Rotator Cuff Tear Size Classification Explained

Rotator cuff tears are classified by size using a system first described in the 1980s that groups them into four categories based on the longest dimension of the torn area: small (under 1 cm), medium (1 to 3 cm), large (3 to 5 cm), and massive (over 5 cm). This framework, known as the Cofield classification, remains the most widely referenced in clinical practice. But tear size alone captures only one dimension of the problem, and surgeons today rely on several overlapping classification systems to describe how deep a tear goes, how far the torn tendon has pulled back, how much muscle has degenerated, and which tendons are involved.

The Cofield Size Classification

In 1984, DeOrio and Cofield proposed measuring rotator cuff tears by their maximum diameter and sorting them into four groups. A small tear measures less than 1 cm, a medium tear spans 1 to 3 cm, a large tear falls between 3 and 5 cm, and a massive tear exceeds 5 cm in either the side-to-side or front-to-back dimension.¹ This system is simple, memorable, and has held up remarkably well for four decades, which is why it still appears in surgical notes and MRI reports.

That said, reducing a three-dimensional tear to a single measurement has obvious limits. A tear that is 4 cm wide but only partially detached from the bone is a very different clinical problem from a 4 cm tear with full tendon retraction and advanced muscle wasting. The Cofield classification tells you the footprint of the damage, not how the shoulder is actually functioning. Surgeons recognized this early on, and a handful of additional classification systems emerged to fill the gaps.

Partial-Thickness Tears and the Ellman Classification

Not all rotator cuff tears go all the way through the tendon. A partial-thickness tear involves only part of the tendon’s depth, and these tears have their own grading system, described by Ellman. It assigns a grade based on how deep the tear extends into the tendon and a letter based on where the tear sits:

Grade 1: less than 3 mm deep, or less than 25% of the tendon thickness
Grade 2: 3 to 6 mm deep, or 25% to 50% of the tendon thickness
Grade 3: more than 6 mm deep, or greater than 50% of the tendon thickness

The letter codes indicate whether the tear is on the joint-facing side of the tendon (articular, or “A”), the bursa-facing side (“B”), or buried within the tendon substance itself (intratendinous, or “C”).² Articular-side tears are the most common type. Bursal-side tears tend to be more painful and are more often seen in overhead athletes or workers. Intratendinous tears can be tricky to detect on standard imaging because the outer surfaces of the tendon look intact.

The practical significance of partial-thickness grading is straightforward. A Grade 1 tear is rarely a surgical problem and is typically managed with physical therapy. A Grade 3 tear involving more than half the tendon thickness starts to behave more like a full-thickness tear, and many surgeons will consider completing the tear and repairing it rather than leaving a dangerously thin tendon to finish tearing on its own. Grade 2 tears occupy the gray zone where clinical judgment, patient age, activity demands, and symptoms all factor in.

One wrinkle worth knowing: surgeons agree reasonably well on whether a partial tear is articular or bursal, but they agree poorly on grading the exact depth. A study examining interobserver reliability found that while surgeons distinguished full-thickness from partial-thickness tears with high consistency, their agreement on partial-tear depth was low.³ In practice, this means two surgeons looking at the same MRI might disagree on whether a partial tear is Grade 2 or Grade 3.

Tendon Retraction and the Patte Classification

When a tendon tears away from the bone, it often retracts toward the muscle belly like a rubber band snapping back. How far it pulls back matters enormously for whether a surgeon can reattach it. The Patte classification describes this retraction in stages. In the original system, Stage I means the tendon stump sits near its bony attachment, Stage II means it has retracted to the level of the humeral head, and Stage III means it has pulled all the way back to the level of the glenoid (the shoulder socket).

More recently, researchers have proposed modified versions of the Patte system that add finer gradations. One modification divides retraction into five stages by combining the original three-stage framework with quantitative cutoff points measured on MRI, achieving better accuracy at predicting whether a tear can actually be repaired and whether it will heal after surgery.⁴ Another version evaluates retraction on two different coronal MRI slices rather than just one, which improved the system’s ability to predict irreparability in large and massive tears.⁵

Retraction matters because a tendon that has been pulled far from its attachment for months or years often cannot be stretched back without excessive tension. Excessive tension at the repair site is one of the strongest predictors of the repair failing. So a massive tear with minimal retraction may actually be repairable, while a smaller tear with severe retraction might not be.

Fatty Degeneration and Muscle Atrophy

Once a rotator cuff tendon tears, the muscle it connects to starts to change. Without normal mechanical loading, the muscle fibers are gradually replaced by fat. This process, called fatty infiltration or fatty degeneration, is graded using the Goutallier classification, a five-stage system originally developed from CT scans and now applied to MRI. Stage 0 means no fat in the muscle, Stage 1 means some fatty streaks, Stage 2 is an equal mix of muscle and fat, Stage 3 is more fat than muscle, and Stage 4 is mostly fat with little remaining muscle.

