| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.88 MB | Adobe PDF |
Authors
Abstract(s)
O tema abordado na presente Dissertação insere-se no domínio da Visão Computacional, mais especificamente, no Processamento e Análise de Imagem por algoritmos computacionais. O processamento de imagem engloba processos fundamentais que permitem corresponder estruturas similares entre duas ou mais imagens e determinar os parâmetros do modelo da transformação geométrica que melhor alinha as estruturas de interesse representadas nas imagens alvo e de referência. Foi efetuado um estudo previamente a esta Dissertação, com 12 doentes sujeitos a tiroidectomia total e posterior ablação por I-131, e a fazerem terapêutica de substituição com Levotiroxina. Destes 12 doentes, 6 efetuaram PET Cerebral ao fim de 1 mês de paragem da Levotiroxina e os restantes 6 doentes efetuaram PET Cerebral após estimulação por TSHr. Todos os doentes efetuaram também, PET Cerebral Basal, no estado eutiroideu e foram avaliados por um neuropsicólogo em todas as sessões de aquisição de imagem. Esta Dissertação teve como objetivo principal desenvolver um algoritmo computacional automático, que permite carregar, processar e analisar imagens FDG-PET do cérebro, para posterior interpretação dos resultados acerca do metabolismo da glicose no SNC, em pacientes com tiroidectomia total com o uso de substitutos da tiroide durante: hipotiroidismo induzido, estado eutiroideu e sob estimulação por TSHr. O algoritmo computacional calcula as diferenças entre os diferentes estados, retornando uma imagem em z-scores, possibilitando a determinação e quantificação das diferenças metabólicas encontradas entre as imagens. A aplicação do algoritmo desenvolvido mostrou ser uma boa metodologia para o propósito desta Dissertação. Em suma, o algoritmo desenvolvido permite avaliar o metabolismo da glicose no SNC.
The topic addressed in this Dissertation is in the field of Computational View, more specifically, in Image Processing and Analysis by computational algorithms. Image processing encompasses fundamental processes which allows matching similar structures between two or more images and determining the parameters of the geometric transformation model that best aligns the structures of interest represented in the target and reference images. It was made a previous study with 12 patients undergoing total thyroidectomy and subsequent I-131 ablation, and undergoing replacement therapy with levothyroxine. Of these 12 patients, 6 had cerebral PET after 1 month of levothyroxine withdrawal and the remaining 6 patients had cerebral PET after stimulation by rTSH. All patients also performed Cerebral Basal PET in the euthyroid state and were evaluated by a neuropsychologist at all imaging sessions. This Dissertation aimed to develop an automatic computational algorithm, which allows to load, process and analyze FDG-PET images of the brain, for further interpretation of the results about glucose metabolism in the CNS in patients with total thyroidectomy with the use of thyroid substitutes during: induced hypothyroidism, euthyroid state and under stimulation by TSHr. The computational algorithm calculates the differences between the images, returning an image in z-scores, allowing the determination and quantification of the differences between the images. The application of the algorithm developed proved to be a good methodology for the purpose of this Dissertation. In short, the algorithm developed allows the evaluation of glucose metabolism in the CNS.
The topic addressed in this Dissertation is in the field of Computational View, more specifically, in Image Processing and Analysis by computational algorithms. Image processing encompasses fundamental processes which allows matching similar structures between two or more images and determining the parameters of the geometric transformation model that best aligns the structures of interest represented in the target and reference images. It was made a previous study with 12 patients undergoing total thyroidectomy and subsequent I-131 ablation, and undergoing replacement therapy with levothyroxine. Of these 12 patients, 6 had cerebral PET after 1 month of levothyroxine withdrawal and the remaining 6 patients had cerebral PET after stimulation by rTSH. All patients also performed Cerebral Basal PET in the euthyroid state and were evaluated by a neuropsychologist at all imaging sessions. This Dissertation aimed to develop an automatic computational algorithm, which allows to load, process and analyze FDG-PET images of the brain, for further interpretation of the results about glucose metabolism in the CNS in patients with total thyroidectomy with the use of thyroid substitutes during: induced hypothyroidism, euthyroid state and under stimulation by TSHr. The computational algorithm calculates the differences between the images, returning an image in z-scores, allowing the determination and quantification of the differences between the images. The application of the algorithm developed proved to be a good methodology for the purpose of this Dissertation. In short, the algorithm developed allows the evaluation of glucose metabolism in the CNS.
Description
Keywords
Processamento e análise de imagem PET Algoritmos computacionais Eutiroidismo Hipotiroidismo Estimulação por TSHr Image processing and analysis Computational algorithms Euthyroidism Hypothyroidism TSHr stimulation