A CT-based method to compute femur remodelling after total hip arthroplasty

Bone remodelling after total hip arthroplasty has been largely observed and investigated. Most studies rely on projective images and only few obtain 3D information with limited spatial resolution. This study proposes a method to provide quantitative, 3D high-resolution data about femur bone density variations, by means of CT volume processing. This would offer a tool for further research and clinical studies. Five patients subjected to primary, cementless total hip arthroplasty were considered. Calibrated CT volumes were acquired before, just after surgery and 1 yr later. Bone remodelling hinders accurate alignment of femur volumes acquired after a year; instead, prosthesis stem remains unchanged. Thus, after metal artefact reduction, prosthesis was segmented, and stem-based accurate alignment was obtained. A test to exclude prosthesis migration was performed by considering specific femur anatomical landmarks. Bone density error due to artefact reduction and realignment was estimated. Quantitative differences in bone mineral density were computed for each voxel, providing a resolution of about 1 mm. Preliminary results showed that the femur underwent consistent remodelling after a year. Widespread bone density losses appeared in those areas where strain-adaptive remodelling is normally expected, particularly about the calcar. Conversely, distal areas with clear stem-bone contact showed considerable density gains.