| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.39 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A Doença de Parkinson (DP) é um transtorno neurodegenerativo progressivo causado pela perda seletiva de neurónios dopaminérgicos localizados maioritariamente na região pars compacta da substância negra (SN), onde há uma perda de axónios que fazem a ligação desta região e se projetam para os núcleos da base, nomeadamente o corpo estriado. Atualmente, é uma doença diagnosticada clinicamente, no entanto, os estudos de imagem médica têm vindo a ganhar destaque e são fundamentais para a deteção da DP. A Single-Photon Emission Computed Tomography (SPECT) e a Ressonância Magnética (RM) são métodos de imagem médica mais utilizados devido à informação molecular e morfológica que apresentam, respetivamente. Esta dissertação tem como objetivo geral avaliar a DP através da quantificação comparativa de estruturas cerebrais profundas, envolvidas nos sistemas dopaminérgicos, em imagens DaTSCAN SPECT e imagens RM T2 em pacientes com DP e em voluntários saudáveis (VS), recorrendo à base de dados Parkinson’s Progressive Markers Initiative (PPMI). Foram avaliados e comparados diferentes métodos de segmentação (manual, ThreeBox, Threshold e atlas anatómico) em imagens DaTSCAN SPECT, assim como a evolução da DP através do método ThreeBox. Em RM T2 foi avaliada a intensidade do sinal e volume da SN, e a integridade e presença do nigrossoma-1. O Binding Potential Index (BPI) foi calculado para todas as segmentações SPECT e calculou-se a correlação Spearman para aferir sobre a relação entre a perda dopaminérgica no estriado e na SN. A amostra selecionada consiste em imagens de dois grupos: pacientes com DP e VS. Em SPECT selecionaram-se 49 imagens de VS e 42 imagens de pacientes com DP. Em RM escolheram-se 40 imagens de VS e 42 imagens de pacientes com DP. Em imagens DaTSCAN SPECT, a variação da média do BPI, entre VS e pacientes com DP, foi de 37,608% (manual), 50,486% (ThreeBox), 16,868% (Threshold) e para a segmentação com atlas anatómico 60,331% (putámen), 46,490% (globo pálido) e 35,895% (núcleo caudado). Para a evolução da DP a 12 meses obteve-se uma variação de 9,969% e, aos 24 meses, a percentagem de variação foi de 17,389%. No estudo das imagens de RM verificou-se que não existe uma diferença expressiva de intensidade ou volume da SN entre VS e pacientes com DP. Sobre o nigrossoma-1, em VS contabilizaram-se oito imagens com presença unilateral, 31 imagens com presença bilateral e uma imagem onde não foi possível identificar os nigrossomas-1. No grupo de pacientes com DP foram assinaladas 12 imagens com presença unilateral, 11 imagens com presença bilateral e 19 imagens onde não se visualizou qualquer nigrossoma-1. Os valores do coeficiente das correlações Spearman entre o volume e a intensidade da SN, e o BPI foram pouco significativos: (volume SN esquerda - BPI: ρ = - 0,084, P > 0,05), (volume SN direita - BPI: ρ = -0,059, P > 0,05), (intensidade SN esquerda - BPI: ρ = -0,018, P > 0,05) e (intensidade SN esquerda - BPI: ρ = -0,134, P > 0,05). Sobre a correlação entre o vi volume da SN e a idade, a correlação Spearman foi negativa, moderada e significativa (SN esquerda: ρ = -0,329, P < 0,05; SN direita: ρ = -0,365, P < 0,05). Da mesma forma, a correlação entre a intensidade da SN e a idade foi também negativa, moderada e significativa, exceto à esquerda (SN esquerda: ρ = - 0,225, P > 0,05; SN direita: ρ = -0,380, P < 0,05). Na correlação do BPI e a idade, a correlação foi negativa, fraca e não significativa (Corpo estriado esquerdo: ρ = -0,036, P > 0,05; Corpo estriado direito: ρ = -0,178, P > 0,05). A metodologia de segmentação ThreeBox apresentou melhores resultados discriminatórios, enquanto que a metodologia por Threshold obteve os piores resultados discriminatórios. O estudo da evolução da DP permite concluir que há uma tendência progressiva e linear de degeneração dopaminérgica no estriado. Os resultados das segmentações de RM T2 apresentaram fraca discriminação entre VS e pacientes com DP, não havendo diferenças expressivas na intensidade e no volume da SN. As correlações entre a intensidade e o volume da SN com o BPI foram fracas, negativas e estatisticamente não significativas, significando que não existe uma associação entre as variáveis. Na análise sociodemográfica, o fator da idade não apresenta diferenças expressivas.
