Errors of computer-assisted migration analysis in conventional radiographsof femoral hip implants - an experimental study

Several methods have been described in the literature in order to analyze m igration of femoral hip implants in conventional radiographs. However, no m easurements were done regarding the potential errors inherent in such unsta ndardized radiographs of daily routine. in order to quantify this lack of r eproducibility, we carried out experimental examinations with radiographs o f a hip revision prosthesis, where different variables of technical Xray co nditions and femoral positions were changed. All radiographs were analyzed with a computer-assisted processing system by different procedures. At firs t, the radiographs were calibrated geometrically, Then, particular referenc e distances between defined. points of the implant and cortical bone struct ures were analyzed quantitatively. The error of the axial migration analysi s was up to 11.11 mm concerning different femoral positions, up to 8.29 mm in the case of different source-to-film distances and up to 2.21 mm due to different lateral localizations of the central X-ray focus. Significant err ors for the migration analysis in the transverse and rotational planes were only found under conditions of varying the femoral position, and not under different technical X-ray conditions. From these experimental results, we derived the following criteria to minimize failure in a quantitative radiog raphic migration analysis: (1) the patient's leg has to be placed into an a ntirotation device to ensure identical object position; (2) the same source -to-film distance has to be used, (3) the central X-ray focus has to be loc alized on the center of the film-cassette; (4) film-screen systems should b e of the same type and size; (5) the object has to be placed in the same po sition as in previous radiographs. As a conclusion, only if these standardi zation criteria are respected in the daily routine of conventional radiogra phs will an effective and meaningful use of migration analysis systems be p ossible to prove or to exclude mechanical failure of femoral hip implants i n prospective longitudinal followup series.