CT based motion correction of dual gated cardiac PET images

Objectives Dual gating is a method to divide the tomographic data into smaller bins according to the respiratory motion and ECG of the patient. It reduces the harmful motion artifacts but at the same time produces several images with a moderate quality. In this study a novel motion correction algorithm is introduced to combine the dual gated PET images into one motion corrected image with a high sensitivity.

Methods The dual-gated PET/CT phantom and patient data were acquired with Discovery VCT PET/CT scanner (GE Healthcare). The phantom data were collected using a realistic heart phantom with simulated respiratory and pulsatile motion including hot targets mimicking coronary plaques. Reconstruction software (Research gating toolbox) provided by GE Healthcare was used to reconstruct all the PET images. Motion correction of the dual gated PET images of the phantom was carried out by using gated low-dose CT images. From the respiratory gated CT data a model was created to correct respiratory motion and from the ECG gated CT data pulsatile motion from PET images. The efficiency of the motion correction methods was tested in three analyses: motion reduction, contrast to noise (CNR) and size of the target.

Results After the motion correction respiratory motion averaged over all hot targets decreased from 12.1 ± 1.5 mm to 6.5 ± 1.5 mm and pulsatile motion from 6.4 ± 0.8 mm to 3.7 ± 1.2 mm. After correcting the respiratory motion the CNR of myocardium uptake increased 50.5% after both corrections. Similar increase of CNR was found in hot spots. The size of hot spots on average decreased 49.7% after the double correction.

Conclusions The method to correct cardiac motion in PET images reduces motion artifacts, increases CNR and decreases the size of hot targets significantly. The feasibility of this method with the patient data will be tested soon.

Research Support The Finnish Funding Agency TEKE