We studied 24 patients up to 51 hours after ischemic stroke using 18F-fluoromisonidazole positron emission tomography to determine the fate of hypoxic tissue likely to represent the ischemic penumbra. Areas of hypoxic tissue were detected on positron emission tomography in 15 patients, and computed tomography was available in 12 patients, allowing comparison with the infarct volume to determine the proportions of the hypoxic tissue volume that infarcted and survived. The proportion of patients with hypoxic tissue and the amount of hypoxic tissue detected declined with time. On average, 45% of the total hypoxic tissue volume survived and 55% infarcted. Up to 68% (mean, 17.5%) of the infarct volume was initially hypoxic. Most of the tissue "initially affected" proceeded to infarction. We correlated hypoxic tissue volumes with neurological and functional outcome assessed using the National Institutes of Health Stroke Scale, Barthel Index, and Rankin Score. Initial stroke severity correlated significantly with the "initially affected" volume, neurological deterioration during the first week after stroke with the proportion of the "initially affected" volume that infarcted, and functional outcome with the infarct volume. Significant reductions in the size of the infarct and improved clinical outcomes might be achieved if hypoxic tissue can be rescued.