Ligamentous structures of the elbow are comprised of focal condensations of the fibrous joint capsule, thus forming a capsuloligamentous complex similar to that encountered in the shoulder. The ulnar collateral ligament (UCL) originates from the inferior aspect of the medial epicondyle slightly posterior to the center axis of elbow rotation; thus, it is under greater tension with elbow flexion (1). The humeral origin of the UCL corresponds to approximately two-thirds of the width of the inferior aspect of the medial epicondyle and has no insertion on the adjacent medial condyle (2). The humeral origin of the UCL has a wider insertion than its ulnar counterpart. In a cadaveric study, Nakanishi et al. (3) showed that this was the result of the combined proximal attachment of both the anterior and posterior bundles.

In general, the collateral ligament complexes are best demonstrated on oblique coronal images. Axial and sagittal images are useful to confirm suspected pathology. As is the case elsewhere in the musculoskeletal system, ligaments are generally uniformly low in signal intensity on all pulse sequences except where magic angle effects or volume averaging may occur.

Three components of the UCL are classically described, the anterior, posterior, and transverse bundles (Fig. 10.1). The largest and most important component of the UCL is the anterior bundle, a discrete focal thickening of the medial capsule arising from the inferior margin of the medial epicondyle that inserts at the sublime tubercle of the proximal ulna (Fig. 10.2). Timmerman and Andrews (4) have demonstrated two histologic components of the anterior bundle, a deep layer between the two fibrous layers of the capsule and a superficial layer external to the medial capsule (Fig. 10.3). The superficial or extracapsular layer may in fact represent deep fibers of the ulnohumeral head of the flexor digitorum superficialis (5). Although the ulnar insertion of the anterior bundle is classically described as being flush with the articular margin of the medial coronoid, there may in fact be some variability of the ulnar insertion at the sublime tubercle. Timmerman and Andrews (6) have reported the ulnar insertion of the anterior bundle is in close proximity (1 mm) of the joint line. In contradistinction, Munshi et al. (5) found the anterior bundle may insert either at the articular margin or approximately 3 mm distally (Fig. 10.4). The more distal insertion may create a small recess on arthrograms simulating a partial undersurface tear of the anterior bundle. Therefore, differentiating partial undersurface tears of the UCL from a normal distal insertion of the anterior bundle may be clinically important.

On oblique coronal images, the anterior bundle of the UCL is seen as a uniformly low-signal structure. The proximal fibers of the anterior bundle may appear indistinct and lax (7). Typically the anterior bundle insertion at the sublime tubercle is flush with the articular margin. However, as previously mentioned, caution is warranted because variations in attachment may occur in this region.

The posterior bundle is a focal thickening of the medial capsule, which, unlike the anterior bundle, lacks
a distinct superficial component. The posterior bundle originates from the inferior aspect of the medial epicondyle and inserts at the posteromedial margin of the trochlear notch, forming the floor of the cubital tunnel (Fig. 10.5). On oblique coronal MR images, the humeral origin of the posterior bundle is seen posterior to a plane bisecting the epicondyles and extends to the posteromedial aspect of the olecranon. MRI identification of the posterior band is most facile in the axial plane when localizing the floor of the cubital tunnel at the joint line.
The transverse bundle, also termed the ligament of Cooper, extends between the ulnar insertions of the anterior and posterior bundles and plays little, if any, functional role. It is not readily visualized on routine MRI.

Variants of the UCL complex have been described and include a strong oblique pattern in which the transverse bundle flares in a fan-like configuration as it inserts on the anterior bundle and coronoid (8). In nearly one-fourth of specimens, Beckett et al. (8) identified a fourth component of the UCL termed the accessory or extra bundle. The accessory bundle originates from the posteromedial aspect of the capsule and inserts on the
transverse bundle. Importantly, there is little or no variation of the anterior or posterior bundles of the UCL.

The visualized components of the UCL complex are seen as thin linear bands extending along the medial aspect of the elbow joint on MR images. For the purposes of MRI, the transverse ligament is not visualized nor is it considered a clinically important structure with regard to medial elbow stability. The MRI appearance of the UCL is characterized not only by its morphology, but also its signal intensity. The normal UCL has uniform low signal intensity attributable to its highly ordered structure and collagenous composition (9, 10). It has also been noted
that the normal UCL should also have fat signal intensity visible just beneath its fibers (Fig. 10.2). This finding represents the intra-articular fat that is seen in all normal elbow joints (3). The fat would be effaced in the setting of surrounding edema, joint effusion, or with contrast introduced into the articulation for arthrography.

