Conventional radiography is and should be the initial imaging examination performed for a patient presenting with shoulder pain. Although radiographs provide limited evaluation of the rotator cuff and glenoid labrum, they can offer important information about the source of the patient’s symptoms. Radiographs depict an assortment of osseous abnormalities, including fracture, arthritis, postsurgical changes, and tumor, and they are frequently complementary to more advanced imaging modalities, such as computed tomography (CT) and magnetic resonance imaging (MRI). A wide variety of radiographic views are available to aid in the evaluation of the glenohumeral joint. Knowledge about the advantages and disadvantages of each view will assist in optimizing imaging protocols, depending on the clinical presentation ( Box 44-1 ).
* The views expressed in this publication are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of Defense, or the U.S. Government.
Glenohumeral “true” anteroposterior (Grashey) view
Axillary lateral view
Scapular Y view
Stryker notch view
Acromioclavicular anteroposterior and posteroanterior views
The anteroposterior (AP) view ( Fig. 44-1, A ) is obtained with the patient in either the upright or supine position. The beam is directed in a true AP direction relative to the body. The glenoid rim is normally tilted anteriorly about 40 degrees, which results in overlap of the humeral head and glenoid rim in the AP view. This view can be obtained with the humeral head in neutral position, internal rotation, or external rotation. When compared with other radiographic views of the shoulder, the AP view provides the best overview of the osseous structures of the shoulder girdle because the projection allows for relatively uniform distribution of soft tissue density across the entire shoulder. As a result, one or more of the AP views is nearly always included in the standard radiographic evaluation of the shoulder. These projections allow for adequate evaluation of the humeral head, glenoid, and body of the scapula, as well as the acromioclavicular (AC) joint and coracoid process. The AP projection is very helpful in the evaluation after acute trauma for evidence of fracture or dislocation and is also of value in assessing the cause of chronic pain from arthritis, impingement, calcific bursitis, tumor, or infection.
Glenohumeral “True” AP (Grashey) View
The true AP view, or Grashey view ( Fig. 44-1, B ), differs from the standard AP view in that the patient is rotated posteriorly 35 to 40 degrees, thus providing a tangential view of the glenohumeral joint. The advantage of the Grashey view is that it provides a superior evaluation of the glenohumeral joint. This view can demonstrate subtle subluxation in the superior or inferior direction and also will show subtle joint space narrowing associated with glenohumeral arthritis. The disadvantage of this view is that a rapid change in soft tissue density occurs from medial to lateral, and as a result, the lateral aspect of the shoulder, including the acromion and AC joint, is difficult to evaluate because of a rapid change in density on the radiograph and loss of anatomic detail laterally. This loss can be decreased with use of a boomerang-shaped filter draped along the lateral shoulder.
Axillary Lateral View
Variations of the axillary lateral view exist, but the projection is most commonly obtained with the patient supine and with the arm abducted 90 degrees ( Fig. 44-1, C ). The beam is then directed from distal to proximal while tilted 15 to 30 degrees toward the spine. This projection is best suited for evaluation of subtle anterior or posterior subluxation or dislocation and can also help detect subtle Bankart fractures of the anterior glenoid rim. The osseous detail on this projection, however, is often quite poor, and its main benefit is in identifying anterior or posterior dislocation. Numerous variations of this projection have been developed, some with the goal of decreasing movement of the arm in the setting of acute trauma, and others with the intention of accentuating certain anatomic features. The West Point view is a variation of the axillary view developed to optimize visualization of an osseous Bankart lesion of the anterior glenoid rim. It is obtained by placing the patient in the prone position on the x-ray table with the arm abducted 90 degrees from the long axis of the body and with the forearm draped over the edge of the top of the table. The beam is directed 15 to 20 degrees in an inferior-to-superior direction and tilted 25 degrees toward the spine. This projection improves the detection of osseous Bankart lesions but is very difficult to obtain in the setting of acute trauma and is best reserved for the patient in the setting of subacute or chronic instability.
Scapular Y View
The scapular Y view is easily obtained in the setting of acute trauma because it can be obtained with the arm immobilized by the side, and little or no movement of the arm is required ( Fig. 44-1, D ). This view can be very helpful in the setting of acute trauma to evaluate for anterior or posterior dislocation. The projection is obtained with the patient upright or prone and rotated approximately 30 to 45 degrees toward the cassette. The beam is directed down the body of the scapula and results in a projection in which the body of the scapula is seen in tangent and the glenoid fossa en face as a Y-shaped intersection of the scapular body, coracoid process, and acromion process. It is also the projection commonly used when “typing” the undersurface of the acromion.
Stryker Notch View
The Stryker notch view is best suited for viewing the posterolateral aspect of the humeral head and is an excellent radiographic view for detecting flattening of the posterior humeral head or a Hill-Sachs deformity ( Fig. 44-1, E ). It is very limited, however, in its evaluation of the glenoid rim for osseous Bankart lesions. The Stryker notch view is obtained with the patient in either the upright or supine position. The arm is positioned vertically overhead with the elbow flexed and the hand supported on the back of the head. The beam is directed toward the mid axilla and is tilted about 10 degrees cephalad.
