Evaluating hip labral tears using magnetic resonance arthrography: a prospective study of comparison between hip arthroscopy and magnetic resonance arthrography diagnosis (SS-92)

Abstract

To evaluate the usefulness and accuracy of magnetic resonance arthrography (MRAr) with true sagittal scout image mapped radial reformation in localizing hip labral tears. We demonstrate a prospective study to compare the clinical result between the preoperative MRAr diagnosis and hip arthroscopic finding. Materials and Methods: During April 2002 to Dec 2002, 30 patients were referred for hip MRAr due to suspected labral tears in the authors’ hospital. They included 17 male and 13 female patients, age ranged from 17 to 62 yr, mean age 41 yr. All patients received standard hip arthrography using 1:200 diluted Gd-DTPA as contrast medium. MR examinations were performed at a 1.5-T scanner (Vision; Siemens, Er LANGen, Germany) using 3D-FLASH pulse sequence (TR/TE/flip angle = 48ms/11ms/40°, FOV 15 cm, matrix 256∗256, effective thickness 1.5 mm for 44 partitions, and 1 acquisition). Multi-planar reformation (MPR) using double oblique technique was performed to produce the radial reformatted rotation images with 10° increment on the acetabular rim. For ease of localizing the lesion during arthroscopic surgery, we adopt the true sagittal reformatted image on the hip joint as scout localizer to map the radial reformatted images. Labral tears were annotated as anterior-superior (AS), anterior-inferior (AI), posterior-superior (PS), and posterior-inferior (PI) quadrant based on the cranial-caudal axis in true sagittal reformatted scout localizing image. MRAr images were interpreted by one M.S. radiologist (L.C. Lien). Patients who had positive MRAr findings were advised to have the hip arthroscopic operation. By supine position, two or three portals (anterior, anterolateral or posterolateral portals) were placed after application of traction to the involved leg. These lesions were debrided and excised by the aid of flexible and curved electrothermal probes or shaver. All the surgery was performed by the same surgeon (Y.S.C.). Results: Five patients had no MRAr evidence of labral tear. Two of them showed MR evidence of avascular necrosis of the femoral head. Of the 25 patients who had MRAr evidence of labral tear, 17 patients (M:F = 10:8, RT: LT = 13:5, age 22–64 yr, mean age 43 yr) received operation. The duration between MRAr examination and the operation ranged from 5 to 44 days (mean 23 days). One patient had only synovial process in the hip joint near the MRAr suspected region. All the remaining 16 operated patients had labral tears with totally 21 quadrant lesions in arthroscope (distributed as AS: AI: PS: PI = 14: 3: 2: 2). Radial reformatted images showed correct mapping with arthroscope findings in 20 of the 21 quadrant lesions. Conclusions: MRAr using radial reformatted images with true sagittal localizer may have high accurate rate in diagnosis for hip labral lesions and guide the arthroscopist in portal selection, making localization of hip labral tears easier during surgery. However, the accuracy of MRAr diagnosis for hip labral tears also depends on the experienced MRAr technique and imaging reading.