Accuracy of dynamic computed tomography to calculate rotation occurring at lumbar spinal motion segments

Reliability study comparing computed tomography (CT) to biomechanics.

To measure the accuracy and precision of such measurements in comparison with a standard method.

Rotations of lumbar spinal motion segments can be measured with dynamic CT imaging. This may be a useful tool to measure intersegmental motion. Validation of its use is lacking.

Human cadaveric lumbar spines were fixed in a rigid rotation device and rotated, whereas rotation at each level was measured with extensiometers. Rotation at each level was calculated as a percent of total rotation. The spines were placed in a CT scanner and imaged after rotation of the spine in each direction. The percent of total rotation that occurred at each level was calculated with a software program. Accuracy of the CT method was calculated as the average difference between methods. Precision was measured as the standard deviation of the CT measurement. Biomechanical testing and CT were repeated after the posterior anulus fibrosus at L3-L4 was incised with a scalpel. The power of the CT method to detect a change in rotation was tested by calculating the difference between the pre- and postsurgery rotation at L3-L4 and testing it for significance with a Student t test of paired samples.

Differences between CT and biomechanical measurements averaged 0.2%. Precision was 6.0% (Table 1). Postanular injury, percent rotation at the L3-L4 level increased, whereas it decreased at the other 4 levels. The change at L3-L4 was statistically significant (P = 0.047).

Dynamic CT measures vertebral rotations sufficiently accurately to study the effect of a radial tear on axial rotation can.