Acromioclavicular Articulation Views
The AC joints are best evaluated with the patient either sitting or standing with the back flat against the cassette. The arms are usually freely hanging by the sides of the body. The patient is sometimes asked to hold weights or sandbags to accentuate AC joint separation, but the value of this measure is questioned. The beam is directed toward the midline of the body and centered over the AC joints. This view is useful in evaluating for pathologic features of the AC joint, including fracture, AC joint separation, arthritis, and osteolysis of the distal clavicle. Comparison with the contralateral AC joint can be helpful in detecting subtle abnormalities.
In the past, conventional arthrography of the shoulder ( Fig. 44-2 ) was considered the gold standard for diagnosis of tears of the rotator cuff. Partial-thickness articular-sided rotator cuff tears, as well as adhesive capsulitis, are also accurately depicted with use of this method. However, partial-thickness rotator cuff tears involving the bursal surface and intrasubstance tears are not demonstrated by conventional shoulder arthrography, and its usefulness in the evaluation of glenohumeral instability is limited. This procedure is no longer widely performed.
Conventional arthrography was replaced by early CT with arthrography (CTA) performed on single-slice scanners, and films were printed in several viewing windows for review. This technique resulted in an effective tool for assessing labral injuries associated with instability, with a similar performance for cuff tears. However, CTA was subsequently replaced by MRI and magnetic resonance arthrography (MRA).
MRI largely replaced CTA as the advanced imaging modality for evaluation of the shoulder, and MRA has become the modality of choice for the evaluation of glenohumeral instability. The primary factors contributing to the shift from conventional arthrography and CTA to MRI are the superb soft tissue contrast, multiplanar capabilities provided by MRI, and lack of exposure to ionizing radiation. MRI allows a global assessment of the painful shoulder, including the rotator cuff, labrum and capsular structures, osseous outlet, acromion, and articular surfaces. Bone marrow edema, muscle edema, and muscle atrophy are all readily evaluated with MRI without exposure to ionizing radiation. However, CTA is effective for patients who cannot be placed in an MRI scanner because of metallic implants or shrapnel. CTA is also very fast, lasting just a few minutes, compared with the time required for an MRI scan.
Use of CTA may become more prevalent in the future as newer reconstruction algorithms decrease the amount of radiation exposure, possibly leading to increased CTA utilization. Current studies favorably compare CTA with MRI for labral imaging and imaging of rotator cuff tears. The improved performance of CTA is due to advances in imaging technologies. Picture archive and communication systems allow review of images on workstations at variable window and level settings. CT scanner technology has also advanced with multidetector scanners, which permit submillimeter-thick scan slices that can be reformatted in any plane.
Although many variations in technique exist, a method for performing arthrography prior to MRI or CT typically begins with a review of prior radiographs or the appropriate scout films of the affected shoulder. This review ensures identification of abnormalities such as soft tissue calcifications, postsurgical changes, and possible evidence of trauma or instability. Standard scout films include AP projections of the shoulder in internal and external rotation and an axillary lateral projection. The patient is then placed supine on the fluoroscopic table with the arm positioned next to the body in slight external rotation. The anterior skin is prepared and draped in a sterile manner, and the site of joint puncture is determined fluoroscopically. A point is chosen overlying the lower third of the humeral head about 0.5 cm lateral to the medial cortex of the humeral head. An alternative anterior approach is through the rotator interval or a posterior approach. The experience of the arthrographer often dictates the approach. Knowledge of all three approaches allows for tailored techniques based on the clinical situation. Although not directly proven in the literature, a posterior approach may permit more consistent visualization of the anterior capsule structures. The skin is anesthetized with 1% lidocaine (Xylocaine) with a 25-gauge, 0.75-inch needle or ethyl chloride spray. A 22-gauge, 3.5-inch or 1.5-inch spinal needle is then used to enter the joint. The correct position is verified fluoroscopically by injecting a small amount of radiopaque contrast material. If the needle is correctly positioned within the joint, the contrast material will outline the medial surface of the humeral head and spill into the subscapularis recess and axillary pouch ( Fig. 44-3 ).
A total injection volume of 12 to 14 mL of material provides adequate distention of the shoulder joint without undue discomfort. Injection of a larger volume often leads to decompression of the joint through a weak point of the capsular insertion along the medial aspect of the subscapularis recess, resulting in leakage of contrast material into the adjacent soft tissues. This phenomenon degrades the quality of the examination and should be avoided. The diagnosis of adhesive capsulitis is made by identifying increasing resistance to the contrast material with a small injected volume, as well as by noting an abnormally small axillary pouch and a small or absent subscapularis recess. If increased pressure is encountered during injection of the contrast agent, fluoroscopic observation should be performed to ensure that the needle tip remains intraarticular. If the contrast agent is intraarticular, the diagnosis of adhesive capsulitis is made, injection of contrast material is ceased, and the needle is withdrawn. Most practices use an iodinated contrast agent for CTA and a dilute gadolinium solution for MRA. The contrast agent can be mixed with 0.1 to 0.3 mL of 1 : 1000 epinephrine to prolong retention of the contrast agent within the joint, thus allowing adequate time for transportation of the patient to the CT or MRI scanner and for imaging. Imaging should be performed as soon as possible after injection of the contrast agent to minimize resorption of contrast material.
