How accurate are shoulder MRIs?
Ross A. Hauser, MD; Danielle R. Steilen-Matias, MMS, PA-C
Among the many controversies in medicine is the controversy surrounding the use of MRIs and their accurate portrayal of what is causing shoulder discomfort. Questioning the accuracy of shoulder MRIs is not a new challenge patients face. Research has been going on for many years talking about how successful or not successful diagnosis are and even if there was a correct diagnosis, was the diagnosed problem causing the patient pain in the first place? Now there are some who say that in a few years, artificial intelligence will be able to clear up images and make better predictions about treatments and outcomes. Well maybe a tear will be more readily recognized and a surgery may be more precise, but the question will remain. Is it the tear that is actually causing the pain, is the surgery correcting the pain problem? In many cases, the answer will still be no. The biggest problm may still remain that the doctor and patient think an MRI must be taken to “solve” the patient’s shoulder problem.
Important in this 2023 study was the patient’s overall desire to have an examination, an MRI, a diagnosis, and a treatment.
Having a diagnosis and a treatment is the goal of all patients who are having pain challenges. The problem has always been getting the right diagnosis, getting the right treatment.
A September 2023 paper from medical universities in Canada, published in the BioMed Central musculoskeletal disorders (1) assessed the current treatment recommendations for patients visiting their primary care physician for shoulder pain. The goals of this research was to understand and help explain the challenging management of shoulder pain considering that 40% of people with shoulder pain remain symptomatic one year after initial consultation, and, to explore patients’ expectations and experiences of their primary care consultation for shoulder pain.
In this small sample study, 13 people, 8 women, 5 men; in the age 50 average range were followed. Eleven of the patients initially consulted a family physician or an emergency physician, and two participants initially consulted a physiotherapist. The reasons they sought medical care and what they expected from their clinicians were:
- 1) I can’t sleep because of my shoulder;
- 2) I need to know what is happening with my shoulder;
- 3) We need to really see what is going on to help me!; and
- 4) Please take some time with me so I can understand what to do!.
Important in this study was the patient’s overall desire to have an examination, an MRI, a diagnosis, and a treatment. For these patients there is an expectation that a shoulder MRI must be done and most clinicians may not order one because the shoulder MRI may show too much. What does that mean?
Were the 13 patients helped? This was the concluding findings: “Several participants waited until they experienced a high level of shoulder pain before making an appointment since they were not confident about what their family physician could do to manage their condition. Although some participants felt that their physician took the time to listen to their concerns, many were dissatisfied with the limited assessment and education provided by the clinician. . . “Several participants reported that their expectations were not met, especially when it came to the explanations provided.”
One “unexpected finding” that emerged from this study according to the researchers, was the longer than anticipated waiting period between the onset of shoulder pain and when patients decided to consult their primary care clinician.
“Several participants reported that their expectations were not met, especially when it came to the explanations provided.” My two MRIs were not in agreement with what was happening in my shoulder.
In the above study we see that people got a diagnosis and people got a treatment. Did that solve their issues? Here is one sample story that may explain why a patient’s expectations were not met by getting an MRI and a treatment.
My doctors were not in agreement with what was happening in my shoulder. My two MRIs were not in agreement with what was happening in my shoulder. One MRI says a small labral tear. The other says, no it is a problem of the acromioclavicular joint. With the doctor’s opinions at odds and the MRIs at odds I cannot get a diagnosis and no one is dealing with my shoulder pain. At present I get physical therapy which is not helpful.
Inappropriate MRI ordered because the machine is there.
Notwithstanding, there is pressure on doctors, exhibited on them by patients wanting an MRI. Sometimes because the MRI is there, an MRI can easily be offered and given. Especially in a hospital. In a May 2021 paper, researchers writing in the journal Health affairs (2) cited that a 20% rise in inappropriate MRIs could be traced to doctors working in hospitals. The researchers write: “Study findings indicate that the odds of a patient receiving an inappropriate MRI referral ( lower back pain, knee pain, and shoulder pain) increased by more than 20 percent after a physician (left private practice and went) to hospital employment. Most patients who received an MRI referral by a (hospital) employed physician obtained the procedure at the hospital where the referring physician was employed. These results point to hospital-physician integration as a potential driver of low-value care.”
