Radiologic evaluation of the painful shoulder in adults

INTRODUCTION — The complex anatomy and physiology of the shoulder, the most mobile and unstable joint, requires choosing proper imaging exams for different clinical scenarios. The imaging modalities and recommendations for selecting appropriate exams in different clinical situations are discussed in this review. The history and physical examination, which are necessary to develop a differential diagnosis prior to the selection of imaging tests, are presented separately. (See "Evaluation of the adult with shoulder complaints".)

A general review of imaging tests that are used in the evaluation of bone and joint pain, including the physical principles underlying their usefulness, is presented separately. (See "Imaging techniques for evaluation of the painful joint".)

IMAGING MODALITIES — A variety of imaging studies are available that are appropriate in different circumstances. These are listed in the table (table 1) and are discussed in more detail below:

Radiography — Radiography, also called plain film, standardly includes internal rotation, external rotation, and axillary and trans-scapular views and is the first imaging modality for virtually all shoulder pathology. Shoulder radiographs are often the only imaging exam necessary for the evaluation of acute shoulder trauma, calcific tendonitis, arthritis, and osteolysis of distal clavicle (in athletes) [1].

Computed tomography — Computed tomography (CT) of the shoulder is usually reserved for evaluation of fracture/fracture-dislocation or for a prosthetic joint. CT can demonstrate fracture complexity, displacement, and angulation. The ability to visualize images in the axial, sagittal, and coronal planes and in three-dimensional format can help in interpretation and preoperative planning [2]. For most indications, intravenous contrast is not administered and its use is reserved for evaluation of soft tissue masses for suspected bone tumor or abscess.

Magnetic resonance imaging — Magnetic resonance imaging (MRI) is the main modality used for evaluation of the soft tissues of the shoulder, such as the rotator cuff, biceps muscles, and tendons, and of the subacromial/subdeltoid bursa (image 1) [3,4]. For most indications, intravenous contrast is not administered and its use is reserved for evaluation of soft tissue masses for suspected bone tumor or abscess. MRI is also a sensitive modality for detecting subtle fractures, erosive changes of the distal clavicle, acromioclavicular joint changes, morphology of the acromion, early avascular necrosis, bone marrow edema, and muscle atrophy. MRI is a second procedure of choice for evaluation of labral tear or shoulder instability when MR arthrography is not performed.

Interpretation of MRI of the shoulder can vary between experts; collaboration between the clinician and radiologist in assessing imaging findings is often of benefit. Abnormalities detected by MRI of the shoulder should be carefully interpreted in the context of the patient's clinical presentation, as abnormal findings do not always explain the patient's symptoms.

Radionuclide bone scan — Technetium-99m bone scan with single-photon emission CT (SPECT) imaging is limited to evaluation of infection after arthroplasty (shoulder imaging) and for suspected metastases (whole body imaging).

Arthrography — A percutaneous puncture of the shoulder and instillation of a contrast agent into the joint are the initial step in any form of shoulder arthrography. Iodinated contrast is used for conventional and CT arthrography, and gadolinium contrast is used for MR arthrography. Contrast injection into the joint space is frequently performed with fluoroscopic guidance to facilitate needle placement. Systemic intravenous contrast is not administered with arthrography.

Once the contrast agent is instilled in the joint, its presence makes the identification of normal or damaged structures within the joint easier. Extravasation of contrast into the subacromial and subdeltoid space, for example, is consistent with a full thickness rotator cuff tear.

Conventional arthrography — In conventional arthrography, radiographs are obtained after instillation of iodinated contrast. Conventional arthrography is the procedure of choice for the diagnosis of frozen shoulder (adhesive capsulitis) [5]. Distention of the shoulder capsule during arthrography may be useful therapeutically. (See "Frozen shoulder (adhesive capsulitis)", section on 'Intraarticular dilatation (distension)'.)

Magnetic resonance arthrography — MR arthrography involves the intraarticular injection of a gadolinium-based contrast agent followed by MRI. MR arthrography is the gold standard for evaluation of shoulder instability or labral tear (image 2) [3,6,7]. MR arthrography is also indicated when there is high suspicion of rotator cuff tear with a normal or inconclusive MRI, and for evaluation of intraarticular small bodies [7]. Intraarticular anesthetic injection at the time of MR arthrography can help determine whether the pain originates from the glenohumeral joint [7].