CT of the shoulder is performed primarily as a means of evaluating the osseous structures after trauma. Multidetector CT examinations with oblique sagittal and oblique coronal reconstructions can accurately detect the extent of humeral head and neck fractures. The precise number of fracture fragments, as well as the extent of displacement and angulation of fracture fragments, is accurately depicted. The scapula is a complex anatomic structure composed of the body, coracoid and acromion processes, and the glenohumeral articular surface. As a result, the full extent of scapular fractures is difficult to describe fully using conventional radiography, whereas the multiplanar capabilities of CT make it ideal for evaluation of complex scapular fractures. After glenohumeral dislocation, CT examination is the best modality to depict the size and position of a glenoid rim fracture fragment, which can be an important part of presurgical planning ( Fig. 44-4 ). CT with three-dimensional reformations and subtraction of the humeral head is also the modality of choice for preoperative assessment of glenoid bone loss after anterior dislocations.
Computed arthrotomography of the shoulder was widely used prior to the development of MRI and for years was considered the gold standard in the imaging of labral abnormalities. The combination of CT with intraarticular injection of contrast material provides a highly sensitive method of evaluating the glenoid labra and articular cartilage. During the past two decades, MRI and MRA largely replaced CTA in the evaluation of rotator cuff abnormalities and glenohumeral instability. Recent reports of CTA performed with use of multidetector CT (MDCT) demonstrated accuracy equivalent to or better than MRA for proximal biceps tendon, anterior labral injury, and superior labral anterior to posterior (SLAP) lesions. It is especially useful in the evaluation of the shoulder in a patient with a contraindication to MRI. Axial, oblique sagittal, and oblique coronal reformatted images are obtained to provide a better evaluation of the rotator cuff and labrum and to provide the same reconstruction planes as MRI ( Fig. 44-5 ).
Although MDCT arthrography clearly depicts full-thickness rotator cuff abnormalities and many of the lesions associated with glenohumeral instability, MRI and MRA improve sensitivity in the detection of interstitial and bursal-sided tears, tendinopathy, and bone marrow edema without using ionizing radiation. Relative to CTA, MRI and MRA provide superior soft tissue contrast. However, MDCT arthrography with submillimeter-thick source slices allows exceptional spatial resolution, which assists in the detection of labral tears and the visualization of cartilage lesions.
Sonography of the rotator cuff ( Fig. 44-6 ) was first popularized in the 1980s as a simple and noninvasive method of evaluating the rotator cuff, and in recent years it has been repopularized as an accurate and cost-effective method for evaluation of the rotator cuff. Ultrasound examination of the shoulder requires a high-resolution transducer (7.5 to 10 MHz). Evaluation of the shoulder is performed with the patient in the sitting position. The examination begins with the arm in the neutral (thumb up) position. In this position, the bicipital tendon (see Fig. 44-6, A ) is clearly visualized at the superolateral margin of the shoulder, and the presence or absence of fluid can be noted in the subacromial-subdeltoid bursa. Before evaluation of the rotator cuff, the shoulder is rotated internally, and the arm is placed behind the back. This maneuver results in retraction of the critical portion of the supraspinatus tendon from beneath the acromion, allowing maximal visualization of this portion of the rotator cuff. The rotator cuff should be evaluated in both the sagittal and coronal planes.
The normal rotator cuff (see Fig. 44-6 ) is sharply defined, uniform in thickness, and homogeneous in echo texture; it measures 4 to 6 mm in thickness anteriorly, and is normally somewhat thinner posteriorly. A thin echogenic band paralleling the upper surface of the cuff characterizes the subacromial-subdeltoid bursa. The overlying deltoid muscle is characterized by a speckled appearance that is distinct from the normal overlying cuff ( Table 44-1 ).
|Type of Disease||Imaging Signs|
|Tendinopathy||Thickened and heterogeneous tendon|
|Partial-thickness tear||Focal change in echogenicity|
|Focal thinning of tendon|
|Full-thickness tear||Nonvisualization of the cuff|
|Focal defect or absence of the cuff|
|Discontinuity of the cuff|
|Subacromial subdeltoid bursitis||Band of decreased signal superficial to cuff|
Although sonography of the rotator cuff has been shown to be as accurate as MRI in the evaluation of the rotator cuff tendons, it is clear that sonographic evaluation of the shoulder has a steep learning curve and is operator dependent. Thus MRI provides a more global evaluation of the shoulder, including the labrum and osseous structures, which has led to the use of MRI in most practices.
