Magnetic resonance appearance of posterosuperior labral peel back during humeral abduction and external rotation.

OBJECTIVE: To describe the magnetic resonance appearance of posterosuperior labral peel back and determine the reliability of MR in the abducted and externally rotated (ABER) position for the prospective diagnosis of arthroscopically proven cases of posterosuperior labral peel back.
METHODS: After approval by the institutional review board (IRB) of the University of Pittsburgh Medical Center, USA, databases of patients who underwent arthroscopy over a 2-year period for one of three clinical diagnoses [suspected type 2 superior labrum anterior to posterior (SLAP) tears, posterior instability, or multidirectional instability] were reviewed after anonymization by an honest broker. Sixty-three cases were selected by the following inclusion criteria: operative report documenting labral peel back in the ABER position, age <40 years, and preceding MR arthrogram evaluations with images in the ABER position (n=34). Inclusion criteria for the control group differed from those for the case group insofar as the operative note documented the absence of posterosuperior labral peel back (n=29). Cases and controls were randomized in one list and evaluated independently by two fellowship-trained musculoskeletal radiologists unaware of the surgical results and using a three-point grading system (0 = posterosuperior labrum normally positioned lateral/craniad to glenoid articular plane in ABER; 1 = posterosuperior labral tissue flush with the glenoid articular plane in ABER; 2 = posterosuperior labral tissue identified medial/caudal to glenoid articular plane in ABER). Only one image in ABER showing abnormal posterosuperior labral position was required for a grade of 1 or 2 to be assigned. Sensitivity, specificity, and positive and negative predictive value were calculated as well as the level of agreement between readers (kappa).
RESULTS: Both readers assigned a grade of 2 to 25 of 34 patients with surgically proven labral peel back. Of the patients with surgically proven SLAP tears with peel back in ABER, reader A assigned a grade of 1 to seven patients and a grade of 0 to two patients, while reader B assigned a grade of 1 to eight patients and a grade of 0 to one patient. In the control group of 29 patients, reader A assigned 28 patients a grade of 0, one patient a grade 1, and no patients a grade 2. Reader B assigned 27 patients a grade of 0, two a grade 1, and no patients a grade 2. After the data had been dichotomized, with grade 1 and 0 cases both being regarded as negative, the MR criteria showed a sensitivity of 73%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 78%. The kappa coefficient of inter-rater agreement was excellent at 0.9, with disagreement in only four of 63 cases. In five of the 34 cases with peel back, a labral tear, defined by imbibition of contrast agent within a gap between labrum and underlying glenoid bone, could not be identified in standard planes in the neutral position.
CONCLUSIONS: The use of the glenoid articular plane as a reference line to evaluate labral peel back in the abducted and externally rotated position is a fairly accurate and highly precise method for detection of posterosuperior labral peel back. Raising the possibility of labral peel back may help alert the arthroscopist to the presence of superior labral instability while the arm is abducted and externally rotated.