Systemic administration of gadolinium-containing MRI contrast agents in patients with moderately to severely impaired renal function (eg, estimated glomerular filtration rate <15 to 30 mL/min) has been associated with increased risk of development of nephrogenic systemic fibrosis (NFS). With small doses and intraarticular administration for arthrography, NSF has never been reported. The approach to the use of gadolinium in such patients is discussed in detail separately. (See "Nephrogenic systemic fibrosis/nephrogenic fibrosing dermopathy in advanced kidney disease".)

Computed tomography arthrography — CT arthrography (image 3) is used when MR arthrography is contraindicated (eg, pacemaker, incompatible vascular clips, claustrophobia) or unavailable, or when a prosthetic joint is being evaluated, which would result in artifact on MRI [8]. With multidetector CT (MDCT) arthrography, tears of the supraspinatus and infraspinatus tendons of the rotator cuff can be detected and characterized with similar accuracy as MR arthrography [9]. In addition, MDCT arthrography is superior to MR arthrography in the detection of osseous, cartilage, and labroligamentous injuries related to anterior shoulder instability and helps select the proper surgical treatment [10]. However, considering the ionizing radiation and the decreased soft tissue and bone marrow resolution, CT is not routinely utilized for initial labroligamentous assessment.

Ultrasonography — Ultrasonography of the shoulder is useful in the evaluation of the rotator cuff [3,11-15], biceps tendon [3,11,13-15], and calcific deposits [3,11,13,16]; in the measurement of subacromial space [17]; and in detection of muscle atrophy [13,18]. Ultrasound is valuable in the dynamic evaluation of shoulder impingement [3,13,15]. Ultrasound can also be used to guide therapeutic interventions [13,19-23] and to detect periarticular soft tissue pathology in the setting of shoulder arthroplasty [24]. (See "Musculoskeletal ultrasound of the shoulder" and "Musculoskeletal ultrasonography: Clinical applications" and "Musculoskeletal ultrasonography: Nomenclature, technical considerations, and basic principles of use".)

Advantages of ultrasound include [22,25]:

●Portability

●Lack of contraindications

●No ionizing radiation compared with radiography or CT

●Real-time diagnosis

●Utility for needle guidance for aspiration and/or injection, and fine-needle fenestration for calcific tendinosis

●Better patient tolerance compared with MRI

Limitations of ultrasound include [11,13,26-29]:

●Limited acoustic window due to presence of bone

●Requirement for extensive training and lack of generalizability, as it is highly operator-dependent

●Limited availability as operator expertise required on site

●Does not image osseous lesions

●Lack of sensitivity for evaluating shoulder instability or labral tear

●Decreased accuracy in small or very large (>3 cm) rotator cuff tears

●Less sensitivity for partial thickness rotator cuff tears than MRI

RECOMMENDATIONS FOR SPECIFIC CLINICAL SETTINGS — The American College of Radiology (ACR) established criteria for selecting the imaging modalities of choice according to the most common clinical scenarios for traumatic shoulder pain and atraumatic shoulder pain [30,31]. Appropriate choice of imaging exam is based upon etiology of shoulder pain as traumatic versus atraumatic, duration of symptoms, age at presentation, and clinical or radiographic suspicion for a specific condition.

Traumatic shoulder pain — The role of various modalities in the workup of traumatic shoulder pain, based upon the ACR Appropriateness Criteria for Shoulder Pain – Traumatic, is discussed below [30].

Radiography — Radiography, also called plain film, of the shoulder should be the initial imaging test. Further imaging should be guided by the individual clinical scenario and by the plain film findings. For example, most patients under 30 years of age that suffer an anterior dislocation (image 4) of the shoulder have an associated labral injury [32]. MR arthrography is indicated in these cases. Anterior shoulder dislocation may also be associated with rotator cuff injury in approximately 50 percent of patients under the age of 40 and in 80 percent of patients over the age of 60 [33]. In such cases, ultrasound may be of value as the initial screening modality [34].

Computed tomography — CT can detect intraarticular extension of fracture and can demonstrate fracture complexity, displacement, and angulation, as well as small loose intraarticular fragments. With helical multidetector CT (MDCT) technology that allows for volumetric acquisition and multiplanar reformation, three-dimensional images can be generated, which can help the surgical planning.