Magnetic Resonance Imaging
MRI has evolved into the imaging modality of choice for the evaluation of shoulder impingement. It offers consistent soft tissue contrast superior to any other imaging modality and provides an excellent overall assessment of the osseous outlet and acromion, tendons, muscles, capsular structures, and labrum. The addition of intraarticular contrast material has also been shown to increase sensitivity in the detection of undersurface partial-thickness tears of the rotator cuff and abnormalities of the labrum. A 2009 metaanalysis concluded that MRA is the most sensitive and specific technique for diagnosing rotator cuff tears when compared with MRI and ultrasound.
The type of magnet and various pulse sequences that are used to evaluate the shoulder differ according to equipment availability and the preference of the imager. Numerous commercial coils are suitable for MRI of the shoulder. Surface coils come in a variety of configurations and are generally adequate for most clinical applications. Quadrature and phased-array coils offer the highest signal/noise ratio and the best overall image quality. The arm is typically positioned next to the side to minimize respiratory artifact. For specialized views, it may be placed over the head. The arm should be placed in the neutral position with the thumb up to provide the most comfort for the patient, thus limiting motion artifact. External rotation tends to be uncomfortable for the patient and tightens the anterior capsule, resulting in poor visualization of the anterior labrum. Internal rotation may obscure the posterior labrum. Internal rotation can also result in pseudo thickening of the subscapularis tendon. The abduction and external rotation (ABER) position reportedly improves visualization of the undersurface of the rotator cuff, as well as of the anterior band of the inferior glenohumeral ligament and the anterior labroligamentous complex. The shoulder is usually imaged in three planes: the axial, oblique sagittal, and oblique coronal planes. A typical imaging protocol includes a T1-weighted sequence in the oblique coronal or sagittal plane ( Table 44-2 ). The T1-weighted images provide the best overall signal/noise ratio and are thus excellent for depicting anatomy. T1-weighted images also nicely demonstrate the osseous anatomy of the shoulder, and because fat is bright, this sequence will demonstrate fatty atrophy of the rotator cuff. Fast spin-echo T2-weighted images with frequency-selective fat saturation are commonly obtained in the oblique coronal and oblique sagittal planes. Fluid is bright on T2-weighted images, thus allowing the most accurate detection of shoulder disease, including rotator cuff abnormalities, labral tears, bone marrow edema, and paralabral cysts. Axial imaging may be performed with use of either fast spin-echo T2-weighted or gradient-echo sequences and typically provides the best opportunity for evaluation of the anterior and posterior labra, as well as of the glenohumeral ligaments and subscapularis tendon.
The use of low field strength magnets has been steadily gaining acceptance in the orthopaedic community as a cost-effective means of performing MRI of the extremities and in particular the shoulder. The advantages of a low field strength magnet include a large-bore opening of the magnet, resulting in better patient acceptance and less difficulty with claustrophobia. As a result, the low field strength systems are less constrained by patient size and body part. The lower strength of the magnet, however, results in decreased signal/noise ratio and longer scan times. The images obtained on a low field strength system are generally considered “less pretty” than the images obtained on a high field strength system, but several studies have demonstrated comparable accuracy for the depiction of rotator cuff disease (see Table 44-2 ). High field strength systems, however, are still considered superior and provide higher diagnostic accuracy for the detection of cartilage and labral lesions in the shoulder and are the most accurate for diagnosing rotator cuff tears.
Magnetic Resonance Arthrography
Although conventional MRI is established as the imaging modality of choice in the evaluation of shoulder impingement syndrome, MRA is accepted as the method of choice for the evaluation of glenohumeral instability. Many of the lesions associated with instability are subtle and may undergo partial healing. This situation, coupled with the fact that the normal anatomic structures of the shoulder lie in close proximity, can make the diagnosis of these lesions difficult even with high-quality conventional MRI. MRA, on the other hand, is performed by distending the joint with fluid (saline solution or dilute gadolinium), which more accurately depicts subtle labral tears, cartilage and ligamentous abnormalities, and partial-thickness tears of the undersurface of the rotator cuff. The standard injection technique for MRA of the shoulder is described in previous sections. About 12 to 14 mL of gadolinium diluted 1 : 200 with normal saline solution or 1% lidocaine is injected. Imaging protocols vary, but a standard set of imaging sequences typically includes T1-weighted images with frequency-selective fat saturation in the axial, oblique sagittal, and oblique coronal planes. The T1-weighted images have a high signal/noise ratio, resulting in the anatomic detail that is critical in detecting the subtle lesions associated with glenohumeral instability. A T2-weighted sequence performed in the oblique coronal plane is the most important sequence for depicting rotator cuff abnormalities. A T2-weighted sequence may also be helpful in detecting other pathologic processes, such as a paralabral cyst or bone marrow edema ( Fig. 44-7, A and B ). The ABER view may be added to the standard imaging protocol for any patient thought to have an anterior labral pathologic process.
Imaging of Specific Shoulder Abnormalities
Interpretation of a shoulder MRI examination requires the evaluation of images obtained in several imaging planes and with various pulse sequences. The following sections provide a systematic approach for MRI evaluation as the imaging basis for recognizing common pathologic processes; when appropriate, complementary imaging modalities are discussed. The important structures that must be thoroughly evaluated on each MRI examination include osseous structures (outlet and acromion); rotator cuff muscles and tendons; the biceps tendon and rotator interval; and the labrum (capsular structures and articular surfaces).
