The aim of this study was to define the accuracy of 99mTc-ethyl cysteinate dimer-single photon emission computed tomography (99mTc-ECD-SPECT) in distinguishing transient ischemic attack from completed ischemic stroke at early stages after the onset of symptoms. In a prospective study we examined 82 patients within 6 hours after the onset of symptoms (neurologic deficit caused by middle cerebral artery ischemia) using both 99mTc-ECD-SPECT and computed tomography (CT). The follow-up was based on Scandinavian Stroke Scale (SSS) 24 hours and 5-7 days, as well as on CT 7 days, after the event. SPECT evaluation was performed both visually and using semiquantitative region-of-interest (ROI) analysis. According to visual SPECT analysis, on admission 59 of 82 patients had activity deficits in the symptomatic hemisphere. After 7 days, all these patients had neurologic symptoms (SSS 28 +/- 12 points), caused by a cerebral infarction as evidenced with CT. Twenty-three of 82 patients displayed no early activity deficit despite clinical symptoms. None of these patients had neurologic symptoms after 7 days (indicating transient ischemic attack or prolonged reversible ischemic neurologic deficit). In the semiquantitative SPECT analysis, all patients had abnormal count densities in the respective ROI (activity < 90% compared with the contralateral side). All patients with transient ischemia (n = 23) had count rate densities more than 70% of the respective contralateral ROI, whereas all patients with subsequent infarction (n = 59) had values < 70%. Use of 99mTc-ECD-SPECT allows transient ischemia to be distinguished from ischemic infarction using relative regional activity thresholds within the first 6 hours after onset of symptoms.