Osteoartrite na articulação temporomandibular: análise integrativa de marcadores por imagem, clínicos e biomoleculares
Carregando...
Data
2019-12-06
Autores
Bianchi, Jonas [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A osteoartrite (OA) é a forma mais comum das artrites. Na articulação temporomandibular (ATM) ela se destaca por causar degeneração da cartilagem de forma progressiva, remodelação do tecido ósseo condilar, quadros agudos e crônicos de dor. Atualmente, o diagnóstico da OA na ATM vem se tornando mais preciso, devido ao desenvolvimento de novas tecnologias, exames por imagem como a tomografia computadorizada de feixe cônico (TCFC) de alta resolução, análises computacionais e inteligência artificial. Dessa forma, esse trabalho teve como principal objetivo avaliar possíveis novos biomarcadores para a OA utilizando marcadores clínicos, por imagem e biomoleculares. Os objetivos secundários foram de comparar diferentes softwares para análises dos marcadores por imagem; avaliar o poder de diagnóstico desses marcadores; e por fim, desenvolver um modelo integrativo utilizando inteligência artificial, dados clínicos, biomoleculares e por imagem para o diagnóstico da OA na ATM. Como metodologia, foram desenvolvidos 3 artigos científicos, apresentados em sequência nessa tese. Nossa amostra foi composta por pacientes com diagnóstico clínico de OA na ATM e pacientes controles, sendo que foram coletados exames de TCFC, sangue, saliva e dados clínicos. Nossos resultados mostraram que o software desenvolvido pelo nosso grupo para análises das imagens é confiável e que os novos marcadores por imagens são capazes de diferenciar pacientes controles e com OA. Também demonstramos que nosso modelo estatístico integrativo de marcadores clínicos, biomoleculares, por imagem e inteligência artificial é capaz de diagnosticar a doença com acurácia de 0.837 ([0.761,0.902]). Por fim, nossos resultados sugerem que diferentes biomarcadores e uma integração estatística e computacional dos dados podem diagnosticar a doença e que estudos futuros devem ser realizados a fim de avaliar o comportamento das variáveis em relação a predição da AO na ATM.
Osteoarthritis (OA) is the most common form of arthritis. In the temporomandibular joint (TMJ) it stands out for causing progressive cartilage degeneration, condylar bone tissue remodeling, acute and chronic pain. Nowadays, the TMJ OA diagnosis is improved by the development of new technologies and imaging exams such as high-resolution cone beam computed tomography (CBCT), advanced computational analysis and artificial intelligence. Thus, this study aimed to evaluate possible new biomarkers for OA using clinical, imaging and biomolecular markers. The secondary objectives were to compare different software for image marker analysis; evaluate the diagnostic power of these markers; and finally, to develop an integrative model using artificial intelligence, clinical, biomolecular and imaging markers for the diagnosis of TMJ OA. As methodology, 3 papers were developed, presented in sequence in this thesis. The sample consisted of patients with clinical diagnosis of TMJ OA and control patients. CBCT, blood, saliva and clinical data were collected. Our results showed that the software developed by our group for image analysis is reliable and that the new imaging markers can differentiate controls and OA patients. We also demonstrate that our integrative statistical model of clinical, biomolecular, imaging markers and artificial intelligence can diagnose the disease with an accuracy of 0.837 ([0.761.0.902]). Finally, our results suggest that different biomarkers and a statistical and computational integration of the data can diagnose the disease and future studies should be performed in order to evaluate the behavior of the variables regarding the prediction of OA in the TMJ.
Osteoarthritis (OA) is the most common form of arthritis. In the temporomandibular joint (TMJ) it stands out for causing progressive cartilage degeneration, condylar bone tissue remodeling, acute and chronic pain. Nowadays, the TMJ OA diagnosis is improved by the development of new technologies and imaging exams such as high-resolution cone beam computed tomography (CBCT), advanced computational analysis and artificial intelligence. Thus, this study aimed to evaluate possible new biomarkers for OA using clinical, imaging and biomolecular markers. The secondary objectives were to compare different software for image marker analysis; evaluate the diagnostic power of these markers; and finally, to develop an integrative model using artificial intelligence, clinical, biomolecular and imaging markers for the diagnosis of TMJ OA. As methodology, 3 papers were developed, presented in sequence in this thesis. The sample consisted of patients with clinical diagnosis of TMJ OA and control patients. CBCT, blood, saliva and clinical data were collected. Our results showed that the software developed by our group for image analysis is reliable and that the new imaging markers can differentiate controls and OA patients. We also demonstrate that our integrative statistical model of clinical, biomolecular, imaging markers and artificial intelligence can diagnose the disease with an accuracy of 0.837 ([0.761.0.902]). Finally, our results suggest that different biomarkers and a statistical and computational integration of the data can diagnose the disease and future studies should be performed in order to evaluate the behavior of the variables regarding the prediction of OA in the TMJ.
Descrição
Palavras-chave
Osteoartrite, Tomografia Computadorizada de Feixe Cônico, Biomarcadores, Osteoarthritis, Cone-Beam Computed Tomography, Biomarkers