ABSTRACT - Parkinson’s Disease (PD) is a progressive neurological disorder characterized by the loss of dopaminergic neurons located on Substantia Nigra pars compacta, where axons connected themselves to the basal ganglia, namely the striatum. Currently, it is a clinically diagnostic disease, however, medical imaging has gained some ground and could be essential to detect PD. Single-photon emission Computed Tomography (SPECT) and Magnetic Resonance Imaging (MRI) are the most common techniques used, thanks to their molecular and morphological information, respectively. This thesis has the primary goal of the evaluation of PD with comparative quantification methods of cerebral structures, involved in dopaminergic systems, in DaTSCAN SPECT and MRI T2 images. Healthy controls (HC) and PD patients will be compared, using the Parkinson’s Progressive Markers Initiative (PPMI) database. Different segmentation methods (manual, ThreeBox, Threshold, and Atlas) will be compared in DaTSCAN SPECT images, as well as the evolution of PD with the ThreeBox method; In MRI T2 will be evaluated the intensity and volume of SN and the integrity and presence of nicrosome-1; Binding Potential Index (BPI) will be calculated for all SPECT methods; It will be used Spearman’s correlation to relate the dopaminergic loss in striatum and SN. In SPECT, 49 HC and 42 DP patients’ images were selected, and in MRI, 40 HC and 42 DP patients’ images were selected. In DaTSCAN SPECT images, the BPI average mean variation was 37,608% (manual), 50,486% (ThreeBox), 16,868% (Threshold), and for the Atlas method it was 60,331% (putamen), 46,490% (globus pallidus) and 35,895% (caudate nucleus). The evolution of DP at 12 months achieved a BPI reduction of 9,969% and, at 24 months, the BPI was 17,389% less. In MRI analysis, it was noted that there is no significant difference in volume or intensity of SN between HC and PD patients. The nigrossome-1 presence in HC was as follows: eight unilateral, 31 bilateral, and one bilateral absence. In PD patients: 12 unilateral, 11 bilateral, and 19 bilateral absences. The Spearman’s correlation was calculated between the volume of SN and BPI, and the intensity of SN and BPI: (left SN volume- BPI: ρ = -0,084, P > 0,05), (right SN volume - BPI: ρ = -0,059, P > 0,05), (left SN intensity - BPI: ρ = -0,018, P > 0,05) e (left SN intensity - BPI: ρ = -0,134, P > 0,05). Also, the Spearman’s correlation was calculated between the volume of SN and age (left SN: ρ = -0,329, P < 0,05; right SN: ρ = -0,365, P < 0,05); between intensity of SN and age (left SN: ρ = -0,225, P > 0,05; right SN: ρ = -0,380, P < 0,05); and between BPI and age (left striatum: ρ = -0,036, P > 0,05; right striatum: ρ = -0,178, P > 0,05). The ThreeBox method presented the best results for distinguishing between HC and PD patients, while the Threshold method was the worst one. The evolution of the PD method proved that there is a progressive and linear tendency of dopaminergic loss in the striatum. For MRI segmentations, there viii was a weak separation in intensity and volume of SN between HC and PD patients. The correlations between intensity and volume of SN with BPI are weak, and negative, meaning that there isn’t an association between them. Also, age isn’t a significant factor between HC and PD patients.
