Aplicação da inteligência artificial na predição de cirurgia ortognática ou ortodontia convencional
Carregando...
Data
2023-01-26
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Quando as alterações esqueléticas, oclusais e estéticas apresentadas são tão severas que não podem ser tratadas com ortodontia compensatória, deve ser realizado tratamento ortodôntico-cirúrgico para que possa restabelecer a harmonia facial e função para o paciente. No entanto, para alguns casos limítrofes, o conhecimento empírico de profissionais experientes pode ser um fator decisivo na tomada de decisão. Hoje, com aumento de dados e melhorias tecnológicas é possível utilizar modelos de inteligência artificial (machine learning) como suporte para tomadas de decisão. O objetivo principal desse estudo foi de avaliar se modelos de inteligência artificial podem ser utilizados como auxílio na tomada de decisão em ortodontia, por meio de dois artigos. No primeiro artigo, o objetivo foi de revisar sistematicamente como a inteligência artificial pode ser aplicada ao diagnóstico e planejamento ortodôntico utilizando exames radiológicos e testar se diferentes modelos de inteligência virtual eram capazes de predizer o tratamento. O segundo artigo, avaliou se um paciente necessita ou não de cirurgia ortognática a partir dos valores obtidos em telerradiografias laterais, previamente ao tratamento avaliando tanto tecido duro e mole em pacientes classe II ou classe III esquelética, por meio de diferentes algoritmos de inteligência artificial. Como resultados, o primeiro artigo resultou em 12 estudos divididos em 5 categorias diferentes de diagnóstico: osteoartrite de articulação temporomandibular, maturação óssea, classificação esquelética, síndrome da apnéia obstrutiva do sono e necessidade de cirurgia ortognática. Em geral, a IA aumentou a acurácia do diagnóstico. Os resultados para o segundo artigo mostraram que o modelo combinado de algoritmos de machine learning apresentou alta performance para todos os grupos (amostra total, classe II e classe III), sendo que os pacientes classe III mostraram valores superiores para acurácia (0.873), F1-score (0.8) e curva AUC (0.923). Concluindo assim que a inteligência artificial pode ser utilizada como auxílio na tomada de decisão para tratamento ortodôntico e cirúrgicos.
When skeletal, occlusal, and aesthetical discrepancies are too severe that cannot be treated with compensatory orthodontics, orthodontic preparation to orthognathic surgery must be done to restore the patient’s facial harmony and function. However, some borderline patients can be a challenge. For these cases, the experience and clinician’s empirical knowledge plays an important role in the decision-making process for planning. The increasing amount of data and technological advancements can make it possible for artificial intelligence (AI) models (machine learning) to support the decision-making process. Thus, this study aimed to evaluate if artificial intelligence models can be used as an auxiliary tool to predict if a patient needs orthognathic surgery, by presenting two papers. The first paper’s objective was to systematically review how artificial intelligence can be applied to orthodontic diagnosis and planning using X-ray-based images and test if different AI models were capable to predict the treatment. The second paper’s objectives were to evaluate the machine learning models’ capacity to predict if class II and class III patients need orthognathic surgery using soft and hard tissue values obtained from lateral radiographs. As results, the first article resulted in 12 studies that were divided in 5 different categories: TMJ osteoarthritis, skeletal maturation/development, skeletal pattern, obstructive sleep apnea, and orthognathic surgery. The results for the second article showed that the combined machine learning prediction model showed top ranked performance in the testing-set for all groups (entire sample, class II and class III) and the best results were observed for class III validation-set (accuracy 0.873, F1-score 0.8, and AUC 0.923). In conclusion, artificial intelligence can be used as an aid for orthognathic surgery decision.
When skeletal, occlusal, and aesthetical discrepancies are too severe that cannot be treated with compensatory orthodontics, orthodontic preparation to orthognathic surgery must be done to restore the patient’s facial harmony and function. However, some borderline patients can be a challenge. For these cases, the experience and clinician’s empirical knowledge plays an important role in the decision-making process for planning. The increasing amount of data and technological advancements can make it possible for artificial intelligence (AI) models (machine learning) to support the decision-making process. Thus, this study aimed to evaluate if artificial intelligence models can be used as an auxiliary tool to predict if a patient needs orthognathic surgery, by presenting two papers. The first paper’s objective was to systematically review how artificial intelligence can be applied to orthodontic diagnosis and planning using X-ray-based images and test if different AI models were capable to predict the treatment. The second paper’s objectives were to evaluate the machine learning models’ capacity to predict if class II and class III patients need orthognathic surgery using soft and hard tissue values obtained from lateral radiographs. As results, the first article resulted in 12 studies that were divided in 5 different categories: TMJ osteoarthritis, skeletal maturation/development, skeletal pattern, obstructive sleep apnea, and orthognathic surgery. The results for the second article showed that the combined machine learning prediction model showed top ranked performance in the testing-set for all groups (entire sample, class II and class III) and the best results were observed for class III validation-set (accuracy 0.873, F1-score 0.8, and AUC 0.923). In conclusion, artificial intelligence can be used as an aid for orthognathic surgery decision.
Descrição
Palavras-chave
Ortodontia, Inteligência artificial, Cirurgia ortognática, Orthodontics, Artificial intelligence, Orthognathic surgery