Abstract
Since the cortical bone has higher elastic modulus compared to the cancellous bone, its geometry is very important for stress analysis of bone structure. During finite element modeling of bone structure, cortical bone is generally determined as the region having higher CT values with respect to specific threshold value. However, it is difficult to determine the thin cortical bone regions by considering a specific threshold value. This study proposes a method to select regions of cortical bone from clinical CT images by considering CT value distributions of cortical and cancellous bone. Applying the method to bovine proximal femur, the mean error in cortical thickness compared to the actual bone was found to be less than one pixel (0.39×0.39 mm). Hence, the proposed method could accurately determine the cortical bone regions from clinical CT images. The method was also applied to develop a finite element model with the precise cortical bone structure.
