Introduction
Radiostereometric Analysis (RSA) is a technique using the concept of digital image correlation (DIC) to determine the position of radiopaque markers in a three-dimensional space. The advantage of RSA over traditional camera-based tracking systems is that no direct optical visibility of the tracked object is necessary, thereby allowing assessment of movements of intraarticular structures. The purpose of this study was to compare a novel RSA setup to a commercial 3D camera system regarding its precision.
Methods
Optical markers (0.4 mm diameter) and tantalum beads (1 mm diameter) were applied to a sawbone knee. The tibia was mounted to the table, the femur was fixed into a six degrees of freedom robotic test setup (KR125; KUKA Robotics), which was optimized to simulate joint movements by using simVITRO (Cleveland Clinic BioRobotics Lab). The robot system has a positional accuracy of 0.05 mm. A protocol consisting of 3 mm of anterior tibial translation (ATT3), followed by 3 mm of medial tibial translation (MTT3), followed by return to the home position, was performed. The protocol was repeated for a total of 10 times. In each step, the position of the femur in respect to the tibia was captured by either a commercial 3D camera system (ARAMIS, GOM GmbH), or the RSA system with one x-ray source and one digital x-ray detector. The changes between positions were calculated and the two systems compared using two-tailed t-est. Correlation between the techniques was performed using Pearson’s correlation.
Results
For ATT3 the measured translation was 3.040 mm (SD 0.005) in the ARAMIS group and 3.039 (SD 0.094) in the RSA group. No significant differences were found between the groups. For MTT3 the measured translation was 2.857 mm (SD 0.020) in the ARAMIS group and 2.798 (SD 0,153) in the RSA group. No significant differences were found between the groups. Correlation between the techniques revealed a near-perfect correlation of 0.97 (95 % CI 0.93 – 0.98).
Conclusion
The presented RSA system can precisely determine motions with an accuracy of < 0.2 mm and is therefore well suited for tracking of joint movements in biomechanical experiments.