Use of computerized CT analysis to discriminate between Alzheimer patients and normal control subjects

A newly developed computerized technique was used to analyze the CT scans of 49 patients with dementia of the Alzheimer type and 31 normal control subjects. Nine brain regions distributed across five CT slices were evaluated for each individual. For the purpose of analysis, the patients and controls were divided into an exploratory set and a test set. Several discriminant functions were conducted on the exploratory set and applied to the test set. The combination of variables that focused on regions in the temporal lobe was most accurate in differentiating Alzheimer patients from controls (94%). This degree of accuracy was achieved only when subjects younger than 65 years old were analyzed separately from those 65 years old and older. The newly developed computer software program was able to discriminate between independently selected groups of Alzheimer patients and control subjects. The program was most effective when the analysis emphasized regions in the temporal lobe and when subjects younger than 65 years old were analyzed separately from those 65 years old and older.