Osseous Outlet and Acromion
The clinical syndrome of shoulder impingement refers to a painful compression of the soft tissues of the anterior shoulder (i.e., the rotator cuff, subacromial bursa, and bicipital tendon) between the humeral head and the coracoacromial arch (i.e., the coracoid process, acromion, coracoacromial ligament, and AC joint). Pain occurs when the arm is elevated forward and internally rotated or placed in the position of abduction and external rotation.
In the normal shoulder, the upward pull of the deltoid on the proximal humerus is resisted by an intact rotator cuff so that the humeral head remains centered on the glenoid in all arm positions. If this stabilizing mechanism becomes weakened as a result of trauma, overuse, or age, the humeral head is pulled upward under the structures of the coracoacromial arch. Impingement is initially followed by subacromial bursitis and rotator cuff tendinopathy. Over time, irreversible cuff trauma occurs with fibrosis and degeneration. In the latter stages, a bony excrescence (subacromial enthesophyte) tends to form at the anteroinferior margin of the acromion where the coracoacromial ligament inserts. Tears of the rotator cuff are frequent in this stage, undoubtedly in part because of direct cuff trauma from the spur. After massive tears of the rotator cuff, bone-to-bone contact may result between the humeral head and the undersurface of the anterior third of the acromion, causing sclerosis and proliferative changes in this area.
Neer introduced the term impingement syndrome and described three stages with the rotator cuff disorder. For more detailed information, see Chapter 52 .
The diagnosis of impingement syndrome is usually made clinically on the basis of appropriate historical and physical examination findings. A thorough history and physical examination by an experienced physician have an 84% to 90% sensitivity and a 75% to 95% specificity for diagnosis of a tear of the rotator cuff. Many imaging modalities are available to assist in the evaluation of the progressively painful shoulder, and their role is both to assess the extent of abnormality of the rotator cuff and to identify configurations of the osseous outlet that may predispose to rotator cuff impingement.
The osseous changes that occur with impingement are seen late and thus offer little in establishing an early diagnosis and preventing progression of the associated soft tissue injuries. The following osseous abnormalities may be associated with the clinical syndrome of impingement :
Enthesophyte formation on the anteroinferior aspect of the acromion
A long anterior portion of the acromion with anterior or lateral downward sloping of the acromion
An os acromiale that is not fused
Hypertrophy of the AC joint
Optimal conventional radiographic views have been described for identifying these variations of the osseous outlet. The AP radiograph at a 30-degree caudal angle ( Fig. 44-8 ) is helpful in visualizing the anterior aspect of the acromion and in detecting inferiorly directed enthesophytes. A modified transcapular lateral view obtained with 10 to 15 degrees of caudal angulation (the supraspinatus outlet view) helps further identify the anteroinferior aspect of the acromion. A high-riding humeral head with remodeling of the undersurface of the acromion and sclerosis of the greater tuberosity are conventional radiographic findings ( Fig. 44-9 ) that are pathognomonic of a chronic rotator cuff tear. Although conventional arthrography, ultrasonography, and CT offer improved visualization of the rotator cuff relative to conventional radiography, they add little in the direct evaluation of the osseous outlet.
Magnetic Resonance Imaging of the Osseous Outlet and Acromion
The multiplanar capabilities of MRI and its ability to demonstrate the relationship of the entire osseous outlet to the underlying rotator cuff allow excellent assessment of the outlet. Bigliani described three different radiographic acromial shapes ( Fig. 44-10 ) and related the configuration of the undersurface of the acromion to the presence of rotator cuff tears. A type I acromion (see Fig. 44-10, A ) has a flat undersurface, a type II acromion (see Fig. 44-10, B ) has a curved undersurface, and a type III acromion (see Fig. 44-10, C ) has an anterior hook. The acromial types II and III have an increased association with rotator cuff tears. A type IV acromion with a convex undersurface has subsequently been described, but no definite correlation has been shown to exist between type IV acromion and impingement. On MRI, the shape of the acromion (see Fig. 44-10 ) is best assessed on the oblique sagittal view just lateral to the AC joint. One study, however, suggests poor correlation of acromial arch shape between conventional radiography and MRI. Another study found no association between rotator cuff tears and the acromial structure. A third study reported poor interobserver agreement between radiographs and MRI scans at categorization of acromial shape.
