Quantitative computer tomography in Alzheimer's disease: a re-evaluation

This investigation addresses the question of usefulness of computer-tomographic (CT) scanning in discriminating patients with Alzheimer's disease (AD) from healthy, aged controls. Quantitative measures of brain volume loss found to be significant by other investigators as well as additional unique variables are used to discriminate 58 longitudinally studied patients meeting NINCDS-ADRDA criteria for the clinical diagnosis of probable AD from 59 controls. The sensitivity and specificity of both single CT scan parameters and multivariate models comprised of such CT parameters are explored. Reasons for diagnostic misclassification are also illuminated.