Detec??o de les?es de esclerose m?ltipla em imagens de resson?ncia magn?tica do tipo Fluid Attenuated Inversion Recovery (FLAIR)

• Ano de defesa: 2016
• Autor(a) principal: Klein, Pedro Costa
• Orientador(a): Pinho, M?rcio Sarroglia
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontif?cia Universidade Cat?lica do Rio Grande do Sul
• Programa de Pós-Graduação: Programa de P?s-Gradua??o em Ci?ncia da Computa??o
• Departamento: Faculdade de Inform?tica
• País: Brasil
• Palavras-chave em Português: ESPECTROSCOPIA DE RESSON?NCIA MAGN?TICA; ESCLEROSE M?LTIPLA; DIAGN?STICO POR IMAGEM; PROCESSAMENTO DE IMAGENS; INFORM?TICA
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/6780

The white matter lesion detection is an important procedure for the diagnostic of Multiple Sclerosis in patients. As important as the detection, monitoring the progression of the disease, by calculating the volumes of the lesions, also shows itself necessary. In clinical practice, this procedure is done manually by a professional or, in many cases, only a qualitative analysis is made. The manual nature of this procedure implies in a series of deficiencies on the procedure, such as variations between diagnoses from different experts on the same subject and variation between diagnostics from the same expert to the same subject at distinct moments. Yet, the manual procedure shows itself time consuming, due the large amount of slices acquired by exam and the need for a careful analysis from the expert to quantify the lesions. In order to avoid these problems, automatic approaches for Multiple Sclerosis lesion detection and quantification using computer aided diagnostic systems are proposed. These methods, mostly, demand for the acquisition of an extra modality of magnetic resonance images, where the anatomy of the brain is evidenced, thus allowing the white matter lesion identification. This additional exam goes beyond the scope of the traditional clinical practice, which implies in additional costs and prevents the method of being applied to old exams, for monitoring the disease progression. This work proposes a method for automatic detection and segmentation of Multiple Sclerosis Lesions that uses only the modality of magnetic resonance exam adopted for clinical practice, using probabilistic atlases spatially aligned to the patient?s exams for identifying the brain structures. The results obtained through the usage of the method into a set of 24 patients and 6 healthy controls of different ages, showed that the developed method is capable of detecting white matter lesions with some precision. However, the quantification of these lesions was impaired mostly due divergences between the white matter probabilistic atlas and the real white matter region of the exams.