Lisa K. O’Brien, DO and Brian R. Waterman, MD
Anterior glenohumeral instability is a common diagnosis among the athletic population, with those competing in contact sports at highest risk.1 Anterior shoulder instability covers a spectrum of injury, from subluxation to frank dislocation, and can result in varying severities of soft-tissue and bony pathology. Recognition of the involved anatomic structures through radiographic and advanced imaging is crucial for directing treatment and obtaining favorable outcomes. Increased attention is being paid to the evaluation of bone loss rather than focusing solely on soft-tissue injury because there is a high failure rate associated with soft-tissue repair alone in the face of significant glenoid width loss. Glenoid bony defects are composed of a continuum of 2 general types: fragment-type (eg, acute glenoid fracture, bony Bankart, glenolabral articular disruption) and attritional-type bone loss. Glenoid bony defects can occur in 5% to 56% of first-time dislocators, and in upward of 90% of patients with recurrent anterior glenohumeral instability.2,3 Likewise, the Hill-Sachs lesion, or impaction fracture of the posterolateral humeral head, can occur in 65% to 88% of first-time anterior dislocators, and in up to 93% of patients with recurrent instability.2,4,5 Furthermore, glenohumeral subluxations, which comprise up to 85% of all glenohumeral instability events, are not benign and can also result in alarmingly high rates of soft-tissue and bony injury.6 Plain-film radiography can be used as a screening tool to identify patients with significant bone loss both of the glenoid and humeral head. Advanced imaging is required to better quantify the amount of bone loss. Although magnetic resonance imaging (MRI) has classically been used for soft-tissue evaluation and computed tomography (CT) for osseous evaluation, the use of MRI for assessment of bone loss has recently gained traction thanks to advancements in its technology. Additional modalities like 3-dimensional (3D) reconstruction and imaging with intraarticular gadolinium have also been proven useful. Various measurement techniques to quantify bone loss and direct treatment algorithms have evolved in response to a previously high failure rate of arthroscopic soft-tissue repair. The development of the glenoid track concept and the on-track off-track classification system has enhanced clinical decision making in addressing bipolar bone loss, or concomitant bone loss of both the glenoid and humeral head.
Imaging for the assessment of an athlete with anterior shoulder instability should be initiated with plain radiography. Orthogonal radiographs are an essential adjunct to confirming glenohumeral reduction and assessing for other abnormalities. Standard radiographs include an anteroposterior (AP) or true AP (Grashey) view, scapulolateral view (ie, scapular-Y view), and an axillary lateral view.7 Advantages of radiography include its relatively low cost, near ubiquitous access, and relative ease in obtaining. A major disadvantage is its user-dependent quality and comparatively low accuracy and reliability as to quantifying bone loss.8,9 Therefore, plain radiographs are often used primarily as a simple screening tool to detect significant deficits in glenoid and humeral head bone.
Anteroposterior and True Anteroposterior Views
A standard AP radiograph of the shoulder is taken with the x-ray beam aiming directly perpendicular to the plane of the thorax with the arm resting in neutral. Owing to the relative position of the scapular plane angled approximately 45 degrees anterior to the thorax, the result of the AP radiograph is an oblique view of the glenohumeral joint. Therefore, overlapping of the glenoid and humeral head is seen with this view, thereby limiting certain assessments for osseous involvement (Figure 7-1A).
A true AP of the shoulder, otherwise known as a Grashey view, is obtained by directing the x-ray beam 45 degrees medial to lateral. With this view in a reduced shoulder, there will be no overlap between the humeral head and glenoid (Figure 7-1B). The only structure that will overlap other structures is the coracoid process. Evidence of glenohumeral overlap is indicative of either a dislocation or an inadequate radiograph. In a normal shoulder, the contour of the anteroinferior glenoid will be clearly visible. Lack of clear visualization of the anterior glenoid sclerotic contour in an otherwise well-executed radiograph suggests bone loss. In one study, Jankauskas et al found the Grashey view to be specific (100%) but not sensitive (54% to 65%) for detecting glenoid bone defects.9
Otherwise known as a scapular-Y, tangential lateral, or Y lateral, the scapulolateral view aims to obtain an image along the axis of the scapular spine. This view is a true lateral of the glenohumeral joint and is helpful in determining displacement of the humeral head in conjunction with orthogonal views. In the case of an anterior dislocation, the humeral head will be found anterior to the glenoid, and vice versa for posterior dislocations. This view also can provide relative assessment of the approximate humeral head and glenoid contour.
The scapulolateral view can be obtained in a variety of ways. The patient is typically positioned upright with the affected shoulder against the image receptor. The patient is rotated 45 to 60 degrees oblique (toward the image receptor) until the body of the scapula is perpendicular both to the image receptor and the x-ray beam. If tolerated, the arm can be placed behind the patient’s back to allow superimposition of the humerus over the scapula. Alternatively, the arm may be allowed to hang free or rest in a sling.
