Quantification of serotonin transporters in nonhuman primates using [(123)I]ADAM and SPECT

We reported recently a highly selective radioligand, 2-([2-([dimethylamino]methyl)phenyl]thio)-5-[(123)I]iodophenylamine (ADAM), for SPECT imaging of serotonin transporters (SERT). In this article we describe the kinetic modeling of [(123)I]ADAM and its ability to quantitatively and reproducibly measure the concentrations of SERT in the nonhuman primate brain. We also investigate simplified models of tracer behavior that do not require invasive arterial blood sampling.

Three female baboons each underwent 3 [(123)I]ADAM SPECT studies. The studies consisted of a dynamic sequence of seventy-two 5-min scans after injection of 330 +/- 50 MBq (mean +/- SD) [(123)I]ADAM. Rapid arterial blood samples were obtained and corrected for the presence of labeled metabolites. Dynamic imaging and metabolite-corrected plasma data were analyzed using graphic analysis to give the distribution volumes (DVs) of different brain regions. DV ratios (DVRs) of target to cerebellum were derived and compared against a kinetic reference tissue model and simple target-to-background ratio.

Averaged over all 9 scans, the mean DV in the midbrain was 4.86 +/- 1.06 mL/mL and the mean DV in the cerebellum was 2.25 +/- 0.48 mL/mL. The mean test-retest repeatability of the midbrain DV was 14.5%. The reference tissue model gave a mean midbrain DVR of 2.01 +/- 0.17 and correlated strongly with the DVR calculated from the full kinetic model (correlation coefficient [R(2)] = 0.94; P < 0.001), but with much improved repeatability (test-retest, 5.4%; intersubject variability, 5.2%). Similarly, the simple ratio method gave strong correlations with the full kinetic model (R(2) = 0.89; P < 0.001) and a test-retest of 7.6%.

Accurate, repeatable quantification of SERT in the nonhuman primate brain is possible using kinetic modeling of dynamic [(123)I]ADAM SPECT scans. Simplified models, which do not require arterial blood sampling, gave accurate results that correlated strongly with the full kinetic model. The test-retest reliability of the simplified reference region models was excellent. Quantification of SERT is possible using full kinetic modeling and also with simpler reference region methods.