Effect of Misalignment Between Transmission and Emission Scans on SPECT Images
Kenya Murase, Shuji Tanada, Takeshi Inoue, Yoshifumi Sugawara, Ken Hamamoto
Abstract
Kenya Murase, Shuji Tanada, Takeshi Inoue, Yoshifumi Sugawara, Ken Hamamoto
Abstract
Attenuation coefficient maps (μ maps) are a useful way to compensate for nonuniform attenuation when performing single-photon emission computed tomography (SPECT). We have previously reported that brain SPECT images could be improved by using a μ map constructed with transmission data and the iterative expectation maximization (EM) algorithm. However, if transmission and emission data are acquired separately, patient movement between the transmission and emission scans may cause misalignment of the μ map and the emission data. The effect of such misalignment on SPECT images reconstructed using the EM algorithm was investigated using both a thorax phantom, incorporating nonuniform attenuation, and clinical brain SPECT data. We found that misalignment between transmission and emission scans produced large errors, with a 2-pixel shift of the μ map causing an error of about 13% in the reconstructed activity level. These findings suggest that precise alignment of the μ map and the emission data is needed, and that simultaneous transmission and emission data acquisition should be strongly recommended to achieve this.
OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Attenuation coefficient maps (μ maps) are a useful way to compensate for nonuniform attenuation when performing single-photon emission computed tomography (SPECT). We have previously reported that brain SPECT images could be improved by using a μ map constructed with transmission data and the iterative expectation maximization (EM) algorithm. However, if transmission and emission data are acquired separately, patient movement between the transmission and emission scans may cause misalignment of the μ map and the emission data. The effect of such misalignment on SPECT images reconstructed using the EM algorithm was investigated using both a thorax phantom, incorporating nonuniform attenuation, and clinical brain SPECT data. We found that misalignment between transmission and emission scans produced large errors, with a 2-pixel shift of the μ map causing an error of about 13% in the reconstructed activity level. These findings suggest that precise alignment of the μ map and the emission data is needed, and that simultaneous transmission and emission data acquisition should be strongly recommended to achieve this.
Key concepts: Single-photon emission computed tomography, Correction for attenuation, Imaging phantom, Attenuation, Emission computed tomography, Transmission (telecommunications), Physics, Iterative reconstruction