@misc{oai:repo.qst.go.jp:00065501,
 author = {田島, 英朗 and 伊藤, 浩 and 山谷, 泰賀 and 田島 英朗 and 伊藤 浩 and 山谷 泰賀},
 month = {Sep},
 note = {We are aiming at developing high sensitivity and accurate brain PET scanner. We can achieve high sensitivity for the cerebrum region if we can arrange PET detectors in a hemispheric shape, because the solid angle coverage of the detector for the peripheral region of the hemisphere highly increases compared with conventional cylindrical PET consisted with the same amount of detectors. However, the sensitivity at the center region is low, where the cerebellum and hippocampus are located during PET measurement. These regions are important for accurate functional brain imaging. The cerebellum region may be used as a reference region in kinetic analysis, and the hippocampus region is targeted for early diagnosis of the Alzheimer’s disease. In this study, therefore, we are proposing a helmet-PET with a jaw detector as an add-on detector to achieve high sensitivity for both center and peripheral regions. The key points of the geometry are the closely positioned detectors and the detector covering the jaw.
To show the effectiveness of the proposed geometry, we analyzed the geometrical sensitivity as the relative coverage of the solid angle where the coincidence measurement was possible. We calculated the geometrical sensitivity for each sampling point in the FOV. As the result, we found that the hemispheric-shaped geometry had high sensitivity for the cerebrum region. In addition, the jaw detector significantly improved the sensitivity at the center region of the geometry. We also calculated geometrical sensitivity for helmet-PET with ear detectors or neck detector as add-on detectors. The hemispheric-shaped detector and the add-on detectors were defined on the common sphere. All of these add-on detectors were equivalently effective for improving the sensitivity at the center region. The geometrical sensitivity at the center region of interest increased to 300% by adding 12% amount of hemispheric-shaped detector as the add-on detector.
We conducted imaging simulations for the geometries assuming 3.0 × 3.0 × 3.0 mm3 cubic detector elements arranged as the hemispheric-shaped detector and the add-on detectors. The diameter of the sphere where the detector arranged was 250 mm. The detector elements were uniformly distributed on each circle with polar angle sampling on sphere. Projection data were generated by forward projection calculation incorporating the detector response function, and then Poisson noise was added. The images were reconstructed by the ordered subset expectation maximization method. The normalized standard deviation was calculated for hot disks placed central and peripheral regions in the numerical phantom. The simulation results showed the proposed geometry increases image quality especially for the central region. In conclusion, our proposed geometry has good potential for high sensitivity and accurate brain PET imaging., World Molecular Imaging Congress},
 title = {Simulation Study of Helmet-PET with Add-on Detectors for High Sensitivity Brain Imaging},
 year = {2014}
}