Using xenon-enhanced computed tomography for the study of cerebral blood flow, simultaneous measurements of end-tidal and arterial blood xenon concentrations using the blood collection method were performed to investigate the validity of substituting the end-tidal for the arterial blood xenon concentration. Simultaneous measurement by both methods was performed 68 times in 27 patients. There was no statistical correlation between the arterial blood accumulation rate constant obtained by arterial blood and end-tidal samples, nor between the arterial blood saturation value obtained by the two methods, even when correction was made for age. In brain tissue, all parameters calculated using the end-tidal concentration were lower than those using arterial blood. We therefore suggest that cerebral blood flow values calculated using end-tidal xenon concentration are useful only for qualitative cerebral blood flow mapping, and not applicable to absolute values of cerebral blood flow.