Wojciech Zbijewski PhD, Presenter: Research grant, Carestream Health, Inc

Andreu Badal-Soler, Abstract Co-Author: Nothing to Disclose

Iacovos Kyprianou PhD, Abstract Co-Author: Nothing to Disclose

Joseph Webster Stayman PhD, Abstract Co-Author: Research grant, Varian Medical Systems, Inc

John A. Carrino MD, MPH, Abstract Co-Author: Research grant, Siemens AG Research grant, Carestream Health, Inc Research Consultant, General Electric Company

Jeffrey H. Siewerdsen PhD, Abstract Co-Author: Research grant, Carestream Health, Inc Research grant, Siemens AG Medical Advisory Board, Carestream Health, Inc License agreement, Elekta AB

New cone-beam CT (CBCT) systems (e.g., extremity and ENT imaging) involve a compact geometry with elevated x-ray scatter compared to other CBCT embodiments (e.g., C-arms) and more challenging scatter correction. We investigate the complex scatter distributions of compact-geometry CBCT via Monte Carlo calculations to elucidate the factors governing scatter magnitude, high-frequency heterogeneity, the performance of antiscatter grids, and strategies for scatter correction.

Scatter was estimated using a GPU-accelerated Monte Carlo simulator (MC-GPU) with a probabilistic x-ray spectrum model (60–120 kVp) and an analytical model of detector response (Varian 3030 FPD with CsI scintillator). A focused grid (10:1 GR, 200 lpi) was included, and two geometries were compared: compact CBCT for extremities (the knee, 434 mm SAD, 554 mm SDD) and a C-arm for intraoperative CBCT (595 mm SAD, 1180 mm SDD). Voxelized models of the knee, head, and other sites were used (1 mm voxels, ~200x200x400 mm3 volume). Simulations involved 1010 photon histories / projection.

Scatter-to-primary ratio (SPR) was significantly increased for compact CBCT compared to C-arm geometry due to the short air gap: for the knee at 90 kVp, mean SPR = 0.6 vs. 0.16, respectively. Compact CBCT exhibited strong kVp dependence (mean SPR increased to 1.2 at 60 kVp) and major benefit from the grid (mean SPR reduced to 0.17), whereas the grid was of marginal benefit to the C-arm. Use of a grid increased the ratio of coherent to incoherent scatter from 0.4 to 1.1 and increased the spatial variation in coherent and total scatter. Simulations revealed new design considerations (e.g., more uniform scatter with the grid oriented along the axis of the knee) and form the basis of accurate, patient-specific scatter correction strategies.

Compact CBCT systems exhibit significantly elevated scatter and non-uniform scatter distributions compared to extended geometries. Antiscatter grids are important in reducing SPR but further increase scatter non-uniformity depending on object and grid orientation. Such effects defy simple corrections that assume smooth distributions and motivate more sophisticated correction strategies.

Compact CBCT systems are increasingly prevalent but involve scatter characteristics that confound image quality. Monte Carlo studies reveal complexities and form the basis of accurate correction.

Zbijewski, W, Badal-Soler, A, Kyprianou, I, Stayman, J, Carrino, J, Siewerdsen, J, Scatter Properties of Compact Geometry Cone-Beam CT Systems.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11016607.html