Author: Craig K Abbey, John M. Boone, Richard E. Colbeth, Andrew M. Hernandez, Vance Robinson, Paul Schwoebel, Jeffrey H. Siewerdsen, Alejandro Sisniega, Wojciech B. Zbijewski, Huanyi Zhou 👨🔬
Affiliation: UC Santa Barbara, University of California, UT MD Anderson Cancer Center, Johns Hopkins University, University of New Mexico Albuquerque, UC Davis Health, Varex Imaging Corporation 🌍
Purpose: A new CT configuration utilizing a multiple X-ray-source array (MXA) is under development to mitigate cone-beam (CB) artifacts in single-rotation wide coverage imaging. We investigate the sampling, resolution, and noise properties of MXA CT as a function of source-axis distance (SAD) and location in the volumetric Field-of-View (FOV).
Methods: MXA CT pairs six individually-pulsed X-ray sources (4.5o cone-angle each) distributed on 4.3 cm centers parallel to the axis of gantry rotation in a single X-ray tube with a 40x30 cm detector having 0.776 mm pixels. The source-to-detector distance is 1000 mm. The system alternates between simultaneous exposures of odd and even sets of three sources for a total of 720 views/360o. Because the angular and CB sampling across the FOV depends on the SAD, we compare SAD=650 mm, where the entire volume is covered by a full rotation of at least one source with SAD=550 mm, where the regions on the interface of two neighboring source beams are covered by only partial rotations of these sources, providing possible skin dose reduction. Evaluations involve simulations of the Corgi phantom. Images are reconstructed using Penalized Weighted Least Squares with low regularization strength to emphasize effects of data sampling.
Results: Relative to SAD=650 mm, with SAD=550 mm and an 8-cm radial distance from the isocenter, there is a 40% reduction in the angular range and number of views contributing to the reconstruction at a midpoint between neighboring sources, and an increase in the cone angle of the contributing rays. Despite less geometric sampling at SAD=550 mm the edge spread of spherical stimuli across the FOV is comparable for both configurations.
Conclusion: MXA-CT with SAD=550 cm provides up to 20 cm axial FOV without any apparent distortions despite the short-scan angular coverage with oblique rays at the interface between neighboring source beams.