Ross Hauser, MD, Danielle Matias, PA-C, Demi Foster, APRN

Many times a patient will tell us that they have surgery planned. But they do not want one. If you are in a similar situation, you like them, are looking at recovery times that a partial rotator cuff tear surgery may require as part or maybe your whole reason to avoid the surgery. But something does need to be done to prevent the partial from becoming a full tear and from you suffering from a worsening condition. But is it surgery? Next question, do you even have a partial rotator cuff tear? One thing for sure is you do have shoulder pain so something is wrong.

Thinking that an MRI that has a good probability of actually showing a partial rotator cuff tear surgery is flawed medical thinking

The concern about surgery for partial rotator cuff tears is that you may be operating on something that is not even there. This is not solely our opinion, many studies are confirming that an MRI may send you to unnecessary shoulder surgery.

Like the stories we hear from patients, one of which we relayed above, you may get an MRI that no one can make heads or tails out of and when that happens you will be recommended to a surgery “to see what is really going on.”

Many patients come into our office with an MRI of their shoulder. These patients have bounced from doctor to doctor, from specialist to specialist looking for a treatment option. The one thing that has remained constant with many of them is that the picture of their MRI is an accurate reflection of what is going on on their shoulder. This is not always the case. The MRI may be leading this patient down a path of treatment that is not accurate to their situation.

This is probably why you are still searching for treatment, you are chasing a partial rotator cuff tear as the sole problem. We are going to present evidence that you may be chasing the wrong target. What you should be chasing is shoulder instability treatment.

The reliable diagnosis of partial-thickness tears of the rotator cuff is still elusive in clinical practice.

We are going to help take you on a journey of MRI understanding that will reinforce our suggestions that the MRI may be leading you astray. We are going to use the help of some recent research and bring that research up to 2026.

This study comes from the Department of Orthopaedics and Orthopaedic Surgery at Saarland University Medical Center in Germany. It was published in the Archives of Orthopaedic and Trauma Surgery. (1) Here are the learning points:

• The reliable diagnosis of partial-thickness tears of the rotator cuff is still elusive in clinical practice.
• The purpose of the study was to determine the diagnostic accuracy of MRI and clinical tests for detecting partial-thickness tears of the rotator cuff as well as the combination of these tests.

There is a problem, partial rotator cuff tears are difficult to diagnose. Your MRI may be wrong. Your diagnosis may be in question.

• Here is a statement from the research and then we will explain what it means regarding how accurate an MRI is for a partial rotator cuff tear.
  • Sensitivity for MR imaging to identify partial-thickness tears was 51.6%, specificity 77.2%, positive predictive value 41.3%, and negative predictive value 83.7%.

What you may notice in those numbers immediately is that there is no “100%.” 

• So first thing, the MRI reading of an image is not accurate every single time.
  • The sensitivity for MR imaging to identify something that MAY be a partial-thickness tear was 51.6%.
    • What does this mean? This means that the MRI chances of accurately detecting the possibility of a  partial rotator cuff tear are there is about 52%. Basically the same accuracy odds as a coin flip.
  • Once this 52% is separated into probable and non-probable, then the probable partial rotator cuff tears are confirmed as partial tears with a specificity of 77.2%. What this means is that there is a 78% chance you do have a rotator cuff tear. You may then have been told the MRI shows a partial rotator cuff tear. Yet, there is, in reality, a 1 in 5 chance that is actually no tear. Are you being sent to surgery on this recommendation of a 78% probability? This answer is probably yes because the surgery, it will be suggested to you, will go in there and see if you really do have a tear and if you do, clean up your shoulder. See below for the research questioning this type of thinking.

A study led by the  University of Nebraska Medical Center (2) reported on problems of treatment: Some learning points are:

• “Partial-thickness rotator cuff tears remain difficult to diagnose with single imaging (An MRI or CT) modality. . . “
• “In the primary care office, the initial evaluation of shoulder pain should include a thorough musculoskeletal evaluation in order to identify the source of the pain (e.g., shoulder, cervical spine, chest wall), as well as the development of an initial treatment plan.” (Please see our article Is neck instability causing your shoulder pain?)
• “Information regarding the risks of oral medications for the management of (partial rotator cuff tears) and the associated pain keeps mounting.”

