The practice of arthrography, which involves the radiographic visualization of a joint after the introduction of a contrast medium, has been an essential technique in musculoskeletal radiology for nearly a century. Its development traces a compelling narrative of medical innovation, reflecting the broader evolution of diagnostic imaging from 2-dimensional shadows to multiplanar, high-resolution anatomical analysis. The principles established in the early 20th century for visualizing the internal structures of the glenohumeral joint laid the groundwork for the sophisticated procedures used in contemporary clinical practice. The trajectory of this technique is not one of simple replacement by newer modalities but rather a story of adaptation and integration, where a foundational concept was repurposed to augment the power of successive technological breakthroughs. Pioneering Efforts and Foundational Concepts In 1933, Oberholzer pioneered the use of shoulder arthrography by injecting air into the glenohumeral joint to visualize the structures of the axillary recess on conventional radiographs. This early work established the fundamental principle that introducing a contrast agent could delineate intra-articular anatomy that was otherwise invisible on plain films. The potential of arthrography for diagnosing specific soft-tissue injuries was quickly recognized and validated by subsequent researchers. The technique saw its widest use and application during the 1960s and 1970s, becoming a mainstay for evaluating joint pathology before the advent of modern cross-sectional imaging modalities. The Paradigm Shift to Cross-Sectional Imaging The introduction of computed tomography (CT) and magnetic resonance imaging (MRI) represented a paradigm shift in medical imaging, resulting in a significant decrease in indications for conventional arthrography. The superior visualization of cross-sectional anatomy led to CT arthrography (CTA) replacing conventional arthrography as the standard procedure during the 1980s. However, it was largely superseded by MR arthrography (MRA) in the 1990s, as MRI offered superior soft-tissue contrast. However, the technique was not rendered obsolete. Instead, its fundamental principles were integrated with the new, more powerful modalities. Clinicians and radiologists recognized that the limitations of noncontrast CT and MRI could be overcome by applying the arthrographic principle of intra-articular contrast injection and joint distention. This led to the development and widespread adoption of CTA and MRA, the modern iterations of the procedure. This evolutionary path illustrates that foundational techniques are not discarded, but rather synergistically combined with new innovations to create a more comprehensive approach. Arthrography continues to provide invaluable anatomical information and an accurate depiction of internal joint derangement, solidifying its role in the modern imaging armamentarium. Foundational Principles of Glenohumeral Arthrography The enduring utility of arthrography, in both its conventional and modern cross-sectional forms, is rooted in 2 fundamental mechanical and radiological principles: the outlining of structures with a contrast agent and the physical separation of those structures through capsular distention. This dual benefit transforms the joint from a space of contiguous, radiographically similar tissues into a clearly delineated anatomical map, allowing for the detection of subtle pathology that would otherwise remain occult. Contrast and distention dual benefits In standard radiographic examinations, soft tissues, eg, articular cartilage, muscles, menisci, and synovial fluid possess very similar x-ray attenuation properties, resulting in a uniform density on the final image. Arthrography directly addresses this limitation by injecting a contrast agent into the joint. For conventional arthrography and CTA, an iodinated, radio-opaque agent is used; for MRA, a dilute solution of a gadolinium-based agent is employed. This injected contrast material flows into the joint space, coating and outlining the intra-articular structures. Simultaneously, the volume of injected fluid produces capsular distention, which provides a distinct mechanical advantage. The inflation of the joint capsule separates structures that typically lie in close apposition, eg, the glenoid labrum from the glenoid rim. This separation is crucial for assessing the integrity of these structures and for detecting small or nondisplaced tears, adhesions, or small intra-articular loose bodies. The combination of contrast enhancement and joint distention afforded by direct arthrography optimizes the evaluation of the full spectrum of intra-articular anatomy. Arthrographic Procedure Techniques Direct arthrography Direct arthrography is an invasive procedure that requires meticulous technique to ensure patient safety and diagnostic quality. The standard procedure involves accurate needle placement into the glenohumeral joint space, which is paramount and is most