Three-dimensional morphologic study of the child's hip: which parameters are reproducible?

BACKGROUND: Biplanar x-ray images obtained with patients in a standing weightbearing position allow reconstruction of three-dimensional (3-D) bone geometries, with lower radiation exposure than CT scans and better bone definition than MRI. QUESTIONS/PURPOSES: We determined the reproducibility of 3-D parameter values of the hips and pelves of healthy children, using biplanar x-ray images.
METHODS: We built 3-D models of the hips of 33 children without musculoskeletal problems: 10 subjects younger than 9 years and 23 who were 9 years or older. Three anatomic landmarks and nine hip and pelvic parameters were computed for each reconstruction. To determine the reliability of these landmarks and parameters, each bone was reconstructed four times by two independent observers, leading to a total of 264 reconstructions, and parameters were studied for the two age groups and compared between dancers and nondancers.
RESULTS: Taking into account all reconstructions, the interobserver reproducibility ranged from 2 to 4 mm for landmark positions or distance parameters, and 2° to 6° for angular parameters. The most reproducible point was the center of the femoral head (range, 0.2-17 mm). The distance between this center and its projection on the plane fitting the edge of the acetabulum, and the pelvic tilt were the most reproducible parameters.
CONCLUSIONS: Reproducible 3-D reconstructions of hips and pelves of children were possible using biplanar x-ray images, regardless of the children's ages. Although we report preliminary values for 3-D parameters in healthy children's hips, further work is needed to obtain direct validation of our parameters using CT reconstructions of cadaveric specimens to avoid high doses of radiation.