Comparison of a Multi-pinhole Stationary SPECT System and a Parallel-hole High Speed Rotational SPECT system

A multi-detector SPECT system with multi-pinhole collimators has been studied to reconstruct an object with a static data acquisition mode. Because a pinhole collimator can acquire photons that enter a detector at many different angles, and if we attach a multi-pinhole collimator to a gamma camera system that has a multi-detector, we can reconstruct the distribution of activities in an organ without any movement of the detectors. However, one of the issues of this data acquisition mode is its limited angles. The size of an effective area of the gamma camera is limited and many pinholes are required to reconstruct an artifact free image. The number of pinholes also affects the data acquisition efficiency and the spatial resolution of a reconstructed image. On the other hand, a high speed data acquisition with a parallel-hole collimator attached to a multi-detector SPECT system may reconstruct a reasonable quality image with a short time data acquisition. In this case, the number of detected photons is very small, but we can use many projection data acquired with equiangular directions. In this paper, we evaluated the quality of images reconstructed with the data measured with a stationary data acquisition with a multi-pinhole collimator and those measured with a high speed data acquisition with a parallel-hole collimator attached to a triple head gamma camera system. The collimators used in this study were a multi-pinhole collimator composed of 11 pinholes with a diameter of 2 or 3 mm, and a parallel-hole collimator with a hole diameter of 2 mm and thickness 4 cm. Data acquisition times were 3 and 4 min. We corrected the absorption of gamma rays with reference to the X-ray CT images, the scattering of gamma rays with the TEW method, and the statistical noise with the appropriate Butterworth filter. The quality of images acquired with the parallel-hole collimator was better than with the pinhole data acquisition despite the very short data acquisition time.