Statistical shape modeling of proximal femoral shape deformities in Legg-Calvé-Perthes disease and slipped capital femoral epiphysis.

INTRODUCTION: The current understanding of morphological deformities of the hip such as femoroacetabular impingement (FAI), Legg-Calvé-Perthes disease (LCPD), and slipped capital femoral epiphysis (SCFE) is based on two-dimensional metrics, primarily involving the femoral head, that only partially describe the complex skeletal morphology.
OBJECTIVE: This study aimed to improve the three-dimensional (3-D) understanding of shape variations during normal growth, and in LCPD and SCFE, through statistical shape modeling.
DESIGN: Thirty-two patients with asymptomatic, LCPD, and SCFE hips, determined from physical and radiographic examinations, were scanned using 3-D computed tomography (CT) at a voxel size of (0.5-0.9 mm)(2) in-plane and 0.63 mm slice thickness. Statistical shape modeling was performed on segmented proximal femoral surfaces to determine modes of variation and shape variables quantifying 3-D shape. In addition, conventional variables were determined for all femora.
RESULTS: Proximal femur shape was described by eight modes of variation and corresponding shape variables. Statistical shape variables were distinct with age and revealed coordinated, growth-associated differences in neck length-to-width ratio, femoral head medialization, and trochanter protrusion. After size and age-based shape adjustment, diseased proximal femora were characterized by shape variables distinct from those of asymptomatic hips. The shape variables defined morphology in health and disease, and were correlated with certain conventional variables of shape, including neck-shaft angle, head diameter, and neck diameter.
CONCLUSION: 3-D quantitative analyses of proximal femoral bone shape during growth and in disease are useful for furthering the understanding of normal and abnormal shape deviations which affect cartilage biomechanics and risk of developing osteoarthritis.