American journal of otolaryngology · 2024
How reliable is artificial intelligence in the diagnosis of cholesteatoma on CT images?
СтатьяСистематический обзорМетаанализНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Emilio Avallone, De Luca Pietro, Max Timm, Agnese Siani, Pasquale Viola, Massimo Ralli, Giuseppe Chiarella, Filippo Ricciardiello, Antonio Salzano Francesco, Alfonso Scarpa
Аннотация
PURPOSE: This study analysed the main artificial intelligence (AI) models for the diagnosis of cholesteatoma on computed tomography (CT), evaluating their performance and comparing them with each other. The increasing application of AI in radiology requires a systematic comparison of available methodologies.
METHODS: A systematic literature review was conducted, selecting relevant articles from the main databases. The included studies had to fulfil specific criteria regarding methodology, use of AI models for cholesteatoma diagnosis and results in terms of sensitivity and specificity.
RESULTS: The meta-analysis included three studies evaluating the MobilenetV2, Densenet201 and Resnet50 AI models. All models demonstrated high levels of sensitivity and specificity in the diagnosis of cholesteatoma at CT. No statistically significant differences were found between the performance of the various models, suggesting a robust common diagnostic capability between the different neural network architectures.
CONCLUSIONS: AI is making rapid progress in imaging, with recent studies already showing remarkable performance in cholesteatoma diagnosis. The speed of technological development is promising. However, to ensure safe and effective implementation in clinical practice, further studies are needed to validate and standardise these AI models. Future research should focus not only on the diagnostic accuracy, but also on the robustness, reproducibility and clinical integration of these emerging technologies.
METHODS: A systematic literature review was conducted, selecting relevant articles from the main databases. The included studies had to fulfil specific criteria regarding methodology, use of AI models for cholesteatoma diagnosis and results in terms of sensitivity and specificity.
RESULTS: The meta-analysis included three studies evaluating the MobilenetV2, Densenet201 and Resnet50 AI models. All models demonstrated high levels of sensitivity and specificity in the diagnosis of cholesteatoma at CT. No statistically significant differences were found between the performance of the various models, suggesting a robust common diagnostic capability between the different neural network architectures.
CONCLUSIONS: AI is making rapid progress in imaging, with recent studies already showing remarkable performance in cholesteatoma diagnosis. The speed of technological development is promising. However, to ensure safe and effective implementation in clinical practice, further studies are needed to validate and standardise these AI models. Future research should focus not only on the diagnostic accuracy, but also on the robustness, reproducibility and clinical integration of these emerging technologies.