A retrospective study of 2,624 patients who underwent CT perfusion for a suspicion of acute ischemic stroke evaluated whether the non-uniform temporal sampling approach can be used to effectively reduce radiation exposure while maintaining a high diagnostic performance in this setting. The distribution of peak positions of the time-concentration curves was evaluated in an artery or in the penumbra area. The results showed a significant age-dependent variance in the contrast bolus arrival time between healthy and ischemic tissue. This reflects a high variability of brain perfusion and non-uniformity in the sampling scheme, which results in a consistent possibility of suboptimal bolus capturing.