commonly achieved under fluoroscopic guidance, allowing the practitioner to visualize the needle's path in real-time. The flow and pattern of contrast filling can immediately highlight abnormalities, eg, abnormal leakage indicative of a tear or abnormal synovial patterns suggestive of synovitis. After confirming proper placement of the needle within the intra-articular space, synovial fluid should be aspirated for analysis. To exclude a potential infection, the fluid must first undergo visual inspection, focusing on clarity, consistency, and odor. At the same time, the soft tissues surrounding the puncture site should be examined for local signs of infection. For a rapid bedside evaluation, a urine strip containing a leukocyte esterase test can be used when available. These strips detect the leukocyte esterase enzyme released by neutrophils in response to inflammation or infection. A positive leukocyte esterase result may occasionally reflect conditions, eg, gout or inflammatory arthropathy, particularly when clinical signs of infection are absent. In cases where clinical findings and laboratory results conflict, the safest approach is to delay the procedure until the infection is definitively excluded. Once safety has been confirmed, the contrast agent can be administered, either alone or in combination with an anesthetic or glucocorticoid for therapeutic purposes. Following injection, the patient proceeds to definitive imaging, most commonly with CT or MRI. Indirect arthrography The standard procedure described above is known as direct arthrography. An alternative, less commonly performed technique is indirect arthrography, where a contrast agent is administered intravenously (IV) and passively diffuses into the joint space. While this technique can be useful, direct arthrography remains the predominant method because it allows for greater control over the degree of joint distention and achieves a higher concentration of intra-articular contrast, which generally provides superior delineation of fine anatomical details. Modern Shoulder Imaging Modalities The diagnostic evaluation of shoulder pain is a multimodal process, relying on a sophisticated and complementary array of imaging techniques. The selection of the appropriate modality is a clinically driven, algorithmic process, guided by a thorough history and physical examination. Understanding the distinct strengths and limitations of each modality is crucial for efficient and accurate diagnosis. Radiography Across the spectrum of shoulder pathology, whether traumatic or atraumatic, conventional radiography serves as the universally accepted first-line imaging modality. Its wide availability and relatively low cost establish it as the most effective initial screening tool. In the evaluation of trauma, a standard radiographic series should include at least 3 views to detect joint malalignment and identify most clinically significant fractures. In numerous clinical situations, eg, advanced osteoarthritis or calcific tendinitis, radiographs alone frequently provide sufficient diagnostic information. In these cases, additional imaging is often unnecessary, as radiography delivers the essential findings required for appropriate management. Computed tomography While radiography is excellent for initial fracture detection, CT is the modality of choice for the detailed characterization of complex osseous pathology. In cases of severe trauma with complex fractures, CT better defines the number, size, and displacement of fracture fragments than radiography, which are details often essential for preoperative planning. Magnetic resonance imaging MRI is the premier imaging modality for evaluating the soft tissues of the shoulder, a role attributed to its outstanding intrinsic soft-tissue contrast capabilities. MRI provides exquisite detail of the rotator cuff tendons, the biceps tendon, bursae, and surrounding musculature. In addition to assessing soft tissues, MRI is highly effective at detecting pathologies invisible on radiographs, such as subtle fractures, bone marrow edema, and early-stage osteonecrosis. Ultrasonography Ultrasound occupies a well-defined role in shoulder imaging as a dynamic, noninvasive, and cost-effective method for evaluating soft tissue structures. Its value is particularly evident in the assessment of the rotator cuff, where diagnostic accuracy for full-thickness tears rivals that of MRI. A major strength of ultrasound lies in its capacity for real-time, dynamic imaging, which allows for direct visualization of anatomic structures while simultaneously assessing functional abnormalities, eg, dynamic impingement. In addition, sonography has become a valuable tool for procedural guidance, including barbotage in patients with rotator cuff tendon hydroxyapatite deposition, thereby extending its clinical utility beyond diagnosis. Radionuclide bone scans Nuclear medicine imaging, specifically technetium-99m bone scintigraphy, plays a more specialized role. Its primary indications include the investigation