Neuro PET imaging reimagined - Next-generation digital PET/CT neuroimaging

1616 Objectives: To examine the feasibility of improving PET neuroimaging by optimizing parameters such as radiopharmaceutical dose administered to the patient, acquisition time on the scanner, and image reconstruction parameters. Methods: 33 neurological PET/CT studies were acquired on a Philips Vereos digital photon counting PET system 40 minutes post-injection of 5 mCi 18F-FDG. PET listmode data were reconstructed into 6 5-minute frames using 3 iterations, 21 subsets, and both Point Spread Function correction and Gaussian filtering applied. Data were also reconstructed with a 50% and 80% reduced acquisition time using 12 frames of 2.5 minutes and 30 frames of 1 minute each, respectively. Reconstruction settings such as iterations and subsets were assessed for optimization of the reduced count density data sets. SUV was recorded for the putamen, thalamus, cerebellum, and frontal regions. Results: A 50% reduction in acquisition time, or equivalently injected radiotracer dose, was accomplished while maintaining image quality and robust quantification, on average only 1.3% different from reference to full data sets at each timepoint over the 30 minute acquisition. An optimized reconstruction protocol utilizing 3 iterations and 13 subsets was found to be most robust. At the level of an 80% reduction, visual image quality was maintained with a reconstruction utilizing 3 iterations and 5 subsets, but suffered slightly quantitatively, on average 4.9% less than reference. Additionally, uptake steady state did not occur at the current perceived >30 min post injection time point, but revealed continued uptake kinetics throughout the acquisition window until 70 min post-injection. Conclusions: Next-generation digital PET can be utilized to reduce radiotracer dose and/or acquisition time by 50% for neuro PET imaging without impacting visual or quantitative quality. Less radiation dose and/or shorter acquisition times can lead to increased patient safety and less motion artifacts, further improving the quality of such imaging data. Research Support: Ohio Third Frontier OSDA TECH 09-028 and 13-060 grants