Abstract
Recent development of the 133Xe inhalation technique has made it possible to measure cerebral blood flow (CBF) noninvasively. Recording of the head curves from the frontal and temporal areas during inhalation of 133Xe, however, is contaminated by the artifact from the air passages. A method based on Fourier transforms was reported to be able to eliminate air passage artifact (APA) effectively. However, it was pointed out that such an algorithm does not give a complete correction if the artifact seen by the head detectors differs in shape from that recorded from the airways at the mouth, which may happen when there is a slow isotope convection in the nasal and sinus cavities. The purpose of this study was to compare the CBF values calculated by the Fourier method with those by the conventional method of Obrist (VM method). Mean hemispheric gray matter flow (F1) calculated by the VM method in 11 subjects, including normal volunteers and patients with various neurological diseases, was 69.2 ± 13.2 mg/100 g brain/min, whereas F1 calculated by the Fourier method in the same subjects was 64.4 ± 13.5, indicating that APA can be effectively eliminated by the Fourier method.
The F1 values calculated by the Fourier method from the frontal and temporal regions were relatively high, and closer to the F1 values calculated by the VM method. The size of the APA was large in these regions. It was concluded that the deformed APA contaminated the results in these regions, and could not be eliminated effectively by the Fourier method. It is suggested that the shape of the head curve and the size of APA should be carefully examined to ensure that CBF data are reliable.
