The Management And Analysis Of Medical Image Data For Alzheimer’s Disease

With the rapid development of medical imaging technologies, andtheir wide applications in clinical diagnosis, basic medical research andclinical trials, there existed huge amount of medical image data. Theinformation contained in them is extremely rich and forms the basicresource for clinical diagnosis, and for disease related study. One importantuse of the medical imaging data is in the area of the computer-aideddiagnosis which becomes more widely accepted.Alzheimer’s disease (AD) databases have been long developed andwell maintained in developed countries. However in China, no widelyaccepted database system is accessible to AD researchers and clinicians. Inorder to share resources and help accelerate new research on AD, we builtan integrated database that systematically stores, organizes, and managesraw and processed AD related data previously and newly collected atmultiple centers. This database not only contains integrated imaging,demographic, genomic, cognitive test, clinical and various other biomarkerdata types for easy access and analysis but also includes analytic tools as apart of the system. Here，as a validation of the tools we implemented andin order to realize the classification of Alzheimer’s disease, we extractinformation from FDG PET images of whole brain first with each ofseveral feature extract methods, and evaluated the classification accuracyvia the use of the support vector machine procedure.Our data management system can provide a convenient and fast datastorage with convenient query. Moreover, the data management system isdesigned for easy data sharing in multiple Alzheimer’s research centers. Atthe same time, the results of classification for Alzheimer’s disease with PET image can achieve a very high accuracy demonstrating the procedurefeasibility in clinical study.