Also known as posterosuperior impingement, the term refers to abutment of the greater tuberosity of the humerus against the posterosuperior aspect of the glenoid as the arm is abducted and externally rotated. A portion of the rotator cuff (the posterior fibers of the supraspinatus and anterior fibers of the infraspinatus tendons) may be entrapped between the bones, often along with the posterosuperior portion of the labrum (Fig. 12.3) (Park et al. 2002a; Meister 2000; Heyworth and Williams 2009). This contact can be seen in normal, asymptomatic throwers but may become symptomatic and produce a constellation of typical injuries (Heyworth and Williams 2009).

Internal impingement has been thought to be secondary to microinstability that results from chronic stretching of the anterior capsular structures during the throwing motion (Heyworth and Williams 2009). Tightening of the posterior capsular tissues, also known as glenohumeral internal rotation deficit (GIRD), has also been implicated as a possible cause (Burkhart et al. 2003; Myers et al. 2006). This tightening leads to a loss of internal rotation, and, in an effort to maintain pitch velocity, the thrower develops an excessive degree of external rotation of the humerus. This excessive rotation leads to an anterior shift of the inferior glenohumeral ligament complex which then results in a posterosuperior shift of the axis of rotation of the humeral head on the glenoid, thereby worsening the osseous and soft tissue impingement along the posterosuperior aspect of the joint (Fig. 12.4).

Associated injuries include superficial undersurface tears of the posterior cuff (posterior fibers of the supraspinatus, anterior fibers of the infraspinatus), fraying or tears of the posterior superior labrum, and subchondral cysts and/or sclerosis in the greater tuberosity or posterosuperior glenoid. Full-thickness rotator cuff tears are not common in this syndrome (Park et al. 2002a).

Radiographic findings that may be seen in patients with symptomatic internal impingement include sclerotic changes of the greater tuberosity, lytic “geodes” of the posterior humeral head, rounding of the posterior glenoid rim, and enthesopathy/exostosis of the posterior inferior glenoid rim (Bennett lesion) which is thought to be related to avulsive forces along the posterior capsule (Heyworth and Williams 2009).

The superficial undersurface cuff tears that occur with internal impingement are often very subtle and extremely difficult, if not impossible, to detect on MRI (Kaplan et al. 2004). MRI is the study of choice in these patients, and arthrography has been shown to be more sensitive than conventional MRI for detecting partial tears of the cuff (Fig. 12.5) (Magee et al. 2004; Tuite 2003). Scanning with the arm abducted and externally rotated (the “ABER” position) has also been advocated since it results in better separation of the cuff tendons from the humeral head, allowing improved visualization of the undersurface tears (Tuite 2003; Tirman et al. 1994b). Findings commonly associated with undersurface cuff tears include abnormalities of the posterior superior labrum and subchondral cystic foci in the greater tuberosity or superior glenoid. The presence of one or more of these imaging findings in an overhead athlete should suggest the diagnosis of internal impingement (Fig. 12.6) (Giaroli et al. 2005).

Impingement of the supraspinatus tendon at the level of the coracoacromial arch (“outlet” of the shoulder) may be secondary to spurring along the anterior acromion or distal clavicle, a “hooklike” acromion, lateral downsloping of the acromion, or a thickened coracoacromial ligament. This type of impingement is the most common cause of cuff pathology in older patients, and while it may occur in older overhead athletes, it is rare in young throwers (Park et al. 2002a; Tuite 2003; Ouellette et al. 2008). The spectrum of cuff pathology ranges from tendinopathy to partial- or full-thickness tears.

Anatomic features that may lead to the development of impingement symptoms such as acromioclavicular joint spurring or abnormal acromial morphology may be evident on radiographs. These features are best demonstrated on MRI examinations in the oblique sagittal and oblique coronal planes (Fig. 12.7). Supraspinatus tendinopathy manifests as thickening of the tendon with diffuse, heterogeneous intrasubstance signal intensity on T1- and T2-weighted images. A partial-thickness tear is diagnosed when focal, fluid-equivalent signal intensity is identified along the articular or bursal surface of the tendon (Fig. 12.8). Fluid-equivalent signal intensity confined to the substance of the tendon indicates an interstitial tear. A full-thickness tear demonstrates fluid (or contrast) extending across the tendon. It is important to assess the dimensions of the tear, the degree of tendon retraction, and any associated atrophy within the muscles involved.

