Evaluation of contemporary software methods used to quantify polyethylene wear after total hip arthroplasty

Radiographic measurements of the wear of total hip arthroplasty implants are indirect measurements based on radiographic determinations of the location of the femoral head relative to the acetabular component. Using the simplest case of zero wear, we assessed the reproducibility and accuracy of two software applications designed to quantify wear from clinical radiographs.

After a cobalt-chromium head was glued into the polyethylene liner of a titanium shell, one cross-table lateral radiograph and three anteroposterior pelvic radiographs were made for twelve permutations of acetabular component angulation. The three anteroposterior radiographs differed only with regard to the cephalocaudal positioning of the prosthesis relative to the x-ray tube. To assess method reproducibility, each anteroposterior radiograph was assumed to be both the initial and the latest follow-up radiograph of a wear analysis. To assess method accuracy, each anteroposterior radiograph was paired in a wear analysis with each of the two anteroposterior radiographs made when the component was in the same angulation but at a different cephalocaudal position relative to the tube (one radiograph was the initial follow-up radiograph while the second was the latest follow-up radiograph). The analyses of reproducibility and accuracy were performed both with and without inclusion of the lateral radiograph made with the component in the same angulation.

Both methods fared well in the reproducibility analyses, with mean linear and volumetric wear values of 0.00 to 0.07 mm and 0 to 24 mm(3), respectively. In the accuracy analyses, the mean linear and volumetric wear values derived with the two methods were 0.26 to 0.40 mm and 78 to 126 mm(3), respectively.

Whereas the results of the reproducibility analyses showed that the methods were consistent in determining the relative positions of the head and shell from a given anteroposterior radiograph or pair of anteroposterior and lateral radiographs, the non-zero wear results obtained in the accuracy analyses proved that these positional determinations were often inaccurate. Thus, while contemporary software methods may yield reproducible results, their accuracy is limited by their inability to correctly determine the position of the head relative to the acetabular component.