Title

9. Physical Performance of a PET Scanner with Adjustable Data-acquisition Parameters

Hideo Murayama and Tomoyuki Hasagawa*
(*Kitasato Univ.)

Keywords: three-dimensional data acquisition, positron emission tomography, nuclear medicine



The three-dimensional (3D) mode of a positron emission tomography (PET) scanner offers a high sensitivity, but undesirable problems have arisen due to the larger axial field-of-view (FOV) with retracted septa. In the 3D mode, scatter fraction and random coincidence increase, and spatial resolution and sensitivity differ from the center to the circumference of the FOV. In one of the latest 3D PET scanners, ECAT EXACT HR+, the data-acquisition parameter defining the largest absolute difference in ring numbers accepted in coincidence detection (maximum ring difference or mrd), which is closely related to the acceptance angle, can be defined. In addition, there are other adjustable data-acquisition parameters which affect the physical performance.

Measurements were performed in principle by the National Electrical Manufacturers Association (NEMA) protocol with minor modifications. An acrylic cylindrical vessel phantom (20 cm inner diameter and 18.5 cm inner length) was set at the center of the scanner. It was filled with water containing F-18 radioactivity of 1.6 kBq/ml. Plane sensitivities were calculated from total counts of sinogram planes. For the 3D mode, the single slice rebinning algorithm was used to define the sinogram planes. The plane sensitivity in the 3D mode for various mrd parameters is shown in Fig.7. The flat region was about one-third of the total axial FOV in the default condition of data acquisition. The total sensitivity, as a function of mrd, showed that the default parameters had higher sensitivities close to the maxima.

A stainless steel needle (1.0 mm inner diameter and 1.5 mm outer diameter) filled with water containing F-18 radioactivity was positioned parallel to the scanner axis at various radial distances to the central axis. Images were reconstructed using a ramp filter with a cut-off at the Nyquist frequency and transverse spatial resolutions were calculated as the full width at the half maxima. A plastic tube (1.4 mm inner diameter and 5 mm long) filled with water containing F-18 radioactivity was positioned vertical to the scanner axis in the tangential direction. Axial spatial resolution was measured at five radial distances of 0, 5, 10, 15, and 20 cm, and at seven axial positions corresponding to plane numbers of 2, 7, 12, 17, 22, 27, and 32. The transverse spatial resolution in the tangential direction was almost independent of the radial position, while that in the radial direction increased with the radial position. There was discontinuous degradation toward the end planes in the 3D mode. The central flat regions were smaller for large radial distances. For larger mrd values, the axial spatial resolution was markedly degraded in the off-center region.

This study clarified the importance of understanding scanner performance dependence on data-acquisition parameters. The results are useful for maximizing the efficiency of PET measurements by selecting better acquisition parameters for each specific clinical application.



fig07

Fig.7. Plane sensitivity as a function of the plane number for various maximum ring difference (mrd) values in the 3D mode.


Publications:
Hasegawa, T., et al.: IEEE Trans. Nucl. Sci., 46, 652-658, 1999.


back