Abstract
Accurate detection of migration of hip arthroplasty stems without the burden of bone markers and stereo-radiographic equipment is of interest. This would facilitate the study of stem migration in an experimental setting, but more importantly, it would allow assessing stem loosening in patients with a painful hip outside a study protocol.
We developed and validated a marker-free automated CT-based spatial analysis method (CTSA) to quantify stem-bone migration in successive CT scan acquisitions. First, we segmented the bone and stem within both three-dimensional images, then we pairwise registered those elements (Fig. 1). By comparing the rigid transformations of stem and bone, we calculated the migration of the stem with reference to the bone and transferred the three translation and three rotation parameters to an anatomic coordinate system. Based on the rigid transformation, we also calculated the point of the stem that presented the maximal migration (PMM).
Accuracy was assessed in a stem-bone model (Fig. 2) by imposing 39 predefined stem rotations and translations, and by comparing those with values calculated with the CTSA tool. In all cases, differences were below 0.20 mm for translations and 0.19° for rotations (95% tolerance interval (95% TI) below 0.22 mm and 0.20°, largest standard deviation of the signed error (SDSE) 0.081 mm and 0.057°). Precision was defined as stem migration calculated in eight clinical relevant zero-migration scenarios. In all cases, precision was below 0.05 mm and 0.08° (95% TI below 0.06 mm and 0.08°, largest SDSE 0.012 mm and 0.020°). The largest displacement of the PMM on the stem was 0.169mm. The precision estimated in five patients was very dependent on the CT scan resolution and was below 0.48 mm and 0.37° (95% TI below 0.59 mm and 0.61°, largest SDSE 0.202 mm and 0.279°, largest displacement of the PMM 0.972 mm). In optimized conditions, the precision in patients improved largely and was below 0.040 mm and 0.111° (largest SDSE 0.202 mm and 0.279°, largest displacement of the PMM 0.156 mm).
Our marker-free automated CT-based spatial analysis can detect hip stem migration with an accuracy and precision comparable to that of radiostereometric analysis (RSA), but without the burden of bone markers and the cost of stereo-radiographic equipment. As such, we believe our tool could make accurate measurement of stem migration available to departments without access to RSA and boost this type of research. Moreover, as CTSA does not rely on bone makers, it is applicable to all-comers with a painful hip arthroplasty. Indeed, in those patients with a reference CT scan after hip replacement, a new CT scan could demonstrate stem migration. If no initial CT scan is available, a reference scan could be taken during a first visit and repeated later. Additionally, a “stress test” of the hip could be performed. During such test, comparing CT images acquired during forced maximal intern and external rotation could demonstrate stem loosening.
