Shoulder Arthrogram

Earn CME/CE in your profession:

Continuing Education Activity

Glenohumeral instability is a common and perplexing clinical issue. Dependent on the mechanism of injury, patient age, and clinical symptoms, these factors can determine what imaging modality is appropriate for a patient to undergo. One such imaging modality is shoulder arthrography. This activity describes the examination and highlights the role of the interprofessional team in evaluating patients who undergo glenohumeral arthrography.


  • Describe the anatomy of the glenohumeral joint.
  • Summarize the technique for performing a shoulder arthrogram.
  • Outline both indications and contraindications for a shoulder arthrogram.
  • Explain normal and abnormal findings on a glenohumeral arthrogram.


Within musculoskeletal radiology, arthrography has served as an essential technique for close to 100 years. Glenohumeral arthrography was described in 1933 when Oberholzer was studying capsular distortion secondary to shoulder dislocation.[1] During this time, he injected air into the shoulder joint to evaluate the structures, including the axillary recess, on a conventional radiograph. In 1934, Codman had suggested that injecting contrast material into the shoulder joint could demonstrate rupture within the rotator cuff. In 1939 Lindbolm and Palmer determined that arthrography was accurate in diagnosing lesions in the rotator cuff in a substantial number of patients.[1]

The use of iodinated contrast, computed tomography, and magnetic resonance imaging naturally came after this time. At this time, magnetic resonance imaging (MRI) is the first-line imaging modality for assessing joints as it has a superior soft-tissue contrast capability. In a patient who is claustrophobic or has any contraindications to undergo an MRI, a computed tomography (CT) arthrogram is a suitable option. Postoperative joints can lead to artifacts for which CT is a good option. Arthrography remains a useful imaging modality with computed tomography, CT scan, and magnetic resonance imaging, MRI, to allow a detailed assessment of articular structures of interest.

Glenohumeral arthrography, shoulder arthrography, is an imaging technique used in evaluating the glenohumeral joint and associated components. During an arthrogram, a joint injection is done typically under fluoroscopic guidance, but ultrasound or CT can be utilized. The process of a direct arthrogram leads to joint distention and separation of the intra-articular structures. This capsular distention allows for the enhancement and visualization of small joint bodies, the labrum, glenohumeral ligaments, rotator cuff undersurface, the structures of the rotator interval, and the long head of the biceps. Direct arthrography in which contrast is injected into the joint has an alternative procedure termed an indirect arthrogram. An indirect arthrogram is a technique that produces arthrographic images without utilizing direct joint injection.[2]

Historically, arthrography was performed with fluoroscopy and plain radiographs only, but today all patients undergo cross-sectional imaging of the shoulder after the injection of contrast. Typically, this is an MRI, but CT can be performed if contraindications to undergo an MRI are present, or there is a high clinical suspicion of a bony abnormality.[3]

Generally, radiographic examinations demonstrate soft tissues like cartilage, muscle, joint fluid, and menisci to be of the same density. Therefore, these structures are not distinguishable from one another. The term arthrography refers to an imaging modality following the injection of contrast into a specific joint, typically performed with fluoroscopic guidance. Utilizing injected contrast outlines the intraarticular structures and differentiates them from other adjacent soft tissues. The injection also allows for distention of the joint, providing better visualizations and separation of structures. During an arthrogram, a sterile technique and local anesthetic are utilized. A needle is introduced into the joint space where synovial fluid can be aspirated if needed for any diagnostic purpose. Contrast like iodinated contrast is injected into the joint.

Additional medication like an anesthetic or a glucocorticoid can also be injected into the joint space for therapeutic purposes during the arthrogram. The arthrogram can aid in facilitating the identification of ligamentous or tendon injuries, intraarticular "loose" bodies, cartilage or synovial abnormalities, loosening of the joint prosthesis, and sinus tracts. The implementation of fluoroscopy allows for real-time tracking of contrast, which will pass into and fill the joint.[4] The contrast pattern can highlight abnormalities like abnormal contrast leakage or synovitis. The arthrogram can be followed with computed tomography or magnetic resonance imaging. Rotator cuff tears and the labrum are more appreciated when there is delineation by contrast.

Glenohumeral instability is common and can be a perplexing clinical issue where both an accurate and a non-operative method of diagnosis is desirable. Shoulder arthrograms can serve as a useful aid in diagnosis. The glenohumeral joint is susceptible to instability and dislocation due to a combination of the bony discrepancy between the humeral head and the glenoid. This discrepancy allows for a larger range of motion. The biomechanics within the shoulder joint is based upon the interaction of both static and dynamic stabilizing systems. The static structures of the joint include the glenoid, glenoid labrum, humeral head, and capsule. The capsule includes the glenohumeral ligaments. The dynamic stabilizing structures include the rotator cuff and muscular structures surrounding the joint.[5] When evaluating for shoulder pathology, it is crucial to remember the relevance of radiographs to assess osseous and joint structure abnormalities. Typically, an MRI study after conventional radiographs have been evaluated is performed.

The glenohumeral joint has both complex anatomy and physiology. It is the most mobile yet unstable joint in the body and requires appropriate imaging exams for varying clinical scenarios. A thorough history and physical examination are necessary to develop a differential diagnosis before selecting which imaging examination. Plain film radiography evaluation typically includes internal rotation, external rotation, axillary and transscapular views, with other specialty views as needed. The plain film is the first-line imaging modality for nearly all shoulder pathology. They are also often the only imaging examination necessary for evaluating calcific tendinitis, arthritis, acute shoulder trauma, and osteolysis of the distal clavicle in athletes.[6]

Computed tomography of the glenohumeral joint is reserved for evaluating a fracture or fracture-dislocation and a prosthetic joint. The CT scan can demonstrate fractures displacement, angulation, and complexity. On CT, the visualized images in the coronal, axial, sagittal planes, and three-dimensional format, can aid in the interpretation and any preoperative planning.[7] Magnetic resonance imaging is the primary imaging modality used to evaluate soft tissues of the shoulder. Soft tissues include the rotator cuff, tendons, biceps muscle, subacromial and subdeltoid bursae. It also has a high level of sensitivity in detecting subtle fractures, acromioclavicular joint changes, erosive changes to the distal clavicle, early avascular necrosis, bone marrow edema, muscular atrophy, and morphology of the acromion.[8]

The MRI is the second-line imaging modality to evaluate shoulder instability and labral tears when MR arthrography is not performed. Radionuclide bone scans, technetium-99m bone scans, are typically used for evaluating an infection post arthroplasty or suspected metastases and whole-body imaging.[9]

Arthrography involves the percutaneous puncturing into the shoulder joint and then instilling a contrast agent. Iodinated contrast is used for conventional arthrography and CT arthrography, whereas gadolinium contrast is used for MR arthrography. With conventional arthrography, radiographs are obtained after injecting the iodinated contrast, which is still the procedure of choice in diagnosing adhesive capsulitis, frozen shoulder.[10] With the injection, there is distention of the shoulder capsule that can be used therapeutically. MR arthrography involves an injection in the intra-articular space, but a gadolinium-based contrast agent is injected with an MRI following. MR arthrography, or MRA, is the gold standard in evaluating a suspected labral tear or shoulder instability.[11]

MRA is also indicated when there is a high suspicion of a rotator cuff tear in a patient with a normal or inconclusive MRI and the evaluation for intra-articular small bodies. The systemic administration of the gadolinium-containing contrast agent in a patient with moderate to severe impaired renal function with a GFR of less than 15 to 30 mL per minute does have an associated increased risk of developing nephrogenic systemic fibrosis. Given that the MR arthrography involves a small dosage of gadolinium and the administration is via intra-articular, nephrogenic systemic fibrosis has never been reported in this procedure.[12] The risk of nephrogenic systemic fibrosis has been reduced further by using Group II gadolinium-based contrast agents, which have proven safe for use even in patients with severe renal impairment.

Computed tomography arthrography is used when there is a contraindication for MR arthrography, like a patient with incompatible vascular clips, claustrophobia, a pacemaker, or when MRA is unavailable. CT arthrography can be used for evaluating a prosthetic joint which would ultimately result in an artifact on MRI.[13] Ultrasonography of the glenohumeral joint is useful when assessing the rotator cuff, calcific deposits, and biceps tendon. It can also measure the subacromial space and detect the presence of muscular atrophy. It serves as a tool in the dynamic evaluation for the shoulder and shoulder impingement.[14] 

The American College of Radiology, ACR, has criteria for selecting the imaging modalities of choice for patients with traumatic or atraumatic shoulder pain. Imaging examination choices are based upon the etiology of the shoulder pain, whether traumatic or atraumatic, the duration of symptoms, the age of presentation, and any clinical or radiographic suspicions for a particular condition.[15] Again, the initial imaging modality of choice for traumatic or atraumatic shoulder pain should be radiography. Indications for conventional radiographs include evaluation for dislocation or fracture following trauma, evaluation of calcific tendinitis, crystal deposition disease, osteoarthritis, suspicion of a bony neoplasm, particularly with a patient with a non-diagnosis of cancer, suspicion for septic arthritis, or suspected humeral head avascular necrosis. The next imaging choice should be guided by the clinical scenario and findings from the plain films.[16]

For example, a patient who is less than 30 years old and suffers an anterior glenohumeral dislocation may have a suspected labral injury. MR arthrography is indicated in these cases.[17] An anterior shoulder dislocation can also have an associated rotator cuff injury in approximately 50% of patients under 40 years old and about 80% of patients over 60 years old.[18] With these patients, ultrasound may be of value as a screening modality.

MR arthrography is indicated in a younger patient with shoulder instability, suspicion of a labral tear, and within the setting of a shoulder dislocation. MRA allows for distinguishing between a partial versus a full-thickness rotator cuff tear and can identify suspected cartilage or labral tear.[19]

When MR arthrography is contraindicated, then CT arthrography should be used. CT arthrography can detect occult fractures, tendinitis, or labral tears when MRI is contraindicated. In a patient with a suspected rotator cuff tear or impingement, MR arthrography has a higher sensitivity and specificity than MRI or ultrasound for diagnosing both partial as well as full-thickness rotator cuff tears. MRA has a specificity and sensitivity greater than 95% for detecting full-thickness rotator cuff tears. In partial-thickness tears, MRA has a specificity and sensitivity are 96 and 86%, respectively.[20] MRA can differentiate between small partial versus small full-thickness tears and tendinosis.

