PURPOSE OF THE STUDY: This study was undertaken to determine the reproducibility of measurements of fatty degeneration of the rotator cuff using computed tomography (CT). MATERIAL AND METHODS: Fifty-six patients who had undergone surgery for rotator cuff tear were included in this retrospective study. The extent of fatty infiltration was evaluated on CT scans with soft tissue windows in all 56 shoulders using a five-stage scoring system described by Goutallier. Five independent observers made the assessments. The same operation was repeated one month later to test intraobserver agreement. Four parameters were recorded: fatty infiltration of three muscles (supraspinatus, infraspinatus, subscapularis), and overall fatty infiltration grading. Interobserver variability was determined for each parameter using the intercorrelation coefficient (a test of reproducibility of quantitative measurements). RESULTS: The most reproducible measurement was the overall fatty infiltration grade. For this parameter, interobserver agreement was good with an intercorrelation coefficient of 0.75. The interval of confidence was +/- 0.5. Intraobserver agreement depended on the observer's level of experience. It was good for overall fatty infiltration grade assess by three senior observers (r=0.78) and moderate for two junior observers. CONCLUSION: The overall fatty infiltration grade is a reproducible parameter that should be used to evaluate the degree of fatty infiltration as the safety margin of this value (graded 0 to 4) is about 0.5. Fatty infiltration of a torn cuff would not be the only criterion to improve indications for treatment of rotator cuff tears.
PURPOSE OF THE STUDY: This study was undertaken to determine the reproducibility of measurements of fatty degeneration of the rotator cuff using computed tomography (CT). MATERIAL AND METHODS: Fifty-six patients who had undergone surgery for rotator cuff tear were included in this retrospective study. The extent of fatty infiltration was evaluated on CT scans with soft tissue windows in all 56 shoulders using a five-stage scoring system described by Goutallier. Five independent observers made the assessments. The same operation was repeated one month later to test intraobserver agreement. Four parameters were recorded: fatty infiltration of three muscles (supraspinatus, infraspinatus, subscapularis), and overall fatty infiltration grading. Interobserver variability was determined for each parameter using the intercorrelation coefficient (a test of reproducibility of quantitative measurements). RESULTS: The most reproducible measurement was the overall fatty infiltration grade. For this parameter, interobserver agreement was good with an intercorrelation coefficient of 0.75. The interval of confidence was +/- 0.5. Intraobserver agreement depended on the observer's level of experience. It was good for overall fatty infiltration grade assess by three senior observers (r=0.78) and moderate for two junior observers. CONCLUSION: The overall fatty infiltration grade is a reproducible parameter that should be used to evaluate the degree of fatty infiltration as the safety margin of this value (graded 0 to 4) is about 0.5. Fatty infiltration of a torn cuff would not be the only criterion to improve indications for treatment of rotator cuff tears.