The short answer: A shoulder MRI shows whether a rotator cuff tendon is frayed, partly torn, or torn all the way through. It also shows how far the tendon has pulled back and whether the muscle has shrunk or turned to fat. A 2013 Cochrane review found MRI caught 94% of full-thickness tears.
What can a shoulder MRI see in the rotator cuff?
The rotator cuff is four muscles whose tendons wrap the ball of the upper arm bone. Most tears start in the supraspinatus tendon, on top of the shoulder, per AAOS OrthoInfo. An X-ray cannot show these tendons. An MRI can.
On MRI, a healthy tendon looks dark and smooth. A damaged tendon looks brighter, thicker, or thinner. A tear shows up as a gap, often filled with bright fluid. The scan also shows the bursa (a small fluid sac over the cuff), the biceps tendon, the joint cartilage, and the bones.
AAOS notes one thing MRI does that ultrasound does less well: it shows the quality of the cuff muscles. That helps your doctor judge whether a tear is new or has been there a long time.
If you have not had imaging yet and want to know which test usually comes first, see our guide to X-ray, ultrasound, or MRI for shoulder pain.
What is the difference between a partial and a full-thickness tear?
This is the single most important line in most reports. AAOS defines the two types this way:
- Partial-thickness tear: the tear goes only partway through the tendon. The tendon is still attached to bone but is thinner than normal.
- Full-thickness tear: the tear goes all the way through, so part of the tendon detaches from the bone. If only a small part detaches, it is a full-thickness incomplete tear. If the whole tendon detaches, it is a full-thickness complete tear, with essentially a hole in the tendon.
Partial tears are also described by which surface they start on. "Articular-sided" means the joint side. "Bursal-sided" means the side facing the bursa. "Intrasubstance" means inside the tendon. Surgeon Harvey Ellman published a 1990 grading system for partial tears based on location, depth, and area. He built it for findings seen during arthroscopy, and he noted that MRI has trouble defining partial tears precisely.
How accurate is MRI for a rotator cuff tear?
MRI is very good at finding full-thickness tears and less reliable for partial ones. The Cochrane review by Lenza and colleagues pooled 20 studies that checked imaging against surgery. Its main numbers:
- Full-thickness tears: MRI sensitivity 94% and specificity 93% (7 studies, 368 shoulders).
- MR arthrography for full-thickness tears: sensitivity 94% and specificity 92% (3 studies).
- Ultrasound for full-thickness tears: sensitivity 92% and specificity 93% (10 studies).
- Any tear, partial or full: MRI sensitivity 98% and specificity 79%.
The authors found no significant difference among the three tests for full-thickness tears. They warned that MRI and ultrasound may both miss many partial-thickness tears. Most included studies were small and had design flaws. When a partial tear or a labral problem needs a closer look, some doctors order an MRI arthrogram, which puts contrast inside the joint.
What do the common terms on a shoulder MRI report mean?
Radiology reports use shorthand. Here is a plain-language checklist of the terms you are most likely to see, in roughly the order they matter for treatment:
- Tendinosis or tendinopathy: the tendon is worn or degenerated but not torn. AAOS notes that many tears begin as fraying.
- Partial-thickness tear: look for the side (articular or bursal) and how much of the tendon thickness is involved.
- Full-thickness tear: look for which tendons are involved and the size, often given in millimeters or centimeters. AAOS lists a tear larger than 3 cm with good surrounding tissue as one sign surgery may be a good option.
- Retraction: how far the torn end has pulled back toward the shoulder blade. Daniel Patte's 1990 classification describes tears by their extent, their position in the sagittal and frontal planes, and the quality of the muscle.
- Atrophy: the muscle belly has shrunk.
- Fatty infiltration (Goutallier grade): fat has replaced some muscle. Goutallier's original 1994 study graded this in five stages on CT, not MRI. Radiologists later adapted it to MRI. A 1999 comparison found MRI grading was reproducible, but it matched CT only fairly to moderately. More fatty degeneration went with more atrophy.
- Bursitis or subacromial fluid: irritation or fluid in the bursa, often seen with cuff disease.
A 2025 radiology review stressed that atrophy, fatty infiltration, and where the retracted tendon end sits are key details for surgical planning. They are sometimes left out of reports. If yours does not mention them and surgery is on the table, ask. Our general guide to reading MRI results covers the rest of report structure.
Can an MRI show a tear that is not causing your pain?
Yes, and this happens often. In a 1995 study of 96 people with no shoulder pain, MRI showed a rotator cuff tear in 34% of them. Among those over 60, 54% had a tear: 28% full-thickness and 26% partial. Among people aged 19 to 39, only 4% had any tear.
The authors warned against deciding on surgery from MRI alone, without matching exam findings. AAOS also notes that most tears come from slow, age-related wear and are often fairly painless. So a tear on your report may be part of the story, not the whole cause. Your doctor will compare the images with your strength, range of motion, and symptoms.
What happens next after your MRI shows a tear?
The findings guide treatment, but they do not decide it alone. AAOS says your doctor will weigh your age, activity level, general health, and type of tear.
- Tendinosis or a partial tear: usually treated first without surgery. AAOS reports that nonsurgical care relieves pain and improves function in about 80 to 85% of patients. That care includes rest, activity changes, anti-inflammatory medicine, physical therapy, and sometimes a steroid injection.
- Full-thickness tear after a recent injury: AAOS lists an acute injury, a large tear with good tissue, major weakness, and symptoms lasting 6 to 12 months as signs surgery may be a good option.
- Retraction, atrophy, or fatty infiltration: these suggest an older tear. Goutallier found that infraspinatus fatty degeneration had a highly negative effect on supraspinatus repair results. He concluded it is better to repair wide tears before irreversible muscle damage.
Fatty infiltration also matters after surgery. A 2021 meta-analysis of 31 studies found retear rates of 15% to 21% at different follow-up times. Older age, larger tears, and fatty infiltration were among the factors linked to poorer healing. AAOS adds that tears can grow over time, so new weakness after a minor injury is worth reporting.
MRI is one of several imaging tests that change care when read alongside the exam. The same idea applies elsewhere: mammogram.md explains how breast imaging results lead to follow-up decisions.
The bottom line
A shoulder MRI tells you whether a rotator cuff tendon is worn, partly torn, or torn through, plus its size and retraction. It also shows whether the muscle has shrunk or filled with fat. MRI is highly accurate for full-thickness tears and weaker for partial ones. Tears are common in people without pain, so your doctor will match the images to your exam before choosing therapy or surgery.
Last updated: November 2026. This article is for informational purposes only and does not constitute medical advice. Talk with your doctor or orthopedic surgeon about what your shoulder MRI means for you.