A new computed tomography method for evaluation of orthopedic implants : Applied to the acetabular cup

Background: Two major long term complications after total hip arthroplasty (THA) is aseptic loosening and wear of the implant. In the clinical situation a detection level of approximately 1 mm is desirable. However, conventional planar radiographs used in clinical routine cannot reach this level. Other planar radiographic methods have a higher degree of accuracy, but are either invasive or too complicated for routine use. Consequently a new non-invasive method for more accurate routine evaluation of orthopedic implants, is needed. Aims: To develop and test a new computed tomography (CT) method for evaluation of migration and wear of orthopedic implants. And to validat a locally developed image post processing tool for standard orientation. Method: Study I: Two CT scans, ten minutes apart, were obtained from each of ten patients after total hip arthroplasty. A locally developed semiautomated program for volume registration (merging) was used to fuse the pelvis of the two examinations. The acetabular cup was then evaluated visually and numerically to test the method's accuracy for migration detection. Study II: Twenty-six CT scans of a pelvic model were obtained while altering the position of an acetabular cup. Using the same program for volume registration as in Study I, the pelvis from different scans was fused, and the differences in cup position were evaluated visually and numerically by two examiners. The results were compared to direct measurements on the model. Study III: Twenty-four CT volumes from the model study and the 10 pairs of post operative CT volumes from Study I were rotated to a defined pelvic standard orientation and the rotation recorded. Study IV: Eight uncemented cups with femoral components were scanned twice with CT ex vivo. The linear penetration depth of the femoral component into the cup and the thickness of the remaining polyethylene were measured in the CT volumes. Each volume was assessed twice by two examiners. The CT measurement was compared to direct measurements on the cup using a coordinate measuring device and micrometer measurements. Results: Study I: It was shown that the method had a signifcantly higher accuracy than expected from routine radiography in detecting acetabular cup migration, giving both visual and numerical correlate. Study II: The two examiners were able to differentiate between 0 mm, 1 mm and 2-3 mm cup migration with 100% specificity and sensitivity. Numerically intra- and inter-observer 95 % accuracy and repeatability limits were 0.5/0.7 mm respectively. Study III: Reorienting the pelvis in image postprosessing was shown to be accurate and had a precision of 1° for the model and 1.5° for the patients. Study IV: Accuracy of the wear measurements expressed as penetration depth were ±0.6 mm and ±1.0 mm for the two examiners respectively. Accuracy of wear measurements for remaining polyethylene was ±1.3 mm and ±1.0 mm for the two examiners respectively. The method allows for accurate wear detection at a clinically relevant level. Measurements of remaining polyethylene were slightly less reliable than penetration depth, due to metallic artefacts. Conclusions: These four studies present the basic principles of a new CT method, combining volume registration of CT data for migration detection, and landmarking of the prosthetic components to get a numerical correlate to migration and wear detection. It also offers validation of a locally developed image post processing tool for standard orientation. Both migration and wear detection were shown to be improved with this new CT method, when compared to the use of conventional radiographs used in clinical routine. The locally developed tool for reorienting the CT volume in postprocessing was shown to be accurate and precise, enabling measurements relative to the bone or other chosen structure in the volume and minimizing the dependency on patient positioning