Magnetic resonance imaging — MRI is the next imaging exam if radiographs are negative and the shoulder pain is nonlocalized. MRI is the procedure of choice for evaluation of soft tissue injury, such as rotator cuff tear and injury to biceps tendon, ligaments, and muscles [30,31]. MRI can also detect occult fractures. However, MRI is not the initial imaging modality for evaluation of acute shoulder trauma.

Ultrasonography — Ultrasonography is widely used as one of the initial modalities for evaluation of full thickness rotator cuff tear and biceps tendon tear, for which it has high sensitivity and specificity [3,11,13,15,34]. Ultrasound is equally useful in the evaluation of rotator cuff and biceps tendons [3,11,13-15]. Ultrasound has limited usefulness in evaluation of very small [27] or very large (>3 cm) rotator cuff tears [28,29], or in detection of partial thickness rotator cuff tears [11]. Ultrasonography is also recommended when MRI is not available or is contraindicated (eg, vascular clips, pacemaker). (See "Musculoskeletal ultrasound of the shoulder", section on 'Rotator cuff pathology'.)

Arthrography

Magnetic resonance arthrography — MR arthrography should be performed in young patients with suspected labral tear, in the setting of shoulder dislocation, or when shoulder instability is present [3,7]. MR arthrography is recommended for distinguishing between partial and full thickness rotator cuff tears and for identifying a suspected cartilage and labral tear [3,7]. When MR arthrography is contraindicated, CT arthrography can be used [10].

Computed tomography arthrography — CT arthrography can detect occult fractures when MRI is contraindicated. CT arthrography can also detect tendon and labral tears when MRI is contraindicated.

Atraumatic shoulder pain — The workup for atraumatic shoulder pain, based upon the ACR Appropriateness Criteria for Shoulder Pain – Atraumatic [31], is discussed below, based on individual clinical scenarios. The initial imaging modality for evaluation of atraumatic shoulder pain is radiography [31].

Atraumatic shoulder pain, suspected tendinosis or bursitis — Rotator cuff or biceps tendinosis and subacromial or subdeltoid bursitis are common causes of subacute or chronic shoulder pain in primary care practice. Isolated bursitis may be seen in trauma; inflammatory disorders, such as rheumatoid arthritis; and infection. (See "Rotator cuff tendinopathy" and "Biceps tendinopathy and tendon rupture".)

Radiography — Radiography of the shoulder that standardly includes internal and external rotation views is the first exam recommended in suspected tendinosis. Additional radiographic views may be of value. In the presence of calcific tendinosis, radiographic findings include focal, dystrophic calcification within the rotator cuff (image 5). If the radiographs are noncontributory, further workup based upon the ACR recommendations is presented in variant 4 of the ACR Appropriateness Criteria for Shoulder Pain – Traumatic.

Ultrasonography — If local expertise is available, ultrasonography is equivalent to MRI for evaluation of tendinosis or bursitis and is the modality of choice second to radiography for these indications [13]. Ultrasonography is highly sensitive and specific for detection of rotator cuff calcification [11]. The presence of an echogenic focus with definite acoustic shadowing is the most frequent ultrasonographic finding in calcific tendinosis [13,16]. Correlation with radiographs indicates that a hyperechoic focus with minimal or no acoustic shadowing may also be compatible with the presence of rotator cuff or bursal calcific deposits. New treatments for calcific tendonitis utilize ultrasound-guided fine-needle technique [19,20,22]. (See "Musculoskeletal ultrasound of the shoulder", section on 'Sonographic appearance of shoulder pathology'.)

Shoulder ultrasound is also highly sensitive for detection of subacromial bursitis with effusion [3,14]. The exam may include injection of anesthetic and/or glucocorticoid if clinically indicated. (See "Musculoskeletal ultrasound of the shoulder", section on 'Subacromial-subdeltoid bursa'.)

Atraumatic shoulder pain, suspected rotator cuff tear or impingement — Rotator cuff tear is the most frequent disorder in this category after tendinosis. The frequency of rotator cuff tears increases with age. One study reported rotator cuff tears in 34 percent of asymptomatic patients in all age groups and in 54 percent of asymptomatic patients over 60 years of age [35]. (See "Presentation and diagnosis of rotator cuff tears" and "Subacromial (shoulder) impingement syndrome".)