Anterior and lateral downward sloping of the anterior acromion can also narrow the supraspinatus outlet and potentially result in impingement. Anterior downward sloping (see Fig. 44-11, B ) is demonstrated on oblique sagittal MRI; lateral downward sloping (see Fig. 44-11, C ) is best seen on the oblique coronal images. An enthesophyte ( Fig. 44-12 ) extending off the anteroinferior aspect of the acromion can also be clearly demonstrated on MRI. It typically appears as a marrow-containing osseous excrescence, which should have MRI signal characteristics similar to those of the adjacent acromion marrow (bright on T1-weighted images). Potential pitfalls include the attachment of the coracoacromial ligament and the deltoid tendon insertion (see Fig. 44-11, A and C ) on the anterior acromion. These structures may mimic an osseous excrescence, but they can be differentiated from enthesophytes because they lack marrow signal and appear dark on all pulse sequences. The acromion should also be evaluated for os acromiale ( Fig. 44-13 ), which is an accessory ossification center along the outer edge of the anterior acromion. It is normally fused by 25 years of age. An association exists between persistent os acromiale and impingement of the rotator cuff. The deltoid muscle attaches to the inferior aspect of the accessory ossicle, and contraction of the deltoid results in a downward motion of the unstable segment, potentially leading to impingement of the underlying rotator cuff. Os acromiale is demonstrated best on axial images (see Fig. 44-7, A ), but it can also be seen on oblique sagittal or oblique coronal images (see Fig. 44-7, B ), on which it should not be confused with the adjacent AC joint. MRI signs of instability of the os acromiale include fluid signal within the synchondrosis or sclerosis, cystic change, or marrow edema on either side of the synchondrosis ( Box 44-2 ).
Acromion type—assessed on the sagittal imaging plane on the image immediately lateral to the acromioclavicular joint
Type I: Flat undersurface
Type II: Curved undersurface
Type III: Anterior hook
Type IV: Convex undersurface
Lateral downward sloping—assessed on the same image as used to type the acromion
Anterior downward sloping—assessed on the coronal images
Acromial spur—best assessed on T1-weighted images (bright fatty marrow signal within the bony excrescence differentiates a spur from the deltoid tendon slip)
Unfused os acromiale—best assessed on axial images (fluid signal within the synchondrosis suggests instability)
Hypertrophic changes of the capsule and inferiorly directed osteophyte formation of the AC joint (see Fig. 44-13 ) can also be associated with impingement. MRI demonstrates these changes, as well as associated mass effect on the underlying rotator cuff. The coracoacromial ligament ( Fig. 44-14 ) is a soft tissue structure that forms part of the coracoacromial arch. It extends from the coracoid to the acromion and is well seen on oblique sagittal MRI. It normally measures less than 2 mm in thickness and extends across the rotator interval and anterior aspect of the supraspinatus tendon. The role of the coracoacromial ligament in impingement remains controversial; some persons believe that thickening or ossification (see Fig. 44-14, B ) of the ligament may be a potential cause of impingement, whereas others believe that thickening results from impingement.
Coracohumeral impingement, also referred to as subcoracoid impingement, is an uncommon cause of extrinsic impingement that results from a narrowed distance between the coracoid process and the underlying humeral head ( Fig. 44-15 ). Lo and Burkhart suggested using an interval less than 6 mm to make the diagnosis at arthroscopy. On axial images the normal interval is approximately 10 mm. A narrowed coracohumeral distance may result in entrapment of the subscapularis tendon between the coracoid process and the humeral head and can lead to isolated tendinosis and disruption of the subscapularis tendon. Abnormalities isolated to the subscapularis tendon should prompt an investigation of the coracohumeral distance and consideration of this diagnosis. However, after reporting on a small series of patients with subcoracoid impingement at arthroscopy and control subjects, Giaroli et al. concluded that subcoracoid impingement is primarily a clinical diagnosis that may be supported with imaging.
Glenoid dysplasia occurs along a spectrum. The most severe cases can be appreciated on shoulder radiographs, with a hypoplastic glenoid visible on frontal or oblique views of the shoulder ( Fig. 44-16 ). More subtle cases will be seen on MRI as retroversion of the glenoid fossa relative to the scapula or deficiency of the posterior inferior glenoid resulting in a rounded or “S” shape of the posterior inferior glenoid. Harper et al. reported an incidence of 14% in their series of 103 consecutive MRI examinations, as well as an association with posterior labral tears. Weishaupt and collegues reported an association between posterior shoulder instability and glenoid rim deficiency. Although severe cases may be seen with radiographs, milder cases should be sought on lower axial images and by assessing the sagittal images of the posterior inferior glenoid rim (see Fig. 44-16 ). Glenoid dysplasia is associated with increased risk of posterior instability and a posterior labral tear.
The rotator cuff is composed of four tendons: the supraspinatus superiorly, the subscapularis anteriorly, and the infraspinatus and teres minor posteriorly. These tendons are important dynamic stabilizers of the glenohumeral joint, and any review of rotator cuff disease will rapidly expose the supraspinatus tendon as the weak link of the rotator cuff. Most cuff failures originate in the supraspinatus tendon at or near its insertion onto the greater tuberosity of the humeral head. Some investigators attribute this propensity for cuff failure within the supraspinatus tendon to its blood supply. The supraspinatus tendon receives its arterial supply from the anterior humeral circumflex, subscapular, suprascapular, and posterior humeral circumflex arteries. A zone of relative avascularity has been described in the tendon proximal to its attachment site and may represent a “critical zone” for cuff failure. Other authors have found this zone to be vascularized by anastomosing vessels from the tendon and humeral tuberosity. Arterial filling of the cuff vessels in the critical zone depends on the position of the arm; poor filling is present when the arm is adducted. A high correlation has also been shown to exist between rotator cuff tears and subacromial impingement. It is probably a combination of avascularity and subacromial impingement that leads to most rotator cuff abnormalities originating in the critical zone of the supraspinatus tendon.