The Goutallier system is one of the most consequential pieces of the classification puzzle because fatty degeneration is largely irreversible. Even a perfectly executed surgical repair cannot undo advanced fatty changes. Retear rates after surgery are significantly associated with both preoperative tear size and the degree of fatty degeneration.⁶ A tear with Goutallier Stage 3 or 4 changes has substantially worse healing potential, regardless of what the surgeon does with the tendon itself.

The catch is that the Goutallier classification is subjective. It relies on a radiologist or surgeon eyeballing the MRI and deciding how much fat is present. Studies have found it to have moderate reliability at best, with different observers frequently disagreeing, especially in the intermediate stages.⁷ Simplified versions and quantitative MRI techniques that calculate the fat fraction numerically have been proposed, and they do improve consistency.⁸ Muscle atrophy can also be measured using the Thomazeau occupation ratio, which compares how much of the supraspinatus fossa the muscle still fills on a cross-sectional MRI image.⁹

Which Tendons Are Involved

The rotator cuff is made up of four tendons, each connecting a different muscle to the top of the upper arm bone. When a tear gets large enough to involve more than one tendon, the specific combination matters. An alternative definition of “massive” comes from Gerber, who defined it as a tear involving at least two complete tendons, regardless of the measured diameter.¹

The Collin classification groups massive tears into five types based on the specific tendon combination:

Type A: supraspinatus and superior subscapularis
Type B: supraspinatus and entire subscapularis
Type C: supraspinatus, superior subscapularis, and infraspinatus
Type D: supraspinatus and infraspinatus
Type E: supraspinatus, infraspinatus, and teres minor

These distinctions affect what the surgeon can accomplish and how the shoulder will function after treatment.¹⁰ The reason tendon pattern matters so much relates to what is sometimes called the “force couple.” The front-to-back balance between the subscapularis (in front) and the infraspinatus and teres minor (in back) keeps the ball of the shoulder joint centered in its shallow socket. When that balance is disrupted by tears on both sides, the humeral head migrates upward and the shoulder can lose the ability to raise the arm, a condition called pseudoparalysis.¹¹ A Type D or E tear that knocks out the posterior cuff is a different functional problem than a Type A tear that involves the anterior cuff, even if both measure the same number of centimeters.

Why Tear Size Does Not Predict Pain

One of the most counterintuitive findings in shoulder research is that pain has essentially no relationship to how big a rotator cuff tear is. A cross-sectional study of 393 patients with symptomatic full-thickness tears found that no measure of tear severity correlated with pain.¹² Someone with a small tear can be in agony, while someone with a massive tear can report relatively little pain. This disconnect is one reason that imaging findings alone do not drive treatment decisions. A person with a massive tear on MRI who has good function and manageable discomfort may do perfectly well without surgery, while a smaller tear causing significant symptoms and functional loss might warrant repair.

What tear size does correlate with is loss of strength and range of motion, especially when the tear extends to involve more tendons or when fatty degeneration progresses.¹³ The practical message: if your doctor tells you the tear is large or massive, your shoulder may still feel acceptable. The concern with bigger tears is less about today’s pain and more about the risk of progressive muscle degeneration and functional decline over time.

How Tears Are Measured on Imaging

MRI is considered the gold standard for preoperative classification of rotator cuff tears. It provides detailed views of the tendon, the degree of retraction, fatty infiltration of the muscles, and the condition of the surrounding structures.¹⁴ Ultrasound can also detect tears and is cheaper, faster, and more accessible, but it has limitations when it comes to characterizing tear features beyond simple detection.

A meta-analysis comparing the two modalities found no significant difference between MRI and ultrasound for detecting full- or partial-thickness tears.¹⁵ However, when the task shifts from “Is there a tear?” to “How big is it, how far has it retracted, and how much muscle degeneration is there?”, MRI pulls ahead. Ultrasound measurements tend to underestimate both tear size and retraction compared to MRI, and the gap between the two grows as tears get larger. MRI also shows greater interobserver reliability, meaning different readers are more likely to agree on what they see.¹⁶ For surgical planning, particularly for large and massive tears where retraction, fatty infiltration, and tendon pattern all influence strategy, MRI is the preferred tool.