Axillary Lateral and Modified Axillary Lateral Views
The axillary lateral radiograph is the most crucial view required to confirm a reduced glenohumeral joint. This view is taken with the patient either supine or upright and the arm abducted 70 to 90 degrees. In the supine position, the x-ray beam is angled toward the axilla, aiming from caudad to cephalad (vice versa for the upright position). The articulation of the humeral head and glenoid are clearly visualized with this view. An adequate image of a normal shoulder is obtained when there is visible space between the glenoid and humeral head, and the superior and inferior edges of the glenoid are superimposed.10 Static instability is easily identified if the humeral head is displaced anterior or posterior to the glenoid, or if overlapping between the 2 structures is seen. Much like the Grashey view, lack of clear visibility of the anteroinferior glenoid outline suggests bone loss, and humeral head impaction fractures may also be visualized.
An alternative to the axillary lateral is the modified axillary lateral, or Velpeau view. This view can be performed in the event that abduction of the shoulder is painful or poorly tolerated by the patient. The Velpeau view is obtained with the patient’s arm maintained in adduction within a sling. The patient is positioned upright, leaning back 20 to 30 degrees, or with a 20- to 30-degree wedge placed behind the back. The x-ray beam is placed over the top of the shoulder, aiming directly vertical from superior to inferior.
ADDITIONAL RADIOGRAPHIC VIEWS
Additional special radiographs can be obtained to clarify various bony abnormalities associated with anterior shoulder instability. These views include the West Point axillary lateral, apical oblique view, Bernageau glenoid profile view, Stryker notch view, and an AP of the shoulder with humeral internal and external rotation.7
West Point View
The West Point view was initially described by surgeons at the United States Military Academy in West Point, New York.11 This view is performed to identify large glenoid defects and humeral head subluxation. The technique involves the patient lying prone with the affected shoulder abducted to 90 degrees and resting on a pad elevated 8 cm. The x-ray beam is centered at the axilla, angled 25 degrees inferiorly, and 25 degrees medially, resulting in a tangential view of the anteroinferior glenoid rim. A cadaveric study by Itoi et al revealed the West Point view to be better than the axillary lateral view at identifying glenoid bone loss.8
Apical Oblique View
The apical oblique, or Garth view, is used to visualize the anteroinferior and posterosuperior glenoid, as well as the posterolateral and anterior humeral head. This view is obtained with the patient sitting upright and the hand of the affected side resting on the unaffected shoulder. The x-ray beam is angled 30 to 45 degrees from medial to lateral and 45 degrees cephalad to caudad.12
Bernageau Glenoid Profile View
Developed in 1976, the Bernageau glenoid profile radiograph was introduced to obtain a true view of the anteroinferior glenoid.13 The axillary lateral view often results in superimposition of the anteroinferior rim over the anterosuperior rim, making it difficult to assess for bone loss. To perform the Bernageau glenoid profile view, the patient is positioned upright with the shoulder abducted to at least 135 degrees and the hand resting on the head. The x-ray beam is directed along the axis of the scapular plane, angled 30 degrees caudal (Figure 7-2A). However, positioning of the shoulder for this view may not be well tolerated in patients with acute pain or severe instability. Sugaya3 developed a modified Bernageau view that allows the patient to lie in a lateral recumbent position, with the affected shoulder toward the table and the arm abducted in a resting position, known as the “television watching position” (Figure 7-2B).
With the Bernageau view, a normal shoulder will reveal a well-defined anterior osseous triangle with a sharp angle, whereas anterior bone loss can result in a rounded triangle (blunted angle sign) or complete loss of the triangle (cliff sign).14 The same view of the contralateral shoulder has been recommended for comparison.15 Multiple studies have found a high rate of intraobserver and interobserver reliability compared to other plain radiographs, and it was found to have a high rate of reliability when compared to CT.15,16 The Bernageau view has also been used to identify humeral head defects.17
Stryker Notch View
The Stryker notch view is used to assess for humeral head defects, particularly of the posterolateral portion, which is common with anterior glenohumeral instability. This view is especially helpful when Hill-Sachs lesions are not readily visualized on other standard orthogonal imaging. The patient is supine with the hand of the affected side resting palm down on the forehead, with the fingers facing cephalad and the elbow directed anterior. The x-ray beam is centered over the coracoid process and directed 10 degrees cephalad.
Humeral Internal and External Rotation Views
An AP radiograph with the humerus internally rotated is one of the most common plain radiographs used to show impaction fractures of the posterolateral humeral head, or Hill-Sachs lesions.2 The humeral external rotation view can be included for further visualization of the proximal humerus. These views outline the profile of the greater tuberosity. In cases of large Hill-Sachs defects, anterior extension of the compression defect may be seen with this view. The patient is seated upright with the affected humerus externally rotated and the x-ray beam placed as with an AP radiograph.
