A computerized analysis-by-synthesis algorithm improves precision of linear wear measurements in total hip replacements.

Abstract

Precision is important if small polyethylene wear rates are to be detected early and in small sample sizes. Using an automatic, computerized algorithm relying on a synthetic X-ray generated from a computer-assisted design (CAD) model of the implant may significantly improve precision of linear wear measurements. We compared this method to provide improved in vivo precision compared to the widely used Hip Analysis Suite software (version 8.0.1.7). Two anterior-posterior pelvic X-rays were taken on the same day of 18 total hip implants, and wear was measured by three observers using the two methods. Expected real wear was zero and defined as a reference value. Two sources of variability were estimated: one due to the instrument, and one due to different operators. A Wilcoxon Signed Rank Test was used for statistical comparisons. The overall precision (instrument variability and multioperator variability) given a 95% precision limit was significantly higher (p < 0.001) using the proposed method (+/-0.15 mm) when compared to the Hip Analysis Suite (+/-0.84 mm). Due to the availability of the full CAD data of the implant and to further automation of the measuring process, the newly proposed method provided improved precision in measuring polyethylene wear.