Pure tone audiogram classification using deep learning techniques.

Area of Science:

• Artificial Intelligence in Healthcare
• Computational Audiology
• Machine Learning for Medical Diagnostics

Background:

• Pure tone audiometry is a foundational audiology tool for diagnosing hearing loss.
• Accurate interpretation of audiograms is crucial for effective patient management.
• Existing diagnostic methods can be time-consuming and require specialized expertise.

Purpose of the Study:

• To develop a deep learning framework for automated audiogram classification.
• To classify audiograms based on hearing loss degree, type, and configuration.
• To enhance diagnostic accuracy and efficiency in audiological assessments.

Main Methods:

• Retrospective analysis of 12,518 audiograms from 6,259 patients (ages 4-96).
• Development of a deep learning model for classifying audiogram characteristics.
• Performance evaluation using accuracy, precision, recall, and F1-score metrics.

Main Results:

• The deep learning framework significantly outperformed traditional machine learning methods.
• Achieved high accuracy rates (96.75%–99.85%) across all classification tasks.
• Demonstrated strong precision (88.93%–98.41%), recall (89.25%–98.38%), and F1-scores (88.99%–98.39%).

Conclusions:

• Deep learning offers a robust and accurate method for classifying pure tone audiograms.
• The framework can assist healthcare professionals in primary care settings, improving diagnostic accuracy.
• Automated classification aids in large-scale data analysis and supports mobile audiometry interpretation for patients.