MAGNETIC RESONANCE IMAGING
MRI is a common imaging modality used for assessing anterior shoulder instability. It affords the most comprehensive visualization of soft-tissue structures and identification of pathology, with the most commonly involved structures being the anterior labrum, anterior capsule, and the anterior band of the inferior glenohumeral ligament (aIGHL).18 Owing to the small size and complexity of the structures in the shoulder, a high-resolution image is required with various 2-dimensional (2D) sequences, in-plane resolution of at least 0.5 mm, and slice thicknesses of 1 to 3 mm.19 The most commonly available MRI strength is 1.5-Tesla (T), which refers to the strength of the magnetic field. However, more institutions are acquiring 3-T (or greater) MRIs, which operate at twice the strength with a greater signal-to-noise ratio, allowing for higher-quality images.
Musculoskeletal imaging protocols were developed to enhance visualization of this complex anatomy. These protocols vary by institution based on the field strength of the available MRI magnet, radiologist preference, time availability, and patient factors such as anxiety level or presence of metal implants.20 Absolute contraindications to MRI include patients with implanted devices that are not compatible with MRI such as pacemakers, spinal stimulators, and cochlear implants, patients with intraocular metallic foreign bodies, and unstable patients requiring resuscitative equipment. Relative contraindications include patients with severe claustrophobia, agitation, or movement disorders resulting in the inability to lie still. These patients may require assistance with anxiolytics or sedation.
Conventional MRI shoulder protocols consist of variable sequences in the oblique axial, coronal, and sagittal planes. It is important for these planes to be parallel to the glenohumeral joint rather than the thorax to avoid inadequate visualization from poorly cut images. Each plane is distinctively useful for evaluating different structures.20,21 The axial plane displays anterior and posterior capsulolabral anatomy, the subscapularis tendon, and the biceps-labral complex with the long head of the biceps tendon in the bicipital groove. The sagittal oblique plane affords an en face view of the glenoid articular surface, the relationship of the capsulolabral complex, acromial anatomy and its associated ligaments, and the rotator cuff tendons as they travel out to their attachment on the greater tuberosity. In the case of recurrent instability, the sagittal oblique view may reveal attritional bone loss of the anteroinferior glenoid face, resulting in a so-called “inverted-pear” appearance.22 The coronal oblique plane best defines the superior labrum, anchor of the long head of the biceps, and posterosuperior rotator cuff. A patulous inferior capsule may be visualized suggesting underlying hyperlaxity, though this may be difficult to discern without the addition of intraarticular contrast. Coronal images often will underestimate the size of a bony Bankart lesion.21
Various imaging sequences are used to enhance different structures. The sequences are defined by the tissue relaxation times, or the time it takes for excited protons to either return to equilibrium (T1-weighted image) or go out of phase with each other (T2-weighted image) as it reacts to the magnetic field. T1-weighted images have the best spatial resolution and reveal subcutaneous fat and bone marrow as bright white, whereas liquids (joint fluid) and solids (cortical bone) are dark. T2-weighted images reveal fluids such as joint effusions and bone or muscle edema as bright, making them more sensitive to pathology than T1-weighted images, but with less clarity. Subsequently developed was a T2-weighted proton density (PD) fast spin-echo (FSE) sequence coupled with a fat-suppression (FS) technique, such as short T1 inversion recovery (STIR) or a fat-saturation pulse. An FS PD FSE sequence in the axial plane is most sensitive for identifying small paralabral cysts and subtle articular cartilage labral tears (Figure 7-3). FS T1-weighted images are ideal for viewing the labrum at high resolution. Additionally, various chondral mapping techniques are available when indicated, like T1 ρ and T2 mapping with post-processing and color.21
Understanding the nature of the soft-tissue injury is crucial for guiding treatment because more than one structure is typically involved. There are multiple types of anterior labral pathology that vary based on chronicity, degree of displacement, and involvement of the scapular periosteum, aIGHL, bony glenoid rim, or the glenoid articular cartilage.21 Acute bony injuries will reveal increased signal intensity in the glenoid on FS T2-weighted images with a bony fragment, whereas chronic injuries might result in bony attrition without edema or a medialized lesion that is adherent along the glenoid neck. The aIGHL is most frequently detached from its glenoid insertion in the majority of cases with primary anterior instability. However, occasionally it can detach from its humeral insertion, resulting in what is known as a humeral avulsion of the glenohumeral ligament (HAGL). These lesions often do not present alone and can easily be missed on MRI, particularly without intra-articular contrast.23 They require a high index of clinical suspicion particularly after a previously failed labral repair surgery because the instability recurrence rate with a neglected HAGL can reach 90%.24 Additionally, it is important to recognize normal variants of the anterior labrum and glenohumeral ligament complex. These normal variants include a sublabral foramen, sublabral recess, and a thickened middle glenohumeral ligament (MGHL) with absent anterosuperior labrum, otherwise known as a Buford complex.
