Room: AAPM ePoster Library
Purpose: Cone Beam Computed Tomography (CBCT) is used widely in medicine, particularly to verify patient anatomy in radiotherapy. Conventional systems acquire x-ray projections at evenly spaced angles around an isocenter. This work investigates whether alternative angular spacings yield superior image quality, allowing more precise treatment at no additional cost.
Methods: CBCT scans consisting of 50, 100, 150 and 200 projections over 200- and 360-degree arcs were simulated from synthetic thorax and cranial phantoms. Sets of projection angles were calculated such that they would be evenly spaced on the circumference of ellipses of varying major/minor axis ratios e.g. standard even angular separation would be ratio 1. The corresponding projections were then reconstructed using the Feldkamp-Davis-Kress (FDK) and simultaneous algebraic reconstruction technique (SART) algorithms. Image quality was then quantified with root-mean-square error (RMSE) and signal-to-noise ratio (SNR).
Results: Conventional even angular projection spacing (ratio 1) produced suboptimal image quality across all phantoms, exposure quantities and reconstruction methods tested. For the thorax phantom, the mean ratio that produced the best quality was 1.81 (range 1.2-2.5), varying between reconstruction method, number of projections, and arc length. Optimized elliptical spacing improved RMSE by 0.843% (range 0.185%-1.760%) and SNR by 0.508% (range 0.123%-1.02%). For the cranial phantom, the mean optimal ratio was 0.68 (range 0.4-0.9) and improved RMSE 0.525% (range 0.00154%-2.01%) and SNR by 0.26125% (0.000673%-0.111%).
Conclusion: Even angular projection spacing does not provide the highest image quality for CBCT. We have found an alternative elliptical angular spacing that can be optimized for specific anatomical sites, producing improvements up to 2%. This motivates further work to more generally optimize projection location.