Comparison of transmission acquisition approaches for SPECT nonuniform attenuation compensation

This study compared two approaches for acquiring transmission computed tomography (TCT) data for SPECT nonuniform attenuation compensation. One approach, which has been implemented in commercial SPECT systems, acquires the TCT and SPECT data simultaneously using a scanning transmission line source, dual head SPECT system and parallel beam collimation (PB-sim). The other approach acquires the TCT and SPECT data sequentially using long focus, off-set fan beam collimation with a non-scanning line source and a triple head system (FB-seq). The two systems were compared based on: (a) the noise level of the TCT projection data, (b) the spatial resolution of the TCT projection data, and (c) the quality of reconstructed TCT and SPECT images of a thorax phantom. For the thorax phantom data a fast TCT scan (2 min.) was used and total scan time (TCT and SPECT) was the same for the two systems. The results from the TCT noise measurements showed that for the source activities used here (400 mCi for PB-sim, 56 mCi for FB-seq), PB-sim had higher count density in the lung region whereas FB-seq had higher in the heart and liver regions. The measured TCT spatial resolution for the two systems was comparable in the axial direction but was superior with FB-seq in the transaxial direction. The resolution difference was apparent in the reconstructed TCT images. These results suggest that the FB-seq system offers a viable approach for TCT acquisition and one that compares favorably with current commercial approaches based on TCT noise, resolution and reconstructed image quality.