There is a small, well-established procedure that a lot of UK patients get pointed toward without ever really being told what it is, or why the ordinary MRI they just had is being repeated with a needle in the joint. It is called an MR arthrogram. It is not a different scanner and it is not a different scan. It is the same magnet with a small volume of dilute contrast injected directly into the joint capsule, minutes before the pictures are taken.
The reason to bother is straightforward. A joint that has been gently distended by fluid is a joint whose surfaces separate on the images. That separation lets the radiologist see structures that hug each other so tightly on a standard MRI that small tears and delicate flaps blur into the background. When the question is a labrum or a ligament, the extra needle changes what the report can honestly say.
One-line answer
An MR arthrogram is worth the extra needle when the clinical question is a labral tear (shoulder or hip), a triangular fibrocartilage complex (TFCC) tear at the wrist, a partial ulnar collateral ligament (UCL) tear at the elbow, or a suspected loose osteochondral fragment at the ankle. Almost everything else is a standard MRI.
What arthrography actually is
Arthrography, in modern UK practice, is a two-step procedure done inside a single appointment. First, a musculoskeletal radiologist or interventional consultant uses either fluoroscopy (low-dose X-ray) or high-resolution ultrasound to guide a fine needle into the joint capsule. The skin is cleaned with sterile prep, numbed with local anaesthetic, and the needle is advanced under real-time image guidance so that its tip sits inside the joint rather than in the soft tissues around it.
A small test injection of iodinated contrast (for fluoroscopy) or saline (for ultrasound) confirms the tip is intra-articular. Then 10 to 20 millilitres of dilute gadolinium contrast is injected slowly. The joint is gently manipulated to spread the fluid across every recess. The patient then walks (or is wheeled) across to the MRI scanner and the actual imaging begins, usually within 15 minutes of the injection.
The MRI protocol itself is standard - high-resolution T1 and T2 sequences in three planes - but the contrast inside the joint acts like a highlighter pen. Fluid tracks into any tear, undercuts any loose flap, and outlines any structure that ordinarily sits invisibly against its neighbour.
The five joints where arthrogram genuinely wins
Not every joint benefits equally. The evidence, and the day-to-day practice of UK musculoskeletal radiology, points to five clear winners:
- Shoulder - SLAP and labral tears. A superior labrum anterior to posterior (SLAP) tear, a Bankart lesion or a small posterior labral tear can be nearly invisible on standard MRI. Intra-articular contrast lifts the torn edge off the bone and makes it obvious. This is the single most common indication for an MR arthrogram in the UK.
- Hip - acetabular labral tears. The hip labrum is a thin fibrocartilage rim, and small tears in patients with femoroacetabular impingement (FAI) are frequently missed on standard MRI. Sensitivity for labral tears jumps from around 60 per cent on standard MRI to over 90 per cent with arthrography.
- Wrist - triangular fibrocartilage complex (TFCC). For persistent ulnar-sided wrist pain after a fall, a three-compartment wrist arthrogram is often the only way to catch a small TFCC perforation or a subtle scapholunate ligament tear before it drives long-term instability.
- Elbow - ulnar collateral ligament (UCL). In throwing athletes, a partial-thickness UCL tear (the classic "T-sign") is a subtle finding on standard MRI. An arthrogram makes the undersurface of the ligament visible and turns a maybe into a yes or no.
- Ankle - osteochondral lesions of the talus (OCL). When the question is whether an osteochondral fragment is loose, stable or in situ, only intra-articular contrast will tell you if fluid is tracking under the fragment - which is what decides whether the patient needs surgery or conservative care.
When standard MRI is enough
The needle is not free of cost or discomfort, and it is not always the right answer. A well-run musculoskeletal MRI without contrast will answer the question comfortably when:
- The question is the rotator cuff and there is no labral concern. Full and partial-thickness rotator cuff tears are well visualised on standard shoulder MRI. If the referrer is not asking about the labrum, the arthrogram adds nothing.
- The question is a degenerative meniscus. Age-related meniscal tears in the knee are one of the highest-yield findings on any MRI. Knee arthrograms are rarely performed in modern UK practice for this reason.
- The lesion is large and obvious. A large osteochondral defect, an obviously ruptured tendon, a moderate or advanced osteoarthritis picture - none of these need the resolution uplift the arthrogram offers.
A useful test: if the answer the clinician needs is "is there a large structural problem here", standard MRI is enough. If the answer is "is there a small tear at a specific location that will change whether surgery is offered", the arthrogram is usually worth it.
The procedure: what patients experience
Patients often arrive nervous about the injection and leave surprised at how minor it was. A typical UK appointment runs like this. You check in, change into a gown and confirm allergies and kidney function with the radiographer. You lie on the fluoroscopy or ultrasound table for the injection - shoulder patients lie on their back with the arm rotated, hip patients lie flat, wrist and ankle patients sit or lie depending on the joint.
The skin is cleaned and numbed. Most patients describe the local anaesthetic as a small sting, similar to a dental injection. The intra-articular needle itself is rarely painful - the joint capsule is well anaesthetised - and there is a brief sensation of pressure or fullness as the contrast enters the joint. The whole injection takes about 15 minutes.
You then walk to the MRI scanner (this movement helps the contrast spread) and the scan runs for 30 to 45 minutes. Total time on the day, door to door, is around 90 minutes. Most patients drive themselves home afterwards. A mild ache in the joint for 24 to 48 hours as the fluid absorbs is normal, and settles with paracetamol and a quiet day.