of suspected prosthetic joint infections, situations where CT and MRI would be severely degraded by metal artifact, and the evaluation of suspected osseous metastases. Radionuclide imaging reflects functional metabolic activity rather than pure anatomy, making it highly sensitive for detecting infection and metastatic disease. Diagnostic Application and Accuracy in Rotator Cuff Pathology The diagnosis and characterization of rotator cuff tears are among the most common indications for advanced shoulder imaging. While conventional MRI and ultrasound are highly accurate, MR arthrography and CTA serve as powerful problem-solving tools in specific clinical contexts. The primary rationale for using MRA or CTA in the evaluation of the rotator cuff is to enhance the visualization of the intra-articular components of the shoulder. The injected contrast and capsular distention help in distinguishing high-grade, partial-thickness articular-sided tears from small, full-thickness tears. Arthrography is also indicated when there is a high clinical suspicion for a tear, but a prior non-contrast study is inconclusive. The relative diagnostic performance of MRA, CTA, MRI, and ultrasound for rotator cuff tears has been the subject of extensive investigation. Magnetic resonance arthrography MRA consistently proves to be the most accurate imaging test. A 2009 meta-analysis concluded that MRA was statistically more sensitive and specific than either MRI or ultrasound for both full- and partial-thickness tears. A more recent 2020 network meta-analysis reinforced these findings, ranking high-field MRA as having the highest diagnostic value for detecting any type of rotator cuff tear. Computed tomography arthrography CTA serves as a highly effective alternative when MRA is contraindicated. A 2005 study found that CTA demonstrated excellent sensitivity and specificity for diagnosing tears of the supraspinatus and infraspinatus tendons. However, its sensitivity for detecting tears of the subscapularis tendon was markedly lower. Magnetic resonance imaging and ultrasound Conventional MRI and ultrasound are the workhorse noninvasive modalities. Their diagnostic accuracy is generally considered comparable. The 2020 network meta-analysis gave a slight edge to MRI over ultrasound for detecting any tear. However, another meta-analysis concluded that because the accuracy of all 3 modalities (eg, ultrasound, MRI, and MRA) for full-thickness tears was high, ultrasound represents the best overall option when factoring in its lower cost and safety profile. Arthrography in Adhesive Capsulitis and Postoperative Evaluation Beyond the common indications of rotator cuff and labral tears, arthrography plays a specialized role in the diagnosis and treatment of adhesive capsulitis, as well as in evaluating the postoperative shoulder. Adhesive capsulitis Adhesive capsulitis ("frozen shoulder") is a condition characterized by the insidious onset of severe shoulder pain and a progressive loss of range of motion, resulting from fibrotic thickening and contraction of the joint capsule. The diagnosis of adhesive capsulitis is primarily based on clinical findings. While conventional arthrography was historically useful, the current radiological gold standard is now considered to be MRI or MRA, which can directly visualize pathologic changes, eg, thickening of the joint capsule and coracohumeral ligament. Beyond its diagnostic role, arthrography is also a key therapeutic intervention. The procedure of forcefully distending the contracted joint capsule with a large volume of saline, often mixed with a corticosteroid and local anesthetic (a method known as hydrodilatation), can mechanically stretch the fibrotic adhesions, leading to significant improvement in range of motion and pain. One study found that distention arthrography with an intra-articular steroid was superior to physical therapy alone for improving function. Postoperative and prosthetic joint evaluation Evaluating the shoulder after surgery, especially after the placement of a joint prosthesis, presents unique imaging challenges due to the presence of metal hardware artifacts on MRI. In this setting, CT and CTA are often the preferred imaging modalities. When the primary clinical concern is a post-arthroplasty infection, radionuclide imaging becomes the modality of choice, as it reflects functional metabolic activity and is not hampered by the presence of a metal prosthesis. Arthrography Contrast Agents The choice of contrast agent is fundamental to the arthrographic technique. Over the past 2 decades, an intense focus has been placed on the safety profile of these agents, particularly the gadolinium-based contrast agents (GBCAs) used for MRI. Types of contrast agents The primary classes of contrast agents used for direct arthrography are iodinated agents for CTA and gadolinium-based agents for MR arthrography. Notably, the intra-articular administration of GBCAs is technically considered an "off-label" use, as they are formally approved for IV use only. However, this