Viability of the cerebral parenchyma is dependent on cerebral blood flow (CBF). The assessment of cerebral perfusion in patients with acute stroke, in a clinically relevant time frame, could be of utmost importance for patient selection before thrombolytic therapy. In individual patients, quantitative mapping of CBF to indicate the severity and potential reversibility of neuronal damage can be used to predict which brain tissue will be salvaged with reperfusion or die without it (penumbra), as well as which brain tissue is already infarcted. Recent investigations of perfusion CT have shown major advances in the assessment of acute stroke patients. Perfusion CT offers a number of practical advantages over other cerebral perfusion imaging methods as it can be performed immediately after unenhanced CT, and used, in general, to exclude cerebral haemorrhage. It is fast (typical procedure time <5 min) and does not require specialized computer hardware. The accuracy of cerebral perfusion maps has been demonstrated for normal and decreased CBF value by comparison with xenon CT used as a gold standard. Perfusion CT infarct and penumbra maps provide a potential recuperation ratio (PRR) (or Lausanne Stroke Index), defined as PRR = penumbra/ (penumbra + infarct). This index is correlated with the improvement of the National Institutes of Health Stroke Scale (NIHSS) in case of arterial recanalization. Also, the size of the ischaemic area (infarct + penumbra) is correlated with the NIHSS score on hospital admission. Further studies may demonstrate the use of perfusion CT for the assessment of penumbra dynamics in function-specific brain areas. Perfusion CT is now ready to be used in clinical trials as a decision-making tool to tailor more precisely the thrombolytic therapy to the individual patient.