Retrospective identification of subject anthropometry using computed tomography of the leg

Lower extremity injuries from car crashes are associated with decreased quality of life. To better evaluate the forces seen by the lower extremity during car crashes accurate models of the lower extremities must be created. This effort is motivated by a need to identify CT scans of a 5th female and a 50th and 95th male leg for use in finite element model development. Our goal is to outline a method for obtaining retrospective data on skeletal anthropometry and relate this data to the population when subject anthropometry is unavailable. Landmark data was collected from axial slices of lower extremity CT scans without skeletal pathology and CT scout films. The two methods used to collect data were compared and published data was used to create normal distribution curves for the leg lengths of males and females across the adult population. Knowledge of how lower extremity geometry quantitatively relates to subject height was also used to find patient scans representing lower extremity lengths of the 5th female, and 50th and 95th percentile male standard models. From the data collected on the two methods we found that CT 3-D reconstructions are superior for assessing length compared to using CT scouts. This methodology is useful for mining the large database of clinical patient scans retrospectively to create models that can predict injury in humans of all shapes and sizes.