Contrast-Enhanced Joint Imaging for Intra-articular Pathology
Direct MR Arthrography: Joint injection (Gd-DTPA diluted) + MRI. Best for labral tears, partial cuff tears, capsular pathology
Indirect MR Arthrography: IV gadolinium + exercise + MRI. Non-invasive but inferior to direct
CT Arthrography: Joint injection (iodinated contrast) + CT. For bony detail, MRI contraindication
Conventional Arthrography: Joint injection + fluoroscopic radiographs. Largely replaced by MRA/CTA
Key: Direct MRA is the gold standard for intra-articular soft tissue pathology; CT arthrography excels at bony assessment
- Direct MR arthrography (MRA): gadolinium is INJECTED directly into the joint under fluoroscopic or ultrasound guidance, then MRI is performed.
- Indirect MR arthrography: gadolinium is given INTRAVENOUSLY, then the patient exercises to promote diffusion of contrast into the joint.
- Direct MRA is the gold standard for labral pathology — distension of the joint with contrast separates labral tissue from the glenoid, improving tear detection.
- CT arthrography uses iodinated contrast injected into the joint, followed by CT — excellent for bony detail (glenoid bone loss, Hill-Sachs) and in patients with MRI contraindications.
- The key indication for arthrography over non-contrast MRI is the improved detection of partial-thickness tears, labral tears, and capsular pathology.
- “Direct MRA pooled sensitivity for labral tears (about 88%) is significantly higher than non-contrast MRI (about 76%); MRA also has higher specificity (93% vs 87%).
- “Gadolinium for MRA is diluted to approximately 1:200 (2mmol/L) — undiluted gadolinium is TOO CONCENTRATED and causes signal loss on T1-weighted images.
- “Images must be acquired within 30-60 minutes of injection — contrast absorption reduces diagnostic quality rapidly.
- “CT arthrography is the preferred alternative when MRI is contraindicated (pacemaker, metal) and is superior for assessing glenoid bone loss.
- “The sublabral recess and sublabral foramen are NORMAL VARIANTS that mimic labral tears — knowing these prevents surgical over-treatment.
Arthrography is commonly examined in the context of shoulder instability assessment (labral tears, glenoid bone loss), rotator cuff assessment (partial-thickness tears), and hip labral pathology. You must know: the technique of direct vs indirect arthrography, the gadolinium dilution ratio, time constraints for imaging, and the specific advantages of MR vs CT arthrography. Classic traps include confusing normal sublabral variants with tears and not knowing when CT arthrography is preferred over MRA.
INJECTDirect MR Arthrography Technique
Hook:INJECT: the six steps of direct MR arthrography — from needle to imaging.
DISCAdvantages of Direct MRA
Hook:DISC: MRA acts like putting dye in a DISC to find cracks — distension reveals hidden tears.
BUFORDNormal Labral Variants
Hook:BUFORD complex: the most important normal variant to know — absent anterosuperior labrum with a thick cord-like MGHL.
Overview
Arthrography is the introduction of contrast material into a joint to improve the diagnostic imaging of intra-articular structures. It represents one of the most valuable diagnostic techniques for evaluating labral pathology, partial-thickness rotator cuff tears, cartilage lesions, loose bodies, and capsular abnormalities.
The fundamental principle is straightforward: by filling the joint with contrast, intra-articular structures are outlined and separated from one another. A tear that might be invisible on non-contrast MRI becomes conspicuous when contrast material insinuates into the defect, creating a high-contrast interface between the contrast and the adjacent tissue.
Direct arthrography involves percutaneous injection of contrast directly into the joint under image guidance (fluoroscopy or ultrasound). This provides consistent, controlled joint distension and high intra-articular contrast concentration. Indirect arthrography involves intravenous injection of gadolinium, followed by 10-15 minutes of gentle exercise to promote contrast diffusion across the synovial membrane into the joint. Indirect MRA is non-invasive but provides less reliable joint distension and lower intra-articular contrast concentration. Direct MRA is superior and is the technique of choice when arthrography is indicated.
MR arthrography uses dilute gadolinium (bright on T1) and provides excellent soft tissue contrast — ideal for labral tears, partial cuff tears, cartilage defects, and capsular pathology. CT arthrography uses dilute iodinated contrast and provides superior bony detail — ideal for glenoid bone loss quantification, Hill-Sachs measurement, and osteochondral defects. CT arthrography is the preferred alternative when MRI is contraindicated (pacemaker, non-MRI-conditional implants) or when bony detail is the primary clinical question.
