RT Journal Article
SR Electronic
T1 Dynamic FDG brain PET imaging using Helmet_PET: Model corrected blood input function to compute FDG uptake rate during an ambulatory upright task
JF Journal of Nuclear Medicine
JO J Nucl Med
FD Society of Nuclear Medicine
SP 427
OP 427
VO 59
IS supplement 1
A1 Qiao Huang
A1 Chris Bauer
A1 Alexander Stolin
A1 Thorsten Wuest
A1 Samantha Melroy
A1 Stan Majewski
A1 Julie Brefczynski-Lewis
A1 Bijoy Kundu
YR 2018
UL http://jnm.snmjournals.org/content/59/supplement_1/427.abstract
AB 427Objectives: The objective of this work was to compute the rate of cerebral FDG uptake from dynamic PET images obtained using the prototype helmet PET brain imager (Helmet_PET). Methods: In the pilot FDG brain PET imaging experiment, using the small field-of-view (5 cm) wearable Helmet_PET imager [1], dynamic data in 8 subjects were obtained in an upright position in list mode (LM) from the point of injection of ~1-2 mCi of FDG. Each of the subjects followed the repeated two-cycle protocol: walking in place for 30 seconds, stopping for 30 seconds, again walking for 30 seconds, etc, over a period of 6 minutes total. The LM data was reconstructed using MLEM algorithm into twelve 30 sec time bins (24 slices per time frame). The reconstructed spatial resolution was 2x2 mm2 (100 by 100 pixels) with a slice thickness also of 2 mm. Two additional subjects aside from the 8 original subjects were imaged with the imager low enough to capture the carotid arteries in the FOV in LM format from the point of injection to obtain image derived blood input function. These subjects also underwent whole brain PET/CT imaging. The carotid arteries were first identified in representative MR images and then located in corresponding PET slices at the early time points (first 30 second frame) when FDG is in the blood and a region of interest (ROI) drawn to obtain arterial input function (blood time activity curve (TAC)). In addition, an ROI was drawn in the surrounding tissue in the 6 minute time frame to obtain tissue TAC. A model corrected blood input function (MCIF) was generated using a dual output model, wherein, the blood input function corrected for spill-over and partial volume averaging was obtained. The MCIF was then used to compute rate of FDG uptake (Ki) for average TAC obtained from 8 subjects for a region of interest (leg motor cortex) using a FDG transport model correcting for spill-out from the blood and also radioactivity recovery of the tissue of interest [2] . The walking and standing time frames (alternate frames) were also isolated for each of the 8-subjects imaged and average TACs generated. The average rate of FDG uptake for walking in place (Ki-walking) and standing in place (Ki-standing) with standard error was computed. Results: The Ki for the tissue of interest (leg-related motor cortex) averaged over 8-subjects was computed to be 0.13 ml/min/g. The average Ki-standing and Ki-walking was computed to be 0.12±0.014 ml/min/g and 0.14±0.021 ml/min/g respectively (p=0.57). Conclusions: Summarizing this pilot Helmet_PET experiment for a walking-in-place task, the data trend of the two-mode motion protocol suggests that even with a bolus injection, task-related motor cortex activity can be differentiated using the changing rate of FDG uptake. In the future, superior results may be obtained using a continuous infusion ligand delivery, as has been utilized in other FDG studies. Grant support: NIH R24 MH106057 (JBL); Radiology start-up at UVA (BK).