Résumé
PURPOSE: This study was conducted to examine the efficacies of susceptibility weighted images (SWI) for predicting the clinical prognosis of comatose patients following cardiac arrest. METHODS: Thirty-two patients who were resuscitated from cardiac arrest and underwent brain magnetic resonance imaging (MRI) were retrospectively investigated and compared to 32 subjects with normal brain MRI findings who served as controls. The SWI readings were divided into three categories: prominent, diminished, and normal. Comatose patients were divided into two groups: those with a Glasgow-Pittsburgh cerebral performance category (CPC) of 1-2 (good outcome group) and those with a CPC of 3-5 (poor outcome group). RESULTS: Of the 32 patients, 17 (53.1%) showed good neurological outcomes upon hospital discharge. Normal patterns on SWI were mainly seen in the good outcome group (15 patients, 88.2%), while diminished patterns and prominent patterns were frequently found in the poor outcome group (13 patients, 88.7%). The combination of diminished pattern and prominent pattern predicted poor outcome with 86.7% sensitivity (95% confidence interval, 69.5%-100%) and 88.2% specificity (95% confidence interval, 72.9%-100%). CONCLUSION: The SWI findings correlate with the outcome of hypoxic-ischemic encephalopathy and may be a useful adjunct of vegetative state or death in comatose patients after cardiac arrest.