High definition PET for cardiac imaging: Preliminary results

The new Siemens HD.PET is characterized by improved spatial resolution and signal-to-noise on reconstructed PET images. We have tested HD.PET for cardiac imaging. Materials and Methods: We filled the cardiac insert of a RSD Alderson Thorax phantom with ¹⁸F-FDG and acquired it with and without defects. The data were reconstructed with filtered back projection (FBP), 2D and 3D Attenuation Weighted Ordered Subsets Expectation Maximization (AWOSEM), and HD.PET. We calculated the wall/cavity contrast and contrast-to-noise (CNR). Two patients referred for myocardial viability assessment were also acquired. Static and 8-bin ECG-gated images were generated using FBP, 2D and 3D AWOSEM, and HD.PET. Results: In the phantom, the contrast increase with HD.PET ranged from 12.0% to 47.9% as compared to FBP, 2D and 3D AWOSEM. The HD.PET CNR with a 2-mm Gauss filter (73.8) was also increased compared to conventional reconstructions (17.0 for FBP, 37.3 for 2D and 38.5 for 3D-AWOSEM). We found that because of the improved spatial resolution, the estimated defect size increase for HD.PET ranged from 10.0% to 33.7%. On the static images of patient #1’s study, a myocardial ROI on HD.PET images showed a contrast increase of 3.6% with FBP, 16.3% with 2D AWOSEM and 11.8% with 3D AWOSEM. The static CNR was 27.1 for HD.PET, 6.4 for FBP, 16.2 and 17.9 for 2D and 3D AWOSEM. For patient #2, the HD.PET contrast (CNR) was 8.54 (35.0) while it was 5.59 (11.2) for FBP, 6.26 (33.2) for 2D-AWOSEM and 6.82 (30.0) for 3D-AWOSEM. For both patient #1 and #2’s gated images, the myocardial ROI showed a significant contrast and CNR increase with HD.PET as compared to the 3 standard reconstruction methods (all p≪0.05 except for patient #1’s FBP images). Conclusion: Our preliminary work on both phantom and patient ¹⁸F-FDG cardiac studies showed that HD.PET significantly improved defect definition, image resolution, contrast and contrast-to-noise.