Abstract
2016
Objectives Design a multi-segment collimator for a stationary cardiac SPECT scan, using two detectors 90°apart in an L-shape configuration.
Methods We used 14 and 3 view-angles in the transaxial and axial directions, respectively. Each detector was segmented into 7 regions. We stayed with the parallel-beam geometry and used segmented parallel-hole collimator to acquire data. The torso background is truncated, thus this system is more ill-conditioned. To improve the condition, we measured more rays within the FOV of the detector by using a smaller detector bin size. The MLEM algorithm was used to reconstruct the image.We used the summed pixel-to-pixel distance to measure the discrepancy between the standard image and the reconstructed 3D ROI. Effects of limited view-angles, data truncation, etc. were considered.
Results View-angles are sufficient and no image distortion is observed using an NCAT phantom. Our system has a 12-fold sensitivity gain over a conventional single-head system. For background data truncation issue, smaller detector bin-size makes improvement to this ill-condition. Not many changes in ROI reconstruction are observed if the patient body size varies. The activities outside of the common imaging volume do not cause any structural artifacts in the ROI, but cause some image bias within the ROI. The noise affects the image quality and further investigation is needed. Attenuation and blurring effects are the same as rotational system.
Conclusions Our system measures sufficient data for cardiac imaging and has a higher sensitivity gain. Patient positioning is much easier for this system.
Research Support Siemens Medical System
Image comparison
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.