Artificial Intelligence Revolutionizing Meniere Disease Screening

Meniere Disease Screening Enhanced by AI

A recent study published online demonstrates how a machine learning model, utilizing pure-tone audiometry features, can effectively diagnose Meniere disease (MD) and predict endolymphatic hydrops (EH). This advancement underscores the role of artificial intelligence in healthcare, particularly in hearing disorders.

Study Overview

Conducted by Xu Liu, M.D., from Fudan University in Shanghai, the research collected both gadolinium-enhanced magnetic resonance imaging (MRI) sequences and pure-tone audiometry data in a retrospective analysis. The collected data informed the training of five classical machine learning models.

Key Findings

• Light Gradient Boosting (LGB) model achieved an accuracy of 87% for diagnosing MD.
• Sensitivity and specificity rates of 83% and 90%, respectively.
• Area under the curve reached 0.95, indicating strong diagnostic capability.
• The LGB model demonstrated 78% accuracy for predicting EH.

This research highlights the important pure-tone audiometry features such as standard deviation and mean hearing levels that are vital for effective MD diagnosis and EH prediction, particularly at low frequency levels.