Noninvasive method for measurement of cerebral blood flow using O-15 water PET/MRI with ASL correlation

Purpose: A noninvasive image derived input function (IDIF) method was applied to estimate arterial input function from brain (H2O)-O-15-PET/MRI images for the measurement of cerebral blood flow (CBF) because of difficulty in arterial blood sampling during PET/MRI scans. To evaluate accuracy and reproducibility of radioactivity in the internal carotid arteries (ICA) for the IDIF method, a new phantom using a skull bone was applied in the cross-calibration process between the scanner and a gamma-well counter. Methods: Eleven healthy volunteers (9 males, 43.9 +/- 10.9y) underwent PET/MRI studies with a 3-min (H2O)-O-15-PET and several MRI scans including arterial spin labeling (ASL) perfusion MRI. PET images were reconstructed as dynamic data using two sets of reconstruction parameters, which were determined by basic assessment of radioactivity concentration reproducibility in the tubes of the phantom. The IDIF method extracted the time-activity curves of the ICA from several image slices in the PET data. CBF images were calculated using the autoradiographic (ARG) method and a one-tissue compartment model (1-TCM). Results: The global means of CBF from the ARG, 1-TCM, and ASL-MRI were 44.8 +/- 4.3, 47.9 +/- 5.9 and 57.9 +/- 8.6 (mL/min/100 g), respectively. CBF from ASL-MRI was significantly greater compared with CBF from (H2O)-O-15-PET (P < 0.001). However, these CBF values were significantly correlated with each other in the scatter plots (P < 0.05). Conclusions: Noninvasive measurement of CBF using (H2O)-O-15-PET/MRI and IDIF with the cross-calibration method with a skull phantom experiment provided reasonable quantitative values. The IDIF method allowed reliable estimation of arterial radioactivity concentration, which is useful for clinical application. The ASL-MRI perfusion image from the simultaneous acquisition tended to overestimate CBF.