Computed tomography versus magnetic resonance imaging of the brain. A collaborative interinstitutional study.

A retrospective study (1983-1984) of magnetic resonance imaging (MRI) and computed tomography (CT) examinations in 471 patients with known pathology in the brain and craniocervical junction was conducted in order to determine the relative efficacy of MRI versus CT. All MRI examinations involved slice thickness greater than 10 mm, and only single-slice single-echo or multislice single-echo imaging techniques were available. These studies were evaluated independently by two neuroradiologists from a panel of six for anatomic abnormalities, lesion contrast, and radiologist's impression. Results, which excluded microadenomas of the pituitary and approximately 9% of studies in which consensus was not achieved by the readers, were as follows: (1) 14% of the studies were positive on MRI but normal on CT; (2) in 55% of the studies, MRI was better than CT; (3) MRI was equal or better than CT in 95% of the studies; and (4) CT was better than MRI in 5% (21/421) of the examinations. There were no patients in this series where CT was positive but MRI missed the abnormality.