Comparison of Computed Tomography-Based Artificial Intelligence Modeling and Magnetic Resonance Imaging in Diagnosis of Cholesteatoma

• DOI: 10.5152/iao.2023.221004
• Publication Date: 2023-03-31T10:17:38Z
• Keywords: Otitis Media; Diffusion Magnetic Resonance Imaging; Otorhinolaryngology; RF1-547; Cholesteatoma, Middle Ear; Artificial Intelligence; Humans; Reproducibility of Results; Original Article; Magnetic Resonance Imaging; Retrospective Studies

In this study, we aimed to compare the success rates of computed tomography image-based artificial intelligence models and magnetic resonance imaging in diagnosis preoperative cholesteatoma.The files 75 patients who underwent tympanomastoid surgery with chronic otitis media between January 2010 2021 our clinic were reviewed retrospectively. The classified into group without cholesteatoma (n=34) (n=41) according presence at surgery. A dataset was created from images patients. dataset, determined by using most frequently used literature. addition, MRI evaluated compared.Among architectures paper, lowest result obtained MobileNetV2 an accuracy 83.30%, while highest DenseNet201 90.99%. specificity 88.23% sensitivity 87.80%.In showed that can be similar reliability cholesteatoma. This is first study that, knowledge, compares for purpose identifying cholesteatomas.