Does this research’s learning points sound familiar to you?

• You had an imaging study, MRI, or CT scan
• You may have or have not had a physical examination. Some patients report that their examination is a simple reading back of the MRI evaluation report.
• You were given stronger prescriptions for painkillers until your doctors could figure out what to do.

An email we receive sounds something like this:

“I felt something tear in my shoulder. This is not the first time this has happened, I have had tears in both shoulders before. I am scheduled for an MRI but I can tell that I will be on conservative care because the injury will not be considered severe enough for surgery. So I will be in a limbo period either waiting for my shoulder to heal or get worse enough to need the surgery.  I had partial rotator cuff repair surgery before. It was very successful but the rehab was almost a year long and very demanding. By the time I wait for my shoulder to get worse and get the surgery, I could be in this for a year – to 18 months. I need to work.”

The findings of MRI have calculated odds, you could have a partial rotator tear and be told you don’t, you may not have a partial tear and be told you do.

So, if you had 100 patients who were sent to an MRI with suspected partial rotator cuff tears. If the negative predictive value was 83.7% it means that the MRI’s ability to successfully predict that a partial rotator cuff tear is NOT THERE is almost 84% of the time. So there is a 16% chance that you do have a rotator cuff there and you will be told you do not. Equally, there is a 78% chance that you do have a tear, be told you have a tear, and there is a 1 in 5 chance that you do not. This is why there is no 100% in MRIs. This is why you may be chasing the wrong problem.

The researchers of this study took this one step further. They combined the MRI findings with a physical examination to see if the physical examination correlated with the MRI findings and if it helped to accurately diagnose the partial tear. The physical examinations were standard Jobe-Test and Sensitivity for the Impingement-sign which try to predict the possibility of a muscle tear in the rotator cuff.

For the combination of MR imaging, Jobe-test, and Impingement-sign, the results were as follows:

• Sensitivity was 46.9%, which means that an MRI and physical examination’s combined chances of accurately detecting the possibility of a partial rotator cuff tear as being actually there are about 47%.
• Once this 47% is separated into probable and non-probable, then the probable partial rotator cuff tears are confirmed as partial tears with a specificity of 85.4%. So following the physical examination and an MRI in the best, most absolute scenario, you can still only be told it is only likely you have a partial rotator cuff tear.  There is no 100% in MRIs.

The comparative assessment of one hundred patients who underwent ultrasound and MRI for suspected supraspinatus tendinopathy.

In our office, we use ultrasound as a primary means of assessing possible rotator cuff tears.

A September 2020 paper in The Ultrasound Journal (3) reported on the comparative assessment of one hundred patients who underwent ultrasound and MRI for suspected supraspinatus tendinopathy. The supraspinatus tendon was observed in these patients as being intact, suffering from a partial tear, or full-thickness tear. The researchers then looked for agreement between the ultrasound image and the MRI image in the patient’s diagnosis.

• There was an agreement between the ultrasound image and the MRI image in 48/100 patients
• The consistency of agreement varied with the type of pathology:
  • intact tendons by ultrasound had 55.8% agreement with MRI,
  • partial sided bursal tears 50%,
  • partial sided articular tears 25%,
  • and full-thickness focal tears 33.3%.
• Full-thickness complete-width tears had a much better agreement with MRI at 90.9%.
• Age was also significant, with increased disagreement between ultrasound and MRI in patients over 50 years old.

The outcome of this paper was a suggestion to primary care physicians that they use ultrasound as the initial test in younger patients and patients with suspected full supraspinatus tears, based on clinical exam, with MRI as an option for further evaluation to quantify supraspinatus muscle atrophy.

A February 2022 paper (4) from Brazilian researchers found that “the measurement or classification of the (rotator cuff) tear size can be accurately performed. Where there is disagreement (between radiologist and surgeon), it is unclear whether the tear size is either underestimated on the (MRI) scan or overestimated during (arthroscopic) surgery.” The researchers noted that a supraspinatus tear is usually detected by scan, infraspinatus and subscapularis tears correct diagnosis in MRI is less accurate.