CT arthrography is indicated if a patient cannot undergo MRI or ultrasound expertise is unavailable. Both the sensitivity and specificity are over 90% for CTA in detecting supraspinatus and infraspinatus tears, but low sensitivity for detecting subscapularis tears.[21]

CTA can also aid in preoperative evaluations by demonstrating the extent of fatty degeneration and tendonous retraction of the corresponding muscle. MRA is also the imaging of choice for labrocapsular structures. It is the most accurate imaging modality for evaluating sports injuries to the shoulder. MRA can evaluate the dynamic stabilizers of the glenohumeral joint that are extra-articular, the static stabilizers that are intra-articular, and the morphology of the capsule. The sensitivity and specificity of MRA in detecting labral tears range from 88 to 100% and 88 to 96%, respectively.[22][23]

Using the abduction external rotation technique increases MRA sensitivity for labral tear diagnosis to close to 100%; hence this modality is indicated for suspected lesions to the rotator cuff and glenoid labrum in athletes. CTA is accurate in delineating anatomic derangement, including the glenoid labrum, but soft tissue evaluation is limited. The sensitivity for CTA for labral tears is between 73 to 76% and a specificity of 92%.[24]

In a patient with atraumatic shoulder pain with suspected adhesive capsulitis, the diagnosis is mainly based on clinical findings. Adhesive capsulitis, also known as frozen shoulder, is due to the contraction and thickening of the glenohumeral joint capsule and synovium. This process results in a progressive limitation in the joint's mobility with associated significant pain. Conventional arthrography is the imaging modality of choice for both diagnosis and treatment. There is a decreased capacity for injecting contrast into the joint, which is diagnostic for adhesive capsulitis in these cases. The distention of the capsule during the arthrogram can serve as a therapeutic tool.[25]

Anatomy and Physiology

The glenohumeral joint is a ball and socket joint located between the humerus and scapula, which serves as the major joint that connects the upper limb to the trunk. In terms of movement, like a ball and socket joint, the glenohumeral joint allows for extension, flexion, abduction, adduction, internal rotation, external rotation, and circumduction. This is one of the most mobile joints within the body, which is at the cost of joint stability.[5]

The glenohumeral joint is formed by the articulation of the humeral head to the glenoid fossa of the scapula. As a synovial joint, hyaline cartilage covers the articular surfaces. Compared to the glenoid fossa, the head of the humerus is larger in size, which allows for a wide range of motion of the joint but leads to instability. Anatomically to reduce this disproportion of the articulating surfaces, the glenoid fossa has a fibrocartilage rim known as the glenoid labrum to deepen the surface. The joint capsule is a fibrous sheath and serves to enclose the joint structure. The capsule extends from the anatomic neck of the humerus to the border of the glenoid fossa. The joint capsule has laxity, allowing for greater mobility. Lining the inner surface of the joint capsule is the synovial membrane which produces the synovial fluid and reduces the friction between the articulating surfaces.[26]

Synovial bursae are also present to reduce friction. A bursa is a sac filled with synovial fluid that cushions the tendons between the other joint structures. Clinically significant bursae include the subacromial and subscapular bursae. The subacromial bursa is located deep to the acromion and the deltoid. It is superficial to the joint capsule and the supraspinatus tendon. This bursa aids in reducing the friction below the deltoid, which promotes more free motion of the rotator cuff tendons. The subscapular bursa lies between the scapula and the subscapularis tendon. The bursa aids in reducing wear and tear to the tendons during the movement of the glenohumeral joint. Vascularly, the glenohumeral joint is supplied by both the anterior and poster circumflex humeral arteries. Both are branches off of the axillary artery. Branches off the suprascapular artery arising from the thyrocervical trunk also contribute to the vascular supply. Innervation is provided primarily by the suprascapular, axillary, and lateral pectoral nerves.

The origination of the musculotendinous rotator cuff is from the scapula with insertion onto the humeral tuberosities, which almost envelops the humeral head. The four muscles that comprise the rotator cuff are the supraspinatus superiorly, the subscapularis anteriorly, and both the infraspinatus and teres minor posteriorly. The four muscle tendons form a broad aponeurosis that fuses with most of the glenohumeral joint capsule. Within the intertubercular groove between the attachments of the subscapularis and supraspinatus lies the long tendon of the biceps. The long tendon of the biceps passes through the glenohumeral joint from a synovial sheath within the bicipital groove. It attaches to the superior margin of the rim of the glenoid.[27]

Regarding the glenohumeral arthrogram, it is important to note the anatomy of the subcoracoid and subacromial bursa for interpretation purposes. The subcoracoid bursa lies anteriorly, whereas the subacromial bursa superiorly. These bursae tend to combine into one large bursa.[28] The thick, musculotendinous rotator cuff separates the bursa from the glenohumeral joint capsule. The rotator cuff is separated from the deltoid muscle and the coracoacromial arch by the bursa. Most rotator cuff ruptures occur at the critical area in the supraspinatus portion of the cuff. The critical area is just proximal to the insertion point onto the greater tuberosity of the humerus. A complete thickness tear produces an abnormal communication between the bursa and the glenohumeral joint. When contrast is injected into the joint, it will then flow into the bursa, which is identified on the arthrogram.

The static stabilizers of the shoulder include the osseous and joint anatomy. The clavicle is an S-shaped bone that articulates with the sternoclavicular joint medially and the acromioclavicular joint laterally. They can have varying appearances, from flat to curved. The lateral half of the clavicular shaft has an irregular osseous ridge for insertion of the trapezoid and deltoid muscles.[29]

The scapula is a triangular bone with three parts: the body, spine, and neck.[30] The scapula has three structures that articulate with the clavicle and the humeral head. These structures include the glenoid fossa, coracoid process, and acromion. The glenoid fossa is an articular surface for the glenohumeral joint. It has a supraglenoid tubercle that serves as the origin of the long head of the biceps tendon and an infraglenoid tubercle that serves as the origin of the long head of the triceps muscle. The glenoid has a distinctive pear-shaped morphology and measures approximately 6 to 8 cm². The cavity lies in retroversion with an average of 7 degrees.[31]

The pear-shaped morphology of the glenoid allows for the maintenance of shoulder stability. The presence of anterior bone loss is typically secondary to anterior shoulder dislocation, which results in an inverted pear-shaped appearance to the glenoid. The glenoid has hyaline cartilage on its articular surface that is thinner in the center but thicker on the periphery, with an average thickness measuring 1.8 mm.[32] The contact areas to the glenoid change with motion. There is a bare area of cartilage that covers a thickened subchondral bone and is known as the bare spot or the tubercle of Assaki.[33]

In the coracoid process, there's an entire process of the scapula that arises from the anterolateral aspect. The morphology of the coracoid varies, but it serves as the origin of the pectoralis minor muscle and short head of the biceps. Orientation of the coracoid can impact the subcoracoid space and lead to subsequent subcoracoid impingement with tendinosis of the subscapularis muscle. The posterolateral extension of the scapular spine that articulates with the clavicle is the acromion. Acromion morphology varies, but Bigliani has classified its morphology on radiographs into three types. A type one acromion is flat, a type two is curved, and type three is hooked. A type three acromion is thought to be acquired, not congenital.[30]

Normally, the articular surface of the humeral head is retroverted by approximately 30 degrees.[34] The head of the humerus is covered by cartilage thicker in the center but thinner on the periphery. The average thickness of the cartilage here is 1.24 mm.[35] There is a bare area on the posterior aspect of the humeral head, which is between the insertion of the synovial membrane and joint capsule with adjacent articular cartilage. Also another required bare area is between the insertion of the supraspinatus muscle in the greater tuberosity and the adjacent articular cartilage. Of note, this bare area increases in any undersurface supraspinatus muscle tears and cartilaginous lesions. The coracohumeral interval is the space between the lesser tuberosity and the coracoid process. This interval varies with external and internal rotation and tends to be smaller in females, normally measuring anywhere between 10 to 11.5 mm.[36] The interval measurement alone is not diagnostic for subcoracoid impingement syndrome but also requires evidence of clinical symptoms for diagnosis.

The acromioclavicular joint, AC joint, is a synovial joint between the medial surface of the acromion and the lateral surface of the clavicle. The stabilization of the AC joint depends on static structures like the ligaments and dynamic structures like the musculature. A small fibrocartilage disc exists between the two bones. Both the inferior and superior AC ligaments reinforce the AC capsule. The deltoid muscle attaches to the lateral border of the clavicle and the acromion anteriorly, whereas the trapezium attaches to the posterior and superior aspects. Both of these muscles provide dynamic stabilization for the AC joint. The primary stabilizer of this joint is the coracoclavicular ligament.[37]

The coracoclavicular ligament consists of two parts: the conoid and the trapezoid ligaments, which form a V-shaped structure. These attach under the surface of the clavicle at the junction of the middle and outer thirds. The conoid is located more medially and prevents anterior and superior displacement and rotation. The trapezoid ligament is located laterally and prevents posterior displacement while limiting rotation. The coracoacromial ligament is triangular and attaches to the lateral, horizontal portion of the coracoid with insertion on the tip of the acromion in front of the acromioclavicular joint.[37]

The glenoid labrum acts as a passive stabilizer to the joint and adds depth to the glenoid cavity by approximately 50%.[38] The glenoid labrum also serves as the primary attachment site for glenohumeral ligaments, the long head of the biceps tendon, and the joint capsule. This fibrocartilage labrum attaches posteriorly and inferiorly to the glenoid. On the anterior aspect of the glenoid labrum, it blends with the anterior band of the inferior glenohumeral ligament. Superiorly, the glenoid labrum blends with the biceps tendon and the superior glenohumeral ligament. Two types of chondrolabral attachments occur, including type A and type B. Type A has an abrupt transition, whereas type B has a transitional zone, shown on MRI as a linear band of increased signal intensity located between the hyaline cartilage and the labrum. The average width of the labrum is approximately 4 mm and a thickness of 3 mm.[39] The glenoid labrum typically has a triangular shape, but other morphologies are described as a rounded, cleaved, flat, notched, or even absent labrum. When evaluating the labrum, magnetic resonance angiography (MRA) is the preferred method.

