Intra- and interobserver variability of magnetic resonance imaging for quantitative assessment of abductor and external rotator muscle changes after total hip arthroplasty

Purpose: Magnetic resonance imaging (MRI) is one of the most widely used noninvasive diagnostic modalities for musculoskeletal evaluation. We conducted a retrospective study to determine whether MRI of the hip joint abductor and external rotator muscles yields reproducible findings in patients after total hip arthroplasty (THA). Materials and methods: MR images were obtained 12 months after THA in 10 patients and were analyzed by two independent, blinded observers. The images were analyzed on workstations with three-dimensional analysis capabilities. The readers evaluated the gluteus medius, gluteus minimus, and obturator externus muscles on the nonoperated side (NOS) and the THA side. For each of the three muscles, the readers analyzed the cross-sectional area, cross-sectional signal intensity, intensity in a region of interest (ROI), and volume on both sides. Results: Unlike intraobserver variability, interobserver variability showed significant differences between the measurements for some abductor and external rotator muscles, especially for volume (p < 0.001–0.02) and cross-sectional area (p < 0.001–0.05) analysis of the gluteus medius and minimus muscles on the THA side but also on the NOS. There were no differences in signal intensity in the ROI (p = 0.29–0.83). The coefficients of variation (CoV) of all 4 parameters quantified were lower on the NOS (average, 5.9%) than on the THA side (average, 7.3%). Analysis of the gluteus minimus muscle volume showed both higher intraobserver (18.8%) and interobserver (13.9%) variability on the THA side. Conclusion: MRI is an accurate and reliable test for evaluation of soft tissue structures in patients who have undergone THA. MRI evaluation of abductor muscles yields reproducible results on the THA side and on the NOS. Prosthesis artifacts limit evaluation of the gluteus minimus muscle on the THA side and cannot be recommended. Since interobserver variability was always higher than intraobserver variability, follow-up examinations should ideally be performed by the same reader in order to obtain clinically meaningful results.