Are radiographic trochanteric surface irregularities associated with abductor tendon abnormalities?

Abstract

To assess the association between trochanteric surface irregularities seen on conventional radiographs and magnetic resonance (MR) evidence of abductor tendon abnormalities. A total of 150 consecutive patients were evaluated in this retrospective study (age range, 21-88 years; mean age, 58.7 years ± 16.1 [standard deviation]; 57 men, 93 women). Because patients' rights are protected by a procedure in which they are asked to provide general approval for their records and images to be reviewed for scientific purposes, specific approval by the institutional review board was not required. Two readers independently analyzed conventional radiographs and MR images of the hip. Trochanteric surface was graded on conventional radiographs as normal, osseous irregularities extending 1-2 mm, or osseous irregularities extending more than 2 mm. On MR images, the gluteus minimus and gluteus medius tendons were classified as normal or as having tendinopathy or a partial- or full-thickness tear. Logistic regression analysis and the Fisher exact test were used for statistical analysis. Sensitivity, specificity, accuracy, positive and negative predictive values, and positive likelihood ratio were calculated. To assess interobserver agreement, a κ statistic was used. The positive predictive value of surface irregularities larger than 2 mm for MR tendinopathy or a partial- or full-thickness tear was 90% (37 of 41 patients). The sensitivity of radiographic changes was 40%; the specificity, 94%; the accuracy, 61%; the negative predictive value, 49%; and the positive likelihood ratio, 5.8. Interobserver agreement for detection of trochanteric surface irregularities on conventional radiographs ranged from 0.28 to 0.76. Pronounced (>2 mm) surface irregularities of the greater trochanter on conventional radiographs were associated with abductor tendon MR abnormalities.