Parametrically defined cerebral blood vessels as non-invasive blood input functions for brain PET studies
Parametrically defined cerebral blood vessels as non-invasive blood input functions for brain PET studies
复制标题
DOI：
10.1088/0031-9155/49/6/013
复制
发表时间：
2004-03-21
影响因子：
3.5
通讯作者：
Nahmias, C
中科院分区：
文献类型：
--
作者：
Asselin, MC;Cunningham, VJ;Nahmias, C
提出了一种非侵入性的替代动脉血液采样的血液输入功能的生成脑正电子发射断层扫描（PET）研究。该方法旨在直接从PET图像中提取血管的尺寸，并同时校正部分体积和溢出的放射性浓度。这涉及断层成像过程的模拟，以生成不同血管和背景几何形状的图像，并在最小二乘意义上选择最适合所采集的PET图像的图像。进行体模实验以验证该方法，然后将其应用于注射6-[F-18]fluoro-L-DOPA的8名受试者和注射[C-11] CO标记的红细胞的1名受试者。在体模研究中，以半个像素（1 mm）的精度估计填充有11 C溶液并插入到充水圆柱体中的注射器的直径。在最接近上级矢状窦的8.7 mm直径注射器中，放射性浓度回收至100 +/- 4%。在人体研究中，与相同受试者的磁共振静脉造影测量值相比，该方法系统性高估了上级矢状窦口径23 mm。发现与血管解剖结构相关的差异来源并不是该方法对人类受试者适用性的根本限制。这种方法有可能提供准确的定量血液放射性浓度的PET图像，而不需要血液样本，校正延迟和分散，共配准的解剖图像，或手动定义的感兴趣的区域。
A non-invasive alternative to arterial blood sampling for the generation of a blood input function for brain positron emission tomography (PET) studies is presented. The method aims to extract the dimensions of the blood vessel directly from PET images and to simultaneously correct the radioactivity concentration for partial volume and spillover. This involves simulation of the tomographic imaging process to generate images of different blood vessel and background geometries and selecting the one that best fits, in a least-squares sense, the acquired PET image. A phantom experiment was conducted to validate the method which was then applied to eight subjects injected with 6-[F-18]fluoro-L-DOPA and one subject injected with [C-11]CO-labelled red blood cells. In the phantom study, the diameter of syringes filled with an 11C solution and inserted into a water-filled cylinder were estimated with an accuracy of half a pixel (I mm). The radioactivity concentration was recovered to 100 +/- 4% in the 8.7 mm diameter syringe, the one that most closely approximated the superior sagittal sinus. In the human studies, the method systematically overestimated the calibre of the superior sagittal sinus by 23 mm compared to measurements made in magnetic resonance venograms on the same subjects. Sources of discrepancies related to the anatomy of the blood vessel were found not to be fundamental limitations to the applicability of the method to human subjects. This method has the potential to provide accurate quantification of blood radioactivity concentration from PET images without the need for blood samples, corrections for delay and dispersion, co-registered anatomical images, or manually defined regions of interest.