MRI and ultrasound are comparable in both sensitivity and specificity for the detection of rotator cuff tears [3,36]. MR arthrography is more sensitive and specific than either MRI (without arthrography) or ultrasound for diagnosing both full and partial thickness rotator cuff tears [7,36].

Radiography — Radiographs are usually normal or may show degenerative or osseous changes of the coracoacromial arch, including osteophytes or syndesmophytes. In the late stages of rotator cuff tear with complete loss of subacromial space, radiography may be diagnostic, and no further imaging is needed (image 6). The ACR recommendations for imaging of patients over age 35 with suspected rotator cuff tear or impingement and with normal radiographs or radiographs that demonstrate coracoacromial arch abnormalities are presented in variant 5 of the ACR Appropriateness Criteria for Shoulder Pain – Traumatic.

Magnetic resonance imaging — MRI of the shoulder without contrast is recommended as the initial imaging test when rotator cuff tear or impingement is suspected [4,6]. If subacromial injection were performed, MRI should be delayed a minimum of 24 hours to avoid misinterpretation of presence of fluid [37].

MRI demonstrates both osseous and soft tissue abnormalities associated with impingement syndrome [4,38,39]. Osseous abnormalities such as shape of the acromion process, humeral head abnormalities, and osteophyte formation of the acromioclavicular joint can be demonstrated [6]. Soft tissue findings include tendinosis (tendon degeneration or repeated chronic tendon injury without disruption), partial thickness rotator cuff tear (image 7), full thickness tear (image 1), biceps tendon abnormalities, thickened coracoacromial ligament, and subacromial bursitis [4]. With large full thickness chronic rotator cuff tear, muscle atrophy is usually seen. Some of the above osseous and soft tissue abnormalities can be seen in asymptomatic individuals, and clinical correlation is required [35,40].

Reported sensitivity and specificity of MRI for detecting full thickness rotator cuff tears are 92 percent [36]. Reported sensitivity and specificity of MRI for detecting partial thickness rotator cuff tears are 64 and 92 percent, respectively [36]. The lower sensitivity for MRI detection of partial thickness rotator cuff tears is due to different imaging techniques, criteria used for detection of these tears, and limitation of MRI in detecting articular surface partial tears of the rotator cuff relative to MR arthrography [7].

MRI is also helpful in the preoperative assessment of the feasibility of rotator cuff repair by measuring the size of the tear and the degree of muscle atrophy [41].

Ultrasonography — Ultrasonography is the preferred initial tool in the evaluation of the rotator cuff when specific expertise in shoulder sonography is available (image 8). Ultrasonography has comparable accuracy to MRI for identifying and measuring the size of full thickness and partial thickness rotator cuff tears when performed by an experienced examiner using high-quality equipment [3,13,36].

Sensitivity and specificity of ultrasound for complete rotator cuff tears are reported to be over 92 percent [36]. For partial thickness rotator cuff tears, sensitivity and specificity of ultrasound are 67 and 94 percent, respectively [36]. Diagnostic performance varies with the operator.

Dynamic ultrasound can also help to confirm the clinical diagnosis of impingement, with sensitivity between 70 and 80 percent and specificity over 95 percent [15,42]. Ultrasound is also highly sensitive in detection of effusion in the subacromial bursa as an indirect sign of rotator cuff tear [3,14]. In addition to evaluation of the rotator cuff tendon, ultrasonography can also detect the coexisting muscle atrophy, which influences the decision and outcome of surgery [13,18].

Limitations of ultrasonography in this setting include decreased accuracy in very small [27] or very large (>3 cm) rotator cuff tears [28,29] and in partial thickness rotator cuff tears [11]. (See "Musculoskeletal ultrasound of the shoulder", section on 'Rotator cuff pathology'.)

Magnetic resonance arthrography — MR arthrography is recommended if rotator cuff tear is suspected and MRI is normal, or if glenohumeral instability is suspected as a cause of impingement syndrome [3]. MR arthrography has greater sensitivity and specificity for detection of full thickness and partial thickness rotator cuff tears involving the articular surface compared with MRI and has greater sensitivity and specificity than ultrasound [7,36]. MR arthrography has sensitivity and specificity of over 95 percent for detection of full thickness tears [36]. For partial thickness tears, sensitivity and specificity are 86 and 96 percent, respectively [36]. MR arthrography differentiates between small partial tears, small full thickness tears, and tendinosis (image 9).

