Investigation of Megavoltage Digital Tomosynthesis using a Co–60 Source
Digital Tomosynthesis , Cobalt-60 imaging
The ability for megavoltage computed tomography patient setup verification using a cobalt-60 (Co–60) gamma ray source has been established in the context of cobalt tomotherapy. However, it would be beneficial to establish improved cobalt imaging that could be used on more conventional units. In terms of safety and efficiency, this imaging technique would provide the patient with less exposure to radiation. Digital tomosynthesis (DT) is an imaging modality that may provide improved depth localization and in-plane visibility compared to conventional portal imaging in modern Co–60 radiation therapy. DT is a practical and efficient method of achieving depth localization from a limited gantry rotation and a limited number of projections. In DT, each plane of the imaging volume can be brought into focus by relatively displacing the composite images and superimposing the shifted dataset according to the acquisition geometry. Digital flat-panel technology has replaced the need for multiple film exposures and therefore the speed of imaging and capabilities for image processing has put DT in the forefront of both clinical and industrial imaging applications. The objective of this work is to develop and evaluate the performance of an experimental system for megavoltage digital tomosynthesis (MVDT) imaging using a Co–60 gamma ray source. Linear and isocentric acquisition geometries are implemented using tomographic angles of 20-60° and 10-60 projections. Reconstruction algorithms are designed for both acquisition geometries. Using the backprojection approach, the data are shifted and added to reconstruct focal planes of interest. Depth localization and its dependence on tomographic angle and projection density are visualized with an anthropomorphic head phantom. High contrast resolution at localized depths is quantified using the modulation transfer function approach. Results show that focal-plane visibility is improved for larger tomographic angles and that focal-plane visibility has negligible dependence on projection density. Lastly, the presence of noise and artifacts in the resulting images are quantified in terms of the signal-to-noise ratio and the artifact spread function. The work presented here is expected to provide the justification required to proceed with a prototype Co–60 MVDT system for patient set-up verification in modern Co–60 radiation therapy.