Tissue Classification After Bone-Anchored Hearing Implant Surgery: A Machine Learning Approach to Monitoring Skin...

Area of Science:

• Medical technology
• Artificial intelligence in healthcare
• Soft tissue analysis

Background:

• Soft tissue health is a critical outcome following percutaneous bone-conduction device (BCD) surgery.
• Current methods for classifying soft tissue health are subjective, impacting treatment decisions and resource management.
• Machine learning (ML) offers a potential solution to standardize tissue classification.

Purpose of the Study:

• To evaluate ML techniques for classifying soft tissue health after BCD surgery.
• To explore the development of ML-driven applications for tissue monitoring.
• To provide evidence supporting the use of AI in audiology and soft tissue management.

Main Methods:

• A convolutional neural network (CNN) model was developed and tested.
• Image samples (N=398) of soft tissues (green, yellow, red) were collected and preprocessed.
• Feature extraction was performed to train the advanced CNN model.

Main Results:

• The model achieved 87% accuracy in classifying healthy versus unhealthy tissue.
• The model demonstrated 94% accuracy in differentiating between minor and severe unhealthy tissue.
• These results indicate high performance of ML in objective tissue assessment.

Conclusions:

• Objective soft tissue classification remains a challenge for clinicians managing BCD users.
• ML-based classification systems could serve as valuable technological aids for tissue monitoring.
• This approach has significant implications for improving patient management and clinical workflows.