An Automatic Threshold-Based Scaling Method for Enhancing the Usefulness of Tc-HMPAO SPECT in the Diagnosis of Alzheimer's Disease

Functional imaging of the brain can aid in the diagnosis of Alzheimer’s disease. Tc-HMPAO SPECT is widely available and relatively inexpensive to use. Combined with computer-based analysis of images, SPECT is a powerful tool in detecting decreases in brain perfusion caused by Alzheimer’s disease. However, analysis can falsely elevate the perfusion of normal areas and diminish the perfusion of atrophic areas in the Alzheimer’s brain when used with conventional scaling methods. In this paper, we present a technique for scaling images that overcomes the problems associated with conventional scaling methods. Our technique was successful in eliminating or attenuating the false increases in perfusion shown in probable Alzheimer’s patients in over 90% of cases (n=17), and in enhancing the sensitivity of detection of degenerative changes by Statistical Parametric Mapping.