You have shoulder pain that will not settle, the GP has said the words "rotator cuff", and the internet is telling you contradictory things. One clinic wants to send you for an ultrasound this week for £300. Another wants an MRI for £750. A third says do the ultrasound first and only go to MRI if you need surgery. All three can be right, depending on the shoulder in front of them. This is how to pick.
Rotator cuff pathology is one of the most common musculoskeletal problems in UK adults and one of the most over-scanned. The imaging question is not really "which scan is better", because they are answering slightly different questions. The real question is which scan answers your question, at the right price, with the fewest wasted steps.
The one-line answer
For most people with straightforward cuff pain and no red flags, a musculoskeletal ultrasound done by an experienced operator is the correct first scan. It is faster, cheaper, and in the right hands as accurate as MRI for full-thickness tears. MRI becomes the right first scan when surgery is likely, when the labrum or biceps anchor is in question, or when the shoulder is too stiff and painful for ultrasound to work well. Frequently the best pathway is ultrasound now, MRI only if it changes management.
What ultrasound shows reliably
Modern musculoskeletal ultrasound, performed by a Fellowship-trained radiologist or an accredited musculoskeletal sonographer, is genuinely excellent for the rotator cuff. In the right hands its sensitivity and specificity for full-thickness supraspinatus tears matches MRI in most published series. What it does well:
- Full-thickness rotator cuff tears. Retraction, tear size and location can be measured directly. Supraspinatus, infraspinatus and subscapularis are all accessible.
- Partial-thickness tears on the bursal side, and moderate to large articular-sided tears.
- Calcific tendinopathy. Ultrasound is arguably better than MRI at showing calcific deposits and their consistency, which matters for barbotage treatment.
- Long head of biceps tendon pathology - subluxation, tenosynovitis, rupture.
- Subacromial-subdeltoid bursitis and effusion volume.
- Dynamic impingement. The patient can move the arm while the probe is on the shoulder. That is something MRI cannot do. Real-time impingement of the cuff under the acromion is visible and often obvious.
Ultrasound is also fast (typically 15 to 25 minutes), does not involve a magnet or contrast, and lets the radiologist explain findings on screen while the patient watches. For anyone claustrophobic, or with implants that make MRI difficult, that is a real advantage.
What MRI adds
MRI does what ultrasound cannot. Sound waves show soft tissue superbly but stop at cortical bone and cannot look inside the joint from all angles. A shoulder MRI, especially a dedicated musculoskeletal MRI, shows the entire architecture of the joint in three planes and answers the questions a surgeon actually needs answered before operating:
- Labrum. SLAP tears, Bankart lesions and posterior labral pathology are essentially invisible on ultrasound. MRI, especially MR arthrography, is the definitive test.
- Biceps anchor. Where the long head attaches to the labrum. Only MRI shows this reliably.
- Muscle atrophy and fatty infiltration. A torn supraspinatus with Goutallier grade 3 fatty replacement is likely not repairable. MRI grades this. Ultrasound cannot.
- Bone oedema and cystic change. Greater tuberosity oedema, subacromial cysts, and Hill-Sachs lesions from prior dislocation all live on MRI.
- AC joint and acromial anatomy. Bony spurs, os acromiale and the shape of the acromion - all part of the surgical picture.
- Cuff repairability. The combination of tear size, retraction, muscle quality and bone health tells the surgeon whether to repair, augment or reverse. This is an MRI conversation.
None of this is theoretical. If the answer to your imaging is going to be "let us book surgery", you want an MRI in the file, because the surgeon will ask for one anyway.
| Clinical question | Better first test |
|---|---|
| Is there a full-thickness rotator cuff tear? | Ultrasound |
| Is there a labral or SLAP tear? | MRI (or MR arthrogram) |
| Is the tear repairable? Fatty infiltration? | MRI |
| Is there calcific tendinopathy needing barbotage? | Ultrasound |
| Is there subacromial impingement in motion? | Ultrasound (dynamic) |
The cost gap, honestly
All-in UK private pricing in 2026 sits in two well-defined bands. A diagnostic musculoskeletal shoulder ultrasound with a consultant radiologist is £250 to £450. A dedicated shoulder MRI is £550 to £1100, with the top end reserved for central London clinics and MR arthrograms with contrast injection.
The temptation is to view the ultrasound as the cheap option and the MRI as the "proper" one. That is not the right frame. A well-targeted ultrasound that answers your clinical question is a better spend than a bigger MRI that answers questions nobody asked. Equally, a £320 ultrasound that misses a labral tear is not a saving if you end up paying for the MRI two weeks later anyway.
