MRI findings of traumatic and degenerative rotator cuff tears and introduction of the "cobra sign".

Background: A rotator cuff tear (RCT) is a common shoulder diagnosis and its etiology may be acute, traumatic, or chronic degenerative. Differentiation between the 2 etiologies may be important for multiple reasons, but remains difficult based on imaging. Further knowledge about radiographic and magnetic resonance findings to distinguish traumatic from degenerative RCT is needed.

Methods: We analyzed magnetic resonance arthrograms (MRAs) of 96 patients with traumatic or degenerative superior RCT, which were matched according their age and the affected rotator cuff muscle into the 2 groups. Patients older than 66 years of age were excluded from the study to avoid including cases with pre-existing degeneration. In the case of traumatic RCT, the time between the trauma and MRA had to be less than 3 months. Various parameters of the supraspinatus (SSP) muscle-tendon unit were assessed (tendon thickness, presence of a remaining tendon stump at the greater tubercle, magnitude of retraction, layer appearance). The retraction of the 2 SSP layers were individually measured to determine the difference of retraction. Additionally, edema of the tendon and muscle, the tangent- and kinking-sign as well as the newly introduced Cobra-sign (bulging of the distal part of the ruptured tendon with slim configuration of the medial part of the tendon) were analyzed.

Results: Edema within the SSP muscle (sensitivity 13%, specificity 100%,  = .011) or the tendon (sensitivity 86%, specificity 36%,  = .014) are more frequent in traumatic RCT. The same association was found for the kinking-sign (sensitivity 53%, specificity 71%,  = .018) and the Cobra sign (sensitivity 47%, specificity 84%,  = .001). Even though not statistically significant, tendencies were observed toward thicker tendon stumps in traumatic RCT, and greater difference in retraction between the 2 SSP layers in the degenerative group. The cohorts had no difference in the presence of a tendon stump at the greater tuberosity.

Conclusion: Muscle and tendon edema, as well as tendon kinking appearance and the newly introduced cobra-sign are suitable MRA parameters to distinguish between traumatic and degenerative etiology of a superior RTC.