Full text
| |
| Author(s): |
Gabriela Felix Persinoti
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Ribeirão Preto. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Medicina de Ribeirão Preto (PCARP/BC) |
| Defense date: | 2009-03-25 |
| Examining board members: |
Silvana Giuliatti;
José Augusto Baranauskas;
Geraldo Aleixo da Silva Passos Junior
|
| Advisor: | Silvana Giuliatti |
| Abstract | |
Rheumatoid Arthritis (RA) is a chronic, inflammatory, autoimmune disease, whose manifestations are observed, mainly in the joints. One point of extreme importance in this disease is the early diagnosis and the begging of the treatment as quickly as possible. These procedures try to avoid joints damage and bone erosions, since these damages are often irreversible. This project aimed to develop a computational system to aid diagnosis of rheumatoid arthritis patients. Artificial intelligence techniques, such as Artificial Neural Networks, Decision Trees and Genetic Algorithms, and the content management system Drupal were employed to the system development. The system goal is to provide a structured environment in which clinical and gene expression information about RA patients are registered and categorized. From the information stored in the system database, the machine learning algorithms can be trained to select clinical information or genes which are relevant to predicting the patients response to treatment with Disease Modifying Anti-Rheumatic Drugs (DMARDs). After the training, the generated models generated can be used for predicting the response to DMARDs of new RA patients. The system aims to select characteristics to early identify patients who might have more aggressively disease outcome and thus, it provides a new tool to physicians in the analysis of the best treatment for each patient individually. The system developed was called ARIA System and it is available at the web site http://gbi.fmrp.usp.br/artrite. Only doctors and collaborator researchers have permission to access it. The contents available in the system are the following: Patients Register, Feature Selection Tool and Prediction the Response to DMARDs Tool. Besides these contents, the system offers different kinds of displaying and interaction with the content. There are, until now, 126 patients registered in the ARIA System, 106 of them are used for training the features selection tool and twenty of them are used to test the tool for predicting patients response to DMARDs. (AU) | |