Evaluation of errors in measurements of infantile hip radiograph using digitally reconstructed radiograph from three-dimensional MRI.

Abstract

Plain hip radiograph is commonly used for the diagnosis of infantile acetabular dysplasia. Many infants are unable to maintain adequate position during radiography. Besides, the infantile hip is much smaller and has a higher cartilage component in the acetabulum and proximal femur compared with the adult hip. In this study, we developed a digitally reconstructed radiograph synthesized from magnetic resonance imaging (MRI) and investigated errors of hip radiographic measurements in different pelvic positions. MRI of both hips was performed in 10 patients (mean age 3.9 years). Three-dimensional (3D) bone models were created from MRI data. We tilted 3D pelvic bone models between 10° anteversion and retroversion and through 10° rotation on the affected and contralateral sides using 3D axes. Following this, we created digitally reconstructed radiographs in each pelvic position and calculated the acetabular index (AI), center-edge angle (CEA), migration percentage (MP), and teardrop distance (TDD). AI tended to increase with pelvic retroversion and did not change with pelvic rotation. CEA tended to decrease with pelvic retroversion and rotation on the contralateral side. MP increased with pelvic retroversion and rotation on the contralateral side. TDD did not change significantly with pelvic tilt and rotation. Radiographic measurements of hip in infants were highly influenced by pelvic movement. AI was influenced by pelvic tilt; CEA and MP were influenced by both pelvic tilt and rotation. We need to keep in mind that infantile hip radiographs could have about ±5° errors in AI and CEA.