Clinical Imaging
Imaging Atlas
Systematic Approach
Systematic Arthrogram Assessment
| Structure | Normal Appearance | Pathological Findings |
|---|---|---|
| Labrum | Triangular or rounded, firmly attached to glenoid rim, smooth margins | Tear: contrast extending into or beneath the labrum. Bankart: inferior labral detachment. SLAP: superior labral tear with or without biceps anchor involvement |
| Rotator cuff | Intact tendon without contrast extension into the tendon substance | Full-thickness tear: contrast extends from joint into subacromial space. Partial tear (articular surface): contrast insinuates into undersurface of tendon without full extension through |
| Biceps tendon | Located in the bicipital groove, intimately associated with the rotator interval | Biceps subluxation: tendon displaces medially over the lesser tuberosity. Sheath tear: contrast extends into the bicipital sheath |
| Capsule and ligaments | Capsule smoothly lines the joint. Glenohumeral ligaments visible as thickenings | HAGL: contrast extending beyond the humeral attachment site. Capsular redundancy: excessive volume in inferior pouch (MDI) |
| Articular cartilage | Smooth, uniform thickness articular surface without contrast undercutting | Chondral defect: contrast replaces or undercuts the cartilage surface. Grading by depth and area |
| Loose bodies | No filling defects within the contrast-filled joint | Filling defect: contrast surrounds a dense rounded opacity (chondral or osteochondral loose body). Best seen on CT arthrography |
Joint-Specific Applications
Shoulder MR Arthrography
The shoulder is the most common joint for arthrography and the most frequently tested in the fellowship exam. Direct MR arthrography of the shoulder is the gold standard for:
- Labral tears: Bankart (anterior-inferior), reverse Bankart (posterior), SLAP (superior), and HAGL lesions
- Partial-thickness rotator cuff tears: Particularly articular-surface partial tears (Ellman classification) that are difficult to see on non-contrast MRI
- Capsular pathology: Multidirectional instability (increased capsular volume), adhesive capsulitis (decreased volume)
- Loose bodies: Detected as filling defects within the contrast-filled joint
Injection technique: Under fluoroscopy or ultrasound guidance, a 22-gauge spinal needle is advanced into the glenohumeral joint (anterior approach, targeting the junction of the middle and lower thirds of the glenoid). A small test injection of iodinated contrast confirms intra-articular position on fluoroscopy. Then 10-20mL of dilute gadolinium (1:200) is injected.
Key sequences: T1-weighted fat-suppressed images in three planes (axial, coronal oblique, sagittal oblique) are the mainstay. T2-weighted sequences complement by showing oedema. The ABER position (abduction and external rotation) is obtained in the axial plane to improve visualisation of the anterior-inferior labrum and anterior band of the inferior glenohumeral ligament.
Sensitivity for labral tears: In a meta-analysis of 4,667 shoulders, direct MRA achieved a pooled sensitivity of 88% and specificity of 93% for glenoid labral lesions, compared to 76% sensitivity and 87% specificity for non-contrast MRI.
Evidence Base
MR Arthrography vs MRI for Glenoid Labral Injury
- Sixty studies pooling 4,667 shoulders from 4,574 patients, verified against arthroscopy or open surgery.
- MR arthrography pooled sensitivity 88% and specificity 93% for glenoid labral lesions, versus non-contrast MRI sensitivity 76% and specificity 87%.
- On summary ROC analysis MRA was marginally but consistently superior to MRI for detecting glenohumeral labral lesions (Level 2a evidence).
MRA vs Non-Contrast MRI for Rotator Cuff
- Meta-analysis of 65 studies with surgical (open or arthroscopic) reference standard comparing MRI, MR arthrography and ultrasound.
- MR arthrography was significantly more sensitive AND more specific than either MRI or ultrasound for both full- and partial-thickness rotator cuff tears.
- Area under the summary ROC curve was greatest for MRA (0.935), then ultrasound (0.889), then MRI (0.878); MRI and ultrasound did not differ significantly.
Why Gadolinium Is Diluted — Concentration & Iodine Effect
- Phantom study of gadopentetate diluted across 0.625–40 mmol/L in saline, albumin and iodinated contrast, scanned at 1.5T and 0.2T.
