Registration Accuracy In Dynamic CT of the Wrist Joint across Multiple CT Vendors: A Phantom Study
Abstract
Purpose
To evaluate the registration accuracy across acquisition and reconstruction protocols of different CT systems in 4D CT of wrist bones.
Methods
A rotating wrist phantom with three 3D-printed bones (scaphoid, lunate, capitate) was scanned on four CT systems from two manufacturers (Aquilion ONE PRISM and ONE Vision, Canon Medical; single-source and dual-source SOMATOM Force, Siemens Healthcare). One static 3D scan and multiple axial 4D scans were acquired at different phantom rotation speeds, each lasting 10 s. Single-source covered 0.5–3 motion cycles, and dual-source 1–6 cycles (one cycle: radial–ulnar–radial; 0.5 cycle: radial–ulnar). Dose dependence was evaluated on two systems (80 kV/40 mA and 120 kV/100 mA). Images were reconstructed in full and partial modes, segmented, and registered using point-to-image registration. Registration accuracy was calculated as translation and rotation errors of the scaphoid and capitate relative to the lunate, referenced to the static scan, and reported per wrist cycle. Effects of CT system, bone type, reconstruction method, and rotation speed were analyzed using a linear mixed model (p < 0.05).
Results
Registration accuracy was unaffected by dose. For one cycle, the errors of the best single-source system were median 0.29 mm (interquartile range: 0.21–0.36 mm) and 1.37° (0.82°–2.18°) for full and 0.21 mm (0.16–0.25 mm) and 0.70° (0.52°–0.95°) for partial reconstructions for the capitate; the dual-source system showed the highest accuracy (errors of 0.14 mm (0.12–0.17 mm) and 0.48° (0.39°–0.62°)). Registration accuracy showed statistically significant difference among all parameters (p<0.001), except between the single-source scanner types (p=0.12).
Conclusion
Registration accuracy was dose-independent and primarily influenced by motion speed, reconstruction method, and bone type, with dual-source CT performing best. These findings provide baseline motion-induced registration errors, supporting differentiation between true kinematics and apparent displacement from motion blur.