IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 44, NO. 4, AUGUST 1997 1499 Sub- Millimeter Planar Imaging with positron emitters: EGS4 code simulation and experimental results D.Bollinit, zyxwvutsr A. Del Guerratv2, zyxwvuts Senior Membel; ZEEE, G. Di Domenicot, M.Galli*, M. Gambaccinit and G.Zavattinit TDipartimento di Fisica, and INFN Sezione di Bologna (Italy) SDipartimento di Fisica, and INFN Sezione di Ferrara, Via Paradiso 12,I-44100 Ferrara (Italy) *ENEA INN, Bologna (Italy) Abstract Experimental data for Planar Imaging with positron emitters (pulse height, efficiency and spatial resolution) obtained with two matrices of 25 crystals ( 2 x 2 ~ 3 0 zyxwvutsr mm3 each) of YAP:Ce coupled with a Position Sensitive PhotoMultiplier (Hamamatsu R2486-06) have been reproduced with high accuracy using the EGS4 code. Extensive simulation provides a detailed description of the performance of this type of detector as a function of the matrix granularity, the geometry of the detector and detection threshold. We present the Monte Carlo simulation and the preliminary experimental results of a prototype planar imaging system made of two matrices, each one consisting of 400 (2x2x30mm3) crystals of YAP:Ce. I. INTRODUCTION We have already presented zyxwvutsrq [ 1, 21 the experimental results obtained for a new scintillator detector system for application in Positron Emission Tomography, made of a bundle of small YAP:Ce crystals closely packed, coupled to a position sensitive photomultiplier tube (PSPMT). We have now carried out an extensive Monte Carlo simulation by using the EGS4 code [3] so as to study the performance of this type of detector as a function of the bundle geometry and detection threshold. The experimental results are reproduced extremely well by the Monte Carlo simulation. This has allowed us to optimize the YAP matrix geometry for the detector for planar imaging with positron emitters. The prototype consists of two YAP:Ce matrices of 400 crystals (2x2~30 mm3 each) coupled to a Hamamatsu R2486-06 PSPMT. The preliminary experimental results together with the Monte Carlo simulation are presented. 11. MONTE CARLO SIMULATION We have implemented the general purpose EGS4 code so as to simulate the standard PET situation for two opposite detectors, and in particular: zyxwvutsr e radionuclide sources with their positron spectrum, positron slowing and annihilation; 'This work has been partially supported by MURST 40% 1994 and 'Corresponding author 1995 zyxwvutsrqp 0 angular distribution of the two 5 11 keV photons around e transport of the photon within the phantom, if any; interaction of the photon within the matrix of YAP, with a complete simulation of the center of mass position reconstruction made by the PSPMT e simple ray tracing of the interaction points of the two photons, if both produce a signal above the threshold; e presentation of the image on the central plane between the two detectors. 180'; HI. COMPARISON WITH EXPERIMENTAL RESULTS We first reproduced the experimental set-up used for our previous measurements [2]: two detectors, each made of a matrix of 25 match-like crystals ( 2 x 2 ~ 3 0 mm3) coupled to a PSPMT Hamamatsu R2486-06, placed 13 cm apart, with a 22Na radioactive source in the central plane (0.8 mm diameter and embedded in plastic). Figure 1 shows the experimental measurements (fig. la) compared with the Monte Carlo results (fig. lb) for a coincidence pulse height spectrum, with the same electronic threshold. The measured energy resolution (18% at FWHM) was used for the Monte Carlo simulation. Figure 2 shows the spatial resolution distribution as obtained with the measurements (fig. 2a) and with the Monte Carlo (fig. 2b). By taking into account the finite dimensions of the source (0.8 mm diameter) a spatial resolution of 0.9 mm FWHM can be estimated. Figure 3 shows the results for three equal activity spherical "Nu sources, 0.8 mm diameter, spaced by 2 mm and placed on the central plane (fig. 3a) as reconstructed by the measurements (fig. 3b) and by the Monte Carlo simulation (fig. 3c). The profile along the x direction of the reconstructed image is in figure 3d (experimental data) and in figure 3e (Monte Carlo data). The agreement is extremely good. There are also reproduced the smaller peaks in between two sources, which are due to the contribution of the incorrect reconstructed events because of the multiple Compton interactions of photons. The efficiency comparison was already presented in a previous paper [2]. The agreement between Monte Carlo and experimental results made us confident in using the Monte Carlo simulation for the choice of the final geometry of the YAP matrix for the prototype detector. 0018-9499/97$10.00 zyxwvut 0 1997 IEEE