Ultrasound outperformed MRI in diagnosing partial-thickness subscapularis tear

A February 2022 study in the journal Arthroscopy (5) found ultrasound outperformed MRI in diagnosing partial-thickness subscapularis tear. In this study, a comparison between ultrasound diagnosis and MRI diagnosis was made by confirming these imaging readings against what doctors found during an arthroscopic procedure. The significance of this study is that of 437 patients who were scheduled to undergo arthroscopic rotator cuff repair, preoperative ultrasound and MRI of the shoulder was performed, and ultrasound correctly diagnosed 122 of 157 patients with subscapularis (SSC) tears, with an overall sensitivity of 77.7% which was significantly greater than that of MRI.

• Conclusions: Ultrasound is a reliable and accurate diagnostic method for subscapularis (SSC) tears, especially in easily missed partial-thickness tears.

A word about artificial intelligence and shoulder MRIs

Artificial intelligence is a computer’s ability to present accurate prediction models that can help in the diagnosis of a patient. We present this small section, which will get larger as the years go by, to help people who may have heard the term used in their radiological assessments and may hear the term more often as they continue their medical journeys toward shoulder pain treatment.  We also present this section because occasionally we get asked about “AI.”

A June 2023 paper in the journal Arthroscopy (6) looked to see if artificial intelligence could make MRI and ultrasound reading “more accurate.” What they found was that “AI models based on ultrasound images demonstrated a diagnostic performance similar to that of radiologists. Another comparative study demonstrated that AI models using MRI images exhibited greater accuracy and specificity compared to orthopedic surgeons in the diagnosis of rotator cuff pathology, albeit not in sensitivity (the correct assessment that a tear is present.)

A September 2025 study in the journal Academic radiology (7) looked to suggests ways to develop and validate a deep learning system (a more human decision making process in AI) for grading partial-thickness supraspinatus tendon (SST) tears and to compare its performance with experienced radiologists. In this study of 1150 patients the proposed deep learning system provided a reliable and efficient method for grading partial-thickness SST tears, achieving radiologist-level accuracy with greater consistency and faster evaluation speed. In other words ,the AI was just as good as the radiologists. 

Related articles

Prolotherapy Research Shoulder Osteoarthritis, Tears and Chronic Pain Disorders

Prolotherapy injections for rotator cuff tears and rotator cuff tendinopathy

References

1 Brockmeyer M, Schmitt C, Haupert A, Kohn D, Lorbach O. Limited diagnostic accuracy of magnetic resonance imaging and clinical tests for detecting partial-thickness tears of the rotator cuff. Archives of orthopaedic and trauma surgery. 2017 Dec 1;137(12):1719-24. [Google Scholar]
2 Mathiasen R, Hogrefe C. Evaluation and management of rotator cuff tears: a primary care perspective. Current reviews in musculoskeletal medicine. 2018 Mar 1;11(1):72-6. [Google Scholar]
3 Dalai A, Langford L, Beavis C, Obaid H. Development of supraspinatus imaging guidance for primary care physicians with a focus on patient selection. The Ultrasound Journal. 2020 Sep 3;12(1):40. [Google Scholar]
4 Gill KP, Bateman M, Mazuquin B, Littlewood C. A comparison between preoperative and intraoperative measurement and classification of the size of rotator cuff tears. Revista Brasileira de Ortopedia. 2023 Aug 11;58:356-60. [Google Scholar]
5 Zhu S, Pu D, Li J, Wu D, Huang W, Hu N, Chen H. Ultrasonography outperforms magnetic resonance imaging in diagnosing partial-thickness subscapularis tear. Arthroscopy: The Journal of Arthroscopic & Related Surgery. 2022 Feb 1;38(2):278-84. [Google Scholar]
6 Zhan H, Teng F, Liu Z, Yi Z, He J, Chen Y, Geng B, Xia Y, Wu M, Jiang J. Artificial Intelligence Aids Detection of Rotator Cuff Pathology: A Systematic Review. Arthroscopy: The Journal of Arthroscopic & Related Surgery. 2023 Jun 23. [Google Scholar]
7 Ni M, Jiesisibieke D, Zhao Y, Wang Q, Gao L, Tian C, Yuan H. Improving Deep Learning-Based Grading of Partial-thickness Supraspinatus Tendon Tears with Guided Diffusion Augmentation. Academic Radiology. 2025 May 20. [Google Scholar]