Magnetic resonance imaging of pelvic entheses--a systematic comparison between short tau inversion recovery (STIR) and T1-weighted, contrast-enhanced, fat-saturated sequences.

OBJECTIVE

To assess the contribution of contrast material in detecting and evaluating enthesitis of pelvic entheses by MRI.

MATERIALS AND METHODS

Sixty-seven hip or pelvic 1.5-T MRIs (30:37 male:female, mean age: 53 years) were retrospectively evaluated for the presence of hamstring and gluteus medius (GM) enthesitis by two readers (a resident and an experienced radiologist). Short tau inversion recovery (STIR) and T1-weighted pre- and post-contrast (T1+Gd) images were evaluated by each reader at two sessions. A consensus reading of two senior radiologists was regarded as the gold standard. Clinical data was retrieved from patients' referral form and medical files. Cohen's kappa was used for intra- and inter-observer agreement calculation. Diagnostic properties were calculated against the gold standard reading.

RESULTS

A total of 228 entheses were evaluated. Gold standard analysis diagnosed 83 (36%) enthesitis lesions. Intra-reader reliability for the experienced reader was significantly (p = 0.0001) higher in the T1+Gd images compared to the STIR images (hamstring: k = 0.84/0.45, GM: k = 0.84/0.47). Sensitivity and specificity increased from 0.74/0.8 to 0.87/0.9 in the STIR images and T1+Gd sequences. Intra-reader reliability for the inexperienced reader was lower (p > 0.05).

CONCLUSIONS

Evidence showing that contrast material improves the reliability, sensitivity, and specificity of detecting enthesitis supports its use in this setting.