Computed tomography arthrography — CT arthrography is indicated if the patient cannot undergo MRI or if ultrasonography expertise is not available [8]. Sensitivity and specificity of over 90 percent for CT arthrography for detection of supraspinatus and infraspinatus tear have been reported, with low sensitivity for subscapularis tear detection [9]. CT arthrography is also valuable in preoperative evaluation by demonstrating the extent of tendon retraction and fatty degeneration of the corresponding muscle [8].

Atraumatic shoulder pain, suspected instability and/or labral tear — Subacute shoulder pain in patients under 35 years of age is often related to sports injuries. Injuries of the upper extremity account for more than 25 percent of all sports-related injuries. As much as 30 percent of adolescents and children participate in school athletics. In sports that involve repetitive overhead actions, such as baseball, football, basketball, tennis, and swimming, the most common injuries involve the posterosuperior labrum, supraspinatus tendon, infraspinatus tendon, humeral head, glenoid cavity, acromioclavicular joint, anteroinferior capsulolabral complex, biceps tendon, and subscapularis tendon [43,44]. The ACR recommendations for imaging of a patient with shoulder pain in whom glenohumeral joint instability or a tear of the glenoid labrum is suspected are presented in variant 3 of the ACR Appropriateness Criteria for Shoulder Pain – Traumatic and are discussed below.

Radiography — Radiography should be the initial imaging test. In cases of shoulder dislocation, plain film demonstrates humeral and glenoid fractures such as Hill-Sachs and Bankart lesions (image 10 and image 11). Further evaluation of the associated soft tissue injuries, including glenoid labral, capsular, and tendon tears, should be performed with MR arthrography.

Magnetic resonance arthrography — MR arthrography is the method of choice for imaging the labrocapsular structures [3,45]. MR arthrography is the most accurate imaging modality for sports injuries of the shoulder. For evaluation of anterior shoulder instability, MDCT arthrography has been shown to be more accurate than MR arthrography for osseous, cartilage, ligamentous, and labral injuries [10,44]. However, considering the ionizing radiation and the decreased soft tissue and bone marrow resolution, CT is not routinely utilized for initial labroligamentous assessment. MR arthrography evaluates the dynamic stabilizers (rotator cuff muscles and tendons), which are extraarticular; the static stabilizers (glenohumeral ligaments and glenoid labrum) and the long head of the biceps tendon, which are intraarticular; and capsular morphology (image 12) [3,44,45]. Sensitivity and specificity of MR arthrography for detection of labral tears range from 88 to 100 percent and 88 to 96 percent, respectively [46,47].

The use of abduction and external rotation (ABER) technique increases the sensitivity for labral tears to close to 100 percent [48]. This was illustrated in a study that compared radiography, arthrography, and CT arthrography with MR arthrography and the use of arthroscopic findings to confirm the diagnosis [43]. MR arthrography with ABER yielded the highest sensitivity and specificity for all lesions with the exception of bone sclerosis and enthesophytes. Therefore, MR arthrography is indicated in all suspected lesions of the rotator cuff and glenoid labrum in athletes [3].

Magnetic resonance imaging — With high field or high resolution and appropriate expertise, MRI is a good alternative to MR arthrography. MRI demonstrates bony and soft tissue injuries, including muscle, tendon, and labral tears [3]. A meta-analysis of MRI for detection of glenoid labral injury reported sensitivity of 76 percent and specificity of 87 percent in the detection of labral tears [49,50]. Evaluation of the capsular structures is limited.

Computed tomography arthrography — CT arthrography is accurate in delineating the anatomic derangement including the glenoid labrum (image 3); however, evaluation of soft tissues is limited. Thus, it is usually performed only when MR arthrography is contraindicated (as with shoulder prostheses, pacemakers, and MR-incompatible vascular clips) [8]. Reported sensitivity and specificity of CT arthrography for labral tears is between 73 and 76 percent and 92 percent, respectively [49].

Atraumatic shoulder pain, suspected adhesive capsulitis (frozen shoulder) — Adhesive capsulitis is caused by thickening and contraction of the joint capsule and synovium resulting in progressive limitation of the mobility of the glenohumeral joint, accompanied by significant pain [23]. Diagnosis is based mainly on clinical findings. Imaging may be used to exclude intraarticular or rotator cuff abnormalities, to confirm the diagnosis, and to guide therapeutic injections. (See "Frozen shoulder (adhesive capsulitis)".)