Conventional radiography plays only a limited role in the direct evaluation of the rotator cuff, although it is frequently the initial imaging study performed for patients with the clinical syndrome of impingement. Radiographs allow identification of associated pathologic change, especially of the osseous outlet and acromion. Conventional radiography findings associated with cuff pathology include the following :
A high-riding humeral head (<7 mm between the humeral head and the undersurface of the acromion) is associated with a chronic rotator cuff tear
Scalloping and loss of normal acromion undersurface convexity (see Fig. 44-9 )
Magnetic Resonance Evaluation of the Rotator Cuff
The normal anatomy of the rotator cuff is accurately depicted with MRI. The supraspinatus muscle originates along the posterosuperior portion of the scapula above the level of the scapular spine. A single tendon arises out of the muscle and extends superiorly above the humeral head to insert onto the greater tuberosity of the humeral head. The supraspinatus tendon is best evaluated in the oblique coronal and oblique sagittal MRI planes. The subscapularis is located anteriorly and has multiple tendon slips. It has a broad origin along the anterior aspect of the scapula and attaches to the lesser tuberosity on the anterior aspect of the humeral head. An extension of the subscapularis tendon known as the transverse ligament extends across the intertubercular groove and helps to stabilize the long head of the biceps tendon within the intertubercular groove. The subscapularis muscle and tendon are best evaluated on axial and oblique sagittal MRI. The infraspinatus is located posterosuperiorly and has a broad origin along the posterior aspect of the scapula inferior to the scapular spine. The teres minor is located posteroinferiorly, below the level of the infraspinatus, originating along the axillary surface of the scapula and inserting on the most inferior aspect of the greater tuberosity of the humeral head. The infraspinatus and teres minor are best evaluated on oblique coronal and oblique sagittal images.
The normal tendons demonstrate uniformly dark signal intensity on all pulse sequences; increased signal on T1-weighted images is a nonspecific finding that may represent a wide array of conditions ranging from artifact to a complete tear. T2-weighted images with fat suppression have a lower signal/noise ratio and provide less anatomic detail, but they are both sensitive and specific for depicting the full range of rotator cuff abnormalities.
Increased signal within the rotator cuff tendons on T1-weighted images has many causes. When increased signal is noted on T1-weighted images, the morphologic features of the tendon and the T2-weighted images should be evaluated. A normal appearance of the tendon and normal signal intensity on T2-weighted images suggest that the tendon is probably normal. Potential causes of increased signal within a normal tendon on T1-weighted images include partial volume averaging and the “magic angle” phenomenon. This MRI artifact results in increased signal in normal tendons that are angled 55 degrees relative to the direction of the main magnetic field. It occurs primarily on T1-weighted images, and as echo time lengthens, as occurs with T2 weighting, the signal intensity again decreases. This MRI artifact is particularly likely to occur in the critical zone of the supraspinatus tendon and therefore occasionally causes confusion about the integrity of the cuff.
A wide spectrum of rotator cuff pathology can be accurately depicted on MRI, ranging from tendinosis to a full-thickness tear of the rotator cuff ( Table 44-3 ). Tendinosis ( Fig. 44-17 ) is noted as intermediate signal on T1-weighted MRI that persists on T2-weighted MRI. The signal abnormality does not reach the same brightness as fluid on T2-weighted images (see Fig. 44-17, B ). The tendon may also demonstrate diffuse or focal thickening, but no evidence of tendon disruption will be seen. Tendinosis represents a degenerative process of the tendon, which histologically represents a combination of inflammation and mucoid degeneration. On arthroscopic examination, the tendon may appear to be edematous and hyperemic, with occasional fraying, roughening, or degeneration of the surface of the tendon, and it may be difficult at times to differentiate tendinosis from early partial-thickness tearing both at arthroscopy and on MRI.
|Cuff Pathology||Magnetic Resonance Imaging Appearance|
Partial-thickness tears ( Figs. 44-18, 44-19, and 44-20 ) can occur on either the articular surface or the bursal surface or within the substance of the tendon. Articular surface tears are the most common type to occur (see Fig. 44-18 ). A partial-thickness tear is seen on MRI as an area of increased signal on T1-weighted images that increases to fluid signal intensity on T2-weighted images. The fluid signal intensity extends only partially through the thickness of the tendon from superior to inferior. These tears may partially heal with granulation tissue, making them difficult to distinguish from tendinopathy on MRI. MRA is more sensitive than conventional MRI in detecting articular surface tears; however, it does not increase sensitivity for detecting partial-thickness tears of the bursal surface. It has been suggested that the sensitivity for detecting partial-thickness undersurface tears can be further improved by adding the ABER view to the MRA protocol.
