Respiratory gating of cardiac PET data in list-mode acquisition

Purpose: Respiratory motion has been identified as a source of artefacts in most medical imaging modalities. This paper reports on respiratory gating as a means to eliminate motion-related inaccuracies in PET imaging. Methods: Respiratory gating was implemented in list mode with physiological signal recorded every millisecond together with the PET data. Respiration was monitored with an inductive respiration monitor using an elasticised belt around the patient's chest. Simultaneous ECG gating can be maintained independently by encoding ECG trigger signal into the list-mode data. Respiratory gating is performed in an off-line workstation with gating parameters defined retrospectively. The technique was applied on a preliminary set of patient data with (CO)-O-15. Results: Motion was visually observed in the cine displays of the sagittal and coronal views of the reconstructed respiratory gated images. Significant changes in the cranial-caudal position of the heart could be observed. The centroid of the cardiac blood pool showed an excursion of 4.5-16.5 mm (mean 8.5 +/- 4.8 mm) in the cranial-caudal direction, with more limited excursion of 1.1-7.0 mm (mean 2.5 +/- 2.2 mm) in the horizontal direction and 1.3-3.7 mm (mean 2.4 +/- 0.9 mm) in the vertical direction. Conclusion: These preliminary data show that the extent of motion involved in respiration is comparable to myocardial wall thickness, and respiratory gating may be considered in order to reduce this effect in the reconstructed images.