Accuracy and reproducibility of computer-assisted measurement of polyethylene wear in knee arthroplasty.

Abstract

When using computer-assisted methods to evaluate polyethylene wear in knee arthroplasty (TKA), variations in the inclination of the X-ray beam and lack of standard calibration may affect accuracy and reproducibility. To address these issues, we evaluated the polyethylene thickness of unimplanted specimens of known dimensions using the Imagika software. Radiographs were taken with small controlled variations in the inclination of the X-ray beam. Reproducibility was studied based on triplicate measurement of 132 fluoroscopic images by three observers. Calibration was tested against a reference based on a spherical metal ball with a known diameter. The mean differences between the measured and true values ranged from 0.6 mm to 0.8 mm. The repeatability coefficient revealed a maximum variation of 0.43 mm for the same observer, and 0.39 mm between observers. There were significant differences between the measurements of polyethylene thickness performed using two different calibration methods. The variance of measurements was lower with digitized images than with fluoroscopic images. Imagika was not efficient to measure wear in TKA.