Cochlear Implantation in Postlingually Deaf Adults is Time-sensitive Towards Positive Outcome: Prediction using...

Area of Science:

• Audiology
• Biomedical Engineering
• Machine Learning in Healthcare

Background:

• Age-related hearing loss is a growing public health concern, often requiring cochlear implants (CI) for severe-to-profound cases.
• Current prediction models for post-operative CI outcomes have limitations, necessitating improved approaches.
• Predicting cochlear implant success is crucial for managing patient expectations and optimizing surgical outcomes.

Purpose of the Study:

• To develop and validate a machine learning model for predicting cochlear implant (CI) outcomes in postlingually deaf adults.
• To identify key preoperative factors influencing CI performance, specifically word recognition scores (WRS).
• To compare the predictive accuracy of a Random-Forest Regression (RFR) model against traditional linear models.

Main Methods:

• A Random-Forest Regression (RFR) model was developed to predict postoperative word recognition scores (WRS).
• Key predictors included duration of deafness (DoD), age at CI operation (ageCI), duration of hearing-aid use (DoHA), and preoperative hearing metrics.
• Model performance was assessed using mean absolute error (MAE) and Pearson's correlation coefficient (r), with cross-hospital validation.

Main Results:

• The RFR model achieved high prediction accuracy (r=0.96, MAE=6.1) significantly outperforming a linear model (r=0.7, MAE=15.6).
• Cross-hospital validation confirmed the RFR model's reliability (r=0.91, MAE=9.6).
• Duration of deafness (DoD) was the most influential factor, followed by age at CI operation (ageCI) and duration of hearing-aid use (DoHA). Patients with shorter DoD (<10 years) showed better outcomes.

Conclusions:

• Machine learning, specifically RFR, offers a robust method for predicting CI outcomes in postlingually deaf adults.
• Preoperative factors, particularly the duration of deafness, significantly influence CI success.
• Early CI intervention and continued hearing aid use are recommended for patients with severe-to-profound hearing loss.