Purpose of the study: The mean rotation center (MRC) characterizes the movement of two solids in relation to each other. This parameter has been proposed for the cervical spine to describe the motion of vertebral segments. Two lateral views (flexion and extension) are required to draw the necessary lines and establish the centers of rotation. The process is rigorous but time-consuming. We validated a computerized analysis system for automatic determination of the cervical MRC and study the localizations observed in healthy subjects.

Material and methods: Validation of the computerized system. Accurate angle measurements: nine cervical spines were harvested from anatomic specimens. A K-wire was inserted sagittally into each vertebra. Lateral images were obtain in flexion and extension. The measurements of mobility made by the software were compared with manual measurements. Reproducibility tests (intra- and interobserver): six pairs of flexion and extension views in healthy subjects. Two different observers made fifteen successive measurements of each MRC for each spinal segment. Frequently encountered positions of the MRC in healthy subjects: stress films were obtained in 51 healthy subjects aged 18–40 years. For each spinal segment, the MCR was determined with the computerized system.

Results: Accuracy of the angle measurements: the precision was 1.4° for a 95% interval of confidence. Reproducibility: variability of the position in X and Y for the MRC (expressed in percent of the size of the vertebral body) was: 19.6 and 24.5 for C2–C3; 112 and 15.3 for C3–C4; 7.7 and 9.4 for C4–C5; 9.1 and 9.4 for C5–C6; 13.1 and 11.8 for C6–C7. Positions frequently encountered in healthy subjects: the most frequent position of the MRC varied from one segment to another. There was a frequent position for each segment. These frequent positions were situated in the posterosuperior quadrant of the subjacent vertebra for C2–C3, C3–C4, C4–C5, and C5–C6. For C6–C7, the frequent positions for MRC were at the level of the intervertebral space, behind the center of the disc.

Discussion: The software tested here appeared to provide good measurements for cervical spine from C3 to C7. At these levels, the measures were accurate and reproducible, as were the coordinates for the MCR of each segment. The frequent positions of the MRC found in this study are the same as reported by other authors. This method is easy to apply in routine practice.