MRI Does Not Improve Inter- or Intrarater Reliability for Hip Arthritis Grading Systems

Abstract

Background: Magnetic resonance imaging (MRI) scans and radiographs are often utilized in assessing for preoperative osteoarthritis in patients undergoing hip preservation surgery. Purpose: To determine if MRI scans improve inter- or intrarater reliabilities over radiographs for findings of hip arthritis. Study Design: Cohort study (Diagnosis); Level of evidence, 3. Methods: Anteroposterior and cross-table lateral radiographs as well as a representative coronal and sagittal T2-weighted MRI scan were reviewed for 50 patients by 7 experienced subspecialty hip preservation surgeons, with a minimum experience of 10 years. Radiographs and MRI scans were assessed for joint space narrowing, subchondral cysts, osteophytes, subchondral sclerosis, Likert osteoarthritis grade (none, mild, moderate, or severe), and Tönnis grade. MRI scans were also evaluated for bony edema, heterogeneous articular cartilage, and chondral defects. Inter- and intrarater reliabilities were calculated utilizing the Fleiss method with a 95% CI. Results: The scans of 50 patients (28 female and 22 male) with a mean age of 42.8 years (SD, 14.2 years; range, 19-70 years) were reviewed. Radiographs revealed fair agreement for joint space narrowing (κ = 0.25 [95% CI, 0.21-0.30]), osteophytes (κ = 0.26 [95% CI, 0.14-0.40]), Likert osteoarthritis grading (κ = 0.33 [95% CI, 0.28-0.37]) and Tönnis grade (κ = 0.30 [95% CI, 0.26-0.34). Radiographs revealed moderate agreement for subchondral cysts (κ = 0.53 [95% CI, 0.35-0.69]). MRI scans demonstrated poor to fair agreement for joint space narrowing (κ = 0.15 [95% CI, 0.09-0.21]), subchondral sclerosis (κ = 0.27 [0.19-0.34]), heterogeneous articular cartilage (κ = 0.07 [95% CI, 0.00-0.14]), Likert osteoarthritis grade (κ = 0.19 [95% CI, 0.15-0.24]), and Tönnis grade (κ = 0.20 [95% CI, 0.15-0.24]). MRI scans demonstrated substantial agreement for subchondral cysts (κ = 0.73 [95% CI, 0.63-0.83]). Intrarater reliabilities were statistically improved compared with interrater reliabilities, but no differences were found between radiographs and MRI scans for joint space narrowing, subchondral cysts, osteophytes, osteoarthritis grade, or Tönnis grade. Conclusion: Radiographs and MRI scans had substantial limitations and inconsistency between raters in evaluating common markers of hip osteoarthritis. MRI scans demonstrated strong reliability in evaluating for subchondral cysts but did not improve the interobserver variability of grading hip arthritis.