It is called “low-value care because the MRI in many cases, was not needed to be helpful to the patient’s condition and as we will see, will confuse the issue of treatment.
Inappropriate MRI because of doctor experience
Previously the same lead researchers wrote in the Journal of general internal medicine (3) that many inappropriate MRIs (23% of claims were classified as inappropriate) could be traced in some instances to “physicians with 10 or less years of experience had significantly higher odds of ordering inappropriate MRIs. Primary care physicians were almost twice as likely to order an inappropriate MRI as orthopedists. ”
Elite rock climbers have a lot of damage seen on MRI, yet have no shoulder pain.
The demonstration that an MRI can show too much damage and my lead to inappropriate surgery can be seen in the shoulders of elite rock climbers. A February 2022 paper in the Orthopaedic journal of sports medicine (4) looked for how prevalent significant degenerative findings were on shoulder MRI in asymptomatic athletes of overhead sports. The researchers “hypothesized that glenoid labrum, long head of the biceps tendon, and articular cartilage pathology would be present in more than 50% of asymptomatic athletes.” To test their theory the researchers recruited 50 elite level rock climbers (age range, 20-60 years), and performed physical examinations and MRIs.
- MRI evidence of:
- tendinosis of the rotator cuff was diagnosed in 80% of patients,
- subacromial bursitis was diagnosed in 79% of patients,
- and long head of the biceps tendonitis was diagnosed in 73% of patients.
- Labral tears were present in 69% of shoulders, with discrete labral tears identified in 56%.
- Articular cartilage changes were also common, with humeral pathology present in 57% of shoulders and glenoid pathology in 19% of shoulders.
MRI of the shoulder can be misleading.
As we have seen in the studies above, MRI of the shoulder can be misleading. While an MRI can be a helpful too, we do not order them frequently for patients. As documented, numerous studies done on asymptomatic patients regularly show abnormalities on MRI, including glenoid labrum abnormalities, supraspinatus tendinopathy, and rotator cuff tears. Also documented are the thousands of surgeries based primarily on MRI readings without a correlating physical exam. But it happens all the time that patients are signed up for surgery and say the doctor barely touched them.
Even in cases where the MRI reading notes a “complete tear” of the rotator cuff, a non-surgical consultation maybe helpful. In our clinic we have seen, following an examination, that the tear is not a “complete” tear.
1 Lowry V, Desmeules F, Zidarov D, Lavigne P, Roy JS, Cormier AA, Tousignant-Laflamme Y, Perreault K, Lefèbvre MC, Décary S, Hudon A. “I wanted to know what was hurting so much”: a qualitative study exploring patients’ expectations and experiences with primary care management. BMC Musculoskeletal Disorders. 2023 Sep 26;24(1):755. [Google Scholar]
2 Young GJ, Zepeda ED, Flaherty S, Thai N. Hospital Employment Of Physicians In Massachusetts Is Associated With Inappropriate Diagnostic Imaging: Study examines association between hospital employment of physicians and diagnostic imaging. Health Affairs. 2021 May 1;40(5):710-8. [Google Scholar]
3 Young GJ, Flaherty S, Zepeda ED, Mortele KJ, Griffith JL. Effects of physician experience, specialty training, and self-referral on inappropriate diagnostic imaging. Journal of general internal medicine. 2020 Jun;35:1661-7. [Google Scholar]
4 Cooper JD, Seiter MN, Ruzbarsky JJ, Poulton R, Dornan GJ, Fitzcharles EK, Ho CP, Hackett TR. Shoulder Pathology on Magnetic Resonance Imaging in Asymptomatic Elite-Level Rock Climbers. Orthopaedic Journal of Sports Medicine. 2022 Feb 11;10(2):23259671211073137. [Google Scholar]