ABSTRACT - Parkinson’s Disease (PD) is a progressive neurological disorder characterized by the loss of dopaminergic neurons located on Substantia Nigra pars compacta, where axons connected themselves to the basal ganglia, namely the striatum. Currently, it is a clinically diagnostic disease, however, medical imaging has gained some ground and could be essential to detect PD. Single-photon emission Computed Tomography (SPECT) and Magnetic Resonance Imaging (MRI) are the most common techniques used, thanks to their molecular and morphological information, respectively. This thesis has the primary goal of the evaluation of PD with comparative quantification methods of cerebral structures, involved in dopaminergic systems, in DaTSCAN SPECT and MRI T2 images. Healthy controls (HC) and PD patients will be compared, using the Parkinson’s Progressive Markers Initiative (PPMI) database. Different segmentation methods (manual, ThreeBox, Threshold, and Atlas) will be compared in DaTSCAN SPECT images, as well as the evolution of PD with the ThreeBox method; In MRI T2 will be evaluated the intensity and volume of SN and the integrity and presence of nicrosome-1; Binding Potential Index (BPI) will be calculated for all SPECT methods; It will be used Spearman’s correlation to relate the dopaminergic loss in striatum and SN. In SPECT, 49 HC and 42 DP patients’ images were selected, and in MRI, 40 HC and 42 DP patients’ images were selected. In DaTSCAN SPECT images, the BPI average mean variation was 37,608% (manual), 50,486% (ThreeBox), 16,868% (Threshold), and for the Atlas method it was 60,331% (putamen), 46,490% (globus pallidus) and 35,895% (caudate nucleus). The evolution of DP at 12 months achieved a BPI reduction of 9,969% and, at 24 months, the BPI was 17,389% less. In MRI analysis, it was noted that there is no significant difference in volume or intensity of SN between HC and PD patients. The nigrossome-1 presence in HC was as follows: eight unilateral, 31 bilateral, and one bilateral absence. In PD patients: 12 unilateral, 11 bilateral, and 19 bilateral absences. The Spearman’s correlation was calculated between the volume of SN and BPI, and the intensity of SN and BPI: (left SN volume- BPI: ρ = -0,084, P > 0,05), (right SN volume - BPI: ρ = -0,059, P > 0,05), (left SN intensity - BPI: ρ = -0,018, P > 0,05) e (left SN intensity - BPI: ρ = -0,134, P > 0,05). Also, the Spearman’s correlation was calculated between the volume of SN and age (left SN: ρ = -0,329, P < 0,05; right SN: ρ = -0,365, P < 0,05); between intensity of SN and age (left SN: ρ = -0,225, P > 0,05; right SN: ρ = -0,380, P < 0,05); and between BPI and age (left striatum: ρ = -0,036, P > 0,05; right striatum: ρ = -0,178, P > 0,05). The ThreeBox method presented the best results for distinguishing between HC and PD patients, while the Threshold method was the worst one. The evolution of the PD method proved that there is a progressive and linear tendency of dopaminergic loss in the striatum. For MRI segmentations, there viii was a weak separation in intensity and volume of SN between HC and PD patients. The correlations between intensity and volume of SN with BPI are weak, and negative, meaning that there isn’t an association between them. Also, age isn’t a significant factor between HC and PD patients.
Description
Mestrado em Engenharia Biomédica
Keywords
Doença de Parkinson RM T2 Ressonância magnética DaTSCAN SPECT Corpos estriados Substância negra Parkinson disease MRI T2 Magnetic resonance Striatum Substantia nigra
Citation
Mesquita LP. Avaliação do corpo estriado e da substância negra na doença de Parkinson através de imagens de ressonância magnética [dissertation]. Lisboa: Escola Superior de Tecnologia da Saúde de Lisboa – Instituto Superior de Engenharia de Lisboa/Instituto Politécnico de Lisboa; 2023.
Publisher
Instituto Politécnico de Lisboa, Escola Superior de Tecnologia da Saúde de Lisboa