Though less common in younger athletes, damage to the rotator cuff can occur with higher prevalence in certain populations. This may include patients older than 40 years, contact or overhead athletes, or those with a nerve injury after dislocation.25 The most commonly involved portion of the rotator cuff is the subscapularis and posterosuperior rotator cuff.
Positioning of the shoulder during the MRI procedure is crucial because internal rotation of the shoulder can result in the anterior structures appearing lax and ill defined. Conversely, extreme external rotation makes it difficult to assess the path of the biceps tendon. Ideal positioning of the shoulder is in neutral or slight external rotation.21 In some instances, positioning the shoulder in abduction and external rotation (ABER) during the MRI can help to reveal nondisplaced labral tears such as with Perthes lesions. This technique is often used in conjunction with MRA for better visualization, although it has also been described without the use of gadolinium with a sensitivity and specificity of 94% and 82%, respectively.26 The ABER position tensions the aIGHL and peels the torn labrum away from the glenoid. This view can also be useful for identifying partial-thickness rotator cuff tears.
Although conventional MRI is superior for visualization of soft-tissue structures, it traditionally has not been favored for the assessment of bony pathology. Many practitioners have recommended CT imaging to characterize and quantify bone loss because CT is considered the gold standard in assessing osseous deficits. However, MRI has been gaining traction as advancing techniques have emerged. One such technique creates 3D MRI reconstructions that are similar in quality to 3D reconstructions from CT.20,27,28 The 3D reconstruction is formed from a 3D, dual-echo time, T1-weighted fast low-angle shot (FLASH) sequence coupled with a fat-suppression technique known as the Dixon method. This sequencing technique creates a water-only image that is then processed using subtraction software to isolate only the osseous structures. Another sequencing technique developed to formulate a 3D image is an isotropic volumetric interpolated breath-hold examination (VIBE) with a water excitation sequence using a dedicated shoulder coil and subsequent post-processing software.29 Comparison studies have shown the 3D FLASH Dixon sequence to be as accurate as CT.20,27 Quantitative measurements for quantifying glenoid bone loss have been developed for 3D MRI and are similar to techniques used for 3D CT (see further details in the section “Computed Tomography and Magnetic Resonance Imaging Measurements”).20,27,30 Recently, this sequencing technique has been used to assess osseous pathology of other joints, such as the hip in femoroacetabular impingement.31
Advantages of 3D MRI are elimination of the time, cost, and radiation required for obtaining a CT scan. Disadvantages of 3D MRI include the separate software and specialized skill set required by radiologists, which may not be feasible or readily available at some institutions.
MAGNETIC RESONANCE ARTHROGRAPHY
There continues to be debate about whether intra-articular contrast enhancement is required for appropriate visualization of capsulolabral detail during MRI. Magnetic resonance arthrography (MRA) involves a diluted paramagnetic gadolinium-based contrast agent that is injected into the glenohumeral joint by a radiologist and used to outline the labrum, capsule, and rotator cuff with increased sensitivity compared with MRI.32 Saline has also been described as an alternative injection agent.33 Furthermore, a hemarthrosis in the setting of an acute injury may be sufficient alone to elucidate capsulolabral pathology. The distension afforded by the additional fluid provides improved detail of the glenohumeral ligaments, and it can place the labrum under tension so that chondrolabral separation is more apparent. In a comparison of conventional 3T MRI to MRA, one study revealed statistically significant increased sensitivity for the detection of partial-thickness rotator cuff tears, anterior labral tears, and superior labral tears with MRA.34 Subtle HAGL lesions are also more readily visible with the addition of intra-articular contrast.18 However, some authors argue that MRA is not necessary if a properly optimized MRI with high signal-to-noise and good spatial resolution is obtained.21
Notably, none of the commercially available gadolinium-based contrast agents are approved for intra-articular use by the US Food and Drug Administration. Therefore, MRA is still considered to be an off-label procedure.35 Despite this, MRA with gadolinium is generally considered safe within the musculoskeletal radiology community. Adverse effects of MRA with gadolinium, based primarily on case reports, include local pain, reaction to the contrast, septic arthritis, synovitis or adhesive capsulitis, extra-articular placement or extravasation, improper needle placement resulting in local neurovascular injury, or poor visualization due to improper dilution.19,36–38 Most systemic adverse effects have been described from intravenous administration rather than intra-articular use; these include nephrogenic systemic fibrosis, tissue gadolinium deposition, and anaphylactoid systemic reactions.35