The effects of lead shielding of background organ activity in pinhole Single Photon Emission Computed Tomography (SPECT) breast tumor imaging

The objective of this study was to investigate the effect on contrast and SNR (Signal-to-Noise Ratio) of shielding the patient bed with lead. Shielding reduces other-organ photon contamination, but may result in incomplete sampling of other organs and increased data inconsistency, if shielding attenuation is not modeled during reconstruction. A Trionix XLT scanner with one pinhole collimator was used to obtain experimental data on a breast phantom mounted on a torso phantom. The torso phantom contained the liver, lungs, and myocardium, all filled with varying amounts of ⁹⁹Tc^m-MIBI. The breast phantom contained two spherical lesions, also filled with ⁹⁹Tc^m-MIBI. The bed contained a cut-out for the prone breast attachment. Ten scans were performed with bed shielding, and without, in order to determine the statistical significance of the analysis. The incomplete circular orbit angular range was 186 degrees and consisted of 94 views with a 2 degree step between each view. The data were then reconstructed by OSEM with 8 subsets, using up to 30 iterations without modeling shielding attenuation. A 3D region-of-interest analysis was performed on a lesion to evaluate shielding effects. Post-filtering was applied for better image quality. The use of shielding slightly increased noise in contrast estimates, from about 0.25 to 0.30. In addition, shielding increased bias in contrast from 0.1+/-0.08 to 0.5+/-0.09. For a given cutoff frequency, SNR is lower with shielding than without. These results suggest that if shielding is to be of benefit, shielding attenuation must be modeled within iterative reconstruction.