect MR arthrography of anterior shoulder instability in the ABER and the apprehension test positions: a prospective comparative study of two different shoulder positions during MRI using intravenous gadodiamide contrast for enhancement of the joint fluid.  Skeletal Radiol. 1998 Sep;27(9):488-94.

3.  Tirman PF, Bost FW, Garvin GJ, Peterfy CG, Mall JC, Steinbach LS, Feller JF, Crues JV 3rd.  Posterosuperior glenoid impingement of the shoulder: findings at MR imaging and MR arthrography with arthroscopic correlation.  Radiology 1994 Nov;193(2):431-6.

4.  Cameron J. Seibold, Thomas A. Mallisee, Scott J. Erickson, Melbourne D. Boynton, William G. Raasch, and Michael E. Timins.  Rotator Cuff: Evaluation with US and MR Imaging.  RadioGraphics 1999;19:685-705.

5.  Tirman PF, Bost FW, Steinbach LS, Mall JC, Peterfy CG, Sampson TG, Sheehan WE, Forbes JR, Genant HK.  MR arthrographic depiction of tears of the rotator cuff: benefit of abduction and external rotation of the arm. Radiology 1994 Sep;192(3):851-6.

6.  Tirman PF, Feller JF, Janzen DL, Peterfy CG, Bergman AG.  Association of glenoid labral cysts with labral tears and glenohumeral instability: radiologic findings and clinical significance.  Radiology 1994 Mar;190(3):653-8.

7.  Beltran J, Bencardino J, Mellado J, Rosenberg ZS, and Irish RD.  MR arthrography of the shoulder: variants and pitfalls.  RadioGraphics 1997 17: 1403-1412.

8.  Park YH, Lee JY, Moon SH, Mo JH, Yang BK, Hahn SH, Resnick D. MR Arthrography of the Labral Capsular Ligamentous Complex in the Shoulder: Imaging Variations and Pitfalls. AJR 2000; 175: 667-672.

9.  Stoller DW.  MR arthrography of the glenohumeral joint. Radiol Clin North Am 1997 Jan;35(1):97-116.

10.  Tirman PF, Palmer WE, Feller JF.  MR arthrography of the shoulder.  Magn Reson Imaging Clin N Am. 1997 Nov;5(4):811-39.

11.  Schweitzer ME.  MR arthrography of the labral-ligamentous complex of the shoulder.  Radiology. 1994 Mar;190(3):641-4.

[Home | Abdominal Aortic Aneurysm Repair | Coronary Artery Screening | MR Arthrography | Percutaneous Radiofrequency Ablation | Vertebroplasty]

SDI Radiologists
4516 N. Armenia Avenue
Tampa, FL 33603