practice has been the undisputed standard of care for decades and is supported by an extensive body of literature demonstrating its safety and efficacy. Nephrogenic systemic fibrosis Nephrogenic systemic fibrosis (NSF) is a rare but devastating systemic disease that occurs exclusively in patients with preexisting, severe renal impairment after exposure to GBCAs. In 2006, a strong association was identified between NSF and GBCAs, particularly older, less stable, and linearly structured agents (Group I GBCAs). This understanding led to a transformation in clinical guidelines. The widespread shift to using safer, more stable, macrocyclic-structured agents (Group II GBCAs) has been remarkably effective, leading to the near-total elimination of new, unconfounded cases of NSF worldwide. A 2020 meta-analysis of over 4,900 patients with stage 4 or 5 chronic kidney disease who received a Group II agent reported no cases of NSF. As a result, recent recommendations from professional societies state that kidney function evaluation is often no longer mandatory before the administration of Group II GBCAs. A critical point pertains to the safety of intra-articular administration. The risk of NSF is associated with systemic (intravenous) administration. For MRA, the dosage of GBCA is very small and is injected directly into the joint. The European Medicines Agency, while suspending the IV use of particular high-risk Group I agents, explicitly permits their use for MRAs, reflecting the distinct safety profile of intra-articular injection. Recommended Approach for Imaging the Painful Shoulder The American College of Radiology (ACR) Appropriateness Criteria represent a cornerstone of evidence-based practice in diagnostic imaging. The ACR guidelines are developed and reviewed by interprofessional expert panels that conduct an extensive analysis of the medical literature and establish recommended imaging approaches for painful shoulder evaluation based on the clinical presentation. The following 2 overarching principles emerge from the ACR guidelines regarding imaging patients with a painful shoulder: 1. Radiography first: The initial imaging modality of choice for virtually any new presentation of shoulder pain, whether traumatic or atraumatic, is conventional radiography. 2. Clinically driven selection: The choice of the subsequent imaging study after radiography is strictly guided by clinical factors, including the patient's history and physical examination findings. Imaging traumatic shoulder pain For patients presenting with shoulder pain directly attributed to a traumatic event, the 2024 ACR guidelines provide an algorithm based on clinical suspicion and imaging results (see Table 1. ACR Guidelines for Imaging Traumatic Shoulder Pain). Imaging atraumatic shoulder pain For patients with chronic or insidious-onset shoulder pain, the imaging algorithm recommended by the 2018 ACR guidelines is subtly different (see Table 2. ACR Guidelines for Imaging Atraumatic Shoulder Pain). Key Points Clinicians should keep in mind the following considerations when performing shoulder arthrography: 1. Most accurate modality: MRA is the most accurate modality for internal derangement. For diagnosing rotator cuff tears and glenoid labral tears, MRA consistently demonstrates the highest diagnostic sensitivity and specificity. However, its status as a "gold standard" must be qualified; arthroscopy remains the definitive reference standard, and the real-world performance of MRA may vary. 2. Imaging selection: The modern approach to shoulder imaging is a tiered, evidence-based algorithm, as codified by the ACR Appropriateness Criteria, which invariably begins with radiography. 3. Optimal imaging choice: The "best test" is context-dependent. The optimal imaging choice is not a fixed property of the technology alone, but is contingent on the patient's clinical state. The appropriateness of an invasive MRA versus a noninvasive MRI for a suspected labral tear, for example, changes depending on whether the injury is acute with a joint effusion or chronic and noneffusive. 4. Contrast agent safety: The safety profile of contrast agents is well-established. The risk of nephrogenic systemic fibrosis from gadolinium-based contrast agents has been effectively mitigated through the implementation of rigorous safety guidelines and the use of modern, stable Group II agents. The risk from a modern MRA is considered negligible. Glenohumeral arthrography has undergone a remarkable evolution, transforming from a foundational radiographic technique into a sophisticated adjunct that enhances the power of modern cross-sectional imaging. Its enduring value lies in the dual principles of contrast outlining and capsular distention, which together provide an unparalleled view of the intricate intra-articular anatomy of the shoulder. While conventional MRI and ultrasound are the primary workhorses for evaluating most soft-tissue shoulder pathologies, MRA and CTA remain indispensable tools for problem-solving and diagnosis.