Deconvolution of surface topology for quantification of initial wear in highly cross-linked acetabular components for...

Evaluation of the surface morphology of short-term retrieved cross-linked acetabular components requires differentiation between the features generated during machining and the smaller-scale morphologies generated during the in vivo wear process. Previously, the distinction between the waviness of machining and the roughness of wear has been related to the grain size of the UHMWPE. Here a low-frequency cutoff is proposed, based on the maximum spectral frequency of machining marks, rather than on the grain size of the bulk UHMWPE material, as a reliable method for deconvolving machining marks from in vivo wear following short-term implantation. To this end, as-machined articulating surfaces of conventional (GUR 1050) and two groups of highly cross-linked UHMWPE acetabular components were examined to determine whether they exhibited a periodic surface morphology with a well-defined spatial frequency. The surface frequency spectra revealed low-frequency peaks associated with the machining marks, which were unique to each type of implant. Furthermore, the surface frequency spectra appeared uniform within a single group of implants. Statistically significant differences in the surface roughness and waviness were observed between the three groups of new implants. Our research suggests that machining marks can be effectively deconvolved from the articulating surface with the use of a Fourier transform algorithm with a single cutoff frequency of 0.08 1/microm, corresponding to a wavelength of 12.5 microm. The results of this study provide a unified conceptual framework for discriminating between waviness and roughness of the articulating surface for machined orthopedic components. The distinction between waviness and roughness is expected to be crucial for the comprehensive evaluation of wear surfaces after short-term implantation, when machining marks may be partially worn away or plastically deformed in vivo.