Evaluation of elbow injuries in athletes begins with conventional radiographs. Standard radiographic views of the elbow include anteroposterior, lateral, and oblique projections ( Fig. 63-1 ). The anteroposterior and oblique views are obtained with the elbow extended and the forearm supinated, permitting visualization of the proximal radioulnar and radiocapitellar articulations, medial and lateral epicondyles, and trochlea. In this anatomic resting position, the carrying angle of the elbow typically measures 12 to 15 degrees valgus. The lateral view, obtained with the elbow flexed 90 degrees and the forearm in a neutral position, permits evaluation of the radiocapitellar and ulnotrochlear articulations, distal humerus, and olecranon and coronoid processes. An appropriately acquired lateral elbow radiograph clearly shows the ulnotrochlear articulation with minimal overlap of osseous structures. An anatomically positioned radial head is collinear with the capitellum. A dark radiolucent area is normally seen just anterior to the distal humerus, representing the anterior fat pad. Similarly, the posterior fat pad sign is a well-demarcated lucency posterior to the distal humeral metaphysis, and its presence is always pathologic. Displacement of these fat pads, produced by an intraarticular effusion or hemarthrosis, resembles the spinnaker on a sailboat and is referred to as the sail sign .
Conventional radiographs are usually satisfactory for developing treatment plans for fractures and dislocations of the elbow without further workup. Fractures of the radial head are frequently encountered, and a specific radial head (radiocapitellar) view may better characterize the fracture pattern and aid in classification. With the elbow flexed to 90 degrees, the beam is angled 45 degrees anteriorly from a true lateral. This projection provides an unobstructed view of the radial head without overlap of the proximal ulna. Other common fractures include those involving the epicondyles (especially the medial epicondyle in skeletally immature patients), distal humerus, and olecranon. The “terrible triad” injury includes fractures of the radial head and coronoid process, along with dislocation of the elbow joint. Conventional radiographs are usually sufficient for confirming concentric reduction of the elbow after closed manipulation.
Any patient presenting with non–activity-related pain or a mass at the elbow is first evaluated with conventional radiographs. Occasionally, such radiographs reveal abnormal subchondral lesions of the capitellum consistent with Panner disease or osteochondritis dissecans. Atypical imaging findings, such as a calcified soft tissue mass or a lytic osseous lesion, may prompt further evaluation by an upper extremity or oncologic orthopaedic specialist.
Many injuries to the elbow involve a combination of osseous and soft tissue structures that surround the elbow. When conventional radiographs do not reveal a problem or provide incomplete information, the choice of additional studies is guided by the history and physical examination. These additional studies may include computed tomography (CT), magnetic resonance imaging (MRI), fluoroscopy, and a bone scan. Some relevant injuries and their classic imaging findings are highlighted in this chapter to aid the general orthopaedic surgeon or sports medicine specialist.
The elbow features complex three-dimensional anatomy. Although some fractures are easy to recognize on standard radiographs, they may be notoriously difficult to interpret confidently. For comminuted, intraarticular injuries, traction radiography may improve visualization of the individual fragments, but these images are difficult to obtain in an acutely injured, awake patient. CT, with sagittal, coronal, and possibly three-dimensional reconstruction, is an excellent modality for defining these complex osseous injuries. Therefore intricate fractures of the distal humerus and medial coronoid facet are commonly evaluated by CT scan for preoperative planning.
Elbow dislocations are evident on conventional radiographs and are characterized by the position of the radius/ulna relative to the humerus. Simple dislocations without fracture require no additional imaging beyond postreduction radiographs, once concentric anatomic relationships are verified. Complex fracture-dislocations occasionally require advanced imaging with a CT scan to characterize the fracture pattern, degree of displacement, and/or the presence of associated loose bodies, which may block concentric reduction and lead to persistent joint subluxation. Failure to identify fractures of stabilizing osseous structures such as the coronoid or radial head will lead to a poor functional outcome. A dynamic live fluoroscopic examination of the elbow after induction of general anesthesia may be necessary for injuries with equivocal stability after closed reduction. Finally, MRI is the best option for determining the pattern of soft tissue disruption in acutely unstable injuries or those that present with delayed complaints of subjective instability, which may require collateral ligament repair or reconstruction.
Persistent tenderness over a suspected injury site despite negative findings of conventional radiographs may be an indication for further evaluation in select cases. The posterior fat pad sign corresponds to an occult fracture in more than 75% of patients. The radial head is the most frequent site of an occult fracture. When a radial head fracture is suspected, a radiocapitellar radiographic view may be helpful. When clinical suspicion is high but radiographs are unremarkable, CT or MRI may detect an osseous injury. Fat-suppressed T2-weighted (or short tau inversion recovery) pulse sequences are the most sensitive for detecting radiographically occult fractures. Fractures show a linear pattern of signal change, with decreased signal on T1-weighted images or increased signal on T2-weighted images ( Fig. 63-2, A ). In contrast, osseous contusion produces a nonspecific diffuse increase in signal on T2-weighted images without a discrete fracture line ( Fig. 63-2, B ).
The olecranon is the most common stress fracture site, especially among baseball players. When the history and physical examination suggest this condition, a CT, MRI, or bone scan may be diagnostic. A CT scan typically reveals a subtle fracture line in the olecranon. Findings on MRI resemble those of occult fractures described previously. A bone scan would simply reveal nonspecific abnormal uptake at the site of repetitive stress.
