Positron Emission Tomography (PET) is a nuclear medicine imaging equipment that enables noninvasive visualization of metabolic and molecular processes in vivo, and has become increasingly important for the early diagnosis of neurological diseases. A P...
Positron Emission Tomography (PET) is a nuclear medicine imaging equipment that enables noninvasive visualization of metabolic and molecular processes in vivo, and has become increasingly important for the early diagnosis of neurological diseases. A PET detector is composed of a scintillation crystal coupled to a photosensor, and Geiger-mode avalanche photodiode (GAPD) cells are widely employed due to their high gain, compactness, scalability, and compatibility with magnetic resonance (MR) environments. Depending on the application, PET systems range from organ-dedicated scanners incorporating several thousand GAPD cells to total-body scanners comprising tens of thousands of GAPD cells.
When these GAPD cells are individually read out which generally provides the best detector performance, the number of required data acquisition (DAQ) channels increases drastically, resulting in high system complexity and increased data processing cost. Therefore, multiplexing that reduces the number of the readout channels techniques are essential in PET systems. However, sharing signal paths among multiple GAPD cells inevitably introduces performance degradation. Thus, achieving a high multiplexing ratio while minimizing performance degradation is a key challenge in multiplexed PET detector design.
Among various approaches, the resistive charge division (RCD) network is commonly adopted due to its simplicity, low cost, and its ability to provide an N:4 multiplexing ratio. Nevertheless, in multi-channel PET detectors, the combination of GAPD cell capacitance and the large total resistance of the RCD network leads to long RC time constants, causing pulse pile-up and distorted energy and timing performance. A conventional approach to mitigate this problem is to add additional RCD networks to reduce the resistance connected each GAPD cell. However, this solution increases the number of readout channels.
The objective of this study is to address the inherent trade-off between performance degradation and the increase in readout channels by developing and evaluating a polarity-based multiplexing circuit for multi-channel PET detectors. The proposed method is compared with two conventional configurations: a 32:8 RCD network constructed from two 16:4 RCD networks and a 32:4 RCD network. The timing resolution of the proposed detector was also measured.
A 32-channel PET detector was constructed by arranging two 16-channel detector modules in a linear configuration. Each 16-channel detector comprised a 4 × 4 GAPD cell array (ARRAYJ-30035, SensL, Cork, Ireland) with an active area of 3.07 × 3.07 mm², optically coupled one-to-one to a 4 × 4 array of Lutetium yttrium oxyorthosilicate (LYSO) crystals (3.21 × 3.21 × 20 mm³, EPIC, Shanghai, China) using optical grease (BC-630, Saint-Gobain, Hiram, USA).
In the proposed polarity-based multiplexing scheme, the 4-corner outputs from one 16-channel module are processed by two-stage inverting amplifiers to generate four positive-polarity signals, while the corresponding outputs of the other module are processed by two-stage of inverting and non-inverting amplifiers to generate four negative-polarity signals. These eight signals are then combined using four summing amplifiers, achieving 2:1 multiplexing ratio, and subsequently acquired by the DAQ system.
To evaluate the performance of each multiplexing method, the energy resolution, flood histogram quality, distance-to-width ratio (DWR), and peak-to-valley ratio (PVR) were measured. The experimental results show that all LYSO crystals were clearly distinguished in the flood histograms for all three multiplexing configurations. The energy resolution and the DWR/PVR values in the row and column directions were as follows: for the polarity-based multiplexing, 13.3%, 14.6/37.8 (row direction), and 16.0/30.9 (column direction); for the 32:8 RCD network, 13.5%, 14.8/41.6 (row direction), and 16.8/33.1 (column direction); and for the 32:4 RCD network, 14.2%, 13.8/29.9 (row direction), and 10.6/29.0 (column direction). The single timing resolution (STR) of the polarity-based multiplexing detector was measured to be 1.61 ns FWHM.
The proposed polarity-based multiplexing method achieves a 32:4 multiplexing ratio while maintaining an energy resolution of 13.3%, demonstrating performance comparable to that of the 32:8 RCD network in terms of DWR and PVR. The results demonstrate that the proposed polarity-based multiplexing method enables channel reduction without considerable performance degradation, making it a feasible option for scalable PET readout systems and related radiation detection applications.