Andrew Primak PhD, Presenter: Research grant, Siemens AG

Rainer Raupach PhD, Abstract Co-Author: Employee, Siemens AG, Forchheim, Germany

Lifeng Yu PhD, Abstract Co-Author: Research grant, Xoran Technologies, Inc

Thomas G. Flohr PhD, Abstract Co-Author: Employee, Siemens AG, Forchheim, Germany

Cynthia H. McCollough PhD, Abstract Co-Author: Research grant, Siemens AG Research grant, RTI Electronics AB Research grant, Bayer AG

Partial (180 + fan angle) scan reconstructions improve temporal resolution but increase CT number variations. This work evaluates a new approach to reduce partial scan artifacts.

An anthropomorphic cardiac phantom was positioned 7 cm off-isocenter and scanned using a dual-source cardiac perfusion mode (stationary table, x-ray on continuously for 85 s, simulated 60 bpm ECG). The central portion of the phantom was replaced with a water-filled tank containing a syringe with an iodine solution. Partial scan (83 ms) images were reconstructed at 5 different phases of the cardiac cycle: center phase (C), C±42 ms, and C±125 ms. Full scan (330 ms) images at the center phase were also reconstructed. Every partial scan image was decomposed into multiple spatial frequency bands (FBs) and a temporal convolution performed separately for each FB using a band-specific filter width. The temporal filter was broad (narrow) in the low (high) FBs in order to preserve temporal resolution at high spatial frequencies while minimizing temporal variations in CT number at low spatial frequencies. Thus, 360 degree symmetry could be achieved for low spatial frequencies. The final multi-band filtered (MBF) images were derived by recombining the FBs. The mean CT number variations within a circular ROI inside the syringe were recorded vs. time for all 3 datasets (full, partial and MBF-processed partial scan images). In addition, previously acquired in-vivo animal myocardial perfusion data were retrospectively processed using the MBF approach.

Partial scan images of the phantom data displayed a periodic CT number variation of up to 21.7 HU at a time interval consistent with the time it takes for the partial scan source trajectory to make one full rotation. MBF processing of the same data reduced the maximum CT number variation to 6.9 HU. The full scan images had a maximum CT number variation of 3.2 HU. The visual quality of the animal myocardial perfusion curve was dramatically improved after the MBF processing.

This new approach uses multi-band spatio-temporal filtering of partial scan image data to reduce the magnitude of partial scan related CT number variations by more than a factor of 3.

Artifactual variations in CT number due to the use of partial scan reconstructions compromise quantitative applications of cardiac MDCT, but can be significantly reduced using MFB filtration.

Primak, A, Raupach, R, Yu, L, Flohr, T, McCollough, C, A Novel Approach to Reduce Partial Scan Artifacts in Cardiac Multidetector Row CT (MDCT).  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6017037.html