The clinical usefulness of routine coronal and sagittal reconstructions in cranial computed tomography.

In order to test a hypothesis that the routine use of one midsagittal and four preselected coronal reconstruction images would disclose lesions unsuspected on standard cranial computed tomography (CT) scans, 204 patient examinations were subjected to such multiplanar reconstruction manipulation; the original CT diagnoses, reconstruction diagnoses, and final clinical diagnoses were compared. The two CT diagnoses agreed in 144 cases. Lesions well shown on the standard CT examination were not seen on the reconstruction images in 23 cases (11% of the total), and reconstruction revealed brainstem or cerebellar atrophy not detected on standard CT in 37 cases (18% of the total). Applications of these findings to daily neuroradiologic practice are suggested.