Interstitial tears of the rotator cuff represent tears that occur within the substance of the tendon but do not involve the bursal or the articular surface ( Fig. 44-19 ). These tears are demonstrated on MRI as a focal linear area of fluid signal that is contained within the substance of the tendon, and they typically occur at the footprint of the supraspinatus tendon as it attaches to the greater tuberosity of the humeral head.
A partial articular-sided supraspinatus tendon avulsion lesion is a subset of partial-thickness tears. The tear represents a partial-thickness articular-sided avulsion of the supraspinatus tendon at its most anterior attachment site. This type of tear deserves special attention and should be accurately described on MRI because the recommended treatment for this subset of tendon tears differs from the standard partial-thickness tears described previously. On MRI, a small articular-sided avulsion is seen as fluid signal extending into the articular surface of the supraspinatus tendon at its anterior attachment site with partial avulsion of the tendon at this level, and this lesion represents a subset of the articular surface partial-thickness tears ( Fig. 44-20 ).
A full-thickness tear of the rotator cuff tendon is defined as a tear that extends through the complete thickness of the tendon from superior to inferior. This tear allows communication between the joint space and the subacromial-subdeltoid bursa. MRI criteria for establishing the diagnosis of a full-thickness tear include high (fluid) signal completely traversing the tendon from superior to inferior on T2-weighted images, a gap or absence of the tendon, and retraction of the musculotendinous junction ( Fig. 44-21 ).
Rotator cuff muscles can undergo atrophy after tendon tears or after denervation. Denervation results in muscle edema followed by progressive fatty infiltration. Chronic rotator cuff tendon tears also result in muscle atrophy even if no nerve injury is present. The terminology with regard to atrophy of the rotator cuff musculature is confusing. The term fatty infiltration is typically used to describe actual fatty infiltration or replacement of the muscle, whereas muscle atrophy is used to describe a loss of muscle bulk. Muscle atrophy is seen on MRI as loss of muscle bulk (best depicted on sagittal images through the cuff musculature), whereas fatty infiltration is seen as streaks of high T1-weighted signal within the substance of the muscle ( Fig. 44-22 ). When they are normal, the four muscles of the rotator cuff demonstrate symmetric bulk and signal characteristics. Loss of bulk indicates muscle atrophy, whereas streaks of high T1-weighted signal within the substance of the muscle indicate fatty infiltration. Goutallier and colleagues reported a system for grading fatty infiltration of the cuff musculature on the basis of CT imaging, and although this system has been widely adapted for use with MRI, a study by Fuchs and associates demonstrated poor correlation between CT and MRI with regard to grading fatty infiltration. Fatty infiltration is typically graded as mild, moderate, or severe based on the extent of fatty infiltration (high T1-weighted signal) with the belly of the muscle. Muscle atrophy can be graded separately as mild, moderate, or severe on the basis of muscle bulk depicted on sagittal images at the level of the supraspinatus fossa. Regardless of the terms used to describe muscle atrophy and fatty infiltration, they are associated with poor functional outcomes after rotator cuff tendon repair and increased risk of a repeat tear. Assessment of preoperative MRI scans should include evaluation of muscle atrophy.
Calcific tendinitis (hydroxyapatite crystal deposition disease) can also be diagnosed with MRI ( Fig. 44-23 ). The crystalline deposits are typically within the critical zone of the rotator cuff and appear as areas of low signal intensity on both T1- and T2-weighted images. Calcific deposits within the rotator cuff may be difficult to identify on MRI because both the tendon and the calcific deposit appear dark on all pulse sequences. Ancillary MRI findings that may aid in the identification of calcific tendinitis include globular thickening of the involved tendon and high signal within the tendon and surrounding tissues as a result of associated inflammation. Gradient-echo imaging may improve conspicuity of the deposits as a result of the “blooming” artifact associated with local magnetic field heterogeneity. Use of gradient-echo imaging and comparison with conventional radiographs improve the likelihood of detecting calcific tendinitis.
Most cuff tears originate in the supraspinatus tendon; however, large tears may extend into either the infraspinatus or subscapularis tendon. Isolated tears, although less common, occasionally occur in either the infraspinatus or subscapularis ( Fig. 44-24 ) tendon. An isolated tear of the infraspinatus tendon is usually associated with the internal impingement syndrome (discussed further in the section on glenohumeral instability). An isolated tear of the subscapularis tendon may result from shoulder dislocation or in association with coracohumeral impingement and is best demonstrated on axial MRI as high signal traversing the tendon with retraction of the tendon away from its normal attachment site on the lesser tuberosity. An extension of the subscapularis tendon known as the transverse ligament holds the long head of the biceps tendon in the intertubercular groove, and a tear of the subscapularis tendon may result in disruption of the transverse ligament, leading to medial subluxation or dislocation of the long head of the biceps tendon. Axial MRI is well suited not only for evaluating the integrity of the subscapularis tendon but also for demonstrating medial subluxation of the biceps tendon out of the intertubercular groove. The biceps tendon may be displaced superficial to the subscapularis tendon, into the subscapularis tendon ( Fig. 44-25 ) or intraarticular ( Fig. 44-26 ).