Purpose or Case Report: Diagnostic performance of arthrographic and non-arthrographic MRI in detecting hip labral tears has been well studied in adults but remains relatively unexplored in pediatric patients. Particularly, it is unclear if one protocol is better suited for individuals with pre-existing hip pathology (defined in this study as hip dysplasia, femoroacetabular impingement, or Perthes) that predisposes to labral abnormalities, compared to acquired or traumatic hip pathology. Thus, we aimed to compare performance of non-arthrographic and arthrographic MRIs for identifying labral tears in children with and without underlying hip abnormalities.
Methods & Materials: Hip MRIs completed at our institution between 2015-2025 were retrospectively reviewed in this IRB-exempt study. Inclusion criteria were age <18 years, presenting complaint of “hip pain,” performance of either MR hip arthrogram or non-arthrographic hip MRI, and clinical concern for labral pathology. Exclusion criteria included ipsilateral hip surgery. MRI reports were analyzed for radiological determination of labral pathology, with arthroscopy findings used as a “gold standard.”
Results: 181 hip studies met inclusion criteria. 146 studies were of traumatic hips; arthroscopic correlation was available for 28 studies (19 non-arthrographic, 9 arthrographic). Non-arthrographic MRIs had 86% sensitivity, 20% specificity, 75% PPV, 33% NPV, and 68% accuracy for detecting tears; arthrographic MRI performance was 100% / 20% / 50% / 100% / 56%. 35 studies had pre-existing hip pathology; arthroscopic correlation was available for 19 of these studies (8 non-arthrographic, 11 arthrographic). Non-arthrographic MRI performance was 57% / 100% / 100% / 25% / 63%, while arthrographic MRI performance was 100% / 0% / 82% / N/A / 82%.
Conclusions: In both cohorts, MR arthrography demonstrates higher sensitivity and lower specificity compared to non-arthrographic MRI. Arthrography thus may provide superior joint-space characterization to detect labral tears (i.e., few false negatives), with a tradeoff of identifying normal labral sulci as tears (i.e., more false positives). Compared to non-arthrographic accuracy, arthrographic accuracy is lower in patients with traumatic hips, but higher in patients with pre-existing hip pathology. This knowledge could play an important role in determining the best MRI hip study for each unique patient population.