High Resolution TOF-DOI Prism-PET Brain Scanner: Experimental Results From The First Prototype Scanner

Introduction: We have developed a one-ring positron emission tomography (PET) prototype scanner for human brain imaging based on our unique single-ended readout Prism-PET depth-encoding detector modules. The prototype has a decagon geometry conforming to the human head, leading to a cost-effective and high-sensitive scanner because of the limited number of detectors and larger solid angle coverage. The parallax error (PE) associated with the conformal geometry can be overcome by the 2.57 mm full width at half maximum (FWHM) depth-of-interaction (DOI) resolution of Prism-PET detectors.

Methods: The one-ring prototype has a long diameter of 38.5 cm, a short diameter of 29.1 cm, and an axial field-of-view (FOV) of 25.5 mm (Figure 1a). It consists of 40 DOI Prism-PET detector modules, each with an array of 16 x 16 lutetium yttrium oxyorthosillicate (LYSO) scintillator crystals. Each crystal size is 1.5 x 1.5 x 20 mm³ and the pitch is 1.6 mm. The crystal array is coupled 4-to-1 to an 8 x 8 readout array of 3.0 x 3.0 mm² silicon photomultipliers (SiPM) pixels (Hamamatsu, Japan) on one side and a segmented prismatoid light guide array on the opposite side. The configuration and performance of the prototype are summarized in Figure 1b. The spatial resolution and image quality were qualitatively evaluated by an ultra-micro hot spot phantom and a 3D Hoffman brain phantom (Data Spectrum Corporation, USA). The ultra-micro hot spot phantom with different rod diameters arranged in 6 segments (0.75, 1.0, 1.35, 1.7, 2.0, and 2.4 mm) was filled with 37 MBq (1 mCi) of F-18 at the start and scanned for 60 min. The Hoffman brain phantom was also filled with 37 MBq (1 mCi) of F-18 initially and scanned for 310 min. Both phantoms were placed at the center of the FOV. The images were reconstructed by CASToR (Customizable and Advanced Software for Tomographic Reconstruction) using OSEM with 16 iterations and 12 subsets for the hotspot phantom, and 7 iterations and 8 subsets for the brain phantom. The image matrix size was 360 x 360 x 128 with 0.5 x 0.5 x 0.5 mm³ voxels. For image reconstruction, normalization, attenuation correction, inter-crystal scatter (ICS) rejection, and DOI rebinning were performed. The direct normalization was implemented using a 10-inch uniform cylindrical phantom. The Hoffman phantom filled with F-18 was scanned by Siemens Biograph TruePoint PET/CT for comparison.

Results: Our scanner achieved 264 ps, 2.57 mm, and 10.66 % (all in FWHM) TOF, DOI, and energy resolutions, respectively. The reconstructed image of the ultra-micro hotspot phantom at the center of the FOV is shown in Figure 1c. Spots with a diameter of 1.35 mm and larger can be clearly resolved. Figure 1d shows the experimental setup of the 3D Hoffman brain phantom. Reconstructed images of one slice are shown in Figure 1e where the CT image serves as the ground truth for the anatomical structure. The fine details of the cortical grey-white border are perfectly visualized in the Prism-PET image when compared to the Biograph Truepoint PET image.

Conclusions: We thank the Department of Radiology and Center for Biotechnology (CFB) at Stony Brook University for their financial support through the NIH REACH program.

• Download figure
• Open in new tab
• Download powerpoint