BACKGROUND AND PURPOSE: An accurate measure of the severity of ischemic insult and the resulting prognosis is needed to assess the effectiveness of new treatments for acute stroke. We studied the reproducibility and accuracy of measurements of infarct volume with MRI and correlated the measurements with outcome. METHODS: Infarct volume was measured on T2-weighted images with the Analyze image analysis software. This technique was found to be highly accurate and reproducible. RESULTS: Measurements of infarct volume were found to be highly accurate and reproducible. Twenty-one patients (mean age, 66.5 years; range, 28 to 90 years) with cortical middle cerebral artery territory infarcts in whom adequate data could be obtained were studied within 72 hours from onset (mean delay to MRI, 27.5 hours; range, 5 to 72 hours). The Scandinavian Stroke Scale was used to calculate a prognostic score, and clinical outcome was assessed at 3 months. Infarct volume was found to significantly predict outcome. Mean infarct volume in the independent patients was 35.7 +/- 29.7 cm3 compared with 88.3 +/- 71.3 cm3 in dependent patients and 166.5 +/- 65.9 cm3 in dead patients (F = 10.52, P < .001). Patients with an initial infarct volume less than 80 cm3 were found to have a better outcome than those with larger infarct volumes. Secondary hemorrhage visible on MRI also predicted a poor outcome. In contrast, the Scandinavian Stroke Scale did not significantly predict outcome. CONCLUSIONS: The results demonstrate that measurement of the size of middle cerebral artery infarction with MRI is a useful tool in assessing prognosis and will have a valuable role in assessing new therapeutic agents.
Saunders et al. (1995) studied this question.