Conventional arthrography — Conventional arthrography is the procedure of choice for the diagnosis and treatment of frozen shoulder (adhesive capsulitis) [23]. Decreased capacity to contrast injection in the joint is diagnostic of frozen shoulder. Distention of the shoulder capsule during arthrography may be useful therapeutically. (See "Frozen shoulder (adhesive capsulitis)", section on 'Intraarticular dilatation (distension)'.)

Magnetic resonance imaging — MRI with intravenous contrast administration can show signal and thickness abnormalities of the shoulder joint capsule and synovial membrane and thickening of the glenohumeral ligaments [23]. Obliteration of the rotator interval fat pad is characteristic [23].

Radiography — Radiography has limited use and is nonspecific.

Ultrasound — Diagnostic ultrasound shows thickening of the coracohumeral ligament and rotator interval synovitis; however, it is not routinely performed [23]. Ultrasound can guide therapeutic injections into the shoulder joint, with improvement in clinical outcomes [22].

Chronic arthritis — Most shoulder arthritides can be accurately diagnosed with plain film radiography.

Radiography — Findings on radiography of osteoarthritis include joint space narrowing, hypertrophic bone formation (osteophytes), subchondral sclerosis, and subarticular cysts formation (image 13). Rheumatoid arthritis in the shoulder is characterized by "high-riding" humeral head. However, "high-riding" humeral head can also be seen with pseudogout and chronic rotator cuff tear. The acromioclavicular joint in patients with rheumatoid arthritis may reveal erosive change of the distal clavicle (image 14).

Magnetic resonance imaging — MRI is indicated when there is clinical suspicion for soft tissue abnormalities, such as rotator cuff involvement, subacromial bursitis, and muscular atrophy [51,52]. In addition, MRI will detect synovial hypertrophy and bone abnormalities such as bone erosions, small subchondral cysts, and bone marrow edema. MRI can help in diagnosis of synovial osteochondromatosis, amyloid arthropathy, and pigmented villonodular synovitis [52].

Ultrasonography — The most frequent sonographic finding in rheumatoid arthritis is the presence of an effusion in the subacromial or subdeltoid bursa and/or the glenohumeral joint; however, this is a nonspecific finding [53]. Additional findings include tenosynovitis, synovial proliferation, and bone erosion. (See "Musculoskeletal ultrasound of the shoulder", section on 'Glenohumeral joint pathology'.)

Septic arthritis — Nongonococcal infectious arthritis should be suspected as a cause of acute arthritis in an infant or an older patient. In other age groups, septic arthritis is often associated with medical conditions such as diabetes, liver diseases, alcoholism, malignancy, chronic arthritis, shoulder prosthesis, or surgery and associated infections [54]. In sexually active individuals, gonococcal arthritis should be suspected. (See "Septic arthritis in adults" and "Disseminated gonococcal infection".)

Radiography — Radiographs may demonstrate moderate to large joint effusion in the initial phases and, subsequently, joint space narrowing and erosive changes. Shoulder effusions are difficult to detect unless they are large enough to cause inferior displacement of the humeral head. ACR recommendations for further evaluation of septic arthritis are presented in variant 8 of the ACR Appropriateness Criteria for Shoulder Pain – Traumatic.

Ultrasound arthrocentesis — Shoulder joint aspiration is the procedure of choice. Echogenic material in synovial fluid in the glenohumeral joint has been noted in bacterial infections. Ultrasonography can be helpful for joint aspiration, especially in difficult cases, by providing real-time guidance or by allowing marking of the shoulder for joint aspiration [14,55]. (See "Musculoskeletal ultrasonography: Guided injection and aspiration of joints and related structures", section on 'Shoulder joint and subacromial/subdeltoid bursa'.)

Magnetic resonance imaging — MRI is highly sensitive but is not specific for the diagnosis of a septic joint [56]. The combination of bone erosions and bone marrow edema on MRI is highly suggestive of the diagnosis. Synovial and soft tissue findings are also well demonstrated on MRI. The addition of diffusion-weighted imaging (DWI) to routine MRI increases the specificity of MRI for differentiating effusion due to inflammation versus infectious arthritis and also for distinguishing reactive bone edema from osteomyelitis.