The long head of the biceps tendon can originate from four structures, including the supraglenoid tuberosity, labrum, and coracoid process. This tendon makes an oblique intra-articular course, traverses through the rotator interval, and makes a 30 to 45° turn along the anterior aspect of the humerus before exiting the joint in the bicipital groove between the greater and lesser tuberosities.[40]

The region where the tendon attaches to the superior labrum is known as the bicipital-labral complex or the biceps anchor. There are three different types of attachments of the long head of the biceps tendon to the supraglenoid tubercle. A type one is where the biceps has a firm attachment to the glenoid rim, and no free edge is present. A type two is where the attachment of the biceps to the glenoid is in a more medial position where a small sulcus is located between the labral free edge and the cartilage with the synovial lining. A type three attachment is where there is a deep recess and the superior labrum projects into the joint. The large synovial recess located within the undersurface of the bicipital labral junction is known as the sublabral recess or sulcus. This sulcus is located between the 11 and 1 o'clock positions. The type two and type three attachments are frequent anatomical variants.[35] 

The joint capsule for the glenohumeral joint is composed of three layers. These layers include the synovial lining and the articular surface, composed of a loosely packed collagen and overlayer and a thicker bursal surface layer made of dense collagen. Bands that consist of collagenous thickening within the capsule form the glenohumeral ligaments. These ligaments and the coracohumeral ligament reinforce the thin capsular tissue. They are also a part of the static stabilizers of the shoulder joint and are best visualized on computed tomography angiography (CTA) or MRA.[41]

The superior glenohumeral ligament has a variable origination which includes the supraglenoid tubercle, long head of the biceps tendon, superior labrum, middle glenohumeral ligament, or a combination. This extends in a nearly perpendicular plane to the middle glenohumeral ligament and parallel to the coracoid process close to the biceps tendon. The insertion site is into the small depression above the lesser tuberosity of the humerus known as the fovea capitis, where this blends with the medial band of the coracohumeral ligament. This occurs within the rotator interval and helps form a sling around the long head of the biceps.[42]

The middle glenohumeral ligament has a variable size and constancy. It can be absent in about 20 to 30% of shoulders.[43] Its insertion is seen at the level of the superior anterior labrum. This ligament runs oblique and posterior to the superior margin of the subscapularis and subsequently blends with the anterior capsule. Its distal attachment is to the anterior proximal humerus below the insertion site of the superior glenohumeral ligament. The Buford complex is a variant and represents a second cordlike structure of the middle glenohumeral ligament with an absent anterosuperior portion of the labrum.[44]

There is a normal communication channel between the glenohumeral joint capsule with the subscapularis bursa. This communication is between the superior and middle glenohumeral ligament known as the foramen of Weitbrecht. Just below this is another communication channel between the middle and inferior glenohumeral ligament known as the Forman of Rouviere.

The inferior glenohumeral ligament consists of the axillary pouch and the anterior and posterior bands. The origination of the anterior band is along the inferior two-thirds of the anterior aspect of the glenoid labrum and is usually more prominent and thicker. The inferior glenohumeral ligament has an insertion onto the anatomical neck of the humerus.[43]

The coracohumeral ligament is an extra-articular structure that extends in an anterolateral fashion from the base of the coracoid process with fanning out to insert onto the greater tuberosity and the bicipital sheath. The distal aspect of the ligament divides into two functional bands, including a smaller medial band and a larger lateral band.[41]

The medial band blends with fibers of the superior glenohumeral ligament to form a ligamentous complex that surrounds both the medial and inferior aspect of the intra-articular portion of the long head of the biceps tendon before insertion onto the lesser tuberosity of the humerus as well as merges with the rotator interval capsule along the superior fibers of the subscapularis tendon. In contrast, the lateral band of the coracohumeral ligament surrounds the superior and lateral aspect of the long head of the biceps tendon intra-articularly before insertion onto the greater tuberosity of the humerus and also on the anterior aspect of the supraspinatus.[45] This creates a sling-like band that surrounds the long head of the biceps, also known as the biceps pulley or the biceps reflection pulley. The function of this pulley is to limit any medial subluxation of the bicipital tendon and increase efficiency during the contraction of the muscle.

The rotator interval is a triangular anatomical space in the anterosuperior portion of the shoulder. This interval is defined by the coracoid process at the base, the anterior margin of the supraspinatus tendon superiorly, and the superior margin of the subscapularis tendon inferiorly. This interval allows the biceps tendon to become intra-articular as it courses to its insertion onto the superior aspect of the labrum. As the bicipital tendon enters into the space, it becomes covered by the biceps pulley. Both fibers from the supraspinatus and subscapularis cross over the interval and merge with the superior glenohumeral ligament and the coracohumeral ligament.[41]

The active stabilizers to the glenohumeral joint include both the muscles and rotator cuff. The components of the rotator cuff include the supraspinatus, infraspinatus, teres minor, and subscapularis muscles which originate from the scapula and are fused onto their respective attachment sites. The origin of the supraspinatus muscle is from the supraspinous fossa and superior aspect of the scapular spine. It runs laterally and inserts upon the highest impression of the greater tuberosity of the humerus. The supraspinatus tendon has two portions with an anterior half that is both thick and long and a posterior half that is thin and short.[46]

The infraspinatus muscle originates from the inferior aspect of the spine of the scapula and the infraspinatus fossa. The infraspinatus runs in a superolateral position, occupying most of the greater tuberosity of the humerus. The infraspinatus tendon has a fan-like appearance, with its superior half being thick and long and the posterior half being thin and short.[47] The subscapularis muscle arises from the anterior aspect of the scapula and has about nine tendonous bands that merge laterally into the flattened tendon.[48]

The majority of the tendonous portion of the muscle inserts upon the lesser tuberosity of the humerus, but the superior aspect has tendon fibers that cross over the bicipital groove, which covers the biceps tendon and extend across to the greater tuberosity. There it blends with the supraspinatus tendon, which helps in creating the suspension pulley for the biceps tendon. The inferior third of the subscapularis muscle attaches to the inferior portion of the lesser tuberosity and the anterior aspect of the humeral shaft. The superior and inferior subscapularis muscular components have innervations by the upper and lower subscapular nerves. The coracohumeral interval is the space that lies between the humeral head and the coracoid process where insertion of the subscapularis muscle occurs.[49]


MRI shoulder arthrography is typically ordered to assess anatomical structures poorly visualized without the use of intra-articular contrast. These structures include the biceps-labral complex, glenohumeral ligaments, labrum, rotator interval, and the joint capsule.[50] 

Injury to one or more of these structures can result in shoulder instability. Patients with shoulder instability complain of pain and a feeling of looseness of the joint. They often experience repeated dislocations. In a patient with shoulder instability, MRI arthrography allows for the accurate assessment of the glenoid labrum, glenohumeral articular cartilage/glenoid bone stock, capsular tears, glenohumeral ligament tears, and rotator cuff tears, particularly following prior rotator cuff repair.[51] 

Again, plain radiographs of the glenohumeral joint should be performed before an arthrogram, as discussed above. Arthrography can also be utilized for suspected adhesive capsulitis of the glenohumeral joint or in patients with persistent symptoms postoperatively from orthopedic shoulder surgery.


Contraindications for glenohumeral arthrography include active infection of the joint of interest, overlying soft tissues, or skin adjacent to the needle path. The presence of cellulitis overlying the glenohumeral joint is a contraindication as infection can be spread into the joint space when the needle is traversing its course from the overlying skin into the joint capsule.[10]

Reflex sympathetic dystrophy, RSD, is a contraindication as it can be reactivated post joint injection. RSD is a form of complex regional pain syndrome, a chronic condition characterized by severe and burning pain, typically affecting the extremities. There can be pathological changes to the skin and bone, tissue swelling, excessive sweating, and allodynia. RSD is triggered by a tissue injury with no underlying associated nerve injury and is thought to be secondary to a malfunctioning of the sympathetic nervous system.[52]

Relative contraindications include a patient who has a history of a contrast allergy. In such patients, a prophylaxis protocol should be followed or consider using normal saline to achieve shoulder distention for MR arthrography. If an allergic reaction occurs with MR arthrogram, it is typically related to either the local anesthetic or the iodine solution rather than the gadolinium solution. In anticoagulated patients, arthrography can typically be performed when the INR is less than 1.5 to 2.0.[53] There is a risk-to-benefit ratio to consider withholding anticoagulation in patients undergoing this procedure. Using a smaller gauge needle may be considered in patients who use an anticoagulant.

CT arthrography should be considered an alternative technique when MRI is neither feasible nor possible, including when an MRI is contraindicated or no MRI is available. If a patient is claustrophobic or has surgical hardware, CT arthrography should be considered. Movement artifacts also promote the usage of CT over MRI. CT can also serve as a salvage procedure when MRI proves to be inconclusive.[54]

An additional advantage of CT over MRI includes a submillimeter resolution in most standard multidetector scanners. CT is typically more accessible and allows for fast acquisition. CT can also better characterize osseous structures like a bony Bankart or a Hill Sachs lesion. It also better demonstrates pathological calcifications like calcific tendinitis or a Bennett lesion. CT is less susceptible to artifacts.[21]

The use of gadolinium-based contrast agents in pregnancy should be considered. These contrast agents can cross the placental barrier, enter the fetal circulation, and pass through the kidneys into the amniotic fluid. No definitive adverse effects of the gadolinium-based contrast agents on the human fetus have been documented, but the potential for bioeffects upon exposure is not understood well.[55]

The usage of gadolinium-based contrast agents should be avoided in pregnancy unless there is no alternative imaging and the benefits outweigh the potential risks to the fetus. A small fraction of this contrast administered to a lactating woman is excreted into their breast milk. A small portion of excreted milk is absorbed by the infant's gut. The minute amount absorbed by the nursing infant's gastrointestinal tract is unlikely to be harmful. If there's concern by the referring physician, radiologist, or the patient, it can be advised for the mother to discard breastmilk for 24 hours post-gadolinium contrast administration. IV administration of the contrast and neonates and infants is safe and performed routinely. With this information and the fact that temporary disruption of breastfeeding is stressful for both infant and mother, the recommendation for suspension of breastfeeding for 24 hours is unnecessary.[55]


An arthrogram procedure involves the injection of contrast into the glenohumeral joint space to maintain a sterile field; aseptic technique materials are necessary, including skin cleansers, sterile drapes, and sterile swabs.[56] Contrast should be prepared as noted below. Syringes are utilized: a 20 mL syringe for preparation, a 10 mL syringe for injection, and a 1 mL syringe for the gadolinium contrast.[57]

A 25 gauge needle for injection of local anesthetic and a 20 to 22 gauge spinal needle for the arthrogram itself are utilized. Image guidance using fluoroscopy, ultrasound, or computer tomography is utilized. 1% lidocaine can be used to anesthetize the skin, subcutaneous tissue, and projected tract that the needle will traverse. For CT arthrography, a non-ionic iodinated contrast is required. The injected solution is diluted to a one-to-one ratio with normal saline or lidocaine. For MR arthrography, a gadolinium contrast is utilized.

A variety of contrast mixture protocols exist. A common protocol is included here. In a 20 mL syringe, a mixture of 5 mL of iodinated contrast with 15 mL normal saline and gadolinium contrast at approximately two millimoles per liter should be made. The usage of epinephrine can be considered 0.2 mL, 1 in 1000 solution if there is a delay expected before the MR arthrogram. An option could include using a long-acting anesthetic for a diagnostic exam or a corticosteroid. A therapeutic test can aid in determining if the pain the patient is experiencing is caused by an intra-articular source where a long-acting anesthetic such as ropivacaine 0.2% or bupivacaine 0.5% can be utilized. Some reports have determined a lower in vitro toxicity with ropivacaine instead of bupivacaine or lidocaine.[58]

There are reports of cartilage toxicity from these long-acting anesthetics. The corticosteroid can be injected after the injection of contrast, so it does not compromise the diagnostic exam. A corticosteroid should not be injected if an infection is suspected. CT or MRI follows the arthrogram to capture further imaging post-injection.


