Title :
Optimal camera placement for cardiac imaging using rotating multi-segment slant-hole single photon emission computed tomography
Author :
Wang, Yuchuan ; Tsui, Benjamin M W ; Baird, William H.
Author_Institution :
Dept. of Radiol., Johns Hopkins Med. Institutions, Baltimore, MD, USA
Abstract :
Rotating multi-segment slant-hole (RMSSH) collimators provide much higher detection efficiency and allow fully 3D cardiac SPECT using a much lower number of camera positions as compared to conventional low-energy high-resolution parallel-hole collimators. In this study, the effects of number and placement of camera positions in RMSSH SPECT for cardiac imaging were studied, to achieve the best possible reconstruction image quality with minimal image artifacts. Both simulation and phantom experiments were performed. A modified SIMIND Monte-Carlo code and the realistic 3D NURBS-based Cardiac-Torso (NCAT) phantom were used in simulation studies. A GE 400AC/T camera equipped with a 4-segment slant-hole collimator with a 32.5° slant-angle and a Data Spectrum anthropomorphic torso phantom with cardiac insert were used in phantom experiments. Ectomography (single camera position) and RMSSH (two to three camera positions) projection data of the heart were simulated and acquired from different camera positions, and then reconstructed using an iterative OS-EM algorithm for comparison. For incomplete angle RMSSH SPECT of the heart, Ectomography gives noticeable image artifacts and distortions at any camera position. RMSSH SPECT images obtained using two camera positions 90° apart show significant improvements as compared with only one camera position used in Ectomography, and those placed symmetrically with respect to the long axis of the heart give the best image quality. For complete angle RMSSH SPECT acquisition of the heart using three camera positions, all combinations of camera positions that satisfy Orlov´s condition show similar images that are of higher quality and quantitative accuracy than that from incomplete angle acquisitions. We conclude that the number and placement of camera positions in cardiac RMSSH SPECT should be chosen carefully for best reconstructed image and quantitative accuracy.
Keywords :
Monte Carlo methods; cardiology; diagnostic radiography; image reconstruction; medical computing; medical image processing; optimisation; single photon emission computed tomography; 3D NURBS-based Cardiac-Torso phantom; Orlov condition; SIMIND Monte-Carlo code; SPECT; acquisition; camera positions; cardiac imaging; ectomography; heart; image quality; optimal camera placement; parallel-hole collimators; reconstructed image; rotating multisegment slant-hole single photon emission computed tomography; Anthropomorphism; Cameras; Collimators; Heart; High-resolution imaging; Image quality; Image reconstruction; Imaging phantoms; Single photon emission computed tomography; Torso;
Conference_Titel :
Nuclear Science Symposium Conference Record, 2002 IEEE
Print_ISBN :
0-7803-7636-6
DOI :
10.1109/NSSMIC.2002.1239436