Avascular necrosis of bone — The presence of risk factors such as sickle cell disease or chronic glucocorticoid use may suggest avascular necrosis (osteonecrosis) of the humeral head. (See "Treatment of nontraumatic hip osteonecrosis (avascular necrosis of the femoral head) in adults".)

Radiography — Radiography is not sensitive in the early stages, when only a small joint effusion is present. Sclerosis, crescent sign, or humeral head collapse may be visible weeks to months after the onset of pain.

Magnetic resonance imaging — MRI is the most sensitive and specific modality for detecting the earliest changes when plain radiographs and radionuclide bone scans are normal [52].

Radionuclide bone scan — Radionuclide technetium-99m bone scan of the shoulder is not as sensitive as MRI; however, abnormal uptake appears earlier than radiographic changes on plain films.

Postoperative shoulder pain — Imaging of the postoperative shoulder is indicated most often due to persistence of pain or functional limitation. Rotator cuff repair, capsulorrhaphy, or acromioplasty can produce symptoms, which may be indistinguishable from rerupture or inadequate surgical results [57].

Radiography — Radiography is the initial imaging modality indicated after surgical procedures for fractures, tumors, acromioplasty, and soft tissue repair including rotator cuff, labrum, and ligaments. Radiography can confirm the correct position of metallic fixation devices or osseous tunnels and detect recurrent humeral head dislocation, subluxation, or tumor recurrence [57].

Magnetic resonance imaging — MRI is indicated for symptoms of persistent impingement or suspected reinjury of the rotator cuff and biceps tendon. MRI can evaluate the morphology of the glenoid and acromion, rotator cuff or biceps tendon retear, deltoid detachment, muscle edema and atrophy, and labral pathology due to residual impingement [57].

Magnetic resonance arthrography — MR arthrography is most useful for evaluation of labral pathology, optimal delineation of the rotator cuff, capsulolabral structures, and tendon tears [3,57]. MR arthrography is more accurate than MRI in the detection of partial rotator cuff retears [57].

Computed tomography — When MRI cannot be performed due to metallic artifact, CT is valuable for evaluation of bony and soft tissue abnormalities. Metal artifact-reduction (MAR) techniques should be used if available. MDCT with multiplanar reformation images is useful for evaluation of bony structures, including glenoid morphology, non-union, loosening, and fracture [57].

Computed tomography arthrography — MDCT arthrography is useful in the evaluation of the postoperative rotator cuff, biceps tendon, and labrum, especially when adjacent metallic hardware produce extensive artifacts on MRI. CT arthrography identifies the retear in the reconstructed rotator cuff, biceps tendon, and labrum [57].

Shoulder arthroplasty — Shoulder arthroplasty is increasingly common in orthopedic surgery practice. Successful evaluation of these different types of shoulder reconstruction requires an understanding of the complications and the role of various imaging modalities in assessment of prosthesis complications.

Radiography — Radiography is the initial modality of choice for evaluation of arthroplasty. Plain film demonstrates complications of the prosthesis such as fracture, subluxation or dislocation, glenoid or humeral component loosening, and periprosthetic bone resorption due to particle disease [58]. Plain film also evaluates the integrity of the bone surrounding the prosthesis.

Ultrasound — Patients with shoulder pain after arthroplasty and negative radiographs should undergo ultrasound [12]. Ultrasound has been shown to be very accurate in the evaluation of the rotator cuff in patients who have undergone arthroplasty. If the ultrasound scan is equivocal, CT arthrography is indicated. Ultrasound may also guide joint aspiration in suspected infected prosthesis. (See "Musculoskeletal ultrasound of the shoulder", section on 'Shoulder anatomy for ultrasound evaluation'.)

Computed tomography — When MRI cannot be performed in the presence of a shoulder prosthesis due to significant metallic artifact, CT can be valuable for evaluation of soft tissue or bony abnormalities. MAR techniques should be used if available. MDCT identifies abnormal alignment, version alteration, articular surface or bone stock deficiency, or other pathology compromising fixation and function [58].

Computed tomography arthrography — MDCT arthrography is useful in the assessment of the prosthetic and periprosthetic abnormalities, as well as rotator cuff and labral-capsular abnormalities [8]. Joint fluid can be aspirated during the intraarticular administration of contrast medium and sent for culture and sensitivity testing. Image postprocessing to produce volume-rendered three-dimensional CT images can be an important tool for assessment of hardware integrity.