A referring physician typically orders the exam. This physician should know the indications and contraindications for the arthrogram procedure and any risks or benefits for the patient to undergo the examination. The physician should discuss this with the patient. A physician/radiologist is responsible for all aspects of the study, which include reviewing indications for the arthrogram with MRA or CTA, specifying the performed pulse sequences, the interpretation of images acquired, making the final reports, and assuring both the quality of the images and the interpretation of such.[2]

The radiologist should apply the current knowledge about the gamut of MRI contrast agents to include the composition, choice, risks, and benefits of the agent to be used and appropriate dosing. The radiologist will perform the arthrogram at hand. A physician assistant, nurse practitioner, or registered radiologist assistant who is certified as an advanced level radiographer can perform patient assessment, management, and selected examinations, including arthrograms, under radiologist supervision. The radiology technologist participates directly in assuring the patient's comfort and safety, preparing and positioning the patient to undergo the MRI, and obtaining the MRI data for interpretation by the radiologist.[3]


As with any procedure, informed consent, including the risks, benefits, and alternatives, is discussed with the patient and is obtained before the procedure. The joint, as well as the laterality, should be confirmed. A history of allergies, as well as reactions, should also be reviewed. If a patient is to undergo an MRI after the arthrogram, it must be confirmed that the patient is safe to undergo such a procedure prior. This includes asking about pacemaker status or spinal stimulator presence.[59]

For an elective arthrogram, management of a patient’s anticoagulation is typically performed by the referring physician. This involves holding the anticoagulation medication for an appropriate number of days and achieving a periprocedural INR less than 2.0 for a patient on warfarin.[53] If the patient is on an anticoagulant, the increased risk of bleeding is rare, but the possibility of a resulting hematoma with the need for evacuation should be discussed with the patient. 

If a patient is undergoing direct arthrography, an injection into the glenohumeral joint is performed under fluoroscopy, CT, or ultrasound guidance. The contrast will depend on whether a patient is undergoing a CT or an MRI arthrography. The CT injectate is an iodinated contrast with a concentration of no greater than 240 mg of iodine per mL.[58]

The MRI injectate is a diluted solution, 1/200 to 1/250, of gadolinium containing an MRI contrast agent used in sterile saline, which reaches a concentration level of 0.0020 to 0.0025 mmol/mL. Some institutions use a mixture of local anesthetic with normal saline to reduce artifact and periprocedural pain. A typical solution includes 10 mL normal saline, 10 mL local anesthetic like 0.5% ropivacaine, and a 0.1 mL gadolinium MRI contrast agent. While the glenohumeral joint has a capacity of approximately 8 to 15 mL, a 12 mL intra-articular injection of the typical solution is adequate for joint distention.[60]

If the injection is fluoroscopically performed or under CT guidance, a small volume of 1 to 3 mL of dilute iodinated contrast can be injected first to confirm the intra-articular positioning. The MRI scanning should be performed within 90 minutes post-intra-articular injection to optimize imaging quality.


A shoulder arthrogram is typically performed to assess the rotator cuff complex and detect labral tears. When performed in conjunction with MRI, this provides a detailed evaluation of the cartilage and structures within the glenohumeral joint itself. The imaging technique should both be accurate and atraumatic for the patient. If contrast media is improperly placed, this can simulate pathologic conditions, and an improper needle placement can cause injury. A systematic approach should be utilized when performing a shoulder arthrogram to minimize injury to articular structures and perform the technique effectively. A variety of methods of fluoroscopically guided shoulder arthrograms have been described since their first utilization in 1933.[61]

An initial report describes positioning the patient obliquely to place the glenohumeral joint space in profile while the needle is advanced into the space. In 1975, that was replaced by a supine position for the patient to avoid injuring the glenoid labrum.[62] In 1978, a report suggested that the needle be advanced with the stylet apparatus removed and local anesthetic placed within the needle hub.[63]

Utilizing this technique, a drop in the fluid level of the anesthetic would indicate that the tip of the needle was in the correct placement within the intra-articular space. In 1980, Neviaser suggested that the shoulder be internally rotated once the needle tip was at the humeral head to facilitate intra-articular placement.[64]

A posterior approach for shoulder arthrograms has been proposed in the more recent literature. Aspects of the previously mentioned techniques are incorporated within the fluoroscopically guided anterior approach and will be further discussed.

Radiographic images are essential for performing an arthrogram. The Scout radiograph includes anteroposterior radiographs of the shoulder on internal and external rotation. These views allow for assessing for calcium hydroxyapatite deposition in the rotator cuff complex. If calcium deposits are not identified before injection of the contrast medium, these depositions could be interpreted as an extension of the contrast agent into the rotator cuff and be diagnosed as a tendon tear. Also, depositions within a tendon may not necessarily be visualized on MRI because both the tendon and the calcification appear as regions of low signal intensity.[65]

Patient positioning is essential to ensure successful arthrogram examination and minimize complications. Ideal patient positioning for fluoroscopically guided arthrography of the shoulder from an anterior approach is with the patient in the supine position and the shoulders slightly externally rotated, using a rotator interval approach.[66] Next, the target is positioned in the center of the field of view, magnified and collimated. With the rotator interval approach, fewer anatomical structures are transgressed, and the usage of shorter needles can be implemented. The needle path is shorter at the level of the rotator interval with easy avoidance of the long head of the biceps tendon with the arm placed in an external rotation. With the alternative, the Schneider technique, the anteroinferior labrum, and subscapularis tendon can be perforated by the needle or even impregnated by the contrast agent used.[67] The Schneider technique utilizes a straight anteroposterior approach to the glenohumeral joint targetting the junction of the middle and lower thirds of the joint in a vertical approach under fluoroscopy.[68]

The Schneider technique has the same preparation as the rotator interval. It targets the caudal third of the glenohumeral joint. In this approach, the needle may pass through important anterior structures such as the subscapularis tendon and the anteroinferior labroligamentous complex. This may distort these structures and lead to difficulty interpreting the images or misinterpretation.[67] 

A posterior approach for a fluoroscopically guided shoulder arthrogram has been described. In this approach, the patient is in the prone position, and the ipsilateral shoulder is raised off of the table with a pad which allows the glenohumeral joint to be visualized tangentially. The posterior approach has been suggested for a patient for whom the anterior approach should be avoided. This can be secondary to a pathologic condition to avoid any iatrogenic injury and any interpretive errors with the MRI. In the posterior approach glenohumeral arthrogram, the patient is in the prone position with their arm placed into external rotation. The palm is facing downwards.[69] The fluoroscopy tube is then perpendicular to the shoulder joint. Skin is marked lateral to the medial articular cortex of the humerus at the level of the coracoid. In this approach, the needle traverses the infraspinatus tendon. 

The rotator interval approach is an effective, rapid, and easily performed technique, which will be further discussed. With this positioning, the coracoid process is easily avoided.[66] The supine positioning of the patient creates an oblique orientation of the glenoid articular surface. On the frontal radiograph, the posterior aspect of the glenoid projects over the humeral head, whereas the anterior aspect of the glenoid lies medial to the humeral head.[68]

With the patient in the supine position, a needle tip that is advanced in the anteroposterior direction towards the humeral head will thus not come in contact with the glenoid labrum. The oblique positioning is important. Using a posterior oblique position or a Grashey view will place the glenohumeral joint in profile, but it also causes rotation of the glenoid labrum in the direct path of the advanced needle tip when the radiologist aims for the clear space of the glenohumeral joint. The labrum in this circumstance could be penetrated with needle advancement. Theoretically, if the needle was continually advanced between the humerus in the glenoid labrum without contacting bone, the needle tip could be advanced and penetrate the posterior aspect of the rotator cuff. Utilizing external rotation of the glenohumeral joint exposes more of the articular surface of the humeral head anteriorly but also increases the intra-articular area that is available for insertion of the needle via an anterior approach.[68] 

External rotation aids in ensuring that the biceps tendon is not inadvertently injured during the arthrogram. To assist the patient in maintaining external rotation, a weighted object or sandbag may be placed in the patient's hand. Excessive external rotation of the glenohumeral joint is uncomfortable for the patient and can also increase the tightness of the anterior capsule. This tightness will place the anterior aspect of the joint capsule closer in proximity to the humerus, thus decreasing the joint space within the anterior recess and making intra-articular needle placement more difficult.[68]

Once the patient is in the proper supine position as described, the next focus is needle insertion. The target for the procedure is the medial upper quadrant of the humeral head. A radiopaque instrument and fluoroscopy are used to localize the joint insertion site and to ensure that it is not too lateral nor too medial. The skin is prepped and marked over the humeral head to indicate where the needle insertion will occur. Optimum positioning is located at the level of the coracoid at the superomedial humeral head. Even though the articular space extends laterally, connecting both the greater and lesser tuberosities, the intra-articular space is increased within the anteroposterior dimension located at the medial aspect of the humeral head.

The skin should not be marked medial to the medial cortex as needle insertion in this site could jeopardize the anterior labrum. It should be centered fluoroscopically to avoid any distortion or inaccurate needle placements. Once the overlying skin is marked for the entrance site of the needle, the area is prepped and draped in the usual sterile fashion. Next, the skin and subcutaneous tissues are anesthetized. A skin wheal is created using a buffered 1% lidocaine. The lidocaine should be infiltrated into the deeper tissues along the projected track of the needle for the arthrogram. The needle of choice varies by institution, but a 20 or 22-gauge 3.5-inch spinal needle is typically used.[68]

A stylet is useful during the advancement of the needle as it reduces soft tissue injury and avoids obstruction of the needle by the soft tissues themselves. The needle is inserted by an anterior approach in an anteroposterior position. The needle hub should be centered over the tip to ensure that advancement is perpendicular to the fluoroscopy beam. If a patient experiences any discomfort or pain during the advancement of the needle, more local anesthetic can be utilized.

The intra-articular needle placement is crucial. With intermittent fluoroscopy, the needle is advanced posteriorly until at the site of the humeral head. Once there is contact with the articular cartilage and the underlying cortex of the humeral head, the needle has a typical appearance one of two ways at fluoroscopy. In one circumstance, the needle is obliquely angled toward the glenoid due to the needle initially being aimed at a portion of the humeral head that curves away from the inserted needle. The needle has a natural deviation in this direction once contact is made. This needle position is likely to occur if the initial skin mark is placed close to the medial cortex of the humeral head. Another circumstance is where the needle remains in the anteroposterior direction as the tip of the needle did not slip down the curvature of the humeral head.[66]

Once the intra-articular needle is in proper placement, the stylet is removed, and the syringe plunger is drawn backward to exclude the intravascular placement of the needle. A couple of drops of the contrast can be dripped into the needle hub to prevent any injection of air bubbles into the joint space. Extension tubing is then attached. A test injection of local anesthetic is made of approximately 1 to 2 mL of anesthetic not only anesthetizes the joint but also determines the needle tip positioning within the compartment.[66]

The utilization of an anesthetic within the injected solution can also have a functional purpose as a diagnostic procedure. This anesthetic ensures that the patient's pain does come from the glenohumeral joint if the pain is alleviated with its usage. Low resistance to injection will be present if the needle tip is within an adjacent bursa or the glenohumeral joint. When the needle is oriented obliquely, the test injection will also have a low level of resistance.[67]

If the needle is oriented in the anteroposterior direction, low resistance to this tester injection may not occur. The needle tip in this circumstance may be within the hyaline cartilage. The needle can be gently manipulated by slightly moving it away from the humerus or rotating it.