During the deceleration phase of the throwing motion, the muscles of the rotator cuff contract to slow the arm and, in so doing, place the cuff tendons under tension. These tensile forces can result in superficial undersurface tears involving the distal supraspinatus and infraspinatus tendons (Wilk et al. 2009; McConville and Iannotti 1999; Vinson et al. 2007). These lesions are known as “rim-rent” or “PASTA” (partial articular surface tendon avulsion) tears (Tuite et al. 1998) and are diagnosed on MRI when foci of fluid-equivalent signal intensity are noted on T2-weighted images along the undersurface of the distal supraspinatus and/or infraspinatus tendons very near their insertions (Fig. 12.9) (Vinson et al. 2007; Tuite et al. 1998; Ostlere and Marmery 2007). In some cases, retracted fibers can be identified along the undersurface of the tendon just proximal to the tear. These tears, however, can be very difficult to detect, especially if the arm is internally rotated (Vinson et al. 2007), and MR arthrography and scanning in the ABER position have been advocated for improved detection (Tirman et al. 1994b).

Injuries of the subscapularis tendon are much less common in the throwing athlete than are those involving the supraspinatus and infraspinatus tendons. Impingement of the undersurface of the subscapularis tendon against the anterior superior glenoid with flexion and internal rotation of the arm has been suggested as a cause of undersurface tears of the tendon in overhead athletes (Gerber and Sebesta 2000). Partial disruption of the distal fibers of the subscapularis allows for subluxation or dislocation of the long head of the biceps tendon from the bicipital groove and may be a source of anterior shoulder symptoms (Braun et al. 2009).

The subscapularis tendon is best evaluated on axial MRI as it courses anterior to the joint to insert on the lesser tuberosity. Tears of the distal subscapularis tendon may be quite subtle and only noticed because of associated medial subluxation of the biceps tendon (Fig. 12.10).

Injuries to the superior labrum at the biceps-labral anchor (superior labrum anterior to posterior – “SLAP” – tears) are common in throwers. Chronic traction on the long head of the biceps tendon, primarily during the deceleration phase, has been implicated as a causative factor (Park et al. 2002b). SLAP tears in throwers may also result from a “peel-back” phenomenon in which the tendon assumes a more vertical and posterior orientation with abduction and external rotation of the arm. Simultaneous twisting of the tendon at its labral attachment acts to separate the posterosuperior labrum from the glenoid just posterior to the biceps-labral anchor (Fig. 12.12) (Park et al. 2002b; Bedi and Allen 2008).

Injuries of the superior labrum are best demonstrated on MRI in the oblique coronal plane. Fluid or contrast, extending between the labrum and superior glenoid, is diagnostic, although it can be difficult, if not impossible, to distinguish a true SLAP tear from a normal sublabral sulcus that occurs in that same region (Waldt et al. 2006; Tuite et al. 2005). Helpful MRI findings that support the diagnosis of a true SLAP tear include lateral extension of abnormal signal intensity into the substance of the labrum, irregularity of the labral margin, and extension of the signal into the proximal biceps tendon (Fig. 12.13) (Jin et al. 2006; Tuite et al. 2000). Scanning the patient in the ABER position has also been recommended in an attempt to provide a semi-dynamic assessment of stability of the biceps-labral anchor (Borrero et al. 2010).

As described above, impingement of the greater tuberosity of the humerus against the posterior superior glenoid is common in throwers and may become symptomatic. In addition to the superficial fraying and tears noted along the undersurface of the infraspinatus tendon, labral degeneration and tearing along the posterosuperior glenoid are also common associated findings (Kaplan et al. 2004). Other pathologic features include subchondral cysts in the posterior portion of the greater tuberosity, chondral abnormalities involving the humeral head, and subchondral cysts or sclerosis in the posterosuperior glenoid (Giaroli et al. 2005).

The posterosuperior labrum is best assessed on oblique coronal and axial images. An oblique axial scanning plane oriented perpendicular to the true long axis of the glenoid may provide a better assessment because it produces images more perpendicular to the oblique orientation of the labrum at this level. MR arthrography has also been advocated given its increased sensitivity for detecting labral pathology (Tuite 2003).