Magnetic resonance imaging — Advances in MAR techniques have led to an increased use of MRI in the evaluation of the painful failed shoulder arthroplasty [59]. MRI assesses the integrity of the implant and the supporting soft-tissue envelope. MRI identifies component loosening and implant failure, rotator cuff and deltoid integrity, infection, subtle fractures, and nerve pathology.

Nuclear medicine imaging — Radionuclide technetium-99 bone scan with single-photon emission CT (SPECT) imaging of the shoulder is the imaging modality of choice for suspected joint replacement infection [60]. Bone scintigraphy is sensitive for identifying the failed joint replacement; however, it is nonspecific. Combined leukocyte with bone marrow scintigraphy has an accuracy of approximately 90 percent and is the imaging procedure of choice for diagnosing prosthetic joint infection [60].

SUMMARY AND RECOMMENDATIONS

●The choice of imaging modality depends upon the differential diagnosis generated based upon the history and physical examination. Each imaging technique has particular advantages and disadvantages for addressing different clinical questions (table 1). (See 'Introduction' above and 'Imaging modalities' above.)

●Radiography, also called plain film, is the first imaging modality for virtually all shoulder pathology. Radiographs are often the only imaging exam necessary for the evaluation of acute shoulder trauma, calcific tendonitis, arthritis, and osteolysis of distal clavicle. (See 'Radiography' above.)

●CT of the shoulder is usually reserved for evaluation of fracture or fracture-dislocation or for a prosthetic joint. For most indications, intravenous contrast is not administered and its use is reserved for evaluation of soft tissue masses and for suspected bone tumor or abscess. CT can demonstrate fracture complexity, displacement, and angulation. The ability to visualize images in the axial, sagittal, and coronal planes and in three-dimensional format can help in interpretation and preoperative planning. Use of intravenous contrast is reserved for evaluation of soft tissue and for suspected bone tumor or abscess. (See 'Computed tomography' above.)

●MRI of the shoulder is the main modality used for evaluation of the soft tissues such as the labrum, rotator cuff, biceps muscles, and tendons, and for the subacromial or subdeltoid bursa. For most indications, intravenous contrast is not administered and its use is reserved for evaluation of soft tissue masses and for suspected bone tumor or abscess. MRI is also a sensitive modality for detecting subtle fractures, erosive changes of the distal clavicle, acromioclavicular joint changes, bone marrow edema, and muscle atrophy. Abnormalities detected by MRI of the shoulder should be carefully interpreted in the context of the patient's clinical presentation, as abnormal findings do not always explain the patient's symptoms. (See 'Magnetic resonance imaging' above.)

●The use of technetium-99m radionuclide bone scan is generally limited to evaluation of infection after arthroplasty (shoulder imaging) and for suspected metastases (whole-body imaging). (See 'Radionuclide bone scan' above.)

●Arthrography of the shoulder involves the instillation of a contrast agent into the joint, often using fluoroscopic guidance to facilitate needle placement. Once the contrast agent is instilled in the joint, its presence makes the identification of normal or damaged structures within the joint easier, usually on radiography or MRI. CT arthrography is used only when MR arthrography is contraindicated or unavailable and when shoulder prostheses, which cause MR artifact, are being evaluated. Intravenous contrast is not administered. (See 'Arthrography' above and 'Conventional arthrography' above and 'Magnetic resonance arthrography' above and 'Computed tomography arthrography' above.)

●The availability of shoulder ultrasonography varies. It is useful in the evaluation of the rotator cuff, biceps tendon, subacromial-subdeltoid bursa, and calcific deposits; in the measurement of subacromial space; in the detection of muscle atrophy; and in the dynamic evaluation of shoulder impingement. It can also be used to guide therapeutic interventions and to detect periarticular soft tissue pathology in the setting of shoulder arthroplasty. (See 'Ultrasonography' above and "Musculoskeletal ultrasound of the shoulder".)

●The American College of Radiology (ACR) has established criteria for selecting the imaging modalities according to the most common clinical scenarios for traumatic shoulder pain and atraumatic shoulder pain. Appropriate choice of imaging exam is based upon duration of symptoms, age at presentation, and clinical suspicion for a specific condition. (See 'Recommendations for specific clinical settings' above.)