Abstract
Introduction
Aseptic acetabular component failure rates have been reported to be similar or even slightly higher than femoral component failure. Obtaining proper initial stability by press fitting the cementless acetabular cup into an undersized cavity is crucial to allow for secondary osseous integration. However, finding the insertion endpoint that corresponds to an optimal initial stability is challenging. This in vitro study presents an alternative method that allows tracking the insertion progress of acetabular implants in a non-destructive, real-time manner.
Materials and Methods
A simplified acetabular bone model was used for a series of insertion experiments. The bone model consisted of polyurethane solid foam blocks (Sawbones #1522-04 and #1522-05) into which a hemispherical cavity and cylindrical wall, representing the acetabular rim, were machined using a computer numerically controlled (CNC) milling machine (Haas Automation Inc., Oxnard, CA, USA). Fig. 1 depicts the bone model and setup used. A total of 10 insertions were carried out, 5 on a low density block, 5 on a high density block. The acetabular cups were press fitted into the bone models by succeeding hammer hits. The acceleration of the implant-insertor combination was measured using 2 shock accelerometers mounted on the insertor during the insertion process (PCB 350C03, PCB Depew, NY, USA). The force applied to the implant-insertor combination was also measured. 15 hammer hits were applied per insertion experiment. Two features were extracted from the acceleration time signal; total signal energy (E) and signal length (LS). Two features and one correlation measure were extracted from the acceleration frequency spectra; the relative signal power in the low frequency band (PL, from 500–2500Hz) and the signal power in the high frequency band (P Hf, from 4000–4800 Hz). The changes in the low frequency spectra (P Lf, from 500–2500 Hz) between two steps were tracked by calculating the Frequency Response Assurance Criterion (FRAC). Force features similar to the ones proposed by Mathieu et al., 2013 were obtained from the force time data. The convergence behavior of the features was tracked as insertion progressed.
Results
Differences were noted visually between the acceleration data recorded at the beginning of insertion and towards the end, both in the time domain (fig. 2A) as well as in the frequency domain (fig. 2B). These differences were also captured by the proposed features. Fig. 3 shows a typical representation of how the time (A), frequency (B) and force (C) features evolved during insertion. Based on a simple convergence criterion, the insertion endpoint could be determined.
Conclusions
The convergence behavior, and the insertion endpoint thus identified, of the force-based and acceleration based features correlated well. The different features capture the changes in damping and stiffness of the implant-bone system that are occurring as the insertion progresses and combining them improves the robustness of the endpoint detection method.
For any figures or tables, please contact the authors directly.
