Frozen shoulder is one of the few orthopaedic diagnoses in 2026 where a good clinical examination still beats a £500 scan. It is defined by a very specific pattern - the shoulder loses movement in every direction, active and passive, and it hurts most at night. If you have that pattern and it has lasted more than three months, you probably do not need an MRI to diagnose it. You need physiotherapy, patience, and sometimes a steroid injection.
But there are situations where an MRI genuinely earns its place. When the story does not fit. When treatment is not working. When surgery is on the table. And when the real problem might be a rotator cuff tear pretending to be a frozen shoulder. This piece walks through what an MRI actually shows in adhesive capsulitis, when the scan is worth doing, and what the report will and will not tell your consultant.
The one-line answer
Frozen shoulder is usually a clinical diagnosis. Most patients do not need an MRI. When one is warranted - atypical presentation, no response to treatment, or considering surgery - MRI shows thickening of the rotator interval and coracohumeral ligament, obliteration of the axillary recess, and helps rule out a rotator cuff tear that has been mimicking the picture.
When frozen shoulder needs no imaging
The classic textbook picture is unmistakable if you look for it. Global restriction of movement, meaning the shoulder will not rotate, abduct or reach behind the back, and crucially the restriction is the same whether the patient tries themselves or the clinician tries to move it. This is what separates it from a cuff tear, where passive movement is usually preserved.
If a patient walks in with more than three months of the following, MRI adds very little:
- Insidious onset shoulder pain without a specific injury, worst at night, waking them up when they roll onto that side.
- Loss of external rotation - they cannot rotate the arm outward with the elbow tucked to the side. This is the single most sensitive sign.
- Loss of movement in every plane, both active and passive.
- Risk factors - being female, aged 40 to 60, having diabetes, thyroid disease, a recent period of immobilisation, or previous frozen shoulder on the other side.
NICE, the British Elbow and Shoulder Society and most UK shoulder surgeons agree: a plain X-ray to exclude arthritis is reasonable in the first appointment. An MRI at that point is generally not.
When MRI is warranted
There are four clean indications where a shoulder MRI genuinely changes management.
- Atypical presentation. Sudden onset after a fall, weakness disproportionate to the pain, a lump, systemic symptoms, or a pattern that just does not fit adhesive capsulitis. Any of these should prompt imaging.
- Failure to respond to a steroid injection and structured physio at 3 to 6 months. If a well-diagnosed frozen shoulder is not thawing on the expected timeline, something else may be going on. Cuff pathology and calcific tendonitis both mimic the picture.
- Considering surgery. Before a manipulation under anaesthesia (MUA) or arthroscopic capsular release, most UK surgeons want an MRI to confirm the capsular thickening and rule out an unexpected cuff tear that would change the operation.
- Ruling out a cuff tear. A partial-thickness cuff tear can produce night pain and stiffness, especially if the patient has been guarding the shoulder for weeks. The two conditions look similar clinically. MRI settles it.
Outside those four situations, an MRI is usually reassurance for the patient, which is not nothing, but is not what the NHS pays for and is not the strongest reason to spend privately.
MRI signs of adhesive capsulitis
The radiologist is looking for a small set of findings that, together, are highly specific for adhesive capsulitis. No single sign clinches the diagnosis, but the pattern is usually clear.
- Thickening of the coracohumeral ligament. Normal is 2 to 3mm. In frozen shoulder it is often 4mm or more. This is one of the most reliable signs.
- Rotator interval capsule thickening and oedema. The rotator interval is the triangular gap between the supraspinatus and subscapularis tendons at the front of the shoulder. In adhesive capsulitis the tissue in this gap is thickened, inflamed, and often obliterates the normal fat triangle beneath the coracoid.
- Axillary recess capsular thickening. Normal capsule in the axillary pouch is 2 to 3mm and is a thin dark line on T2. In frozen shoulder it is thickened to 4mm or more, hyperintense on fluid-sensitive sequences, and this is where the “capsular oedema” pattern is most obvious.
- T2 hyperintensity and post-contrast enhancement. The thickened capsule lights up on T2 and, when contrast is given, enhances brightly. This reflects active inflammation in the freezing and frozen phases.
- Subcoracoid fat obliteration. The normal fat pad under the coracoid is squeezed out by inflamed synovium.
- Reduced joint volume. The joint capsule is contracted. On an arthrogram this is dramatic - normal shoulders take 12 to 15ml of contrast, frozen shoulders often refuse anything past 5 to 8ml.
The report you want back from the radiologist is not just “findings consistent with adhesive capsulitis”. It is that plus a clean statement on the rotator cuff, the labrum, the acromioclavicular joint and the glenohumeral joint. Those are the differentials that change treatment.
Differentials MRI helps rule out
The reason surgeons ask for the MRI is rarely to confirm the frozen shoulder - clinical examination usually does that. It is to rule out the four conditions that can wear the same clinical costume.
- Rotator cuff tear. Full-thickness and large partial-thickness tears cause pain, weakness and secondary stiffness. MRI is the gold standard for diagnosing them and will show the exact tendon, tear pattern, retraction and muscle quality that the surgeon needs.
- Calcific tendonitis. Calcium deposits in the supraspinatus tendon can cause sudden, severe, movement-limiting pain that mimics a frozen shoulder in its early phase. MRI (and an X-ray) will show the calcium clearly.
- Glenohumeral osteoarthritis. Older patients with joint-space narrowing, osteophytes and stiffness can look identical to adhesive capsulitis. An X-ray usually catches this first, but MRI adds detail on cartilage and bone oedema.
- Subacromial bursitis and impingement. Painful arc, weakness on resisted testing and stiffness from guarding. MRI shows bursal fluid, cuff tendinopathy and any acromial spurring.
