Click here to


Are you sure ?

Yes, do it No, cancel

Variation of Tracer Kinetics and Cerebral Metabolic Rate During Dynamic 18FDG PET Acquisition

S Deng1*, J Vasquez1, P Fox1,2, G Clarke1,2, (1) Research Imaging Institute, (2) Dept of Radiology, University of Texas Health Science Center at San Antonio, San Antonio TX


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

The cerebral metabolic rate of glucose (CMRGlc) is traditionally estimated via kinetic modeling of 18F-FDG uptake assuming metabolic steady-state on the whole time series. In contrast, recent functional PET results are indicating indicate non-stational glucose metabolism, even at rest. Aims: (i) to compare CMRGlc calculated in different time-windows during the same acquisition, (ii) investigate the consistency of CMRGlc variation across subjects.

6 healthy young subjects (over-night fasting) received PET scans (ECAT HR+) at rest, with an arterial catheter placed in the radial artery for quantification purposes. Fasting plasma glucose was measured before and between scans. After I.V. injection of 5mCi(±10%) FDG, images were acquired in 3D-mode and 1-min frame from 1 min to 70 min. CMRGlc of 0-30min, 15-45min, 30-60min and 45-60min in the dynamic acquisition were calculated with convention FDG kinetic models, with lump constant of 0.52. The coefficients of variation (CoV) of inter-window, inter-subject, and inter-kinetic-models were calculated to quantify the variation of metabolic rate and kinetic coefficients during the dynamic acquisitions.

For 2-compartment model (2C), CMRGlc calculated with the whole time-series (24.99±1.8 uMol/100g/min) is different with CMRGlc in 0-30min (29.6±4.1), 15-45min (27.5±2.1), 30-60min (23.2±1.4) and 45-60min (21.4±2.67) (p=5.6E-06). For 3 compartment model(3C), similar variation is identified: 31.0±4.6 for whole time series, 31.7±3.6 for 0-30min, 29.6±3.4 for 15-45min, 27.8±3.1 for 30-60min and 25.8±2.8 for 45-60min(p=5.19E-04). For CMRGlc, inter-window CoV is lower than inter-subject CoV for 3C model (10.20% vs 14.8%), while higher for 2C model(14.46% vs 7.59%). For kinetic parameters, inter-window and inter-subject CoV are all higher than that of CMRGlc.

Within-scan variation of CMRGlc has been identified, consistent across subjects. Given that the CoV of kinetic coefficients is higher than that of CMRGlc, and that within-subject inter-window CMRGlcs are highly correlated, the physiological origin of CMRGlc variation during scans should be consideredevaluated.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email