PURPOSE: To determine the accuracy of magnetic resonance (MR) arthrography in the detection and staging of lesions of the acetabular labrum. MATERIALS AND METHODS: Fifteen hips of 14 patients with chronic hip pain and clinical suspicion of labral lesions were examined with direct or indirect MR arthrography and underwent subsequent arthroscopic evaluation. The study population consisted of six women and eight men aged between 40 and 59 years. Nine arthrograms were obtained intra-articular administration of gadolinium solution, and six involved articular motion exercise after intravenous administration of gadolinium solution. In 14 cases a phased-array pelvic coil was used, and a shoulder coil in one. Fat-suppressed T1-weighted images were obtained in the coronal, oblique coronal and oblique axial planes, and T1-and T2-weighted images were obtained in the axial plane. Labral lesions were graded according to the Czerny classification and evaluated on the basis of presence or absence of a tear and location (anteroinferior, anterosuperior, superior, posterosuperior, posteroinferior). The findings were then correlated with the arthroscopic findings. RESULTS: After direct and indirect MR arthography, the findings-based on the Czerny classification -were as follows : stage IA:3; IB:1; IIA:8; IIB:1; IIIA:4; IIIB:0. MR arthrography also showed that seven cases were located in the anterosuperior portion of the labrum, seven in the superior portion, and three in the posteroinferior portion. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MR arthrography for diagnosis of the labral tear was 92%, 75%, 92%, 75 %, and 88 %, respectively. CONCLUSION: In this preliminary study, MR arthrography appears to be a promising imaging modality for accurate diagnosis and useful for screening patients with chronic hip pain.