TY - JOUR T1 - A simplified brain-shaped phantom to evaluate O-15 image quality of digital photon counting PET-CT JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 329 LP - 329 VL - 61 IS - supplement 1 AU - Kenji Hirata AU - Keiichi Magota AU - NAOTO NUMATA AU - Michiaki Endo AU - Mao Kusuzaki AU - Daiki Shinyama AU - Ronee Asad AU - Kentaro Kobayashi AU - Tohru Shiga AU - Kohsuke Kudo Y1 - 2020/05/01 UR - http://jnm.snmjournals.org/content/61/supplement_1/329.abstract N2 - 329Background: O-15 labeled compounds, such as water, carbon dioxide and oxygen, are useful PET tracers for visualizing brain perfusion and metabolism. Although O-15 enables radiolabeling of physiological molecules without modifying their chemical characteristics, the shortcomings of O-15 include short half-life (2 minutes) and long positron range (maximum = 9.1 mm, mean = 3.0 mm), which is considerably longer than those of F-18 (maximum = 2.6 mm, mean = 0.7 mm). Positron range affects spatial resolution of the acquired images, and thus the image characteristics should be evaluated with a brain-shaped phantom filled with O-15 solution. However, due to short half-life of O-15, phantom experiments are difficult, especially in filling radioactive solution quickly into complicated cavities, such as Hoffman phantom or a recently published brain-shaped phantom (Iida et al., Ann Nucl Med 27:25-36,2013) with no air bubble left in the cavity. Therefore, we designed a simplified brain-shaped phantom to enable quick preparation, and performed an experiment using O-15 labeled water. To demonstrate the difference between O-15 and F-18, we used a state-of-the-art PET-CT scanner equipped with digital photon counting detectors that achieved excellent time-of-flight measurements. Methods: The newly designed phantom is shown in Fig A-D. The phantom consists of 3 parts: 1) brain slice, 2) common cavity, and 3) cylinder. The brain slice is 173 mm in AP, 148 mm in RL, and 25 mm in thickness, which was designed based on CT images of Iida’s phantom (above mentioned). The brain slice is connected to the common cavity (5mm in thickness). The cylinder is connected to the common cavity and used as the inlet hole for liquid (50 mm in diameter and 28 mm in length). A latest model of 3D printer and photo polymer resin (CT value = ~140 HU) were used to create the phantom. The cavity corresponds to gray matter, and the resin corresponds to white matter and CSF space. O-15 labeled water was produced with in-house cyclotron. 350 mL of radioactive solution of 4600 kBq/mL at PET acquisition start was filled in the cavity. Emission scanning was performed for 60 minutes in list mode. A total of 21 images (starting at 0, 2, 4,,, 40 minute, each duration = 1 minute) were reconstructed with voxel size being 1x1x1 mm (iteration 3, subsets 17). Point spread function correction was not applied to avoid possible adverse effects. In comparison, F-18 FDG (55 kBq/mL) was filled and scanned (10 minutes). Image-based FWHM was estimated by finding best gaussian filter applied to gold standard binary image. Results: It took 5 minutes from the beginning of preparation for O-15 water infusion to the initiation of PET scan. No air bubble was observed in the brain slices based on CT. Visually, 4600 to 600 kBq/mL images of brain slice were equivalently high-quality images, but 300 kBq/mL and lower-dose images were noisier (Fig. E). In the cylinder part, signal-to-noise ratio (SNR = mean/SD) was stable (~13) until 300 kBq/mL images and then decreased, while FWHM was stable (5.2 - 5.7 mm) until 2.6 kBq/mL image (Fig. F). F-18 had higher SNR (19) and narrower FWHM (4.7 mm). Next, brain slice was used to estimate FWHM, which was 6.6 to 6.8 mm for 4600 kBq/mL to 20 kBq/mL O-15 image, and 5.7 mm for F-18 image. Conclusions: We created a simplified brain-shaped phantom using a 3D printer to enable rapid filling of short-half-life tracer. With digital photon counting PET-CT, the spatial resolution was insensitive to low count rate, while signal-to-noise ratio degraded earlier. In terms of spatial resolution, O-15 image was slightly inferior to F-18 image, probably due to positron range. Such image characteristics will contribute to image interpretation in neurology, especially to precise partial volume correction. $$graphic_60272216-FF44-4923-B46C-57043B50EC92$$ ER -