Regional cerebral blood flow measurement with intravenous [15O]water bolus and [18F]fluoromethane inhalation.

In 20 patients with ischemic cerebrovascular disease, classic migraine, or angiomas, we compared paired dynamic positron emission tomographic measurements of regional cerebral blood flow using both [15O]water and [18F]fluoromethane as tracers. Cerebral blood flow was also determined according to the autoradiographic technique with a bolus injection of [15O]water. There were reasonable overall correlations between dynamic [15O]water and [18F]fluoromethane values for cerebral blood flow (r = 0.82) and between dynamic and autoradiographic [15O]water values for cerebral blood flow (r = 0.83). We found a close correspondence between abnormal pathologic findings and visually evaluated cerebral blood flow tomograms obtained with the two tracers. On average, dynamic [15O]water cerebral blood flow was 6% lower than that measured with [18F]fluoromethane. There also was a general trend toward a greater underestimation with [15O]water in high-flow areas, particularly in hyperemic areas, probably due to incomplete first-pass extraction of [15O]water. Underestimation was not detected in low-flow areas or in the cerebellum. Absolute cerebral blood flow values were less closely correlated between tracers and techniques than cerebral blood flow patterns. The variability of the relation between absolute flow values was probably caused by confounding effects of the variation in the circulatory delay time. The autoradiographic technique was most sensitive to this type error.