BACKGROUND: MRI is believed to be more sensitive than CT scanning for imaging ischemic stroke, but is considered insufficiently sensitive for acute hemorrhagic stroke.
METHODS: In this single-center study at a suburban hospital in Bethesda, MD, and coordinated at the National Institutes of Health, 356 patients with suspected acute strokes underwent MRI and CT scanning of the brain. The imaging studies were later reviewed by a panel of two neuroradiologists and two stroke neurologists blinded to the patients' clinical findings.
RESULTS: The final diagnosis was acute ischemic stroke in 190 patients (53%), acute intracranial hemorrhage in 27 (8%) and transient ischemic attack in 14 percent. Twenty-five percent of the patients were not ultimately diagnosed with cerebrovascular disease. The sensitivity of CT scanning was 26% for acute stroke overall, 16% for acute ischemic stroke and 89% for acute intracranial hemorrhage (corresponding specificities 98%, 98% and 100%, respectively). The sensitivity of MRI was 83% for acute stroke overall, 83% for acute ischemic stroke and 81% for acute intracranial hemorrhage (corresponding specificities, 97%, 96% and 100%, respectively). In the subgroup of patients imaged within three hours of symptom onset, sensitivities for acute stroke were 27% for CT scanning and 76% for MRI, and specificities were 100% and 96%, respectively.
CONCLUSIONS: Although the authors suggest that their findings are consistent with a superiority of MRI over CT scanning for the imaging of acute stroke, MRI missed more patients with acute hemorrhagic stroke, which could have significant implications when selecting patients for thrombolytic therapy.--