SU-E-J-132: Quantitative Assessment of CT Number Variation Induced by Respiratory Motion Artifacts in CT and...

CT number variations due to image artifacts induced by respiratory motion in CT and cone-beam CT (CBCT) imaging were investigated. A model was developed to predict CT number variations observed in CT and CBCT images.

Three target volumes of differing sizes were constructed of tissue-equivalent gel material and embedded into an artificial lung phantom. This mobile phantom mimicked respiratory motion with eight different amplitudes of respiratory motion in the range 0-20 mm at a frequency of 15 cycles/min. A model was developed allowing for simulation of the observed CT number variation as a function of the target volume, phantom motion frequency and amplitude and speed of CT scanning. The simulation results were compared quantitatively with CT profiles that were obtained from CT and CBCT imaging.

The size and distribution of CT numbers of well-defined homogenous targets varied with respiratory motion. In CBCT, CT number variations depends mainly on the target size and motion amplitude. Targets with a dimension smaller than the motion amplitude show highest CT number densities at the edges of the elongated volume. Conversely, targets larger than motion amplitude showed greatest CT number densities in the middle of the target. In CT imaging, CT number variations were more complicated and further depends on the speed of CT scanning where some targets were elongated or shrunk with varying CT number gradients depending on the above parameters.

The sizes and CT number distributions varied with motion for the homogenous targets. The CT number variations were modeled and simulation provided comparable distributions to those observed in CT and CBCT images based on target size, amplitude, frequency, and scanning speed. The results here might provide useful tools to correct image artifacts and consider CT number variation due to motion in treatment planning and dose delivery.