Inteligência artificial para o diagnóstico por imagem da instabilidade crônica do tornozelo
Introduction: the ankle sprain is one of the most common traumatic injuries in the human body, the Anterior Fibulo-talar Ligament (ATFL) rupture results from an ankle sprain and causes Chronic Ankle Instability (CAI) and represents progressive ankle joint morbidity. Ankle magnetic resonance (MRI) is...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2023 |
| País: | Brasil |
| Institución: | Universidade Federal do Ceará (UFC) |
| Repositorio: | Repositório Institucional da Universidade Federal do Ceará (UFC) |
| Idioma: | portugués |
| OAI Identifier: | oai:repositorio.ufc.br:riufc/78493 |
| Acceso en línea: | http://repositorio.ufc.br/handle/riufc/78493 |
| Access Level: | acceso abierto |
| Palabra clave: | CNPQ::CIENCIAS DA SAUDE::MEDICINA::CIRURGIA Ligamentos Laterais do Tornozelo Traumatismos do Pé Instabilidade Articular Inteligência Artificial Lateral Ligament, Ankle Foot Injuries Joint Instability Artificial Intelligence |
| Sumario: | Introduction: the ankle sprain is one of the most common traumatic injuries in the human body, the Anterior Fibulo-talar Ligament (ATFL) rupture results from an ankle sprain and causes Chronic Ankle Instability (CAI) and represents progressive ankle joint morbidity. Ankle magnetic resonance (MRI) is the most used exam for CAI diagnosis, buy it is considered to has low accuracy for diagnosis because it is a static exam. Artificial Intelligence (AI) is been used to the diagnosis of diverse lesions in ligaments, tendons and cartilage. Objective: this study compares the medical CAI diagnosis using MRI and the diagnosis made by IA. Methods: this is a prospective study where 321 patients were divided into two groups: with CAI and without CAI. Axial T2 weighted ATFL cuts were analysed by two doctors and AI. Results: the accuracy of the medical diagnosis was 26% for an Intraclass Correlation Index was 0.66. The best strategy for extracting data by the software is a combination of Local Binary Patterns (LBP) with Gray Level Competition Matrix (GLCM) with a classification between groups using Random Florest (FR). The accuracy of the IA analysis reached 84.34%. Conclusion: the use of IA for diagnosis of CAI can be an important tool for decision making in clinical practice. |
|---|