Target localization accuracy in a respiratory phantom using BrainLab ExacTrac and 4DCT imaging

Abstract

This study evaluated the accuracy of measuring the motion of an internal target using four-dimensional computed tomography (4DCT) scanning and the BrainLAB ExacTrac X-ray imaging system. Displacements of a metal coil implanted in a commercial respiratory phantom were measured in each system and compared to the known motion. A commercial respiratory motion phantom containing a metal coil as a surrogate target was used. Phantom longitudinal motions were sinusoidal with a 4.0 second period and amplitudes ranging from 5-25 mm. We acquired 4DCT and ExacTrac images of the coil at specified respiratory phases and recorded the coordinates of the coil ends. Coil displacement relative to the 0% phase (full-inhale) position were computed for the ExacTrac and 4DCT imaging systems. Coil displacements were compared to known displacements based on the phantom's sinusoidal motion. Coil length distortion due to 4DCT phase binning was compared to the known physical length of the coil (31 mm). The maximum localization error for both coil endpoints for all motion settings was 3.5 mm for the 4DCT and 0.8 mm for the ExacTrac gating system. Coil length errors measured on the 4DCT were less than 0.8 mm at end inhale/exhale phases, but up to 8.3 mm at mid-inhalation phases at the largest motion amplitude (25 mm). Due to the fast image acquisition time (100 ms), no coil distortion was observable in the ExacTrac system. 4DCT showed problems imaging the coil during mid-respiratory phases of higher velocity (phases 20%-30% and 70%-80%) due to distortion caused by residual motion within the 4DCT phase bin. The ExacTrac imaging system was able to accurately localize the coil in the respiratory phantom over all phases of respiration. For our clinic, where end-respiration phases from 4DCT may be used for treatment planning calculations, the ExacTrac system is used to measure internal target motion. With the ExacTrac system, planning target size and motion uncertainties are minimized, potentially reducing internal target volume margins in gated radiotherapy.