Objective: To quantify the 3-dimensional translation and rotation components of head motion during computed tomography and to analyze the influence of such motion on perceptible artifacts and distortion of volume image data sets.
Methods: Using high-precision optoelectronic motion-capture technology, changes in patient head position during axial CT scanning were registered in 20 cases and 2 phantoms with a spatial relative resolution better than 0.003 cm. Statistical analysis was performed on a base of 6-dimensional measurement-vectors, each with 3 translation and 3 rotation values. Because of the recording frequency of the tracking system, more than 80000 values were included in a statistical analysis.
Results: All 20 patients had head motion during the CT scanning, with only 4 of 20 patients showing perceptible motion artifacts. The frequency, the extent, and the direction of the movements did not correlate with either the observations made by the radiologic staff or with the patient's subjective estimation of comfort. Translation movements of the head during CT accounted for a maximum of 0.5 cm and rotations of more than 2 degrees without perceptible motion artifacts. The extent of positional changes of the head was found to correlate with the duration of scanning (Pearson's correlation coefficient: 0.647 for translation shifts, 0.453 for rotation shifts). The mean direction of head motion could be characterized predominantly as a rotation around the longitudinal axis of the body (xy plane) at a significance level of 0.01.
Conclusion: Computed tomography evaluations of the head performed without rigid fixation suffer a spatial distortion of the volume image data sets, caused by interimage motion. The absence of motion artifacts is not correlated with the absence of motion.