Influence of acquisition parameters on MV‐CBCT image quality

Abstract

The production of high quality pretreatment images plays an increasing role in image‐guided radiotherapy (IGRT) and adaptive radiation therapy (ART). Megavoltage cone‐beam computed tomography (MV‐CBCT) is the simplest solution of all the commercially available volumetric imaging systems for localization. It also suffers the most from relatively poor contrast due to the energy range of the imaging photons. Several avenues can be investigated to improve MV‐CBCT image quality while maintaining an acceptable patient exposure: beam generation, detector technology, reconstruction parameters, and acquisition parameters. This article presents a study of the effects of the acquisition scan length and number of projections of a Siemens Artiste MV‐CBCT system on image quality within the range provided by the manufacturer. It also discusses other aspects not related to image quality one should consider when selecting an acquisition protocol. Noise and uniformity were measured on the image of a cylindrical water phantom. Spatial resolution was measured using the same phantom half filled with water to provide a sharp water/air interface to derive the modulation transfer function (MTF). Contrast‐to‐noise ratio (CNR) was measured on a pelvis‐shaped phantom with four inserts of different electron densities relative to water (1.043, 1.117, 1.513, and 0.459). Uniformity was independent of acquisition protocol. Noise decreased from 1.96% to 1.64% when the total number of projections was increased from 100 to 600 for a total exposure of 13.5 MU. The CNR showed a∓5% dependence on the number of projections and 10% dependence on the scan length. However, these variations were not statistically significant. The spatial resolution was unaffected by the arc length or the sampling rate. Acquisition parameters have little to no effect on the image quality of the MV‐CBCT system within the range of parameters available on the system. Considerations other than image quality, such as memory storage, acquisition speed, and individual projection image quality, speak in favor of the use of a coarse sampling rate on the short scan.PACS numbers: 87.57.C‐; 87.57.nf