gms | German Medical Science

Objectives: Roentgen stereophotogrammetric analysis (RSA) accurately measures the early migration of implants in total joint arthroplasty and presents the gold standard to assess implant fixation in vivo (Valstar et al., 2005). The continuous development of RSA, from marker- to model-based approach not only reduces the costs, but also makes the application scope of RSA wider (Kaptein et al., 2003). Accuracy of model-based RSA is comparable to the gold standard maker-based RSA, but less for superior-inferior axis of hip stems (Seehaus et al., 2009). To the authors' knowledge less data exist according intra- and interrater-reliability of available RSA methods. Therefore, the purpose of this study is to carry out a retrospective analysis of clinical RSA for total hip arthroplasty to assess intra- and interrater-reliability of two different RSA approaches.

Methods: A total of n=18 stems (Lubinus, Link, Hamburg, Germany) and n=11 cups (Phoenix, Peter Brehm GmbH, Weisendorf, Germany) components were included. All RSA radiographic image pairs were analyzed by model-based RSA using elementary geometrical shape models (EGS) and marker-based RSA method with MBRSA 4.1 (Medis Specials, Leiden, Netherlands). Analyses was repeated n=3 times by three different observers. Barnhart's method (Barnhart et al., 2007) was used to calculate intragroup and intergroup differences, represented by mean square difference (MSD) and 95% confidence interval (CI).

Results: Overall, the results show that both marker-based and model-based RSA using EGS have low intragroup difference in the measurement of translational migration along all the three (Xt, Yt, Zt) axis and rotational migration around anterior-posterior (Zr) axis (stem <0.028 mm2, cup <0.119 mm2). For the stem component, model-based RSA using EGS indicate an increasing intragroup difference for rotations around cranial-caudal axis (0.994 mm2), which is much smaller in marker-based RSA (0.242 mm²) (Figure 1A [Fig. 1]). For the cup component, marker-based RSA shows the largest intragroup difference for rotations around the medial-lateral axis (0.337 mm²) (Figure 1B [Fig. 1]).

Conclusion: Migration detection for stem components for rotations around cranial-caudal axis indicates a higher intrarater difference for model-based RSA using EGS compared to marker-based RSA, which is possibly related to the limitation of its working principle using pose-estimation technique (Seehaus et al., 2009). Within a next step the interrater reliability analysis should be completed and give further information according reliability of RSA approaches.