论文部分内容阅读
脑局部葡萄糖代谢率(LCMRGlc)是反映脑功能状态的重要指标,18F标记的2-脱氧葡萄糖作示踪剂的动态正电子发射断层扫描(FDG-PET)方法已被用于人脑LCMRGlc定量参数成像,但由于原有的方法需要输入函数,这对人体是有损伤的,因而很少用于临床.本文采用基于参考区的Patlak图形近似模型(rPatlak)和动态FDG-PET成像方法生成LCMRGlc定量参数图,这种方法无需输入函数因而无需采集动脉血样.在对被试进行弹丸静脉注射155MBq的FDG后同步启动脑部动态PET扫描,扫描序列为4×0.5,4×2和10×5min,且扫描期间通过事先植入动脉的导管采集血样,以获得本研究中作为对比的金标准—原始Patlak图形近似方法(oPatlak)所需的输入函数.研究中的模拟数据也采用了同样的扫描序列.两种Patlak图形近似方法采用了最后的10个PET扫描数据.用需要血样数据的oPatlak方法获得相对于参考区的LCMRGlc比值作为金标准,标准摄取值比值(SUVR)也被计算并作为比对.对于实际数据,选择包括白质、灰质、全脑等8个不同脑区作参考区来进行评估.实验结果表明,无论选择哪个脑区作参考区,rPatlak与oPatlak的结果很相似,但SUVR的结果就差得多.模拟研究结果还表明,rPatlak结果的偏差及误差都小于SUVR.最后,用rPatlak方法生成的LCMRGlc定量参数图与oPatlak生成的很相似,但SUVR与oPatlak间就有较大差异.本研究表明,rPatlak方法好于SUVR方法,可以作为oPatlak方法的很好近似,新方法适合用来无损伤地生成LCMRGlc定量参数图.
The brain local glucose metabolism (LCMRGlc) is an important index reflecting the state of brain function. The dynamic positron emission tomography (FDG-PET) method using 18F labeled 2-deoxyglucose as a tracer has been used for LCMRGlc quantitative parameters However, since the original method requires the input function, which is impaired to the human body, it is rarely used in clinic.In this paper, we used the Patlak graph approximation model (rPatlak) and dynamic FDG-PET imaging method based on the reference region to generate LCMRGlc quantitative Parameter map, this method does not need to enter the function and thus does not need to collect arterial blood samples.After the test of intravenous injection of 155MBq of FDG brain synchronous dynamic PET scan, the scan sequence of 4 × 0.5, 4 × 2 and 10 × 5min, Blood samples were collected during catheterization through pre-implanted catheters to obtain the required input function for the gold standard in this study, the original Patlak graphical approximation (oPatlak). The same data was also used for the simulated data in the study . The two Patlak graphical approximations employ the last 10 PET scan data. The LCMRGlc ratio relative to the reference zone is obtained using the oPatlak method that requires blood sample data as the gold standard (SUVR) were also calculated and compared.For the actual data, eight different brain regions including white matter, gray matter and whole brain were selected as reference regions for evaluation.Experimental results show that no matter which brain region is selected The results of rPatlak and oPatlak are similar, but the result of SUVR is much worse.The simulation results also show that the deviation and error of rPatlak results are less than SUVR.Finally, the quantitative parameters of LCMRGlc generated by rPatlak method and oPatlak The results show that the rPatlak method is better than the SUVR method and can be used as a good approximation to the oPatlak method. The new method is suitable for generating quantitative LCMRGlc parameter maps without damage.