MRI findings in the abnormal UCL vary with the severity of the injury. Diffuse increased signal intensity on T1- and T2-weighted images in the periligamentous soft tissues is consistent with the presence of edema and hemorrhage (11). One study also characterized high signal intensity within the periligamentous soft tissues as scar formation (3). With conventional MRI techniques, ligament sprains have been characterized by alterations in morphology, either thinning or thickening, as well as alterations in signal intensity. With regard to the latter, high signal intensity on T1- and or T2-weighted images within the substance of the ligament in the absence of discontinuity have been characterized in the literature as ligament sprain (9, 12–14). Although it has not been addressed in the literature, it is the belief and practice of the author that high T2-weighted signal within the substance of the ligament that parallels the signal intensity of joint fluid is more accurately described as a partial intrasubstance tear of the ligament fibers.

Partial tears of the UCL have been reported, but these reports have focused on pathology of the undersurface, or articular side of the ligament, primarily at either the humeral or ulnar attachments (6, 13). One of the earliest descriptions of this entity compared standard MRI and computed tomographic (CT) arthrography in a series of collegiate and professional baseball players who presented with persistent medial-sided elbow pain with throwing (6, 15). In this small series of patients (n = 7), CT arthrography showed abnormal leakage of contrast around the humerus, and more commonly the ulna, without gross extracapsular extravasation. CT arthrography correctly identified surgically proven lesions in six of seven patients, whereas standard MRI studies identi
fied an abnormality in a single patient. The abnormal collection of contrast extending along the proximal aspect of the ulna, in contiguity with the normal contrast distribution along the ulnar articular surface and along the humeral condyle, were described as the T-sign (Figs. 10.6 and 10.7). Caution should be exercised when making the diagnosis of a partial detachment of the UCL from its attachment to the sublime tubercle, however, because variations in the distal attaching fibers have been described (Fig. 10.4) (5). Correlation with the clinical history, patient presentation, and physical examination findings should prevent falling prey to such a pitfall in imaging diagnosis. When pathologic, it is the belief of the author that this injury may represent a peelback phenomenon in which the ligament is stripped from its osseous attachment as described in anterior labral ligamentous periosteal sleeve avulsion in the glenohumeral joint (16). The role of MR arthrography in the detection of partial undersurface tears of the UCL was subsequently explored (17). The sensitivity, specificity, and accuracy for the detection of undersurface tears of the UCL by MR arthrography were 86%, 100%, and 86%, respectively (Figs. 10.8 and 10.9). Although scarcely addressed in the imaging literature, intrasubstance and superficial partial tears of the UCL should be considered in the imaging assessment of this structure. In applying the general principles for interpretation of signal intensity and morphology abnormalities in tendons and liga
ments, high T2 signal intensity within the substance of the ligament that rivals that of joint fluid would be consistent with an intrasubstance tear, and superficial surface irregularity and high T2 signal that parallels joint fluid would be consistent with a partial superficial ligament tear (Figs. 10.10 and 10.11). The latter should be suspected in cases of common flexor tendon abnormalities.

Full-thickness tears of the UCL are diagnosed by discontinuity along the course of the ligament (Figs. 10.12–10.14) (15, 18). A secondary sign that some authors have used to indicate the presence of full-thickness tear of the
UCL is an abnormal distribution of joint effusion along the course of the ligament with extravasation of fluid into the adjacent soft tissues (19). Whereas standard MRI has a suboptimal performance for detection of partial-thickness tears of the UCL, in the setting of full-thickness tear, most studies show a sensitivity, specificity, and accuracy that approaches 100% (9, 15, 18, 19). Similarly, MR arthrography has been shown to be sensitive, specific, and accurate in the characterization of the anatomy of the UCL in cadaveric studies as well as in the detection of full-thickness tears of this structure (3, 13, 17, 20–23). The imaging criteria for the diagnosis of UCL full-thickness tear are the same as described for standard MRI. The detection of discontinuity of the ligament and abnormal distribution of contrast material at the point of ligament insufficiency is noted.

A description of injury to the UCL complex would be incomplete without localization of the imaging abnormality. In the setting of partial tears of the ligament, reported cases have discussed only the anterior band of the UCL complex (6, 13, 15, 17). Because the fibers of the anterior and posterior bands of the UCL are continuous and fan-like in nature, it can be difficult to localize pathology between these two components of the UCL complex (Figs. 10.15 and 10.16). It can be helpful to carefully assess the axial images in the region of the cubital tunnel, because this anatomic landmark isolates the posterior band. As previously noted, the undersurface as opposed to intrasubstance and superficial pathology predominates (Fig. 10.17). Finally, ulnar attachment, rather than humeral and midsubstance partial tears, is more frequently encountered (6, 13, 15, 17, 24). Full-thickness tears of the UCL occur most frequently in the midsubstance of the ligament (13, 25). Avulsion injuries with failure at the humeral and ulnar attachment have also been described (3, 24). In skeletally immature patients, failure at the humeral side with an avulsion fracture of the medial epicondylar ossification center occurs most frequently (Fig. 10.18). In adults, if avulsion-type injuries occur, they do so at the ulnar attachment to the sublime tubercle (Fig. 10.19). This injury may or may not have a detectable osseous component on radiographs. MRI has been shown to be useful in treatment planning with regard to direct repair or reconstruction in the setting of avulsive injury (25).