The labral findings in internal impingement may consist only of subtle fraying and irregularity of the posterosuperior labrum. Even so, the labral abnormalities may be more readily recognized on MRI than the associated superficial undersurface tears involving the posterior cuff tendons. High-signal subcortical marrow edema and cysts in the greater tuberosity or posterosuperior glenoid are most conspicuous on fat-saturated T2-weighted images and are also helpful secondary signs (Giaroli et al. 2005). Subchondral sclerosis, when present, is best identified on T1-weighted images since its low signal intensity contrasts sharply with the high signal intensity of the adjacent marrow fat.

The anterior labrum and glenohumeral ligaments are placed under tension when the arm is abducted and externally rotated during the late cocking phase of the throwing motion. Over time the accumulation of microdamage in these structures may result in anterior instability (Park et al. 2002a). Similarly, during the deceleration phase, when the forward momentum of the arm is rapidly slowed by both eccentric posterior muscle contraction and static posterior labrocapsular structures, stretching of the posterior soft tissue restraints may result in posterior instability (Park et al. 2002b). Although frank labral avulsions are rare in throwers, other forms of anterior and/or posterior labral pathology may result.

Another potential source of posterior labral pathology occurs with the so-called “batter’s shoulder” in which there is transient posterior subluxation of the lead shoulder during the swing (Wanich et al. 2012). This uncommon injury is thought to be related to the posterior rotational forces that act upon that shoulder as the bat is accelerated through the swing. Patients present with posterior shoulder pain during batting, and most are found to have posterior labral tearing that often requires arthroscopic stabilization for the patient to be able to return to play (Wanich et al. 2012).

The anterior and posterior portions of the labrum are best evaluated on axial MRI. As with labral pathology at other sites, MR findings of a labral tear include abnormal signal within the labrum or interposed between the labrum and glenoid (Fig. 12.14). The morphology of the labrum can be quite variable, so a finding of labral “irregularity” is less specific. With posterior instability, the posterior joint recesses may appear capacious, but since that is a very subjective determination, a more specific finding is that of posterior capsular stripping along the posterior margin of the glenoid. In that region, the normal capsule attaches consistently along the base of the labrum at its junction with the osseous glenoid. Assessment of anterior capsular stripping is much more difficult because of the extreme variability of the normal capsular attachments along the anterior glenoid.

The ulnar collateral ligament (UCL) is composed of three bands: anterior, posterior, and transverse/oblique (Chen et al. 2001). The anterior band is the primary restraint against valgus force at the elbow during the throwing motion, while the posterior band plays a much less significant role. The transverse band plays no significant part in stabilizing the elbow (Chen et al. 2001; Lynch et al. 2008). Repetitive valgus stresses during throwing can cause the anterior band of the UCL to accumulate microdamage within its fibers leading to stretching and partial or complete tears of the ligament (Lynch et al. 2008).

The normal anterior band of the UCL is best seen on coronal and axial MRI as a low signal structure extending from the undersurface of the medial epicondyle to attach to the medial margin of the coronoid process of the ulna at the sublime tubercle. Its humeral attachment is relatively broad and often demonstrates slightly heterogeneous signal compared to the more distal fibers which taper toward their ulnar attachment. The anterior band is well demonstrated in cross section on axial images where the posterior band is also seen as a thin low signal structure forming the floor of the adjacent cubital tunnel.

Following ligamentous sprain, the anterior band appears somewhat thickened, typically with adjacent edema.

Partial tears may involve any portion of the ligament and result in some degree of fiber disruption. An especially subtle lesion involves partial stripping of the distal fibers at their ulnar attachment, often producing the “T” sign on routine or, more commonly, arthrographic MRI (Fig. 12.15) (Timmerman and Andrews 1994; Schwartz et al. 1995; Timmerman et al. 1994; Nakanishi et al. 1996). This may be difficult to differentiate from a reported normal variation of mild separation of the fibers from the sublime tubercle (Munshi et al. 2004); the distinction is usually based on the degree of separation of the ligament from the bone and the amount of fluid or contrast undercutting the ligament at that site. The normally heterogeneous signal within the proximal fibers of the ligament may be difficult to distinguish from a partial tear at its humeral attachment. A partial tear of the posterior bundle is much less common but may clinically mimic an anterior bundle injury. Such a tear is best demonstrated on axial images.