The proposed studies will expand the kinetic modeling of three receptor- binding radiopharmaceuticals: [99mTc]galactosyl-neoglycoalbumin ([99mTc]NGA), [68Ga]deferoxamine-galactosyl-neoglycoalbumin ([68Ga]Df-NGA), and high-density-[99mTc]-galactosyl-neoglycoalbumin ([99mTc]HD-NGA). Each agent is specific for the same receptor, hepatic binding protein, but was designed for a specific modality, and hence, different model structures: [99mTc]NGA, planar imaging and global measurement of receptor biochemistry; [68Ga]Df-NGA, PET imaging and regional measurement of receptor biochemistry; and [99mTc]HD-NGA, SPECT imaging and regional measurement of hepatic plasma flow. We will extend the current kinetic model to accommodate the following structures: A) Account for regional distribution within the liver for hepatic plasma flow F, as well as Ro and ko. Dynamic PET data will be used. B) Account for regional distribution of hepatic plasma flow. A combination of dynamic planar imaging with a single SPECT image will be used. Additionally, we will C) employ signal detection theory to test the clinical accuracy of a new generation of software for the current [99mTc]NGA kinetic model. We will then test the following three hypotheses: A) the dynamic PET model provides accurate and precise measurements of regional F, Ro, and ko; B) the SPECT model provides accurate and precise measurements of regional hepatic plasma flow; and C) higher simulation speed, an appropriate model transform, and a more robust optimizer will provide higher clinical accuracy then the current TcNGA model. Model testing will include the following criteria: a) Goodness-of-fit, b) kinetic sensitivity, c) local identifiability, and d) plausibility. The latter will include the comparison of each model parameter with independently measured values. For the regional models (PET and SPECT), these measurements will include global and regional hepatic plasma flow via indocyanine green extraction and [3H]D-galactose deposition, and regional receptor density and forward binding rate constant measurements via in vitro assay of tissue samples. Testing for clinical accuracy will employ ROC analysis of TcNGA patient imaging data using different optimizers and model transformations as observers. The TcNGA radiopharmacokinetic system displayed kinetic sensitivity to receptor biochemistry and the kinetic model met the criteria of Goodness- of-fit, local identifiability, and plausibility. Consequently, dynamic PET imaging with [68Ga]Df-NGA will permit us to construct a testable radiopharmacokinetic system for regional measurements of receptor density and affinity. Regional quantification of organ function via single photon emission computed tomography (SPECT) has remained an elusive goal. The principal reason for this is a lack of radiotracers with physical and biological properties which are compatible with the requirements of rotating tomographic systems. As the only true single photon-emitting chemical microsphere, [99mTc]HD-NGA will permit us to test SPECT as a method for the absolute measurement of regional hepatic plasma flow. Our proposal incorporates al of the components required to complete a system for quantitative imaging of physiologic process: the radiotracer, biochemistry, tomograph, and image processing algorithms. Failure to treat SPECT as one of many components within a radiopharmacokinetic system will relegate this instrument to morphologic measurements of tissue sizes and shapes.
