The discriminant value of semiquantitative SPECT data in mild Alzheimer's disease.

UNLABELLED In most semiquantitative SPECT studies, overlap between groups of patients with Alzheimer's disease (AD) and age-matched elderly controls is such that single posterior cortical perfusion measurements lack sensitivity. In the present study, the value of a combination of semiquantitative temporoparietal SPECT parameters was examined. METHODS Supratentorial transaxial perfusion measurements were obtained in frontal, anterior temporal, posterior temporoparietal and occipital cortical areas in both hemispheres, in a baseline population of 10 healthy elderly controls and 30 mild to moderately impaired AD patients, as well as in a prospective group of 15 patients with mild cognitive impairment, 12 patients with a diagnosis of probable AD and individual cases of multi-infarct dementia, dementia-frontal type and paranoid psychosis. A linear discriminant function (LDF) was calculated from the baseline subjects' data to classify control and AD subjects individually. RESULTS Highly significant hypoperfusion was noted in both the anterior temporal and posterior temporoparietal regions of interest in the AD group compared with controls, but with significant overlap. Using an LDF incorporating these perfusion measurements in both hemispheres, 10/10 (100%) controls and 26/30 (87%) AD baseline subjects were correctly classified. Using the baseline LDF in the prospective 15 mildly impaired cases, 11/12 new mild AD cases and none of the 3 non-AD cases were classified in the AD group. CONCLUSION These results support the use of a combination of semiquantitative SPECT perfusion estimates from cortical areas with predictable pathological involvement in AD in a linear discriminant format in the clinical assessment of patients with suspected AD.