Single-Scan Scatter Correction in Cone Beam CT Using Stationary Boundary Blockers and Compressed Sensing
B Meng*, L Xing, B Fahimian, Stanford Univ School of Medicine, Stanford, CA, US. H Lee, Yonsei University College of Medicine, Department of Radiation Oncology, Seoul, Korea.TU-A-213CD-5 Tuesday 8:00:00 AM - 9:55:00 AM Room: 213CD
Purpose: In this work, a novel scatter correction algorithm is proposed and implemented that couples the use of horizontal boundary stationary beam blockers with a compressed sensing scatter estimation technique to enable simultaneous acquisition of image and scatter information in a single scan.
Methods: Two how horizontal boundary beam blockers are designed to simultaneously acquiring the image data in the center unblocked region and the scatter data at the boundary region. Hybrid scatter estimation model that combines the scatter interpolation model and our proposed scatter convolution model is pre-computed as the initial scatter estimation for the central region. Based on the low frequency characteristic of scatter signal, compressed sensing optimization penalizes the L1 norm of the discrete cosine transform of scatter signal to reconstruct the scatter map. Estimated scatter is subtracted from the projection data by soft-thresholding, and the scatter-corrected image is obtained by conventional FDK reconstruction algorithm. The proposed method was evaluated with two phantom studies on a Varian TrueBeam STx on-board imaging system.
Results: The scatter shading artifacts were markedly suppressed in the reconstructed images using the proposed method. On the Catphan©504 phantom, the proposed method reduced the constructed error to 13 Hounsfield units, 10% of that without scatter correction, and increased the image contrast by a factor of two in high-contrast regions. On the Rando phantom, the spatial nonuniformity decreased from 10.8 % to 6.8% after correction. The proposed method also outperforms the scatter interpolation correction method in both studies.
Conclusions: A novel scatter correction method is proposed and validated in this work. The method enables image acquisition and scatter correction in a single scan, allows for unobstructed volumetric reconstruction, and avoids the additional requirements such as prior images, multiple scans, moving blockers or full-fan only mode.