Aortic valve imaging using 18F-sodium fluoride: impact of triple motion correction

Figures

• 

  Fig. 1. Overall scheme for simultaneous cardiorespiratory and gross patient motion detection and correction of 18F-NaF aortic valve PET scans. A fixed number of respiratory and electrocardiogram (ECG) gates (4 each) were used. The number of patient repositioning gates depended on the number of repositioning events the patient had during the acquisition. A 3D mesh of gated reconstructions were obtained, which were registered to generate the 3 × MC image set. 3 × MC = triple motion-corrected

• 

  Fig. 2. Segmentation of the aortic valve used for quantification of 18F-NaF uptake (shown as a green polyhedron). The valve was segmented on the end-diastolic image enclosing on the SUVmax observed in the image to ensure high test–retest repeatability. Top row shows the coronal view of the myocardium, while the lower row represents the short-axis view of the aortic valve. ECG-MC = ECG motion-corrected, 3 × MC = triple motion-corrected

• 

  Fig. 3. SUVmax and TBRmax measurements obtained for the three reconstructions. Significantly increased SUVmax and TBRmax values were observed for the 3 × MC datasets when compared to standard and ECG-MC datasets (Pitman–Morgan analyses, all p < 0.05). Standard = end-diastolic imaging, ECG-MC = ECG motion-corrected, 3 × MC = triple motion-corrected. SUVmax = maximum standard uptake value, TBRmax = maximum target-to-background ratio

• 

  Fig. 4. Correlation plots of the aortic valve calcium score and SUVmax and TBRmax. Significantly improved correlations were observed for the 3 × MC when compared to the other reconstructions for both SUVmax (A) and TBRmax (B). Of note, the repeated measurements were averaged for the plots. Standard = end-diastolic imaging, ECG-MC = ECG motion-corrected, 3 × MC = triple motion-corrected. SUVmax = maximum standard uptake value, TBRmax = maximum target-to-background ratio

• 

  Fig. 5. Signal-to-noise ratio (SNR) measurements obtained for the three reconstructions. End-diastolic (Standard) measurements were reported to have significantly lower SNR when compared to the motion-corrected datasets (ECG-MC and 3 × MC). ECG-MC = cardiac contraction motion-corrected, 3 × MC = triple motion-corrected

• 

  Fig. 6. Test–retest repeatability of SUVmax and TBRmax for standard (end-diastolic), ECG-MC, and 3 × MC reconstruction protocols. RC = repeatability coefficient, TBRmax = maximum target-to-background ratio, ECG-MC = cardiac contraction motion-corrected, 3 × MC = triple motion-corrected

• 

  Fig. 7. Test–retest valve PET imaging for a patient before and after motion and background blood pool corrections. Standard and ECG-MC images present less specific uptake patterns at the areas with calcification, while 3 × MC images show reproducible uptake patterns for the two scans. ECG-MC = cardiac contraction motion-corrected, 3 × MC = triple motion-corrected, SUVmax = maximum standardized uptake value, TBRmax = maximum target-to-background ratio, SNR = signal-to-noise ratio

• 

  Fig. 8. Test–retest valve PET imaging for a patient before and after motion and background blood pool corrections. Standard and ECG-MC images present lower TBRmax, while 3 × MC images show higher TBRmax (68% and 30% increase in scan 1 and scan 2, respectively). ECG-MC = cardiac contraction motion-corrected, 3 × MC = triple motion-corrected, SUVmax = maximum standardized uptake value, TBRmax = maximum target-to-background ratio, SNR = signal-to-noise ratio