Estimating polyethylene wear in total hip arthroplasty by using computed tomography and alternative RSA techniques

Abstract: Introduction: Traditionally Total Hip Arthroplasty (THA) components performance, and in particular the performance of the polyethylene liner, is evaluated in 2D by examining plain radiographs or in 3D using Radio Stereo-metric Analysis (RSA). While 2D techniques require only a plain radiograph, they are less accurate. RSA systems, on the other hand, are more accurate but are more complicated to set up and require dedicated equipment. CT scanners are widely available and can measure THA performance in 3D. These techniques fail to estimate wear when tantalum cups are implanted, and alternative methods are needed. Specific Aims: Validate a Computerized Tomography (CT) technique for evaluating THA wear, develop and test algorithms to estimate wear in THA when tantalum cups are implanted. Methods: To validate the 3D CT data and software, a supine hip phantom fitted with a 3-axis micrometer tower was scanned, first (as a feasibilty study) in an experimental ultra-high resolution flat panel CT scanner and then in a multi-detector CT scanner. The micrometers were displaced in the x, y, and z axes, displacement of the micrometers (femoral head) was estimated and then compared to the actual micrometer readings (studies I and II). Wear was also estimated from clinical CT data of patients that had their THA revised: images were analyzed and compared to a coordinate measurement machine (CMM) and a micrometer (study III). For the analysis of liner wear when tantalum cups are used, four methods were tested: standard RSA, Model Based RSA, RSA-Helical axis, and RSA Center of Rotation. To test these methods a supine hip phantom was used and the femur was abducted. The center of the femoral head was calculated using these four methods (Study IV). Results: Study I - Measuring femoral head displacement using a phantom in the high resolution flat panel CT scanner, the mean difference between the actual micrometer displacement and the CT readings was found to be -0.14±0.12 mm (-0.06 to -0.21 mm 95% CI). In study II, similar to study I, a hip phantom was placed in a multi-detector CT and the femoral head displacement was compared between readings from CT data and the actual micrometers displacement. The mean accuracy and precision for the individual axis x, y, and z was 0.159±0.056 mm, 0.113±0.029 mm, and 0.209±0.036 mm respectively, with combined accuracy of 0.285 mm. In study III, we compared CT wear measurement to the actual wear of the same retrieved implant. Ex-planted liners were measured using CMM and micrometer, the average differences and standard deviations were: CMM-CT 0.09±0.29mm, CMM-Micrometer 0.01±0.32, and micrometer-CT 0.11±0.44. In study IV, comparing alternative techniques of calculating femoral head center when tantalum cups are used, the 2D average head-cup distance was calculated by: standard RSA 0.41mm, Model Based RSA 0.38mm, RSA RSA-HAT 0.96mm, and RSA-COR 1.41mm. Conclusion: Under ideal conditions, with no soft tissue or motion artifacts, and with a highresolution flat panel scanner it is possible to record femoral head penetration to a clinicaly acceptable level.When considering the clinical application of current CT technology and measurement techniques, the expected wear measurement accuracy should be 0.3 mm. In cases where CT technology does not provide an adequate solution for wear measurement (when tantalum cups are implanted), model based RSA provides the closest agreement to gold standard RSA and should be considered as a viable solution for wear measurement.