No abstract available

[en] This paper describes present and future situation of both the hardware and the software for PET laying emphasis on technical development since 1990. In the present, some institutes still use 2D-PET, however, most technical researches of PET are aiming at 3D-PET, and several difficulties, such as absorption correction, scattering correction, correction of accidental simultaneous count, have been resolved. The difficulties of data correction and the limitation of high counting rate are largely caused by the properties of the detective element BGO. If we can put LSO to practical use, time resolution and energy resolution would be significantly improved, and moreover, it would also be expected to simplify data correction mentioned above and to improve the accuracy. If the removal-type slice septa become useless, the mechanical correction would be easy, and the diameter of detector would be smaller and it will be profitable for the cost. 3D-PET without the slice septa would be ideal in the future. (K.H.). 54 refs

[en] In obtaining three-dimensional radioisotope (RI) distribution, projection data along with several slices are measured, and two-dimensional reconstruction per slice is performed. However, there are several problems characteristic of RI imaging, such as absorption or scattering of gamma ray in the body and statistical changes in measurements. In addition, the measurement restricted to each slice is of limited value in increasing detection sensitivity. Thus, realization of both three-dimensional measurement and three-dimensional reconstruction is of great concern. In this report, Chapters 2 and 3 provide the outline of slice accumulation method of single photon emission computed tomography (SPECT) and positron computerized tomography (PET), with special attention to image reconstruction and corrections of absorption and scattering. Chapters 4 and 5 review the current status and problems of true three-dimensional imaging, focusing on cornbeam type SPECT and three-dimensional PET. (N.K.) 70 refs

[en] This paper describes the theory and practice of analytical methods of three-dimensional (3D) image reconstruction for PET and SPECT. In 3D PET, the 3D image is reconstructed from parallel-beam projections taken at various slant angles. The main topics are '3D filtered backprojection method' for a constant acceptance angle of projections, '3D reprojection method' used in practical 3D PET scanners, and 'rebinning methods' as the fast approximate reconstruction methods. This paper also describes a sufficiency condition for 3D image reconstruction from cone-beam projection data obtained with cone-beam SPECT, and gives a fundamental reconstruction method on the basis of Radon transform and some practical algorithms. Principle of rotating slant-hole SPECT and its recent topics are mentioned briefly. (author)

No abstract available

[en] It has passed almost 40 years since gamma camera was first developed in Japan. This paper reviews the progress of the physics in nuclear medicine in Japan with emphases on instrumentation and quantification of radio-nuclide imaging. The topics include the progress of gamma camera, single photon emission computed tomography (SPECT), positron emission tomography (PET) and image reconstruction from projections. Various improvements are highlighted including new position arithmetic of gamma cameras, progress of spatial resolution of gamma cameras and PET scanners, analytical methods of attenuation correction in SPECT, move from 2D-PET scanners with septa to fully 3D-PET scanners, advance of the methods of image reconstruction for 3D-PET, efforts to speed up iterative image reconstruction algorithms, etc. Besides technical instrumentation, developments of key components such as scintillators, photosensors, etc. are also described. (author)

[en] In a scintillation camera detector, a pair of outputs from a number of photomultipliers aligned in the X-direction, for example, are applied to associated intermediate taps of two delaylines through two associated groups of attenuators adjusted so that the delay times and attenuations of the delay line outputs correspond proportionally to the positional coordinates, in the X and -X directions, at which the scintillators of the photomultipliers are located. Both outputs of the delay lines are formed, expanded and then fed to ratio circuits for obtaining respective ratio signals to a Z signal. Both ratio signals are fed to a subtraction circuit where a difference signal is obtained for determining the position of a luminous point. Since the contributions of the photomultipliers remote from the luminous point are separated in time by insertion of the delay lines, the resolution of the camera can be improved. (Ohno, Y.)

No abstract available

[en] The advance of computerized tomography has undoubtedly owed to the successful development of the reconstruction algorithm of a section image from its projection (profile). Various approaches to the accurate reconstruction have been made, based on the different mathematical manipulations. These methods and algorithms are roughly divided into three major groups: iteration algebraic method, Fourier conversion method and convolution method (linear supperposition of filtered back-projection). In this paper, these methods are briefly described with particular emphasis on the convolution method, because most of recent CT machines are based on the convolution method. The iterative reconstruction technique includes three major corrective variations of ray-by-ray correction method (ART), point-by-point correction method (SIRT) and simultaneous correction method (ILST). The Fourier conversion method has become practival rather recently by the development of Fast Fourier Transform algorithm and the improvements of high speed computers. The convolution method is similar to the simple back-projection except that the projections are corrected or modified before being back-projected so as to provide more accurate reconstruction. The author has recently pointed out that a generalized correction function can be derived without imposing the frequency roll off by the concept of the ''generalized function'' or ''distribution'', and that the fundamental property of the convolution reconstruction is well understood by this concept. Statistical noise, overall spatial resolution and number of sampling are considered and discussed as the factors affecting image quality. (Wakatsuki, Y.)

[en] Special interest has been shown towards the application of positron-releasing nuclides prepared with cyclotrons to nuclear medicine. In this report, the special features of positron imaging by the simultaneous counting of vanished radiation are described, and 2 kinds of positron imaging devices developed by NIRS are introduced. Both devices belong to opposite two detector type. Positron imaging has a constant resolving power regardless of the position, and can correct for the internal absorption of radiation. In the focus detector type positron camera produced by NIRS for the first time, such method was taken that another focus detector was put opposite to the existing gamma-camera to count simultaneously. Recently, the opposite multicrystal detector type positron camera has also been completed. It was produced by modifying and improving what had been developed by Massachusetts General Hospital. Its detecting sensitivity and maximum counting rate have been improved by ten times as compared with the focus detector type positron camera. (Kobatake, H.)