Kei Takahashi

National Institute of Radiological Sciences, Tiba, Chiba, Japan

Are you Kei Takahashi?

Claim your profile

Publications (7)1.32 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The jPET-D4 is a brain positron emission tomography (PET) scanner that we have developed to meet user demands for high sensitivity and high spatial resolution. For this scanner, we developed a four-layer depth-of-interaction (DOI) detector. The four-layer DOI detector is a key component for the jPET-D4, its performance has great influence on the overall system performance. Previously, we reported the original technique for encoding four-layer DOI. Here, we introduce the final design of the jPET-D4 detector and present the results of an investigation on uniformity in performance of the detector. The performance evaluation was done over the 120 DOI crystal blocks for the detectors, which are to be assembled into the jPET-D4 scanner. We also introduce the crystal assembly method, which is simple enough, even though each DOI crystal block is composed of 1,024 crystal elements. The jPET-D4 detector consists of four layers of 16 x 16 Gd(2)SiO(5) (GSO) crystals and a 256-channel flat-panel position-sensitive photomultiplier tube (256ch FP-PMT). To identify scintillated crystals in the four-layer DOI detector, we use pulse shape discrimination and position discrimination on the two-dimensional (2D) position histogram. For pulse shape discrimination, two kinds of GSO crystals that show different scintillation decay time constants are used in the upper two and lower two layers, respectively. Proper reflector arrangement in the crystal block then allows the scintillated crystals to be identified in these two-layer groupings with two 2D position histograms. We produced the 120 DOI crystal blocks for the jPET-D4 system, and measured their characteristics such as the accuracy of pulse shape discrimination, energy resolution, and the pulse height of the full energy peak. The results show a satisfactory and uniform performance of the four-layer DOI crystal blocks; for example, misidentification rate in each GSO layer is <5% based on pulse shape discrimination, the averaged energy resolutions for the central four crystals of the first (farthest from the FP-PMT), second, third, and 4th layers are 15.7 +/- 1.0, 15.8 +/- 0.6, 17.7 +/- 1.2, and 17.3 +/- 1.4%, respectively, and variation in pulse height of the full energy peak among the four layers is <5% on average.
    Radiological Physics and Technology 01/2008; 1(1):75-82. DOI:10.1007/s12194-007-0014-x
  • [Show abstract] [Hide abstract]
    ABSTRACT: We are developing a small animal PET scanner, “jPET-RD” to achieve high sensitivity as well as high spatial resolution by using four-layer depth-of-interaction (DOI) detectors. The jPET-RD is designed with two detector rings. Each detector ring is composed of six DOI detectors arranged hexagonally. The diameter of the field-of-view (FOV) is 8.8 cm, which is smaller than typical small animal PET scanners on the market now. Each detector module consists of a crystal block and a 256-channel flat panel position-sensitive photomultiplier tube. The crystal block, consisting of 32×32×4 crystal (4096 crystals, each 1.46 mm×1.46 mm×4.5 mm) and a reflector, is mounted on the 256ch FP-PMT.In this study, we evaluated the spatial resolution of reconstructed images with the evaluation system of two four-layer DOI detectors which consist of 32×32×4 LYSO (Lu: 98%, Y: 2%) crystals coupled on the 256ch FP-PMT by using RTV rubber. The spatial resolution of 1.5 mm was obtained at the center of the FOV by the filtered back projection. The spatial resolution, better than 2 mm in the whole FOV, was also achieved with DOI while the spatial resolution without DOI was degraded to 3.3 mm.
    Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment 01/2008; 584(1-584):212-218. DOI:10.1016/j.nima.2007.10.001 · 1.32 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recently, avalanche photodiodes (APDs) are used as photo detectors in positron emission tomography (PET) because they have many advantages over photomultipliers (PMTs) typically used in PET detectors. We have been developing a PET detector with a multi pixel APD for a small animal PET scanner. Previously, we succeeded in the identification of four-layer depth of interaction (DOI) with a position sensitive PMT coupled to the backside of a crystal array by only an optimized reflector arrangement. We adopt this method for the APD detector to replace the position sensitive PMT. While APDs have many advantages as photo detectors, the detector performance with them is affected by electrical noise because of their low internal gain. Since scintillation lights are shared among many pixels by the presented method, weaker signals in pixels far from the interacting crystals are strongly affected by the noise. To evaluate the possible use of the four-layer DOI detector with APDs using the light sharing method, we constructed a prototype DOI detector and tested its performance. The prototype detector consisted of four layers of a 6 x 6 array with Lu<sub>2(1-x)</sub> Y<sub>2x</sub> SiO<sub>5</sub> (LYSO) (Lu: 98 %, Y: 2 %) (Proteus Inc., U. S. A.) crystals, and a pixelized APD with a 4x8 pixel array (S8550, Hamamatsu Photonics K.K. ). The size of each crystal element was 1.46 mm x 1.46 mm x 4.5 mm and all element surfaces were chemically etched. The experimental results indicated that the four-layer DOI detector with the multi pixel APD can be used as a small animal PET detector.