MR arthrography, which involves injecting contrast into the joint before scanning, performs even better than standard MRI for detecting partial-thickness tears, though it is more invasive and usually reserved for cases where the standard MRI is inconclusive.¹⁵

Where Tears Start and How They Grow

Degenerative rotator cuff tears do not appear randomly across the tendon. Research examining 360 shoulders found that tears most commonly begin in a specific zone roughly 13 to 17 mm behind the biceps tendon.¹⁷ This is a region of the supraspinatus tendon that experiences particularly high mechanical stress and has a relatively poor blood supply, making it vulnerable to wear-and-tear damage over time.

Once a tear begins, it alters the mechanical loading on the surrounding tendon. Biomechanical research has shown that as a tear enlarges, strain shifts away from the torn zone and concentrates on the adjacent intact tissue. In a progressive tear model, when the tear extended to involve both the superior portion and half of the middle portion of the tendon footprint, the remaining intact regions showed significantly elevated strain compared to the intact condition.¹⁸ In other words, a tear creates a vicious cycle: the remaining healthy tendon is overloaded, which puts it at risk of tearing too, which would overload the next zone of healthy tendon, and so on. This helps explain why some tears progress from small to massive over time, while also explaining why not all of them do. If the remaining tissue is robust enough to handle the extra load, the tear may stabilize.

The rate at which tears progress varies widely. Conservative treatment succeeds in anywhere from about a third to over 90% of patients, depending on the clinical picture, how long symptoms have been present, and the initial tear size.¹⁹ Full-thickness tears are more common with age: roughly a quarter of people in their sixties and more than half of those in their eighties have full-thickness tears, many of which cause no symptoms at all.²⁰

How Classification Guides Treatment

Classification is not an academic exercise. Each dimension of the tear description feeds into a treatment decision tree. The main factors influencing whether a repair will heal include tear size, fatty infiltration, patient age, and surgical technique.²¹ A small or medium tear with minimal retraction and healthy muscle is usually a straightforward arthroscopic repair. A large tear with moderate retraction may still be repairable but might benefit from patch augmentation, where a biological graft is laid over the repair to share the mechanical load. For large and retracted tears, acellular dermal matrix patches have the strongest evidence base and have been shown to reduce retear rates.²² That said, larger tears augmented with patches can still develop stiffness, so patient selection matters.²³

When a massive tear is deemed irreparable because of severe retraction, advanced fatty degeneration, or both, surgeons have several options. A systematic review comparing these approaches found that arthroscopic debridement (cleaning out the joint), reverse total shoulder arthroplasty, superior capsular reconstruction, and tendon transfers all had similar failure rates, while subacromial balloon spacers and partial repairs fared worse.²⁴ Superior capsular reconstruction, which uses a graft to replace the missing superior cuff and prevent the humeral head from riding upward, has shown promise even in reversing pseudoparalysis in selected patients.²⁵

The ISAKOS Classification and the Push for Standardization

With multiple overlapping systems covering different aspects of a tear, there has been a push to create a single comprehensive classification. The International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS) developed a system that captures location, extension (which tendons are involved), tear pattern, and retraction in one framework. Reliability testing showed nearly perfect agreement among surgeons on tear location, substantial agreement on extension, and fair agreement on pattern.²⁶ The overall reliability is encouraging, though the weaker agreement on tear pattern mirrors a broader challenge in orthopedics: the more granular the classification, the harder it is for different observers to apply it consistently.

More recent proposals have tried to incorporate three-dimensional measurements from MRI to better capture tear geometry, recognizing that a tear’s shape and direction matter as much as its raw diameter.¹ These geometric classifications can alert a surgeon before the operation about the likely repair strategy needed, moving beyond the one-number approach of the original Cofield system. The field is clearly heading toward multi-dimensional descriptions rather than single labels, though simpler systems will likely persist in everyday clinical shorthand for a long time.

Cuff Tear Arthritis

A massive tear left untreated long enough can lead to a specific pattern of joint degeneration known as cuff tear arthropathy. When the rotator cuff no longer keeps the humeral head centered, the ball rides upward and wears against the underside of the acromion. Over time, this produces cartilage loss, bone changes, and progressive arthritis. Hamada and colleagues developed a radiographic classification for this condition, dividing it into five grades based on how much the joint has deteriorated.²⁷ A later modification subdivided Grade 4 to distinguish between cases where the arthritis mainly involves the subacromial space versus those where the glenohumeral joint itself has broken down.

Cuff tear arthropathy represents the end stage of untreated massive rotator cuff disease, and its presence generally rules out a simple rotator cuff repair. By the time the joint itself has developed arthritis, the treatment shifts toward arthroplasty, most commonly reverse total shoulder replacement, which redesigns the joint mechanics so that the deltoid muscle can do more of the work that the cuff no longer can. This is the point at which tear classification gives way to arthritis classification, and the conversation moves from tendon repair to joint replacement.