Lesões cervicais não cariosas e inteligência artificial
Carregando...
Data
2023-11-30
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A prevalência das lesões não cariosas (LNC) de erosão, abfração, abrasão e/ou atrição tem aumento significativamente em todo o mundo e refletem as alterações nos hábitos alimentares, exposição a agentes agressivos e o estresse da vida moderna. Revisões sistemáticas mostram que 70% dos brasileiros apresentam sinais de LNC. Um dos desafios destas lesões é que elas só são percebidas pelos pacientes quando os danos a estrutura dental já são significativos e os tratamentos necessários são caros, extensos, invasivos e desconfortáveis. O uso de chatbots (assistentes conversacionais com inteligência artificial) tem sido utilizado em diversas áreas da saúde para tirar dúvidas sobre patologias, acompanhar a evolução de pós-operatórios, agendar consultas etc. Nesse projeto de pesquisa, a assertividade de um chatbot especialista na classificação do risco de uma pessoa desenvolver LNC foi analisado frente a decisão de clínicos experientes e calibrados. A capacidade do chatbot reproduzir com acurácia a decisão dos cirurgiões dentistas
experts foi confirmada tanto para o examinador TT (n= 44; k=0,70; interpretação do índice: forte ou substancial como para o examinador OJ (n= 44; k=0,83; interpretação: quase perfeita). Conclusão: A o algoritmo baseado em inteligência artificial do chatbot classifica de forma semelhante a dentistas experientes e treinados o risco da pessoa desenvolver LNC.
The prevalence of non-carious lesions (NCL) of erosion, abfraction, abrasion and/or attrition has increased significantly throughout the world and reflects changes in eating habits, exposure to aggressive agents and the stress of modern life. Systematic reviews show that 70% of Brazilians present signs of LNC. One of the challenges of these injuries is that they are only noticed by patients when the damage to the dental structure is already significant and the necessary treatments are expensive, extensive, invasive and uncomfortable. The use of chatbots (conversational assistants with artificial intelligence) has been used in various areas of healthcare to answer questions about pathologies, monitor post-operative progress, schedule appointments, etc. In this research project, the assertiveness of a specialist chatbot in classifying a person's risk of developing LNC was analyzed against the decision of experienced and calibrated clinicians. The chatbot's ability to accurately reproduce the decision of expert dental surgeons was confirmed both for the TT examiner (n= 44; k=0.70; index interpretation: strong or substantial) and for the OJ examiner (n= 44; k= 0.83; interpretation: almost perfect) Conclusion: The chatbot's artificial intelligence-based algorithm classifies the person's risk of developing LNC in a similar way to experienced and trained dentists.
The prevalence of non-carious lesions (NCL) of erosion, abfraction, abrasion and/or attrition has increased significantly throughout the world and reflects changes in eating habits, exposure to aggressive agents and the stress of modern life. Systematic reviews show that 70% of Brazilians present signs of LNC. One of the challenges of these injuries is that they are only noticed by patients when the damage to the dental structure is already significant and the necessary treatments are expensive, extensive, invasive and uncomfortable. The use of chatbots (conversational assistants with artificial intelligence) has been used in various areas of healthcare to answer questions about pathologies, monitor post-operative progress, schedule appointments, etc. In this research project, the assertiveness of a specialist chatbot in classifying a person's risk of developing LNC was analyzed against the decision of experienced and calibrated clinicians. The chatbot's ability to accurately reproduce the decision of expert dental surgeons was confirmed both for the TT examiner (n= 44; k=0.70; index interpretation: strong or substantial) and for the OJ examiner (n= 44; k= 0.83; interpretation: almost perfect) Conclusion: The chatbot's artificial intelligence-based algorithm classifies the person's risk of developing LNC in a similar way to experienced and trained dentists.
Descrição
Palavras-chave
Inteligência artificial, Sistemas de apoio a decisões clínicas, Erosão dentária, Bruxismo, Artificial intelligence, Decision support systems, clinical, Tooth erosion, Bruxism
Como citar
Donvito MAF. Lesões cervicais não cariosas e inteligência artificial [Trabalho de Conclusão de Curso – Graduação em Odontologia]. Araraquara: Faculdade de Odontologia da UNESP; 2023.