GPU-based Fast Cone Beam CT Reconstruction from Undersampled and Noisy Projection Data via Total Variation - Physics > Medical PhysicsReport as inadecuate




GPU-based Fast Cone Beam CT Reconstruction from Undersampled and Noisy Projection Data via Total Variation - Physics > Medical Physics - Download this document for free, or read online. Document in PDF available to download.

Abstract: Purpose: Cone-beam CT CBCT plays an important role in image guidedradiation therapy IGRT. However, the large radiation dose from serial CBCTscans in most IGRT procedures raises a clinical concern, especially forpediatric patients who are essentially excluded from receiving IGRT for thisreason. The goal of this work is to develop a fast GPU-based algorithm toreconstruct CBCT from undersampled and noisy projection data so as to lower theimaging dose. Methods: The CBCT is reconstructed by minimizing an energyfunctional consisting of a data fidelity term and a total variationregularization term. We developed a GPU-friendly version of theforward-backward splitting algorithm to solve this model. A multi-gridtechnique is also employed. Results: It is found that 20~40 x-ray projectionsare sufficient to reconstruct images with satisfactory quality for IGRT. Thereconstruction time ranges from 77 to 130 sec on a NVIDIA Tesla C1060 GPU card,depending on the number of projections used, which is estimated about 100 timesfaster than similar iterative reconstruction approaches. Moreover, phantomstudies indicate that our algorithm enables the CBCT to be reconstructed undera scanning protocol with as low as 0.1 mAs-projection. Comparing with currentlywidely used full-fan head and neck scanning protocol of ~360 projections with0.4 mAs-projection, it is estimated that an overall 36~72 times dose reductionhas been achieved in our fast CBCT reconstruction algorithm. Conclusions: Thiswork indicates that the developed GPU-based CBCT reconstruction algorithm iscapable of lowering imaging dose considerably. The high computation efficiencyin this algorithm makes the iterative CBCT reconstruction approach applicablein real clinical environments.



Author: Xun Jia, Yifei Lou, Ruijiang Li, William Y. Song, Steve B. Jiang

Source: https://arxiv.org/



DOWNLOAD PDF




Related documents