Semi-automatic ROI placement system for analysis of brain PET images based on elastic model: application to diagnosis of Alzheimer's disease.
UNLABELLED: PET with 18F-fluorodeoxyglucose (FDG) is a useful technique to image cerebral glucose metabolism and to detect patients with Alzheimer's disease in the early stage, in which characteristic temporoparietal hypometabolism is visualized. We have developed a new system, in which the standard brain ROI atlas made of networks of segments is elastically transformed to match the subject brain images, so that standard ROIs defined on the segments are placed on the individual brain images and are used to measure radioactivity over each brain region. We applied this methods to Alzheimer's disease. METHODS: This method was applied to the images of 10 normal subjects (ages 55 +/- 12) and 21 patient clinically diagnosed as Alzheimer's disease (age 61 +/- 10). The FDG uptake reflecting glucose metabolism was evaluated with SUV, i.e. decay corrected radioactivity divided by injected dose per body weight in (Bq/ml)/(Bq/g). RESULTS: The system worked all right in every subject including those with extensive hypometabolism. Alzheimer patients showed markedly lower in the parietal cortex (4.0-4.1). When the threshold value of FDG uptake in the parietal lobe was set as 5(Bq/ml)/(Bq/g), we could discriminate the patients with Alzheimer's disease from the normal subjects. The sensitivity was 86% and the specificity was 90%. CONCLUSIONS: This system can assist diagnosis of FDG images and may be useful for treating data of a large number of subjects; e.g. when PET is applied to health screening.