基于Gate与Matlab的PET电子学仿真系统研究

正电子发射断层成像PET （Positron Emission Tomography）是一套庞大、复杂而又极其精密的大型医疗设备，由成千上万个探测单元及相应的信号采集通道组成，其复杂的采集逻辑直接导致系统前期研发迭代次数多、硬件调试困难，且时间与资金成本投入高。本文通过研究PET系统整体结构及前端电子采集模块的功能，设计出基于Gate仿真软件（Geant4 application for tomographic emission）和MATLAB （matrixlaboratory，Mathworks Inc.）数据处理包的PET电子学仿真架构，利用数字化单事件重建DiSER （digitalsingle-event-reconstruction）模型将Gate仿真得到的湮灭事件的能量、时间信息进行综合，转换为光子脉冲信号并由模数转换器ADC （analog-to-digital converter）进行量化，量化数据输入到现场可编程门阵列FPGA （field programmable gate array）和MATLAB协仿真模块中，验证FPGA硬件采集算法的可行性。待各模块的湮灭事件采集完毕，将多模块数据进一步综合输入到FPGA符合处理模块中，完成PET系统符合单元的功能验证。最后，将设计的算法在基于飞行时间技术的TOF （time-of flight）-PET样机上进行了测试，验证了设计的有效性和正确性。该系统仿真平台可应用于PET系统整个研发周期，其搭建思路也可灵活、有效地拓展到其他领域的系统建模与验证之中。

Research on Electronic Simulation System of PET Based on Gate and Matlab

Positron emission tomography (PET) is a technology intensive complex medical product containing thousands of detector and data sampling channels, which leads to difficult debugging, excessive iterations, timeconsuming and high cost in early R&D stage. For solving this problem, a PET electrical simulation system was constructed in the paper based on Gate (Geant4 application for tomographic emission) and MATLAB (matrix laboratory, Mathworks Inc.). Firstly, the annihilation events with TOF (time-of flight) information are recorded by Gate and converted into photon pulse signals using DiSER (digital-single-event-reconstruction) model. The pulses are quantified by ADC (analog-to-digital converter) and transferred to the co-simulation module of FPGA (field programmable gate array) and MATLAB to verify the feasibility of FPGA-based hardware acquisition algorithm. Secondly, the data latched by co-simulation platform is presented to the coincidence processor implementable by FPGA in a pipelined manner to verify the algorithm's validity and optimality. Finally, the proposed simulation model is verified by experiments using a TOF-PET prototype and the Gate-Matlab co-simulation platform presented in the paper can span the development lifecycle of PET, and the basic ideas and principles can be extended to other fields flexibly.