SMC develops AI model for diagnosing Ménière's disease


Samsung Medical Center (SMC) developed the world’s first deep-learning-based artificial intelligence (AI) model for diagnosing Ménière’s disease, one of the exemplary vertigo-causing diseases. The study was published in Scientific Reports.


The research team devised the INner Ear Hydrops Estimation via ARtificial InTelligence (INHEARIT) model based on the CNN algorithm, which is useful for image learning and pattern processing. The model is designed to automatically identify the cochlea and vestibular organs by analyzing the magnetic resonance imaging (MRI) scan of the inner ear using deep learning technology and calculating the ratio of endolymphatic hydrops occupied per area. 


To verify the model’s effectiveness, the team compared the calculation results of radiologists and otolaryngologists and AI from MRI images of 124 patients. According to the research team, the conformity ratio of the results was recorded at 0.971, which is significantly high, showing no notable differences in the results between doctors and AI. By saving time and effort through this model, doctors can reserve more energy for patient diagnosis and treatment.