Develop a computer system that can diagnose Alzheimer's early form

Researchers at the University of Granada (UGR) have developed a new computer system that can diagnose Alzheimer's early form from computer-processed images, reaching 90 percent of success in improving early identification of disease.

This work is part of a larger project in development to create a comprehensive software tool available in clinical practice in hospitals, in collaboration with business and PET-Cartuja PTEC under the project 'PETRI-DENCLASES' winner the 'III Premios Andalucía Information Society' Furthermore, cooperation of this group research project with the ADNI (Alzheimer's Disease Neuroimaging Initiative), makes the project take an international character and to provide one of the most extensive databases worldwide in regard to Alzheimer's disease.

To test this new technique of automatic computer-aided diagnosis, UGR researchers used PET and SPECT tomographic images from three different databases. The first one consists of 97 anonymized and SPECT images labeled by experts, delivered by the close collaboration with Manuel Gomez-Rio and the nuclear medicine department of the Hospital Virgen de las Nieves in Granada.

The second database consists of 60 PET images, with which the company PET-Cartuja de Sevilla contributed to the project. The third and most extensive database, consisting of PET images of 219 subjects provided by ADNI (Alzheimer's Disease Neuroimaging Initiative, USA). These databases consisted of tomographic brain images of elderly subjects affected by Alzheimer's disease or a normal development pattern, so that the developed algorithms identified the regions affected by Alzheimer's and learn to discriminate between those subjects suffering from the disease and the healthy.

The three methods presented in this work will exceed 90 per cent success in identifying Alzheimer's disease using tomographic images of both PET and SPECT. It is currently developing a computer tool in a project with the company PTEC (Malaga), to translate the research results to software designed to be used in any hospital.

Thus, diagnosis is made individually by the neurologist will be assisted by a tool whose accuracy is comparable to estimates of a set of experts.

Part of the results of this research were published in the journals 'Information Sciences, 2010', 'Neuroscience Letters, 2009 "or" Electronics Letters, 2009'.

This work has been carried out by Professor Ignacio Alvarez, a member of SIPB (Signal Processing and Biomedical Applications) TIC-010, University of Granada, and directed by professors Juan Manuel Górriz, Javier Ramirez Perez and Carlos Garcia Puntonet.

Source: EUROPAPRESS

_______________________________
FUNCTIONAL DIVERSITY Web Portal
http://www.diversitatfuncional.org/
Blog ICT AND DISABILITY
http://discapacitat-en.diversitatfuncional.org/