Evaluation of Hip Translation using Three-Dimensional Magnetic Resonance Imaging

INTRODUCTION: Acetabular dysplasia has been implicated as an etiologic factor in the development of secondary osteoarthritis (OA) of the hip . Dysfunction of the hip secondary to dysplasia derives from multifactorial problems such as excessive stresses on the articular cartilage and dynamic hip instability. To our knowledge, however, there have been only a few studies investigating hip instability . Evaluation of dynamic hip instability may not only analyze one of the causes of secondary OA but also unveil the etiology of primary OA of the hip. The purpose of the present study was to elucidate hip instability of the in vivo normal hip joint and acetabular dysplasia by evaluating the three-dimensional (3D) translations using 3D magnetic resonance imaging (MRI). METHODS: Twenty normal hips from young healthy volunteers (mean ages 31.8 years), 10 normal hips from elderly healthy volunteers (mean ages 68.6 years), and 22 dysplastic hips (mean ages 32.9 years) at preor early stages of OA were examined. All subjects were females. Criteria for enrollment as a dysplastic hip were as follows: less than 20 ̊ of center-edge (CE) angle on the mid-coronal MR image; Group I subluxation according to the classification of Crowe et al.; no evidence of joint space narrowing; no deformity of the femoral head; and no previous operation on the hip joint. MR imaging was performed with 3D true fast imaging with steadystate precession (FISP) using a flexible surface body coil on a 1.5-T MR system (MAGNETOM Espree, Siemens, Erlangen, Germany), making images with the hip in four different positions bilaterally: neutral, 45 ̊ of flexion, 15 ̊ of extension, and the Patrick position (Fig. 1). Femoral and pelvic bones were separately extracted at the neutral position and superimposed over the images of each different position using voxel-based registration with the Virtual Place-M software program (Medical Imaging Laboratory, Tokyo, Japan) . The center of the femoral head sphere (head center, FHC) and that of the acetabular sphere (acetabular center, AC) were calculated with a least square method. The distance between AC and FHC at the neutral position was defined as the 3D migration. The 3D translation from a neutral position to each position was measured as the distance of the FHCs between a neutral position and each position (Fig. 2). The MRI data acquired with two days interval were analyzed and compared for inter-observer reproducibility. Two-way repeated measures analysis of variance was performed to consider the dependency between right and left-side data, and multiple linear regression analyses were performed to assess independent relationships. A significance level of 5% was set for all comparisons (p<0.05). RESULTS: The acetabular sphere radius was the determinant for 3Dmigration (β=0.865, p<0.001), and there was a statistical significant difference in 3D-migration between normal female hips and dysplastic hips (p=0.047). From neutral to 45 ̊ of flexion, the FHC of normal young, elderly, and dysplastic hips translated antero-inferiorly by 1.10±0.31 mm (0.60–1.57 mm), 1.70±0.48 mm (1.04–2.45 mm) and 1.30±0.41 mm (0.76–2.39 mm), respectively, and the difference between normal young female hips and elderly hips was statistically significant (p=0.008). Age and acetabular sphere radius were the determinant factors for 3D-translation from neutral to 45 ̊ of flexion (β=0.473, p<0.001 and β=0.351, p=0.003, respectively). From neutral to the Patrick position, the FHC of normal young, elderly, and dysplastic hips translated postero-infero-medially by 1.12±0.39 mm (0.45–1.85 mm), 0.62±0.31 mm (0.28–1.14 mm) and 1.97±0.84 mm (0.95–4.34 mm), respectively, and there were statistically significant differences between normal young female and elderly hips (p=0.010) and between normal young female and dysplastic hips (p=0.005). CE angle was the determinant for 3D-translation from neutral to the Patrick position (β=-0.774, p<0.001). The average root mean square error in 3D-translation was 0.193 mm for interobserver reproducibility. DISCUSSION: This study analyzed 3D morphology and 3D motions of the in vivo native hip with a non-invasive method using 3D MRI voxel based registration technique. The present method had high reproducibility for evaluation of 3D-translation of the hip joint. The interesting findings of the present study were that the elderly female hip translated significantly larger than the normal young female hip from neutral to 45 ̊ of flexion. This discrepancy in 3D-translaton from neutral to 45 ̊ of flexion may be attributed to the age-related curvature difference in the hip joint or muscle imbalance in the elderly subjects. It was also interesting that the dysplastic hip translated significantly larger than the normal female hip in the Patrick position and that CE angle is the major determinant for 3D-translation from neutral to the Patrick position. These results may indicate that hip instability is increased with decreasing CE angle <20 ̊. Besides, the FHC translated postero-infero-medially from neutral to the Patrick position. These findings may be demonstrative of a sole posterior or posteroinferior labral tear in the acetabulum. SIGNIFICANCE: Hip instability was increased in proportion to age and the severity of acetabular dysplasia. 3D-translation from neutral to the Patrick position could be studied in relation with hip joint stability or instability as a result of joint preserving surgery such as osteotomy and labral repair and with disease progression in OA.