Effect of patient motion in a clinical comparison of dual-cone-beam and fan-beam sequential transmission imaging for cardiac SPECT

The objectives of this study were to identify the motions which occurred clinically during the entire imaging sequence, model the motions under controlled conditions using a mathematical phantom, and compare the resulting impact with clinically observed motion. Furthermore, we also endeavored to identify the causes for the discrepancies between the two reconstruction strategies using a patient monitoring system. With institutional approval and written consent, 106 patients underwent dual-cone-beam transmission imaging followed by emission and fan-beam transmission imaging during the rest phase of their cardiac perfusion study. Thirty-six of these also were monitored for motion using a motion-tracking system with 5 near infra-red cameras tracking reflective markers on belts wrapped around the patients. Circumferential profiles were generated from emission short-axis slices for quantitative comparison of the 2 methods of transmission imaging. The mathematical cardiac torso phantom (MCAT) was also used to generate transmission and emission maps with respiratory motion included. Projections were generated after which the transmission map was shifted to simulate the mismatch observed in patient transmission maps. Four different patterns of motion were simulated. Processing was similar to the patient data. The phantom results confirmed our suspicion that a significant portion of the apparent differences in localization we observed in the LV lateral wall of the emission slices attenuation corrected with either the fan or dual-cone beam based attenuation maps are due to positional mismatch between the two transmission maps. This study highlights the presence of patient motion in sequential imaging and hence the potential utility of patient motion-tracking during dual-modality imaging such as SPECT/CT and PET/CT.