Abstract
1166
Objectives: Clinical blood pool imaging is often performed in nuclear medicine facilities with 99m-Technetium-labeled agents, primarily human red blood cells (i.e., the " 99m-Tc tagged RBC scan"). While several PET-based blood perfusion agents have been introduced, these have found limited application, and are seldom used as blood pool agents. Given the high physiologic expression of the GLUT1 transporter on human RBCs, we hypothesized that human RBCs can spontaneously internalize the widely available PET tracer 18F-FDG in vitro and serve as a PET blood pool agent in an immunodeficient mouse model.
Methods: FDG labeling/prep of human RBCs: 250 µl human RBCs (Zen-Bio) were washed with sterile “1X EDTA” solution (140 mM NaCl, 4 mM KCl, 2.5 mM K2EDTA·2H2O), incubated with 370-740 MBq USP grade 18F-FDG (Cardinal Health) 37° C 30’, centrifuged 1000g 10’, and washed twice in 4X volume 1X EDTA solution. FDG-labeled RBCs were resuspended in 250 µl 1X EDTA solution. Sample/wash aliquots were counted in an Atomlab 600 dose calibrator (Biodex) to measure % retained FDG. ECG-gated microPET imaging of NSG mice: 4-6 wk old splenectomized NSG immunodeficient mice (Jax Lab) were fasted O/N and phlebotomized (200-250 µl), prior to microPET imaging with an Inveon PET/CT preclinical scanner (Siemens). Mice were injected with 500 µl FDG-labeled RBC solution via tail vein. ECG-gated raw PET data were acquired 10’ in list-mode format, followed by CT attenuation correction scan. PET images were analyzed with Inveon workstation software (Siemens). Patlak compartment plot option was chosen. For 3D PET and 4D PET data sets, VOIs were selected manually based on corresponding CT images: heart, leg muscle, liver, kidney and brain. Voxel activities were represented in standard uptake values (SUV).
Results: Human RBC incorporation of FDG at 37° C is near max by 30’: The % total FDG incorporation of 250 µl of 1 day old human RBCs incubated with 37-74 MBq FDG at 37° C = 78.3% ± 0.9 at 30’, 85.0% ± 0.0% at 60’, and 85.3% ± 0.3% at 120’. (fig 1; N=3) Minimal unincorporated FDG remains in washed RBCs after 1st wash step: The FDG activity in the washed cells, original incubation solution, and the 3 washes was measured. The relative % total FDG were as follows: Washed RBC fraction (500 µl) = 59% ± 1%; Original supernatant (free FDG after initial incubation) = 36% ± 2%; 1st wash supernatant = 3% ± 1%; 2nd wash supernatant = 2% ± 1%; 3rd wash supernatant = 1% ± 1%. (fig 2) In vivo ECG-gated microPET imaging of NSG mice injected with FDG-labeled RBCs reveal whole body vasculature: NSG mice were injected with 1.7-10.4 MBq FDG-labeled RBCs or 1.4-1.8 MBq of FDG (N=4/group). MicroPET images of FDG-labeled RBC injected mice show that the FDG activity is largely confined to the blood vessels and bone marrow (fig 3b). A small amount (≤5%) of total body FDG activity was seen over the bladder in these mice, suggesting small urinary excretion of residual free/released FDG. By contrast, the biodistribution of free FDG in control mice demonstrates expected intense myocardial and cerebral FDG accumulation, and marked urinary excretion in the kidneys and bladder (fig 3a). While intense myocardial uptake of free FDG in the left ventricle of control mice is clearly visualized on the fused microPET/CT images (fig 4a), FDG distribution in the FDG-labeled RBC injected mice shows activity confined to the cardiac chamber lumens and surrounding pulmonary vasculature (fig 4b).
Conclusion: We show that human RBCs can rapidly incorporate sufficient amounts of FDG to obtain in vivo images of the mouse vasculature using microPET/CT. As modern clinical PET/CT scanners generally possess count detection sensitivities far higher than that of clinical gamma scintigraphic cameras, FDG-RBC PET imaging may achieve comparable results to 99mTc-based blood pool imaging, with a potentially significant overall reduction in patient radiation dose. This technique may have other clinical PET blood pool/perfusion applications as well.