Noninvasive quantitation of cerebral blood flow using oxygen-15-water and a dual-PET system.

UNLABELLED Measurement of the arterial input function is essential for quantitative assessment of physiological function in vivo using PET. However, frequent arterial blood sampling is invasive and labor intensive. Recently, a PET system has been developed that consists of two independent PET tomographs for simultaneously scanning the brain and heart, which should avoid the need for arterial blood sampling. The aim of this study was to validate noninvasive quantitation with this system for 15O-labeled compounds. METHODS Twelve healthy volunteers underwent a series of PET studies after C15O inhalation and intravenous H2(15)O administration using a Headtome-V-Dual tomograph (Shimadzu Corp., Kyoto, Japan). The C15O study provided gated blood-pool images of the heart simultaneously with quantitative static blood-volume images of both the brain and heart. Weighted-integrated H2(15)O sinograms were acquired for estimating rate constant (K1) and distribution-volume (Vd) images in the brain, in addition to single-frame sinograms for estimating autoradiographic cerebral blood flow images. Noninvasive arterial input functions were determined from the heart scanner (left ventricular chamber) according to a previously developed model and compared directly to invasive input functions measured with an on-line beta probe in six subjects. RESULTS The noninvasive input functions derived from this PET system were in good agreement with those obtained by continuous arterial blood sampling in all six subjects. There was good agreement between quantitative values obtained noninvasively and those using the invasive input function: average autoradiographic regional cerebral blood flow was 0.412 +/- 0.058 and 0.426 +/- 0.062 ml/min/g, K1 of H2(15)O was 0.416 +/- 0.073 and 0.420 +/- 0.067 ml/min/ml and Vd of H2(15)O was 0.800 +/- 0.080 and 0.830 +/- 0.070 ml/ml for the noninvasive and invasive input functions, respectively. In addition to the brain functional parameters, the system also simultaneously provided cardiac function such as regional myocardial blood flow (0.84 +/- 0.19 ml/min/g), left ventricular volume (132 +/- 22 mm at end diastole and 45 +/- 14 ml at end systole) and ejection fraction (66% +/- 5%). CONCLUSION This PET system allows noninvasive quantitation in both the brain and heart simultaneously without arterial cannulation, and may prove useful in clinical research.