    Nuclear Science Symposium Conference Record, 2007. NSS '07. IEEE; 01/2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: Positron emission tomography (PET) and optical imaging are two major techniques for molecular imaging. The PET scanner which enables simultaneous detection of optical imaging will be a powerful tool in molecular imaging. Here, we propose a new detector that detects both signals: annihilation photons for PET and photons for optical imaging. It is based on a four layered depth-of-interaction (DOI) PET detector. Since the detector must be set at a close distance to an object for optical imaging, the DOI information becomes important to obtain uniform spatial resolution in PET imaging over the field-of-view. The proposed PET/optical imaging detector consists of a scintillation crystal block and a position sensitive photo multiplier tube (PS-PMT). While the bottom of the crystal block is optically coupled to the PS-PMT, four side surfaces of the crystal block are covered with reflectors. A dichroic mirror is placed on the top of the crystal block. The dichroic mirror allows fluorescence photons (longer wavelength than 600 nm) to pass through while it reflects scintillation photons (wavelength around 450 nm) generated within the crystal block. Owing to the dichroic mirror, both scintillation photons originating inside the crystal block and fluorescence photons coming from outside the block can be detected by the same PS-PMT. To evaluate of the DOI-PET/ optical detector, we first measured the influence of the dichroic mirror. The results show that the dichroic mirror maintains DOI-PET detector performance without significant loss of fluorescence photons for optical imaging.
    Nuclear Science Symposium Conference Record, 2007. NSS '07. IEEE; 01/2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have begun to develop the PET detector which is composed of a scintillation crystal array and a 2times 2 array of inexpensive photomultiplier tubes (PMTs). The detector is intended for use for a low cost commercial whole body PET scanner. Anger type calculation of the four PMT outputs gives the 2-dimensional (2D) position histogram on which response of each crystal element appears and the scintillated crystal is identified on the histogram. The array of inexpensive PMTs however does not provide small sampling intervals like a position sensitive PMT and some crystal responses may overlap in the histogram due to insufficient scintillation light spread to the other three PMT outputs indirectly connected to the light-emitted crystals. To get more light spread and avoid the overlap of the responses, a light guide is generally inserted between the crystal array and the PMT array. We propose a new way to identify crystals which uses no light guides and which removes reflectors between some crystals and inserts some reflectors between the crystal array and the PMT array. The inserted reflector not only promotes light spread but also controls the light path to improve the crystal identification performance. Since the reflector is sufficiently thin, the method can be applied to a depth of interaction (DOI) detector with a new optical structure. By inserting a reflector between DOI crystal layers, it is possible to get a different light distribution between crystals in any DOI layer, which results in discrimination of all crystal responses in the 2D position histogram. We demonstrated the validity of this method and showed that it improved crystal identification ability without degradation of light output, and in turn, energy resolution. We also applied this method to a 2-layer DOI crystal array and verified its effect.
    Nuclear Science Symposium Conference Record, 2007. NSS '07. IEEE; 01/2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: Previously, we proposed an 8-layer depth-of-interaction (DOI) encoding method for a PET detector and proved its validity. The layer of interaction is identified by hybrid method: scintillation light control by the original reflector arrangement for 4-layer DOI encoding and pulse shape discrimination for 2-layer DOI encoding. In the 8-layer DOI detector, four layers then consist of the scintillator of different pulse shape from another scintillator for the other four layers. The two kind crystal layers can be arranged in two ways: stacked alternately (LSLS) or set in the upper and lower four layers (LLSS). Since the two crystal arrangements are expected to show different detector performance, we investigated the difference to understand the characteristics of the DOI detector for its optimization. Gd<sub>2</sub>SiO<sub>5</sub> (GSO) crystals of 0.5 mol% Ce dopant and 1.5 mol% Ce dopant were used for the measurement. The former is in dimensions of 2.9 mm times 2.9 mm times 3.75 mm and the latter is 2.9 mm times 2.9 mm times 3.6 mm, respectively. The experimental results show better performance of the LLSS arrangement in pulse shape discrimination, while inferior in the 4-layer DOI encoding. There was no particular difference between the two crystal arrangements in light output and energy resolution of each layer.
    Nuclear Science Symposium Conference Record, 2006. IEEE; 12/2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: We are developing a small animal PET scanner, "jPET-RD" to achieve high sensitivity as well as high spatial resolution by using four-layer depth-of-interaction (DOI) information of the detector. The jPET-RD is designed with two detector rings each composed of six DOI detector modules arranged hexagonally. The diameter of the field-of-view (FOV) is 8.8 mm, which is smaller than for typical small animal PET scanners. Each detector module consists of a 32 times 32 times 4 LYSO (Lu: 98%, Y: 2%) crystal array and a 256-channel flat panel position sensitive photomultiplier tube. The size of each crystal element is 1.46 mm times 1.46 mm times 4.5 mm. The crystal block consisting of a crystal array and reflector is placed on the central area of the 256ch FP-PMT (49 mm times 49 mm useful area) and coupled with silicone rubber. Previously, we developed a prototype detector for the jPET-RD and successfully implemented crystal identification for the detector module. In the present work, we develop a prototype imaging system for the jPET-RD. The system is composed of two prototype detectors and electrical circuits. We evaluated the spatial resolution of reconstructed images with the one-pair system of the two four-layer DOI detectors for the jPET-RD. The spatial resolution of 1.5mm was obtained at the center of FOV by the filtered back projection algorithm. The spatial resolution of better than 2 mm in the whole FOV was also achieved.
    Nuclear Science Symposium Conference Record, 2006. IEEE; 12/2006

Publication Stats

18 Citations
1.32 Total Impact Points

Institutions

  • 2008
    • National Institute of Radiological Sciences
      • Molecular Imagining Center
      Tiba, Chiba, Japan
  • 2007
    • Chiba University
      Tiba, Chiba, Japan
  • 2006
    • Chiba Institute of Science
      Tiba, Chiba, Japan