Blood sampling devices and measurements.

Quantitative positron emission tomography requires the determination of the tracer concentration in arterial plasma or full blood as a function of time. This defines the experimental input function. The model prediction, with which the positron camera regional time-activity curves are compared, is given by the experimental input function convoluted with the model. This paper reviews different strategies for determining the input function, invasive techniques, such as manual blood sampling and the use of automated blood sampling systems, and non-invasive techniques. The importance of corrections is discussed, such as accurate cross-calibrations of the different detectors used in quantitative PET and corrections for differences in time phase between the regional PET time-activity curves and the input function. We also report on the use of a PET system in non-invasive determinations of the input function. By imaging the neck region the time-activity curve of the carotid arteries can be obtained. The PET time-activity curves of the carotid arteries are in good agreement with a conventional experimental input function determined from the radial artery with an automated blood sampling system. However, PET time-activity curves of the radial arteries can not be used without a deconvolution since the resistance in the intact radial artery causes dispersion compared to the input function obtained by invasive methods.