| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.53 MB | Adobe PDF |
Authors
Abstract(s)
A Doença de Parkinson, um distúrbio neurodegenerativo progressivo que impacta o sistema
motor, apresenta evidências de sintomas não motores manifestando-se até 10 anos antes dos
sinais motores clássicos. Este contexto ressalta a importância da compreensão precoce desses
sintomas não motores para a identificação e tratamento eficazes. Diante desse cenário, propõese neste trabalho o desenvolvimento de um índice de gravidade utilizando uma Rede Neuronal
Artificial (RNA), treinada pelo algoritmo Self-Organizing Maps (SOM), utilizando o banco de
dados FOX Insight, que fornece uma variedade de informações sobre a patologia. Após o préprocessamento dos dados, foram selecionados 41.892 questionários (pacientes), contendo 25
perguntas sobre os sintomas não motores, definidas em colaboração com um neurologista. Os
sintomas foram examinados de forma individual e após o mapeamento foi realizada a divisão
em quatro classes, representando diferentes estágios da doença. Esta abordagem visa fornecer
uma ferramenta eficaz para classificação de pacientes com base nos seus sintomas não motores,
possibilitando uma monitorização mais precisa e intervenções personalizadas ao longo da
progressão da doença. A validação da ferramenta foi realizada utilizando dados de pacientes que
responderam ao questionário em momentos espaçados, simulando consultas médicas. Este
estudo alcançou com sucesso o objetivo de desenvolver um índice de gravidade da Doença de
Parkinson baseado em sintomas não motores. Os resultados destacam a importância dos
sintomas gastrointestinais e do trato urinário em diferentes níveis de gravidade, ressaltando a
associação precoce dos sintomas urinários nos estágios iniciais da DP. A persistência do sintoma
de dificuldade para dormir no grupo 3 sugere atenção especial nas fases iniciais da doença. Estes
resultados realçam a relevância clínica e aplicabilidade prática do índice desenvolvido, embora
seja necessário mais estudos com pacientes reais para validar essas conclusões.
Parkinson's Disease (PD), a progressive neurodegenerative disorder affecting the motor system. Still, PD shows evidence of non-motor symptoms emerging up to 10 years before during a premotor phase. Early recognizing these non-motor symptoms allows effective identification and design possible disease modifiable treatment trials. In this context, this study proposes the development of a severity index using an Artificial Neural Network (ANN), trained by the SelfOrganizing Maps (SOM) algorithm, leveraging the FOX Insight database, providing diverse information on the pathology. After data preprocessing, 41,892 questionnaires (patients) were selected, containing 25 questions about non-motor symptoms, defined in collaboration with a neurologist. Symptoms were examined individually, and after mapping, the division into four classes representing different disease stages was performed. This approach aims to provide an effective tool for patient classification based on non-motor symptoms, enabling more precise monitoring and personalized interventions throughout the disease progression. Tool validation was conducted using data from patients who responded to the questionnaire at spaced intervals, simulating medical consultations. This study successfully achieved the goal of developing a severity index for Parkinson's Disease based on non-motor symptoms. The results highlight the importance of gastrointestinal and urinary symptoms at different severity levels, emphasizing the early association of urinary symptoms in the early stages of PD. The persistence of the difficulty sleeping symptom in group 3 suggests special attention in the early stages of the disease. These findings highlight the clinical relevance and practical applicability of the developed index, although further studies with real patients are needed to validate these conclusions.
Parkinson's Disease (PD), a progressive neurodegenerative disorder affecting the motor system. Still, PD shows evidence of non-motor symptoms emerging up to 10 years before during a premotor phase. Early recognizing these non-motor symptoms allows effective identification and design possible disease modifiable treatment trials. In this context, this study proposes the development of a severity index using an Artificial Neural Network (ANN), trained by the SelfOrganizing Maps (SOM) algorithm, leveraging the FOX Insight database, providing diverse information on the pathology. After data preprocessing, 41,892 questionnaires (patients) were selected, containing 25 questions about non-motor symptoms, defined in collaboration with a neurologist. Symptoms were examined individually, and after mapping, the division into four classes representing different disease stages was performed. This approach aims to provide an effective tool for patient classification based on non-motor symptoms, enabling more precise monitoring and personalized interventions throughout the disease progression. Tool validation was conducted using data from patients who responded to the questionnaire at spaced intervals, simulating medical consultations. This study successfully achieved the goal of developing a severity index for Parkinson's Disease based on non-motor symptoms. The results highlight the importance of gastrointestinal and urinary symptoms at different severity levels, emphasizing the early association of urinary symptoms in the early stages of PD. The persistence of the difficulty sleeping symptom in group 3 suggests special attention in the early stages of the disease. These findings highlight the clinical relevance and practical applicability of the developed index, although further studies with real patients are needed to validate these conclusions.
Description
Keywords
Doença de Parkinson Rede Neuronal Artificial Sintomas Não Motores