This research plan proposes a synchronized-moving-grid (SMOG) system to improve the image quality of kilovoltage cone-beam computed tomography (kV CBCT). CBCT has the advantages over the conventional fan-beam computed tomography (CT) for its ability to acquire multiple slices of images in one rotation, and the spatial resolution and convenience gained by using a flat panel detector (FPD). The introduction of the FPD based CBCT in IGRT has significantly improved the accuracy of radiation treatment and may revolutionize the radiation therapy procedure. CBCT also has potential important applications in image guided surgery and intervention, and in diagnostic radiology for fast imaging of moving organs. However, three major problems-scatter, image lag and gantry flex-significantly degrade the image quality of CBCT, thereby hampering its further applications. Many methods have been proposed to reduce the scatter effect. These methods can be divided into two groups: (1) direct scatter reduction and (2) scatter correction. However, the efficiency o the current direct scatter reduction methods is often limited, and although scatter correction methods can reduce scatter artifacts, they usually further degrade the contrast-to-noise ratio. Various methods proposed to reduce the lag and flex effects have also shown only limited improvements due to the complexity of the issues at hand. The SMOG system aims to address these heretofore unresolved problems. It takes multiple partial- projections at each gantry position during a scan, with the grid moving rapidly in an oscillating pattern synchronized with the gantry motion. A full projection is obtained by merging all partial projections at the same gantry position. This approach not only provides a solution to the scatter problem through simultaneous scatter reduction and correction, it also effectively addresses lag and flex as well, allowing for the simultaneous resolution of all three major problems facing CBCT in one simple device. In addition, the SMOG system includes a software platform comprised of Monte-Carlo (MC)-based simulation and a compressed sensing algorithm that further optimizes the image quality. Preliminary results have demonstrated the applicability and promise of these concepts. For the scatter problem, we used multiple-rotation scans with the grid shifting a distance after each rotation to simulate the SMOG system. Significant improvement in the image quality-as demonstrated by increased CNR and removal of scatter artifacts-has been achieved. A simulation study also shows that the lag in the image region can be corrected by using the lag in the neighboring shadow region blocked by the septa. We have also shown that the grid can be used to directly detect the gantry wobbling distribution in a scan. The SMOG system will be developed and tested in a CBCT test bench, and then transferred to a clinical CBCT system for clinical validation.
This study proposes a synchronized moving grid (SMOG) system to solve scatter, image lag and gantry flex, the three major problems in current clinical cone beam computed tomography (CBCT) with a same device. The SMOG system will vastly improve the image quality toward the second generation CBCT scanner for adaptive radiotherapy. We will build a prototype system and test it in the clinical setting.