Simultaneous Dual Tracer PET Using Generalized Factor Analysis of Dynamic Sequences

Simultaneous dual tracer PET imaging has potential for simultaneous assessment of two physiological processes under identical conditions, but it has not been successfully achieved in the clinical setting because of the lack of potential for energy discrimination. We developed a new approach based on generalized factor analysis of dynamic sequences (GFADS) that exploits temporal differences between radiotracers and applied it to the case of simultaneous imaging of brain metabolism and dopamine transporter density (DAT) using 18FDG and 18FECNT. Monte Carlo simulations of both studies were performed in a population of anthropomorphic phantoms using 5 different specific and non-specific DAT time activity curves (TAC) previously reported in human and monkey studies. 3D acquisitions were simulated while modeling random and scatter coincidences as well as pixelated block detectors, light sharing among crystals elements and dead-time in the PET/CT scanner. Activity distributions of 18FDG as well as specific and non-specific 18FECNT-DAT were simulated separately and 25 dynamic simultaneous dual tracer studies were generated. We found very good agreement between the estimated GFADS factors and the simulated reference TACs, and between the GFADS factor images and the corresponding reference activity distributions with errors less than 7.3 plusmn 1.3%. Biases in estimation of specific DAT binding and relative metabolism activity were within 5.9 plusmn 3.6% compared to the reference values.