Evaluation of localization uncertainty of fiducial markers due to length and position variations induced by motion in CT imaging by measurement and modeling.

PURPOSE: To quantify the variations in the length and position of fiducial markers induced by motion in axial (ACT), helical (HCT) and cone-beam CT (CBCT) imaging and associated uncertainty in image-guided radiotherapy (IGRT) by measurement and modeling.

METHODS: A mobile thorax phantom containing markers of various lengths was imaged using ACT, HCT and CBCT imaging. The phantom was imaged while stationary and moving where it was moved sinusoidally with different motion amplitudes and frequency. An analytical motion model was developed that predicts the localization accuracy of IGRT based on fiducial markers in mobile phantom with ACT, HCT and CBCT.

RESULTS: The apparent lengths of the markers varied with the different motion patterns and CT imaging modalities. In CBCT, the apparent length of the markers increased linearly with the motion amplitude for both half-fan and full-fan modes. In HCT and ACT, the apparent length of the markers increased or decreased non-linearly with motion parameters and speed of the imaging couch. When the marker moved opposed to couch motion the apparent lengths decreased, while they increased when the phantom moved along the direction of the imaging couch as predicted by the motion model. The position of marker centers did not shift and distance between makers did not change in CBCT images. However, in HCT and ACT, the position of marker center and distance between markers varied depending on motion parameters during imaging. The marker center could move superiorly or inferiorly and the distance between markers could increase or decrease depending on the phase of motion as predicted by the motion model.

CONCLUSIONS: The variations of marker length and position due to phantom motion were quantified by measurement and modeling. These variations may lead to large positioning uncertainties in patient setup and tumor localization based on IGRT with fiducial marker registration.