Fully automatic hippocampus segmentation discriminates between early Alzheimer’s disease and normal aging

The hippocampus is among the first structures affected in Alzheimer's disease (AD); hippocampal MRI volumetry is a potential biomarker for AD but is hindered by the limitations of manual segmentation. We propose a fully automatic method using probabilistic and anatomical priors for hippocampus segmentation. Probabilistic information is derived from 16 young controls and anatomical knowledge is modeled with automatically detected landmarks. The results were compared with manual segmentation on data from 16 young healthy controls, with a leave-one-out strategy, and 8 AD patients. High accuracy was found for both groups (volume error 6% and 7%, overlap 87% and 86%, respectively). The resulting volumes were used to discriminate between 25 elderly subjects, 25 early AD patients and 24 patients with amnestic mild cognitive impairment (MCI). The classification proved accurate with 87% of the AD patients and 74% of the MCI patients correctly classified with respect to the elderly controls.