The limitations of 2-dimensional isotope techniques in the study of focal cerebral ischemia were investigated using the intra-carotid 133 xenon injection method and a 254 multidetector scintillation camera. To make sure that the detectors "look" directly on infarcted areas, only patients with infarcts involving cortical surface structures were included in the study. Eleven such patients were found among 43 consecutive patients with completed stroke, all investigated with CT-scan. The blood supply to the infarcted areas was evaluated using 3 different approaches: 1) The first minute washout of 133 xenon (rCBF), 2) the initial distribution of isotope during the first 5 sec and 3) the cumulated counts recorded during 15 min. Compton scatter and the "look through phenomenon" were responsible for the majority of counts recorded from the infarcted areas and the blood flow recorded was found to be grossly overestimated and much more influenced by the blood flow in the surroundings than in the ischemic area itself. However, using the 3 approaches, infarcted areas were always disclosed by our equipment. It is concluded that 2-dimensional isotope technique is not reliable for quantifying focal ischemic lesions. The method should be limited to the qualitative demonstration of the ischemic lesions for which it is fully reliable.