摘要目的 探讨在PET绝对定量分析中,如何在使用无损输入函数的同时,减PET动态显像时间.方法 通过勾画PET图像兔左心室感兴趣区(ROI)得到左心室血池时间活度曲线(PTAC),验证用血池感兴趣区法与标准输入函数模型结合的方法,确定无损输入函数可以替代动脉连续采血.利用输入函数模型拟合PTAC,输出函数模型拟合组织活度时间曲线(TTAC),再利用微分方程数值解法求解葡萄糖代谢房室模型.对该算法在缩短PET动态采集时间方面的应用进行了计算机仿真和动物实验上的验证.结果 (1)无创输入函数验证部分,两者通过非线性最小二乘法得到的ki相关系数r=0.9824(P<0.01).(2)计算机模拟结果表明在中低噪声情况下,可以将PET动态采集时间缩短为25min.动物实验结果表明,当PET动态采集时间缩短为25 min时,ki_(60)=0.9127×ki_(25)-0.0013,相关系数r=0.9182(P<0.01).结论 本方法可以通过选取合适的血池来获得无损输入函数,避免连续动脉采血,同时减少PET动态采集的时间.

abstractsObjective To establish a simplified PET absolute quantification analysis method which avoids arterial blood sampling and reducse the total time of dynamic PET imaging.Methods Plasma time activity curve (PTAC) of left ventricle was obtained by outlining the region of interest (ROI) in PET images.Through a combination of PTAC ROI and standardized input function model, whether definite non-invasive input function may be used as an alternative to serial arterial sampling was further validated.Then, an algorithm using curve fitting [input function model for PTAC, output function model for tissue time activity curve (TTAC)] and numerical solution of differential equation was proposed to obtain a compartment model of glucose metabolism.Computerized simulation and experimental validation of this putative algorithm was completed with regard to its capacity to reduce the total time of dynamic PET imaging. Result (1) Comparison of ki values obtained with the measured and the fitted input curves showed that both procedures gave consistent results using nonlinear least squares (NLS) methods (r=0.9824, P<0.01).(2) Under low or medium noise level condition, computerized simulation results revealed that as short as 25 minutes of dynamic PET sampling could be feasible for FDG-PET full kinetic analysis.The result of animal tumor model showed ki_(60)=0.9127 × ki_(25)-0.0013 when dynamic PET sampling was reduced to 25 min (r=0.9182, P<0.01).Conclusion Non- invasive input function can be obtained with assistance of proper selected plasma, meanwhile, serial arterial sampling may be avoided and time duration of dynamic PET imaging is reduced using the method proposed in this study.