If there is an abrupt decrease in this resistance to injection, this indicates that the tip of the needle is located in a compartment, whether the joint space or bursa. The needle must not be pulled back more than a few milliliters as the tip could enter the subacromial-subdeltoid bursa. If attempting to manipulate the needle does not obtain the desired results, the needle may be medially directed; in this case, it should not be advanced as such could penetrate the glenoid labrum and lead to cartilaginous risk for injury. If a high resistance to the injection is met, the tip of the needle might be tightly opposed to the joint surfaces. Turning the bevel of the needle until the anesthetic easily flows or retracting 1 to 2 mm while injecting can be attempted. If high resistance is met, the needle may be exterior to the joint or in the capsule. The needle may also be plugged.[66] If high resistance cannot be corrected, a different approach can be attempted.

Signs that the needle is intra-articular position include a give and resistance after the capsule has been passed into the joint space. There may be a bony touch from the tip of the needle. Imaging serves as evidence of the intra-articular position. Low resistance with a small amount of injected fluid is a sign of the intra-articular position, but it can also be felt if the tip of the needle is in a tendon sheath or periarticular bursa.[70]

There should be either a spontaneous or aspirated return of joint fluid before the injection of contrast. Using a local anesthetic for the test injections rather than iodinated contrast materials can be utilized as the latter could mask the area and thus make needle visualization more challenging. The gadolinium compound should not be utilized for the tester injection before the MRI, as any inadvertent injection outside of the glenohumeral joint during needle placement can be misinterpreted either as a capsular injury or rotator cuff tear on MRI.[66]

Once the anesthetic test injection indicates that the tip of the needle is within the adjacent bursa or the glenohumeral joint, then iodinated contrast material is used to differentiate between these two. A connecting tube is used between the syringe and the needle. Lower resistance to injection should still be present. The movement of contrast material away from the tip of the needle is consistent with a bursal location or an intra-articular one. To confirm an intra-articular placement of the contrast, a column of contrast agent between the humerus in the glenoid should be observed.[68]

Any free-flowing contrast material that pools over the head of the humerus can be located in the posterior glenohumeral joint recess or anteriorly within the subacromial-subdeltoid bursa. If contrast is not visualized between the humerus and the glenoid, gentle inferior traction upon the patient's arm can be applied to create a negative pressure within the joint space which will pull the contrast agent within the joint space properly and thus confirm the intra-articular placement. The intra-articular contrast material can also be identified within the axillary recess, subcapsular recess, and long head of the biceps tendon sheath, as these are normal communications within the glenohumeral joint.[67]

If the tip of the needle is not within the adjacent bursa or the joint, then an irregular pooling of contrast will occur at the needle tip. Overall, the flow of the injected contrast medium should be away from the tip of the needle and demonstrate opacification of the joint space and that of the subcoracoid recess, which confirms intra-articular positioning.[66] Following the injection, the needle is removed and a bandage placed. The intra-articular position of the needle, as well as the injection with fluoroscopic spot views, should be documented in both internal and external rotation images.

Once there is confirmation that the needle is placed in the intraarticular space, the iodinated contrast material is injected for fluoroscopic imaging or possibly CT. If an MRI examination is to follow, then a mixture of iodinated contrast and gadolinium can be injected. A combination of 0.1 mL of gadopentetate dimeglumine, 10 mL of sterile 0.9% sodium chloride, 10 mL of iohexol, and 0.3 mL of epinephrine within a 20 mL syringe is created. The combination of radiolucent gadolinium with the radiopaque iodinated contrast allows for the continual visualization of the injection and more information during fluoroscopy. A portion of the saline in this mixture can be substituted for a local anesthetic such as ropivacaine or bupivacaine. Also, even if the MRI cannot be completed, valuable information regarding the integrity of the rotator cuff is available from the arthrogram images taken. Iodinated contrast lowers the signaling intensity of low osmolality gadopentetate dimeglumine solutions; the signal intensity intraarticularly on MRI is adequate.[66]

To reduce delayed discomfort or pain for the patient, a non-ionic iodinated contrast agent is utilized. The addition of epinephrine delays the absorption of the contrast agent from the joint cavity. This delaying effect can serve an important role depending on the time interval between the joint injection and the MRI. Notably, the epinephrine intraarticularly can increase arthrographic pain. This pain is either because it is directly irritating the synovium or because its use increases the amount of time the synovium is exposed to the contrast agent used. If a patient is undergoing the MRI, the introduction of air bubbles must be minimized during any part of the injections as they produce artifacts on imaging.

To distend the shoulder, an injection of approximately 10 to 16 mL is typically used. An injection of more than 15 mL does have an increased risk for extra-articular leakage that can be mistaken as a rotator cuff tear and be a source of postprocedural pain.[66] A patient with chronic shoulder subluxation may require a higher volume, in contrast to a patient suffering from adhesive capsulitis who would require less. Approximately less than 7 mL of contrast medium can be used in a patient with adhesive capsulitis. If there is an increase in resistance to injection or retrograde flow of the contrast material from the connection tubing after disconnecting the syringe, this indicates adequate joint distention. The optimal volume of 15 mL of intraarticular fluid for MR arthrography has been described. Regarding MRI quality, exercise after the shoulder arthrogram has neither a beneficial nor detrimental effect.[66]

In a patient with a shoulder prosthetic with a high clinical suspicion for a full-thickness rotator cuff tear, the rotator interval approach is employed. Skin markings are made using the prosthetic humeral head component.[71] For these patients, saline and gadolinium are omitted from the typical solution used in the arthrogram. Only lidocaine and iodinated contrast agent are injected. Post injection, the needle is removed then the shoulder is taken through different ranges of motion to visualize any abnormal extension of contrast into the subacromial-subdeltoid bursa.

In more recent data, modifications to arthrography have been used, such as that of targeting the articular recesses with or without ultrasound guidance have been noted. The purpose of using the articular recess as a target rather than the joint space can be optimum when the joint space is not accessible. A non-accessible joint space can be due to an overlapping normal bone structure or severe degenerative joint changes like osteophytes. This targeting also may aid in avoiding patient manipulation and tube angulation. Utilizing this technique, the needle is advanced until it comes into contact with the bone, which provides the depth limit to insertion. This has the potential to increase the safety of the arthrogram procedure and aid in avoiding the articular fibrocartilage, such as the labrum in the glenohumeral joint. This technique also is transposable to guidance via ultrasound. With this, the needle is best placed tangentially to the transducer rather than vertically. The advantages to utilizing ultrasound over fluoroscopy are the absence of ionizing radiation for the patient and the operator, the imaging of all the soft tissues that surround the glenohumeral joint, and the possibility of performing the ultrasound in an office setting.[57]

Indirect Arthrography

An indirect arthrogram is an MRI technique that produces the arthrogram images without direct injection into the joint.[72] A standard dosage of a gadolinium-containing contrast agent is intravenously injected. Typically, this intravenous administration of gadolinium is comprised of a standard dose of 0.1 mmol per kilogram body weight.[66]

Imaging of the shoulder is performed after a delay of between 5 to 15 minutes has elapsed. During this time, mobilization and exercising of the joint, in this case, the glenohumeral joint, is performed, facilitating the diffusion of contrast into the joint. The advantage of indirect arthrography is that it does not require direct injection into the joint, is less invasive for the patient, and lacks image guidance. Despite a good level of sensitivity and specificity for certain labral and rotator cuff tears, there are drawbacks. These include a lack of predictability of the degree of the intra-articular enhancement. With this technique, there is a diminished signal intensity of the intra-articular fluid, a lack of capsular distention, and a potential for the misinterpretation of imaging due to an enhancement of granulation tissue, synovial structures, vessels, bursa, and tendon sheaths. This diminished signal intensity is dependent on inflammation, the amount of joint fluid, and the synovial surface area. There is an inability to detect any non-anatomical communications between fluid-containing spaces due to simultaneous enhancement with this technique. The third drawback is a lack of joint distention. This joint distention will be present if a joint effusion already exists.[73]

Image Acquisition

For CT scan, axial images should be taken with a 0.625 mm thickness using spiral scanning with isotropic data acquisition. Multiplanar reformats should be oblique, parallel, and perpendicular to the glenoid or the belly of the supraspinatus muscle. CT arthrograms of the glenohumeral joint can adequately detect full thickness and articular surface partial tears to the infraspinatus and supraspinatus. Fatty degeneration and loss of volume to the rotator cuff muscles can also be demonstrated well utilizing CT.[70]

For MRI, the usage of a dedicated shoulder coil should be implemented. A T1 weighted fat suppressed sequence within the axial, oblique sagittal, and oblique coronal plains with a 3 to 4 mm slice thickness. A sagittal, non-suppressed T1 image is needed to detect fatty infiltration. A T2 weighted fat suppression in the coronal plane is used for detecting findings less well demonstrated in T1. The sagittal T2 imaging further optimizes the evaluation of rotator cuff tears in the anteroposterior (AP) plane. This also allows for evaluating signal changes in the bicipital tendon and muscle edema.[20]

Abduction with external rotation, ABER, positions the patient with the palm behind the head.[74] This requires oblique axial, parallel to the shaft of the humerus, has been described as an increased way to detect rotator cuff articular surface tears, the centering of the humeral head, and enhance the visibility of any non-displaced anteroinferior labral tears. With this positioning, the insertion of the supraspinatus muscle is relaxed, allowing the intra-articular fluid to enter more easily. This can improve the delineation of the extent of a lesion to the rotator cuff. Simultaneously, it can improve the evaluation of the inferior glenohumeral ligament and the inferior capsule.

Some Bankart lesions can be visualized where there is the periosteal attachment of the labrum due to the separation of the base of the labrum from the periosteum. This can aid in diagnosing Perthes lesions. This is particularly true in athletes with a superior labral tear from anterior to posterior, SLAP tears. However, some have found that this does not increase accuracy, but it does increase scan times in terms of positioning, using surface coil, and acquiring a new sequence.

Postprocedural care for the patient should be discussed before the arthrogram is performed. It is typically a safe procedure for the patient with discharge following the CT or MRI. A patient should receive education on the signs of infection. A patient should be given instructions for postoperative care and should call with symptoms such as fever or worsening pain. If a patient experiences an increase in joint pain that does not settle within 24 to 48 hours with any associated systemic symptoms like fever, this raises a concern for the possibility of a joint infection, septic arthritis. In this circumstance, the patient needs to be clinically assessed and may require joint aspiration, antibiotic therapy, or surgical lavage.[66]


Risks to the procedure include introducing infection into the joint space, allergic reactions, bleeding, contrast reactions, chondrolysis with a local anesthetic injection into the joint, and damage to any adjacent structures. While infection is rare, there is an increased risk for the elderly population and those with diabetes mellitus, skin infection, or rheumatoid arthritis.[75] A prophylactic antibiotic is unnecessary, but the patient should be informed about the possibility of infection and what signs to look for post-procedure.

