Room: Exhibit Hall | Forum 5
Purpose: Scatter contamination in conebeam CT leads to CT number inaccuracy, image contrast loss and spatial nonuniformity. In our previous work, we propose an effective scatter correction algorithm using stationary beam blocker which directly acquires the scatter samples in a single scan. It is promising due to its simplicity and practicability. Although demonstrated effectively on tabletop system, our previous design fails to achieve good quality image on clinical CBCT system mainly due to gantry wobbling, empirical uniform blocker design, the inserted blocker scatter and simple merging reconstruction algorithm. In this work, we investigate the motion behavior of the beam blocker and design an optimized beam blocker to improve the image quality.
Methods: In this study, we estimate the motion of the beam blocker in each projection using segmentation due to its high contrast in the blocked area. Since the gantry wobbling poses a big challenge to single-scan reconstruction with beam blocker, a blocker distribution objective function which accounts for the data insufficiency is proposed and minimized using generalized pattern search method. The scatter signals induced by the blocker are also estimated using object-free projection with the inserted blocker. The final image is generated using forward projection to compensate for the missing data.
Results: The proposed method is evaluated using clinical head patient on Varian Trilogy OBI system. The CT number error is reduced from 97 HU to 6 HU in the soft-tissue region and image spatial non-uniformity is decreased from 27% to 5% after correction.
Conclusion: We fully explore the potential of the beam blocker method for scatter correction on clinical CBCT system. The proposed blocker design quantifies the image quality and allows for more flexible geometry of the beam blocker in different clinical CBCT system. High performance in patient data processing indicates the clinical applicability of the new method.
Funding Support, Disclosures, and Conflict of Interest: This work is supported by National Key Research and Develop Program of China (2016YFC0105102), the National Natural Science Foundation of China (61471349), zhejiang Provincial Natural Science Foundation of China (Grant No. LR16F010001), National High-tech R&D Program for Young Scientists by Ministry of Science and Technology of China (Grant No. 2015AA020917).