Positron emission tomography (PET) offers an excellent possibility to quantify proteins, such as receptors and transporters, in the living human brain. The gold standard for this quantification procedure is given by the arterial input function with the advantages of fewest modeling assumptions and independence of nonspecific binding.
However, arterial cannulation is a laborious and demanding procedure which includes additional discomfort for the patient. Based on reference regions, several non-invasive modeling techniques have been applied with great success but their usage may be limited due to the underlying assumptions, especially within clinical comparisons and longitudinal designs. Image-derived input functions (IDIF) are extracted directly from PET images, hence, they represent a promising alternative to arterial blood sampling while maintaining independence of a reference region. However, various issues such as a robust definition of the input function and full substitution of manual arterial samples have complicated their application as a tool in clinical routine and research.
Here, we developed a novel combination of image-derived and venous input functions for the quantification of the major inhibitory serotonergic receptor (serotonin-1A, 5-HT1A) with the radioligand [carbonyl-11C]WAY-100635. The implementation of an optimized metabolite analysis showed excellent reproducibility of arterial by venous manual blood samples, enabling full independence of arterial cannulation.
Quantification of 5-HT1A receptor binding potentials (BPP) resulted in strong agreement between the introduced approach and the arterial input function in healthy subjects before (R 2=0.95) and after 8 weeks of hormone replacement therapy (R 2=0.93). Subsequent application of image-derived and venous input functions on a high resolution PET scanner in patients with major depression confirmed a widespread 30% decrease in 5-HT1A receptor binding (BPND and BPP) after electroconvulsive therapy.
To summarize, the combination of image-derived and venous input functions offers quantification of 5-HT1A receptor binding potentials entirely independent of arterial blood sampling. Importantly, reliable results were obtained for different study populations (healthy controls and patients with major depression), treatments (no intervention, pharmacological and electroconvulsive therapy), kinetic modeling approaches (regions of interest- and voxel-based) and PET scanners (standard and high resolution tomographs). The introduced technique offers great potential for application in clinical routine and research protocols and the thorough investigation within different settings encourages for further extension to other radioligands.