A common complication is post-injection pain. This is likely related to synovitis, affecting up to 66% of patients within several hours post-procedure, but typically will resolve in days. Complications are few, but most commonly would include synovial irritation known as chemical synovitis that is induced by the injected iodinated contrast agent. This is a reactive type of synovitis. Chemical synovitis is typically a self-limiting process and is treated with over-the-counter anti-inflammatory medication such as ibuprofen or naproxen.[76] 

Currently, there are no known cases of nephrogenic systemic fibrosis related to arthrography. If there is a failure to access the glenohumeral joint space, this results in extra-articular contrast. Other complications include urticaria and vasovagal reaction/episodes.[75]

Clinical Significance

Image Interpretations

Glenohumeral arthrograms provide unique imaging signs. These include abnormal joint communications with other collections or compartments, labral, tendinous, and ligamentous tears or injuries, synovitis with an irregular capsular margin, cartilaginous abnormalities, and loose bodies. Within the shoulder, there are structures for which an arthrogram provides optimal assessment, including the ligamentous structures, capsule, glenoid labrum, rotator interval, biceps pulley, and biceps-labral complex. A rather thorough review of pathology is beyond the focus scope of this article, but there are some key pathological and anatomical comments to be made.

Normal Arthrogram 

When reading the arthrogram, if there is a normal appearance of the rotator cuff muscles and tendons, an absence of any damage to the glenoid cartilage or humeral head, lack of evidence of any fracture or dislocation to the osseous structures, this would be indicative of a normal shoulder arthrogram.[77]

In a normal glenohumeral arthrogram, thin, halo-like contrast should be seen covering the humeral head superiorly but separated from the bony margin via radiolucent articular cartilage. With external rotation, the thin halo-like contrast material should end abruptly lateral at the humeral anatomic neck. Both the soft tissues above and lateral to the greater tuberosity should be free of the opaque medium. A radiolucent filling defect can be visualized within the opaque-filled recess within the superior aspect of the glenohumeral joint, which is the long tendon of the biceps muscle.[78]

It can be possible to track the course of this tendon through the glenohumeral joint from the superior rim of the glenoid into the contrast-filled synovial sheath in the bicipital groove. The subscapularis recess, which often has a tongue shape, extends immediately from the glenohumeral space, travels under the coracoid process, and deep to the subscapularis. This recess communicates with the glenohumeral joint via a normal opening within the anterior glenoid labrum. This recess will serve as a prominent landmark on internal rotation, but on external rotation, it becomes less prominent as the taut, overlying subscapularis expresses the fluid.[77]

The axillary fold or recess is located at the inferior margin of the glenohumeral space and is a redundant capsule. When the arm is placed into abduction, the axillary fold is obliterated. There is an indentation between the axillary fold and subscapularis recess. Using the axillary view, a thin layer of the contrast medium can be visualized within the glenohumeral space that is separated from the bony margins by the articular cartilages. Both the anterior and posterior margins of the cartilage in the glenoid rim can appear as radiolucent triangular filling defects located within the synovial recesses. Anteriorly, a linear filling defect can be visualized, which represents the long tendon of the biceps muscle. Frequently, this can be visualized anteriorly as this tendon extends into an opaque-filled sheath within the bicipital groove. The subscapularis recess can be seen anterior to the glenoid.[77]

Abnormal Arthrogram

There is an abnormal communication between the subacromial bursa and the glenohumeral joint in the presence of a complete thickness rotator cuff tear. Contrast injected to the joint will then flow into the bursa, which is recognized as a collection of the contrast above and lateral to the greater tuberosity.[79] In a normal shoulder arthrogram, this area should be free of contrast material. On the axillary view, this bursa appears like that of a saddlebag hanging across the humerus at the surgical neck. 

The glenohumeral ligaments are thick bands within the anterior joint capsule. They have attachments to the proximal humerus as well as the glenoid margins. These ligaments are composed of three parts: the superior glenohumeral ligament, middle glenohumeral ligament, and inferior glenohumeral ligament.[80]

The glenoid labrum acts to increase the surface area and depth of the glenoid fossa to better accommodate the humeral head. It also serves as an attachment site for the anterior band of the inferior glenohumeral ligament. In regards to the configuration of the labrum, there is a significant variety in terms of size and shape, particularly in the superior labrum at the junction of the upper to middle thirds of the fossa.[81] Usually, there is a separation of the bursa from the radiolucent joint capsule. This area is occupied by the thick, muscular tendonous rotator cuff. Importantly, at times there may be no clear demarcation between the joint capsule in the bursa, which is likely secondary to either atrophy of the rotator cuff or a large tear of the superior aspect of the cuff.[81]

A close relationship exists between the bursa and the undersurface of the acromion process superiorly. The bursa extends in a downward fashion adjacent to the humeral head to a variable degree. The size or configuration of the bursa does not necessarily correlate to the severity of the lesion or injury to the rotator cuff complex. Instead, this can be an anatomical variation. A complete rupture of the rotator cuff has a distinct appearance and location of contrast filling the subacromial bursa.[82] Notably, there are potential errors to recognize. There can be an inadequate distribution of the contrast medium into the joint space. On the external rotation views, the biceps tendon sheath can fill with contrast and project lateral to the greater tuberosity, but this should not be mistaken as opacification of the subacromial bursa secondary to a rupture of the rotator cuff. On an internal rotation view, the tendon sheath, located within the bicipital groove, will move into medial positioning. Also, the subacromial bursa is larger in size and remains in a lateral position in both the internal and external rotation views.[82]

There may be an inadvertent injection of the bursa inside of the glenohumeral joint. This can cause an appearance of a cuff rupture, but instead is a bursogram and contrast material will not be present within the glenohumeral space or the subscapularis recess. It is important to see contrast located in the bursa and the glenohumeral joint to determine the presence of abnormal communication.

If there is a partial rupture or incomplete tear of the rotator cuff, the glenohumeral joint does not communicate with the bursa; hence the bursa will not opacify using arthrography. If a partial tear is located under the surface of the tendon, this may be visualized as an ulcer-like collection of the contrast medium above the head of the humerus, close to the anatomic neck. An incomplete tear deeper within the actual substance of the tendon can be more difficult to detect.[82]

In a patient with adhesive or retractile capsulitis, moderate pressure will be noticed to inject a small amount of fluid into the glenohumeral joint. In the arthrogram, the axillary and subscapularis recesses are either absent or very small in size. The bicipital tendon sheath will poorly fill or not fill at all. Patients with adhesive capsulitis may have relief of symptoms a few weeks post arthrography.

In a patient with shoulder instability with recurrent anterior dislocation of the shoulder, arthrography can show the location and severity of the soft tissue damage and aid in evaluating the integrity of the rotator cuff complex. In a patient with recurrent dislocations, the humeral head can detach the joint capsule from the glenoid rim and the neck of the scapula between the axillary fold and subscapularis recess. This forms an anterior pouch that can be visualized in the frontal view. This will be seen as a collection of a peak medium that extends from the subscapularis recess into the axillary fold. There's a border that is the normal demarcation between these structures. This pouch and its size can be most appreciated with an internal rotation view.[83] The anterior muscles will be relaxed with an internal rotation view, which is indicative of the degree of detachment of the capsule from the neck of the scapula and the glenoid. There may be evidence of a previous dislocation with an identifiable Hill Sachs compression defect within the humeral head.

There is an area lacking attachment to the glenoid labrum known as the sublabral foramen located in the anterosuperior quadrant. It is never located below the equator, which can serve as a mimic for a SLAP tear.[84] Criteria differentiate the biceps- labral sulcus from a SLAP tear. The following criteria include a smooth, well-defined region with a lack of contrast extension into the labrum. If contrast demonstrates a filling into the subacromial – subdeltoid bursa, this can indicate a full-thickness tear. The contour of the glenoid rim is followed. It must measure less than 5 mm in a medial collateral diameter. It is usually less than 3 mm in diameter. Superior labral anteroposterior lesions are a wide, complex range of lesions centered along with the long head of the biceps and can extend anterior and posterior to the tendon. Snyder described four types of these lesions, but six additional types are described. Type one demonstrates fraying as well as degeneration of the superior labrum which includes an irregular contour to the labrum and an increase in signal intensity in T2.

Type two, which is also the most common, demonstrates detachment of the superior labrum from the glenoid rim. Type three demonstrates a bucket handle tear and an inferior displacement of the central labrum. The tendon and anchor remain intact in type three. A type four is similar to a type three as it demonstrates a bucket handle tear with the same inferior displacement of the central labrum but includes displacement of the tendon and the anchor of the biceps. Types five through ten are subsequently categorized by the associated labral and ligamentous abnormalities demonstrated. A Buford complex is described in approximately 1.5% of shoulder arthroscopies, which is a variant that consists of a cord-like MGHL. The MGHL in this circumstance attaches directly to the superior labrum anteriorly to the biceps with an absent anterosuperior labrum.

A Bankart injury includes an avulsion of the anteroinferior labral ligamentous complex in conjunction with disruption of the scapular periosteum. When visualized with an osteochondral injury, this is known as a bony Bankart. A Perthes injury consists of the detachment of the anteroinferior labral ligamentous complex. The scapular periosteum is intact, although stripped. An anterior labral ligamentous periosteal sleeve avulsion, ALPSA, has similarities to the Perthes injury, but due to the periosteum, this allows a medial displacement and inferior rotation of the torn capsulolabral tissue. The labrum remains attached to the periosteum overlying the glenoid, hence the sleeve categorization. With a glenolabral articular disruption, GLAD, an anterior inferior superficial labral tear, is present and associated with an anterior, inferior articular cartilage injury to the glenoid. In this injury, the MR arthrogram will demonstrate the cartilage in this defect but also contrast that enters into the glenolabral junction.[85]

Enhancing Healthcare Team Outcomes

The radiologist's experience level, assessment in consensus, and the addition of the abduction and external rotation (ABER) view to improving the reproducibility and accuracy of the glenohumeral MRA are factors to consider for traumatic shoulder injury patients.

A prospective evaluation on the influence of the radiologist experience, consensus assessment, and the abduction and external rotation view upon the diagnostic performance of the MRA in patients with a traumatic anterior shoulder instability was assessed. 58 MRI examinations were utilized, 51 of which had additional ABER views. Six radiologists, R1 through R6, and three teams, T1 through T3, with differing experience levels, used a seven lesion standard scoring form; 45 of the 58 MRA exam findings were confirmed surgically. Sensitivity, specificity, Kappa coefficient, and differences in percent agreement or correct diagnosis were calculated per lesion and overall region types to assess the diagnostic accuracy and reproducibility. The results demonstrated that the most experienced radiologist, R1 and R2, and teams, T2 and T3, agreed significantly more in comparison to the lesser experienced radiologist, or T3 and T4 teams.

