ABSTRACT
Two-dimensional (2D) imaging, using either film or electronic portal imaging (EPI), has long been the gold standard for initial patient setup and weekly verification of patient positioning. Weaknesses inherent to 2D imaging for patient setup include the lack of information of tumor position relative to bony anatomy and out of plane rotation. A cone-beam computed tomography (CBCT) image is reconstructed from a number of 2D projections acquired on a flat-panel detector in small angular steps as the source-detector system rotates around the object. The principle of 3D reconstruction of tomographic images from a set of 2D projections is well established and has been described in the context of megavoltage (MV)-CBCT. Image quality is quantified by three main characteristics: uniformity, contrast-to-noise ratio (CNR), and spatial resolution. Spatial resolution represents the ability of a system to distinguish a small object of high contrast from its surrounding. CNR is the most important image quality characteristic in image-guided clinical applications.
