Seungeon Kim, Presenter: Nothing to Disclose

Yongjin Chang, Abstract Co-Author: Nothing to Disclose

Jong Beom Ra, Abstract Co-Author: Research Grant, Samsung Electronics Co Ltd Research Consultant, Samsung Electronics Co Ltd

Coronary artery imaging is important for early detection of cardiac disease. Since coronary arteries are small and move fast, high spatial and temporal resolution is required to get a diagnostic image quality. Due to the limited gantry rotation speed of X-ray CT, however, the reconstructed image usually contains motion artifact or blur. To improve the image quality via the motion correction in the reconstruction process, we use a novel motion estimation scheme based on partial angle reconstruction (PAR) images.

The algorithm aims to reconstruct a motion-artifact-reduced 3D cardiac CT image using projections obtained in a slightly larger angular range than the one needed for a short scan. In the algorithm, two conjugate PAR images are reconstructed from the projections on the small angular range, respectively. Using a pair of conjugate PAR images, we estimate a motion model. The motion correction is then performed by incorporating the estimated motion model into the image reconstruction process. The XCAT phantom and physical dynamic cardiac phantom are used for the feasibility test of the algorithm. The XCAT phantom dataset is generated with a heart rate of 70 bpm and a gantry rotation speed of 300 ms. Two physical dynamic cardiac phantom datasets are also generated by using a slowly rotating X-ray CT system so that the effective heart rate can become 70 and 85 bpm, respectively, if the system rotation speed is assumed to be 300 ms.

The PAR-based motion estimation and correction algorithm is applied to the phantom datasets. The reconstructed images at 20% (rapid motion) and 40% (quiescent motion) of R-R peak of the XCAT phantom show that motion artifact or blur can be significantly reduced by applying the motion correction algorithm; thereby coronary arteries are more clearly visible. Physical dynamic cardiac phantom images reconstructed at 10 phases with two different heart rates, 70 and 85 bpm, also provide the improved temporal resolution.

The PAR-based cardiac motion correction algorithm is proposed for 3D cardiac imaging of high temporal resolution. Its performance is verified by using a digital XCAT phantom dataset and two physical cardiac phantom datasets.

This work improves the accuracy of the cardiac disease diagnosis, by improving the temporal resolution, or reducing the motion blur, in cardiac X-ray CT imaging.

Kim, S, Chang, Y, Ra, J, Cardiac Motion Correction Based on Partial Angle Reconstruction in X-ray CT.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14013124.html