- Gadolinium signal is biphasic: signal rises to a peak then falls at higher concentrations — confirming that undiluted (highly concentrated) gadolinium loses T1 signal.
- Mixing gadolinium with iodinated contrast reduced T1 signal by ~26% at 2 mmol/L and shifted the signal peak to a LOWER gadolinium concentration.
MRA is most valuable for partial tears and labral lesions, and the dilute-gadolinium technique is grounded in the biphasic signal physics above.
Guidelines, Registries & Global Practice
Arthrography is performed worldwide by radiologists or musculoskeletal subspecialty radiologists on referral from orthopaedic surgeons and sports physicians. Direct MR arthrography is the internationally accepted standard for glenohumeral labral assessment and acetabular labral evaluation. The injection is performed under fluoroscopic or ultrasound guidance; a gadolinium-based contrast agent (commonly gadopentetate dimeglumine or an equivalent macrocyclic agent) is diluted to approximately 2 mmol/L (about 1:200) in saline, and imaging is completed within 30–60 minutes before synovial absorption degrades joint distension.
| Body / Society | Position relevant to arthrography | Practical implication |
|---|---|---|
| ACR (American College of Radiology) Appropriateness Criteria | MR arthrography rated 'usually appropriate' for chronic shoulder instability/labral assessment and for suspected hip labral tear/FAI; ultrasound and non-contrast MRI prioritised for many cuff questions | Reserve invasive arthrography for labral and partial-cuff questions, not routine full-thickness cuff or first-line screening |
| AAOS (US) clinical guidance | Imaging is adjunctive; advanced imaging (including arthrography) is directed by the specific surgical question rather than performed routinely | Order the arthrogram when it will change the operative plan (e.g. repair vs reconstruction) |
| BOA / BSSH / UK musculoskeletal radiology practice | MR arthrography or high-quality non-contrast 3T MRI accepted for labral work-up; resource and access drive local choice | Either route acceptable where 3T MRI is available and reported by MSK radiologists |
| ESSR (European Society of Musculoskeletal Radiology) | Consensus/technical guidelines describe direct MRA injection technique, contrast dilution and ABER positioning as standard MSK practice | Standardised injection volumes and dilution improve reproducibility across centres |
| ESR / national contrast-safety guidance (gadolinium) | Use macrocyclic (lower-risk) GBCAs; caution in severe renal impairment; document consent for intra-articular off-label use where required | Contrast-agent choice and consent are governed by general GBCA safety policy, not arthrography-specific rules |
Registry & outcome context. There is no dedicated arthrography registry; the downstream procedures that arthrography informs are tracked instead. Shoulder-instability and labral-repair outcomes feed into arthroplasty/soft-tissue datasets and instability cohorts, and glenoid bone-loss thresholds (the Sugaya en-face method and the glenoid track concept) drive the registry-relevant choice between Bankart repair and bony reconstruction (Latarjet/bone block). Accurate preoperative bone-loss quantification on CT (with or without intra-articular contrast) is therefore the key audited determinant of which instability procedure is performed.
High- vs limited-resource practice variation. In well-resourced settings, 3T MRI and dual fluoroscopic/US-guided injection allow routine direct MRA and 3D CT with en-face glenoid reconstruction for preoperative planning. Where MRI access is constrained, CT arthrography becomes the primary cross-sectional arthrogram because CT is faster, cheaper and more widely available, and still answers bony and many labral questions; conventional fluoroscopic arthrography remains a legitimate fallback for confirming intra-articular position or gross capsular leak. Where advanced imaging is unavailable, diagnostic arthroscopy assumes a larger role. This variation reflects equipment access and cost, not differing biology, and examiners expect candidates to justify the chosen modality by the clinical question and the local resource context.
Pitfalls: Normal Variant vs True Tear
The single most important interpretive skill in shoulder arthrography is distinguishing normal anterosuperior labral variants from true pathological tears. Mislabelling a normal sublabral foramen or Buford complex as a Bankart lesion can lead to inappropriate surgery.
