MRI measurements identify Alzheimer's disease

.
Alzheimer's Donation
Donate Online Now
.

To date, studies attempting to discriminate patients with Alzheimer's disease (AD) from healthy controls on the basis of brain MRI scans have reported a wide range of accuracy rates. Lerch et al. have suggested that in some cases poor experimental design, small sample sizes and failure to use independent validation cohorts have confounded these results. In response, they have evaluated the capacity of fully automated MRI measurements of cortical thickness to discriminate patients diagnosed with probable AD (n = 19) from healthy elderly controls (n = 17).
The discriminatory ability of the cortical thickness measurements varied depending on the region of the brain under scrutiny. Areas of the brain classically associated with AD had the best predictive capacity (>90% accuracy for the medial temporal lobes and other limbic structures), whereas the primary motor cortex had the poorest accuracy. Multivariate discriminant analysis of 25 cortical structures revealed 100% accuracy with 6 different 2-structure combinations, all of which included the parahippocampal gyrus. All validation was performed internally, using a jack-knife or leave-one-out approach to prevent overtraining of the discriminant models.
The authors conclude that automated MRI measurements of cortical thickness could be used to improve the clinical diagnosis of probable AD. They suggest that further evaluation of advanced computational techniques applied to MRI in patients with mild cognitive impairment or very early AD would enable imaging to be used as a diagnostic tool across the clinical spectrum of dementia.
Lerch JP et al. (2006) Automated cortical thickness measurements from MRI can accurately separate Alzheimer's patients from normal elderly controls. Neurobiol Aging [doi: 10.1016/j.neurobiolaging.2006.09.013]