Semiautomated volumetry of the cerebrum, cerebellum-brain stem, and temporal lobe on brain magnetic resonance images

Purpose The aim of this study was to develop an automated method of segmenting the cerebrum, cerebellum-brain stem, and temporal lobe simultaneously on magnetic resonance (MR) images. Methods and materials We obtained T1-weighted MR images from 10 normal subjects and 19 patients with brain atrophy. To perform automated volumetry from MR images, we performed the following three steps: (1) segmentation of the brain region; (2) separation between the cerebrum and the cerebellum-brain stem; and (3) segmentation of the temporal lobe. Evaluation was based on the correctly recognized region (CRR) (i.e., the region recognized by both the automated and manual methods). Results The mean CRRs of the normal and atrophic brains were 98.2% and 97.9% for the cerebrum, 87.9% and 88.5% for the cerebellum-brain stem, and 76.9% and 85.8% for the temporal lobe, respectively. Conclusion We introduce an automated volumetric method for the cerebrum, cerebellum-brain stem, and temporal lobe on brain MR images. Our method can be applied to not only the normal brain but also the atrophic brain.