To determine the effectiveness of magnetic resonance imaging in identifying shoulder pathology in patients with primary traumatic dislocation of the shoulder and to compare these findings with findings at the time of arthroscopic surgery.
Correlation between arthroscopy and magnetic resonance imaging.
Sixteen patients, aged 18 to 30 years, who were randomized to the surgical arm of a study comparing the effectiveness of immediate arthroscopic surgery with immobilization and rehabilitation for primary traumatic anterior dislocation of the shoulder, were included in this study.
Each patient underwent magnetic resonance imaging and a videotaped “tour” of the shoulder prior to any surgical intervention.
Magnetic resonance scans and videotapes were reviewed for the presence or absence of abnormalities in 8 features of the shoulder, and concordant and discordant findings were evaluated.
There was moderate correlation for superior labral lesions (κ = 0.60) and fair agreement for rotator cuff tear (κ = 0.355). When the joint capsule was assessed, there was only fair agreement for both the presence of an abnormality (κ = 0.310) and redundancy and tear (κ = 0.394). Both methods were sensitive for the detection of Hill-Sachs lesions (κ = 1.0), although there was only moderate agreement (κ = 0.44) on estimation of size. There was perfect agreement for the detection of Bankart lesions or equivalent capsulolabral disruption (κ = 1.0).
Magnetic resonance imaging can be considered a valuable tool for the detection of Hill-Sachs and Bankart lesions associated with primary traumatic anterior dislocations of the shoulder. Its ability to detect other pathologic lesions, however, is limited.
*Fowler Kennedy Sport Medicine Clinic and †Department of Radiology and Diagnostic Imaging, London Health Sciences Center, University Campus, London, Ontario, Canada
Received for publication April 2000; accepted February 2003.
Reprints: Robert Litchfield, MD, FRCSC, Fowler Kennedy Sport Medicine Clinic, 3M Center, UWO, London, Ontario, N6A 3K7, Canada. E-mail: firstname.lastname@example.org