Abstract
CT-based radiostereometric analysis (CT-RSA) is a method of measuring implant micromotion using low-dose CT scans. We evaluated the precision, accuracy, and observer reliability of CT-RSA for measuring femoral stem translation in THA in vitro. A cementless femoral stem was implanted in a synthetic femur mounted on a calibrated micrometer platform. Controlled translations were applied along each orthogonal axis across 5 series, generating 90 CT data sets. Only translational micromotions were imposed and analyzed. Precision was calculated from double examinations-repeated scans of the same displacement acquired in a different series-yielding 60 pairs. The femoral stem, head, and bone were segmented, the head's geometric center was defined, and translations were analyzed. Because the platform was fixed 90° to the femoral neck (CCD 135°), data were reported both on native axes, and after a 45° counterclockwise rotation to the conventional RSA frame. Precision was estimated from double examinations (SD × t), accuracy from RMSE versus micrometer readings (RMSE × t), and reliability from intraclass correlation coefficients (ICC) by two blinded observers. On native axes, precision ranged 0.06-0.08 mm and accuracy 0.09-0.13 mm. On transformed axes, precision ranged 0.06-0.07 mm and accuracy 0.03-0.14 mm. Intraobserver reliability was excellent (mean ICC 0.99 for all axes), as well as interobserver reliability (mean ICC 0.91-0.98). CT-RSA achieved sub-tenth-millimeter precision with high repeatability and excellent observer agreement. These findings support CT-RSA as a viable alternative for early migration measurement in THA research. Clinical validation with low-dose protocols and predefined quality criteria remains warranted.