The most experienced radiologists and teams also had more consistent accuracy than the less experienced radiologists. Significant differences were found between radiologists noted in R1-R4 with a p=0.012, R3-R4 with a p=0.03, and teams T2-T3 with a p=0.014. There were significant differences between teams T1 and T2 and radiologists R3 and R4. Significant differences were discovered between T2 and R3, R4.[86]

There was no overall significant difference between the radiologist assessments with or without the ABER view. The addition of this view did not improve the overall diagnostic performance. The radiologist's experience level and consensus assessment contributed to higher accuracy and reproducibility. With this information, to enhance healthcare team outcomes, utilizing an experienced radiologist for assessing a patient with a traumatic shoulder injury can yield a more accurate MRA examination. The addition of the ABER view did not necessarily yield a significant difference. More training and exposure to MRA can likely aid in improving an inexperienced radiologist's accuracy and interpretation of the examination.

Article Details

Article Author

Kaitlyn Crossan

Article Editor:

David Rawson


4/7/2022 1:35:56 PM

PubMed Link:

Shoulder Arthrogram



Peterson JJ,Bancroft LW, History of arthrography. Radiologic clinics of North America. 2009 May;     [PubMed PMID: 19361665]


Rhee RB,Chan KK,Lieu JG,Kim BS,Steinbach LS, MR and CT arthrography of the shoulder. Seminars in musculoskeletal radiology. 2012 Feb;     [PubMed PMID: 22447233]


Helgason JW,Chandnani VP,Yu JS, MR arthrography: a review of current technique and applications. AJR. American journal of roentgenology. 1997 Jun;     [PubMed PMID: 9168710]


Dalinka MK,Kricun ME,Zlatkin MB,Hibbard CA, Modern diagnostic imaging in joint disease. AJR. American journal of roentgenology. 1989 Feb;     [PubMed PMID: 2643256]


Tischer T,Vogt S,Kreuz PC,Imhoff AB, Arthroscopic anatomy, variants, and pathologic findings in shoulder instability. Arthroscopy : the journal of arthroscopic     [PubMed PMID: 21871774]


Lugo R,Kung P,Ma CB, Shoulder biomechanics. European journal of radiology. 2008 Oct;     [PubMed PMID: 18511227]


Haapamaki VV,Kiuru MJ,Koskinen SK, Multidetector CT in shoulder fractures. Emergency radiology. 2004 Dec;     [PubMed PMID: 15517453]


Opsha O,Malik A,Baltazar R,Primakov D,Beltran S,Miller TT,Beltran J, MRI of the rotator cuff and internal derangement. European journal of radiology. 2008 Oct;     [PubMed PMID: 18387766]


Gemmel F,Van den Wyngaert H,Love C,Welling MM,Gemmel P,Palestro CJ, Prosthetic joint infections: radionuclide state-of-the-art imaging. European journal of nuclear medicine and molecular imaging. 2012 May;     [PubMed PMID: 22361912]


Khan AA,Mowla A,Shakoor MA,Rahman MR, Arthrographic distension of the shoulder joint in the management of frozen shoulder. Mymensingh medical journal : MMJ. 2005 Jan;     [PubMed PMID: 15695959]


Thomsen HS, Recent hot topics in contrast media. European radiology. 2011 Mar;     [PubMed PMID: 21136062]


Reimer P,Vosshenrich R, Off-label use of contrast agents. European radiology. 2008 Jun;     [PubMed PMID: 18274752]


Kalke RJ,Di Primio GA,Schweitzer ME, MR and CT arthrography of the knee. Seminars in musculoskeletal radiology. 2012 Feb;     [PubMed PMID: 22447237]


Narasimhan R,Shamse K,Nash C,Dhingra D,Kennedy S, Prevalence of subscapularis tears and accuracy of shoulder ultrasound in pre-operative diagnosis. International orthopaedics. 2016 May;     [PubMed PMID: 26585865]


Expert Panel on Musculoskeletal Imaging:.,Amini B,Beckmann NM,Beaman FD,Wessell DE,Bernard SA,Cassidy RC,Czuczman GJ,Demertzis JL,Greenspan BS,Khurana B,Lee KS,Lenchik L,Motamedi K,Sharma A,Walker EA,Kransdorf MJ, ACR Appropriateness Criteria{sup}®{/sup} Shoulder Pain-Traumatic. Journal of the American College of Radiology : JACR. 2018 May;     [PubMed PMID: 29724420]


Expert Panel on Musculoskeletal Imaging:.,Small KM,Adler RS,Shah SH,Roberts CC,Bencardino JT,Appel M,Gyftopoulos S,Metter DF,Mintz DN,Morrison WB,Subhas N,Thiele R,Towers JD,Tynus KM,Weissman BN,Yu JS,Kransdorf MJ, ACR Appropriateness Criteria{sup}®{/sup} Shoulder Pain-Atraumatic. Journal of the American College of Radiology : JACR. 2018 Nov;     [PubMed PMID: 30392607]


Handoll HH,Almaiyah MA,Rangan A, Surgical versus non-surgical treatment for acute anterior shoulder dislocation. The Cochrane database of systematic reviews. 2004;     [PubMed PMID: 14974064]


Woodward TW,Best TM, The painful shoulder: part II. Acute and chronic disorders. American family physician. 2000 Jun 1;     [PubMed PMID: 10865925]


Singh JP, Shoulder ultrasound: What you need to know. The Indian journal of radiology     [PubMed PMID: 23833420]


de Jesus JO,Parker L,Frangos AJ,Nazarian LN, Accuracy of MRI, MR arthrography, and ultrasound in the diagnosis of rotator cuff tears: a meta-analysis. AJR. American journal of roentgenology. 2009 Jun;     [PubMed PMID: 19457838]


Charousset C,Bellaïche L,Duranthon LD,Grimberg J, Accuracy of CT arthrography in the assessment of tears of the rotator cuff. The Journal of bone and joint surgery. British volume. 2005 Jun;     [PubMed PMID: 15911667]


Waldt S,Burkart A,Imhoff AB,Bruegel M,Rummeny EJ,Woertler K, Anterior shoulder instability: accuracy of MR arthrography in the classification of anteroinferior labroligamentous injuries. Radiology. 2005 Nov;     [PubMed PMID: 16244267]


Applegate GR,Hewitt M,Snyder SJ,Watson E,Kwak S,Resnick D, Chronic labral tears: value of magnetic resonance arthrography in evaluating the glenoid labrum and labral-bicipital complex. Arthroscopy : the journal of arthroscopic     [PubMed PMID: 15525929]


Chandnani VP,Yeager TD,DeBerardino T,Christensen K,Gagliardi JA,Heitz DR,Baird DE,Hansen MF, Glenoid labral tears: prospective evaluation with MRI imaging, MR arthrography, and CT arthrography. AJR. American journal of roentgenology. 1993 Dec;     [PubMed PMID: 8249731]


Zappia M,Di Pietto F,Aliprandi A,Pozza S,De Petro P,Muda A,Sconfienza LM, Multi-modal imaging of adhesive capsulitis of the shoulder. Insights into imaging. 2016 Jun;     [PubMed PMID: 27107871]


Culham E,Peat M, Functional anatomy of the shoulder complex. The Journal of orthopaedic and sports physical therapy. 1993 Jul;     [PubMed PMID: 8348135]


Clark JM,Harryman DT 2nd, Tendons, ligaments, and capsule of the rotator cuff. Gross and microscopic anatomy. The Journal of bone and joint surgery. American volume. 1992 Jun;     [PubMed PMID: 1624486]


Akpinar S,Ozkoç G,Cesur N, [Anatomy, biomechanics, and physiopathology of the rotator cuff]. Acta orthopaedica et traumatologica turcica. 2003;     [PubMed PMID: 14578660]


Rudez J,Zanetti M, Normal anatomy, variants and pitfalls on shoulder MRI. European journal of radiology. 2008 Oct;     [PubMed PMID: 18423935]


Prescher A, Anatomical basics, variations, and degenerative changes of the shoulder joint and shoulder girdle. European journal of radiology. 2000 Aug;     [PubMed PMID: 10963915]


Di Giacomo G,Itoi E,Burkhart SS, Evolving concept of bipolar bone loss and the Hill-Sachs lesion: from     [PubMed PMID: 24384275]


Lecouvet FE,Dorzée B,Dubuc JE,Vande Berg BC,Jamart J,Malghem J, Cartilage lesions of the glenohumeral joint: diagnostic effectiveness of multidetector spiral CT arthrography and comparison with arthroscopy. European radiology. 2007 Jul;     [PubMed PMID: 17186246]


Kralinger F,Aigner F,Longato S,Rieger M,Wambacher M, Is the bare spot a consistent landmark for shoulder arthroscopy? A study of 20 embalmed glenoids with 3-dimensional computed tomographic reconstruction. Arthroscopy : the journal of arthroscopic     [PubMed PMID: 16581456]


Ozaki R,Nakagawa S,Mizuno N,Mae T,Yoneda M, Hill-sachs lesions in shoulders with traumatic anterior instability: evaluation using computed tomography with 3-dimensional reconstruction. The American journal of sports medicine. 2014 Nov;     [PubMed PMID: 25231817]


Dunham KS,Bencardino JT,Rokito AS, Anatomic variants and pitfalls of the labrum, glenoid cartilage, and glenohumeral ligaments. Magnetic resonance imaging clinics of North America. 2012 May;     [PubMed PMID: 22469401]


Giaroli EL,Major NM,Lemley DE,Lee J, Coracohumeral interval imaging in subcoracoid impingement syndrome on MRI. AJR. American journal of roentgenology. 2006 Jan;     [PubMed PMID: 16357410]


Alyas F,Curtis M,Speed C,Saifuddin A,Connell D, MR imaging appearances of acromioclavicular joint dislocation. Radiographics : a review publication of the Radiological Society of North America, Inc. 2008 Mar-Apr;     [PubMed PMID: 18349451]


De Maeseneer M,Van Roy F,Lenchik L,Shahabpour M,Jacobson J,Ryu KN,Handelberg F,Osteaux M, CT and MR arthrography of the normal and pathologic anterosuperior labrum and labral-bicipital complex. Radiographics : a review publication of the Radiological Society of North America, Inc. 2000 Oct;     [PubMed PMID: 11046163]


Major NM,Browne J,Domzalski T,Cothran RL,Helms CA, Evaluation of the glenoid labrum with 3-T MRI: is intraarticular contrast necessary? AJR. American journal of roentgenology. 2011 May;     [PubMed PMID: 21512082]


Schaeffeler C,Waldt S,Holzapfel K,Kirchhoff C,Jungmann PM,Wolf P,Stat D,Schröder M,Rummeny EJ,Imhoff AB,Woertler K, Lesions of the biceps pulley: diagnostic accuracy of MR arthrography of the shoulder and evaluation of previously described and new diagnostic signs. Radiology. 2012 Aug;     [PubMed PMID: 22692037]