In the setting of degenerative joint disease, conventional radiographs may detect intraarticular loose bodies in the coronoid or olecranon fossae. Supplemental oblique or axial projections may be helpful in distinguishing intraarticular and extraarticular calcifications, especially in severe cases of heterotopic ossification. In ambiguous cases, CT is the modality of choice for further information. Singson et al. compared the utility of double-contrast CT-arthrography and conventional radiography in patients with pain, locking, and limited elbow motion. They found that double-contrast CT-arthrography successfully diagnosed 100% of loose bodies and provided precise information regarding the size, number, and location of the lesions. In contrast, conventional radiographs identified only 50% of the intraarticular loose bodies. Zubler et al. arrived at the same conclusion, noting greater accuracy of loose body detection with CT than with conventional radiographs, but particularly in the posterior fossae.
On the other hand, not all investigators agree that CT is necessary for loose body detection. Dubberley et al. found that CT and MRI were no more effective than conventional radiography alone for the detection of loose bodies. Quinn et al. did not advocate CT but recommended MRI for the accurate assessment of elbow intraarticular loose bodies. An advantage of using MRI is that it enables the detection of cartilaginous or osteocartilaginous fragments associated with osteochondritis dissecans of the capitellum in young athletes that may elude characterization by conventional radiography or CT. Furthermore, MRI may distinguish osteophytes and synovial hypertrophy that often mimic loose body formation. The additional diagnostic role of magnetic resonance arthrography (MRA) has not been well investigated to date.
Collateral Ligament Injury
Conventional radiographs are usually insufficient for detecting elbow collateral ligament injury. To detect elbow instability resulting from collateral ligament injury, varus or valgus stress is applied to the elbow during radiography to visualize asymmetric widening of the joint. Stress views that show a more than 0.5-mm increase in joint space are considered abnormal. Other examinations such as CT-arthrography, MRI, and/or MRA offer additional information.
Ulnar Collateral Ligament Injury
The ulnar collateral ligament (UCL) is an important valgus stabilizer of the elbow and is particularly vulnerable to injury in athletes whose sport involves throwing, such as baseball pitchers and javelin throwers. Interpretation of valgus stress radiography in athletes can be challenging. Several investigators have shown increased valgus laxity in the dominant arm of asymptomatic persons. When the index of clinical suspicion for a UCL injury is high, MRI can provide additional information. The UCL is normally seen on MRI as a vertically oriented, uniformly low-signal-intensity structure coursing between the medial epicondyle and the coronoid process on coronal images. The UCL is composed of anterior, posterior, and transverse bundles. The anterior bundle is considered the most important for valgus elbow stabilization. MRI may detect full-thickness tears of the anterior bundle, with increased signal intensity often found within and adjacent to the discontinuous ligament on fat-suppressed T2-weighted images as a result of edema and/or hemorrhage. MRI is less reliable for partial-thickness tear detection. Timmerman et al. found MRI to be 100% sensitive for full-thickness tears but only 14% sensitive for partial-thickness tears. Administration of intraarticular contrast material improves the sensitivity of partial-thickness UCL tear detection ( Fig. 63-3 ). Schwartz et al. reported 86% and 100% sensitivity and specificity, respectively, for partial-thickness UCL tear detection with MRA. Variability in the insertion of the anterior bundle, however, complicates MRI interpretation of partial-thickness tears. The distal ulnar attachment of the anterior bundle may insert anywhere between 1 mm of the articular margin of the coronoid process and 3 mm distal to the sublime tubercle of the ulna. This characteristic can create a small recess on MRA along the medial margin of the coronoid process. Consequently, distinguishing between normal anatomy and a pathologic partial undersurface tear at the distal ligament attachment can be challenging, and clinical correlation is paramount.
Lateral Collateral Ligament Injury
The lateral collateral ligament is consistently composed of the lateral UCL (LUCL), the radial collateral ligament (RCL), and the annular ligament. The LUCL is found posteriorly extending from the lateral humeral condyle to the supinator crest of the ulna ( Fig. 63-4, A ). The RCL arises from the lateral humeral epicondyle, deep to the common extensor tendon, and blends into the annular ligament surrounding the radial head. On MRI, the RCL and LUCL are best seen on sequential coronal images with an image slice thickness of 2 mm or less. The annular ligament is optimally seen on axial MRI images.
Tears of the lateral collateral ligament complex are common in persons with acute elbow dislocations. The LUCL is the primary varus elbow stabilizer. MRI findings of acute LUCL tears include combinations of ligament attenuation, redundancy, and discontinuity evident on coronal and/or axial images ; they are most often seen at the humeral origin of the ligament ( Fig. 63-4, B ). Chronic insufficiency of the LUCL, which can lead to posterolateral rotatory instability, is difficult to identify on MRI. Terada et al. found that MRI of asymptomatic persons revealed inconsistent signal characteristics within the LUCL or frank inability to identify the structure altogether. In contrast, Potter et al. found that abnormalities in the LUCL could be reliably detected with three-dimensional gradient-recalled and fast spin echo MRI in persons with posterolateral rotatory instability compared with asymptomatic control subjects. The role of MRA in this setting is unclear and not currently deemed to be a superior method of characterizing injuries of this ligament complex.