Photons originating from planes adjacent to the slice of interest introduce considerable qualitative and quantitative error in single photon emission computerized tomography (SPECT) images. The extent of out-of-plane contributions was studied, and a method to compensate for these photons was examined. Simulations using accurate Monte Carlo models were performed to acquire projection data. Analysis of these projections showed that as much as 12% of the total counts corresponding to a 1.8-cm-thick slice of interest were due to out-of-plane nonscatter and that out-of-plane scatter could compromise up to 16% of a projection pixel's total count. Reconstructions were obtained by maximum-likelihood expectation-maximization methods, that accounted for the out-of-plane phenomena. These reconstructions gave an improvement in count quantification, noise, and contrast as compared to performing in-plane compensation alone. These findings required 3-D reconstructions with out-of-plane compensation in order to obtain reliable quantitative SPECT images. >
[1]
C E Floyd,et al.
Experimentally measured scatter fractions and energy spectra as a test of Monte Carlo simulations.
,
1987,
Physics in medicine and biology.
[2]
R. Jaszczak,et al.
Inverse Monte Carlo: A Unified Reconstruction Algorithm for SPECT
,
1985,
IEEE Transactions on Nuclear Science.
[3]
K. Lange,et al.
EM reconstruction algorithms for emission and transmission tomography.
,
1984,
Journal of computer assisted tomography.
[4]
L. Shepp,et al.
Maximum Likelihood Reconstruction for Emission Tomography
,
1983,
IEEE Transactions on Medical Imaging.
[5]
D. Rubin,et al.
Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper
,
1977
.