proAD: A web tool for the automatic assessment of prodromal Alzheimer's disease

Abstract

We present the development of a web-based interface (proAD) for the automatic analysis of structural T1-weighted magnetic resonance images (MRI). This web-based tool is meant to be a high level interface to a sophisticated analysis environment aimed at the early diagnosis of Alzheimer's disease patients. The analysis procedure relies on clinical diagnosis and follow up data. It extracts several portions of the temporal lobe, mainly around the hippocampal area, and processes them with no interactive input from the user. The output of the MRI analysis is a classification index which is able to distinguish among: (a) patients with Alzheimer's disease (AD), (b) patients with amnestic mild cognitive impairment (aMCI), (c) patients with amnestic mild cognitive impairment who are likely to develop AD in a time frame of 2 years (aMCIconv), and (c) normal cognitive subjects. The interface is still under development and has been built as easy and user-friendly as possible. Robustness has been considered as essential part for the development. Results are presented to users in a very simple manner, thereby serving as a potential diagnostic tool for neurologists.A working testbed of proAD (still in alpha stage) can be reached at http://magic5gui.ge.infn.it/tesi