We retrospectively examined regional cerebral blood flow values in 78 patients in the stabilized phase of stroke to evaluate the consistency of lateralization compared with computed tomographic imaging. Examinations were performed using the xenon-133 inhalation technique, and the data were processed and displayed by a computer-assisted system of our own design that allows statistical analysis in real time and is suitable for clinical use. The consistency of lateralization correlated with computed tomographic findings was tested for absolute values, percent distribution, and asymmetries. The latter yielded the highest degree of sensitivity (in 83.3% of the patients the affected hemisphere was correctly identified). Absolute values showed at least one hypoperfused area in the affected hemisphere in 48.7% of the patients, and percent distribution did so in 57.7%. Furthermore, the combination of the three methods of data analysis yielded a sensitivity of 91%; the remaining 9% of the patients, however, exhibited some abnormalities of regional cerebral blood flow (bilateral or contralateral hypoperfusion). Although good agreement was found for lateralization by computed tomography and by the combined use of the three methods of analysis, a complete overlap between functional and structural examinations should not be expected. Problems concerning the sensitivity of the xenon-133 inhalation technique under conditions of low blood flow and the statistical analysis of regional cerebral blood flow data are discussed.