Petchprapa CN,Beltran LS,Jazrawi LM,Kwon YW,Babb JS,Recht MP, The rotator interval: a review of anatomy, function, and normal and abnormal MRI appearance. AJR. American journal of roentgenology. 2010 Sep;     [PubMed PMID: 20729432]


Beltran J,Bencardino J,Mellado J,Rosenberg ZS,Irish RD, MR arthrography of the shoulder: variants and pitfalls. Radiographics : a review publication of the Radiological Society of North America, Inc. 1997 Nov-Dec;     [PubMed PMID: 9397454]


Massengill AD,Seeger LL,Yao L,Gentili A,Shnier RC,Shapiro MS,Gold RH, Labrocapsular ligamentous complex of the shoulder: normal anatomy, anatomic variation, and pitfalls of MR imaging and MR arthrography. Radiographics : a review publication of the Radiological Society of North America, Inc. 1994 Nov;     [PubMed PMID: 7855336]


Williams MM,Snyder SJ,Buford D Jr, The Buford complex--the     [PubMed PMID: 8086014]


Vinson EN,Wittstein J,Garrigues GE,Taylor DC, MRI of selected abnormalities at the anterior superior aspect of the shoulder: potential pitfalls and subtle diagnoses. AJR. American journal of roentgenology. 2012 Sep;     [PubMed PMID: 22915393]


Kolts I, A note on the anatomy of the supraspinatus muscle. Archives of orthopaedic and trauma surgery. 1992;     [PubMed PMID: 1389774]


Vad VB,Southern D,Warren RF,Altchek DW,Dines D, Prevalence of peripheral neurologic injuries in rotator cuff tears with atrophy. Journal of shoulder and elbow surgery. 2003 Jul-Aug;     [PubMed PMID: 12934025]


Boon JM,de Beer MA,Botha D,Maritz NG,Fouche AA, The anatomy of the subscapularis tendon insertion as applied to rotator cuff repair. Journal of shoulder and elbow surgery. 2004 Mar-Apr;     [PubMed PMID: 14997093]


Dong Q,Jacobson JA,Jamadar DA,Gandikota G,Brandon C,Morag Y,Fessell DP,Kim SM, Entrapment neuropathies in the upper and lower limbs: anatomy and MRI features. Radiology research and practice. 2012;     [PubMed PMID: 23125929]


Blanchard TK,Bearcroft PW,Constant CR,Griffin DR,Dixon AK, Diagnostic and therapeutic impact of MRI and arthrography in the investigation of full-thickness rotator cuff tears. European radiology. 1999;     [PubMed PMID: 10354876]


Roy EA,Cheyne I,Andrews GT,Forster BB, Beyond the Cuff: MR Imaging of Labroligamentous Injuries in the Athletic Shoulder. Radiology. 2016 Apr;     [PubMed PMID: 26989937]


Genant HK,Kozin F,Bekerman C,McCarty DJ,Sims J, The reflex sympathetic dystrophy syndrome. A comprehensive analysis using fine-detail radiography, photon absorptiometry, and bone and joint scintigraphy. Radiology. 1975 Oct;     [PubMed PMID: 51504]


Patel IJ,Davidson JC,Nikolic B,Salazar GM,Schwartzberg MS,Walker TG,Saad WA,Standards of Practice Committee, with Cardiovascular and Interventional Radiological Society of Europe (CIRSE) Endorsement., Consensus guidelines for periprocedural management of coagulation status and hemostasis risk in percutaneous image-guided interventions. Journal of vascular and interventional radiology : JVIR. 2012 Jun;     [PubMed PMID: 22513394]


Fritz J,Fishman EK,Small KM,Winalski CS,Horger MS,Corl F,McFarland E,Carrino JA,Fayad LM, MDCT arthrography of the shoulder with datasets of isotropic resolution: indications, technique, and applications. AJR. American journal of roentgenology. 2012 Mar;     [PubMed PMID: 22358004]


Elster AD,Link KM,Carr JJ, Patient screening prior to MR imaging: a practical approach synthesized from protocols at 15 U. S. medical centers. AJR. American journal of roentgenology. 1994 Jan;     [PubMed PMID: 8273665]


Kho J,Azzopardi C,Davies AM,James SL,Botchu R, Direct MR arthrography of the shoulder: current practice in the UK. La Radiologia medica. 2020 Jul;     [PubMed PMID: 32026189]


Lungu E,Moser TP, A practical guide for performing arthrography under fluoroscopic or ultrasound guidance. Insights into imaging. 2015 Dec;     [PubMed PMID: 26493836]


Andreisek G,Duc SR,Froehlich JM,Hodler J,Weishaupt D, MR arthrography of the shoulder, hip, and wrist: evaluation of contrast dynamics and image quality with increasing injection-to-imaging time. AJR. American journal of roentgenology. 2007 Apr;     [PubMed PMID: 17377051]


Kanal E,Shellock FG,Lewin JS, Aneurysm clip testing for ferromagnetic properties: clip variability issues. Radiology. 1996 Aug;     [PubMed PMID: 8685361]


Czerny C,Hofmann S,Neuhold A,Tschauner C,Engel A,Recht MP,Kramer J, Lesions of the acetabular labrum: accuracy of MR imaging and MR arthrography in detection and staging. Radiology. 1996 Jul;     [PubMed PMID: 8657916]


Killoran PJ,Marcove RC,Freiberger RH, Shoulder arthrography. The American journal of roentgenology, radium therapy, and nuclear medicine. 1968 Jul;     [PubMed PMID: 5665187]


Schneider R,Ghelman B,Kaye JJ, A simplified injection technique for shoulder arthrography. Radiology. 1975 Mar;     [PubMed PMID: 1118584]


Wills JS,Diznoff SB, A modified technique for needle localization in arthrography of the shoulder. Radiology. 1978 Sep;     [PubMed PMID: 674677]


Neviaser TJ, Arthrography of the shoulder. The Orthopedic clinics of North America. 1980 Apr;     [PubMed PMID: 7432760]


Burk DL Jr,Karasick D,Mitchell DG,Rifkin MD, MR imaging of the shoulder: correlation with plain radiography. AJR. American journal of roentgenology. 1990 Mar;     [PubMed PMID: 2106220]


Dépelteau H,Bureau NJ,Cardinal E,Aubin B,Brassard P, Arthrography of the shoulder: a simple fluoroscopically guided approach for targeting the rotator cuff interval. AJR. American journal of roentgenology. 2004 Feb;     [PubMed PMID: 14736656]


Chung CB,Dwek JR,Feng S,Resnick D, MR arthrography of the glenohumeral joint: a tailored approach. AJR. American journal of roentgenology. 2001 Jul;     [PubMed PMID: 11418432]


Jacobson JA,Lin J,Jamadar DA,Hayes CW, Aids to successful shoulder arthrography performed with a fluoroscopically guided anterior approach. Radiographics : a review publication of the Radiological Society of North America, Inc. 2003 Mar-Apr;     [PubMed PMID: 12640153]


Rutten MJ,Collins JM,Maresch BJ,Smeets JH,Janssen CM,Kiemeney LA,Jager GJ, Glenohumeral joint injection: a comparative study of ultrasound and fluoroscopically guided techniques before MR arthrography. European radiology. 2009 Mar;     [PubMed PMID: 18958474]


Mulligan ME, CT-guided shoulder arthrography at the rotator cuff interval. AJR. American journal of roentgenology. 2008 Aug;     [PubMed PMID: 18647887]


Zanetti M,Hodler J, MR imaging of the shoulder after surgery. Radiologic clinics of North America. 2006 Jul;     [PubMed PMID: 16829248]


Fallahi F,Green N,Gadde S,Jeavons L,Armstrong P,Jonker L, Indirect magnetic resonance arthrography of the shoulder; a reliable diagnostic tool for investigation of suspected labral pathology. Skeletal radiology. 2013 Sep;     [PubMed PMID: 23715887]


Morrison WB, Indirect MR arthrography: concepts and controversies. Seminars in musculoskeletal radiology. 2005 Jun;     [PubMed PMID: 16044381]


Saleem AM,Lee JK,Novak LM, Usefulness of the abduction and external rotation views in shoulder MR arthrography. AJR. American journal of roentgenology. 2008 Oct;     [PubMed PMID: 18806137]


Vollman AT,Craig JG,Hulen R,Ahmed A,Zervos MJ,van Holsbeeck M, Review of three magnetic resonance arthrography related infections. World journal of radiology. 2013 Feb 28;     [PubMed PMID: 23494542]


Newberg AH,Munn CS,Robbins AH, Complications of arthrography. Radiology. 1985 Jun;     [PubMed PMID: 4001360]


Llopis E,Montesinos P,Guedez MT,Aguilella L,Cerezal L, Normal Shoulder MRI and MR Arthrography: Anatomy and Technique. Seminars in musculoskeletal radiology. 2015 Jul;     [PubMed PMID: 26021583]


Kadi R,Milants A,Shahabpour M, Shoulder Anatomy and Normal Variants. Journal of the Belgian Society of Radiology. 2017 Dec 16;     [PubMed PMID: 30498801]


O'Kane JW,Toresdahl BG, The evidenced-based shoulder evaluation. Current sports medicine reports. 2014 Sep-Oct;     [PubMed PMID: 25211618]


Halder AM,Itoi E,An KN, Anatomy and biomechanics of the shoulder. The Orthopedic clinics of North America. 2000 Apr;     [PubMed PMID: 10736387]


Reeves B, Arthrography of the shoulder. The Journal of bone and joint surgery. British volume. 1966 Aug;     [PubMed PMID: 5913134]


Bradley YC,Chandnani VP,Gagliardi JA,Reeves TQ, Partial thickness supraspinatus tears: diagnosis by magnetic resonance arthrography. Australasian radiology. 1995 May;     [PubMed PMID: 7605315]


SAMILSON RL,RAPHAEL RL,POST L,NOONAN C,SIRIS E,RANEY FL Jr, Shoulder arthrography. JAMA. 1961 Mar 4;     [PubMed PMID: 13745910]


De Filippo M,Schirò S,Sarohia D,Barile A,Saba L,Cella S,Castagna A, Imaging of shoulder instability. Skeletal radiology. 2020 Oct;     [PubMed PMID: 32447469]


Ladd LM,Crews M,Maertz NA, Glenohumeral Joint Instability: A Review of Anatomy, Clinical Presentation, and Imaging. Clinics in sports medicine. 2021 Oct;     [PubMed PMID: 34509200]


van Grinsven S,Hagenmaier F,van Loon CJ,van Gorp MJ,van Kints MJ,van Kampen A, Does the experience level of the radiologist, assessment in consensus, or the addition of the abduction and external rotation view improve the diagnostic reproducibility and accuracy of MRA of the shoulder? Clinical radiology. 2014 Nov;     [PubMed PMID: 25218253]