| Feature | Normal variant (do NOT treat) | True labral tear (treat) |
|---|---|---|
| Location | Anterosuperior labrum, 1–3 o'clock (sublabral foramen); superior recess 11–1 o'clock | Anteroinferior (Bankart, 3–6 o'clock) or extending below the equator |
| Margins | Smooth, well-corticated, regular contrast cleft | Irregular, frayed or displaced labral fragment with contrast tracking into substance |
| Orientation of cleft | Smoothly contoured, parallel to glenoid rim (sublabral recess medially angled, smooth) | Laterally extending, irregular, full-thickness separation |
| Associated findings | No paralabral cyst, no bone oedema, intact periosteum | Paralabral cyst, glenoid rim fracture/oedema, capsular stripping (Perthes/ALPSA) |
| Buford complex clue | Absent anterosuperior labrum WITH a thick cord-like middle glenohumeral ligament | Do not mistake the cord-like MGHL for an avulsed labral fragment |
A sublabral foramen or recess at 1–3 o'clock with smooth margins and no paralabral cyst is a NORMAL variant. Contrast tracking below the equator (3–6 o'clock) with irregular margins is a Bankart-type tear. The Buford complex (absent anterosuperior labrum plus a thick cord-like MGHL) is the classic trap — never report the cord-like ligament as a displaced labral fragment.
Controversies & Areas of Uncertainty
As 3T scanners and high-resolution sequences have improved, the incremental benefit of invasive direct MRA over modern non-contrast 3T MRI for labral pathology is debated. Meta-analytic data still favour MRA for sensitivity, but some centres reserve arthrography for equivocal 3T studies, reducing the number of injections. The optimal threshold for proceeding to MRA is not standardised.
3D CT (Sugaya en-face) is the traditional reference for glenoid bone loss, but MRI-based and glenoid-track methods are increasingly used to avoid radiation and to assess bipolar (glenoid plus Hill-Sachs) loss together. The exact bone-loss percentage that mandates bony reconstruction (often quoted as 20–25%) is a continuum modified by the glenoid track concept rather than a single fixed cut-off.
Intra-articular gadolinium is an off-label use of GBCAs. Concerns about gadolinium tissue deposition have driven a shift toward macrocyclic agents, and some groups use saline-only or dilute iodinated CT arthrography to avoid intra-articular gadolinium altogether. Local consent and contrast-safety policy govern practice.
Indirect MRA avoids a joint puncture but gives unreliable distension and contrast concentration. Its role is contested: most experts limit it to situations where direct injection is refused or unavailable, and many MSK radiologists do not offer it at all.
Clinical Decision Scenarios
Practise clinical reasoning and management decisions out loud
“A 25-year-old man has recurrent anterior shoulder dislocations. His non-contrast MRI shows an equivocal anterior labral finding. You are considering further imaging.”
“A 30-year-old woman presents with deep groin pain and mechanical catching in her hip. You suspect a labral tear. Her plain radiographs show cam morphology.”
“An examiner asks you to compare direct MR arthrography, indirect MR arthrography, and CT arthrography. When would you choose each?”
Direct MR Arthrography (Gold Standard)
- Joint injection of dilute gadolinium (1:200 = 2mmol/L) under fluoroscopy/US
- Image within 30-60 minutes (contrast absorbs progressively)
- Pooled sensitivity for labral tears: ~88% (vs 76% non-contrast MRI); specificity 93% vs 87%
- Best for: labral tears, partial cuff tears, capsular pathology
- ABER position improves anterior labral visualisation
CT Arthrography
- Joint injection of dilute iodinated contrast + CT scanning
- Superior bony detail (0.3-0.5mm resolution)
- Gold standard for glenoid bone loss quantification (3D en-face views)
- Alternative when MRI is contraindicated (pacemaker, MRI-unsafe implants)
- More than 20-25% glenoid bone loss = Latarjet rather than Bankart repair
Injection Volumes
- Shoulder: 10-20mL (largest orthopaedic joint injection)
- Hip: 10-15mL (anterior approach, avoid neurovascular bundle)
- Ankle: 5-8mL
- Wrist: 3-5mL (smallest — risk of capsular rupture with overfilling)
Normal Variants (Do NOT Treat)
- Sublabral foramen: opening at 1-3 o'clock (12-18% of shoulders)
- Buford complex: absent anterosuperior labrum + thick cord-like MGHL
- Sublabral recess: superior labral recess at 11-1 o'clock (mimics SLAP tear)
- Key: smooth margins, consistent location, no paralabral cysts
Hip MRA Specifics
- Sensitivity 87% for labral tears (vs 66% non-contrast MRI)
- Alpha angle on radial sequences: more than 55-60 degrees = cam morphology
- Wave sign: chondral delamination (contrast undercuts cartilage)
- Anterior/anterosuperior tears most common in cam-type FAI