TY - JOUR T1 - Feasibility study of factor analysis and kinetic modeling for spatial discrimination of rapid dual-tracer PET JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 2032 LP - 2032 VL - 52 IS - supplement 1 AU - Xiaoyin Cheng AU - Kuangyu Shi AU - Rebekka Kraus AU - Gaspar Delso AU - Sibylle Ziegler Y1 - 2011/05/01 UR - http://jnm.snmjournals.org/content/52/supplement_1/2032.abstract N2 - 2032 Objectives Rapid dual-tracer PET imaging has the advantage of providing complementary physiological information. One key step is to separate tracer time-activity curves (TACs) from the measured signals. Although factor analysis and kinetic modeling have been applied, their ability in discriminating spatially overlapping tracer uptake is still not clear. By designing a simulation framework, this study aims to evaluate the feasibility of these two methods for the differentiation of two highly spatially overlapping tracers [18F]FDG and [18F]FMISO. Methods 2D digital phantoms of 5x5 cm2 tumors with rapid [18F]F-FDG and [18F]FMISO were simulated by solving reaction-diffusion equations on an artificial tumor microenvironment, including various structures of hypoxia, normoxia and necrosis. Dynamic scans for 70 min (139 frames, pixel 1x1 mm2) were modeled and [18F]FMISO was administrated 5 min after [18F]FDG injection. Factor analysis was implemented according to [1,2]. The whole image set was decomposed into 3 factors and 3 factor images, taking into account positive constraints and oblique rotations. A dual-tracer kinetic model based on irreversible two-tissue compartment model [3] was used to compute parametric images. Both factor and parametric images were analyzed and compared to true single tracer images. Results 9 phantoms with different spatially overlapping dual-tracer were considered. Using factor analysis, 8 FMISO and 3 FDG areas were separated and shown in different factor images. It was not possible to further differentiate hypoxia and necrosis. In parametric images, FMISO areas were differentiable from the background, while FDG areas were hard to detect. In particular, necrosis was distinguishable in 8 phantoms. Conclusions Phantom study results show that factor analysis performs better on separating spatially overlapping FDG and FMISO. Nevertheless, kinetic modeling has the ability to distinguish necrosis, which is not possible for factor analysis ER -