Related Experiment Video
Updated: Jan 11, 2026
08:43
A Unified Methodological Framework for Vestibular Schwannoma Research
Published on: June 20, 2017
7.7K
Machine learning model for predicting hearing preservation after vestibular schwannoma surgery: A meta-analysis
Paweł Łajczak1, Emma Ann Finnegan2, Enzo von Quednow3
1Medical University of Silesia, Katowice, Poland; Geneuro - International Research Group in Neuroscience, Vitoria, Espirito Santo, Brazil.
American Journal of Otolaryngology
|November 11, 2025
Summary
Machine learning models accurately predict hearing preservation in vestibular schwannoma (VS) patients undergoing surgery. Further research is needed to improve model generalizability and personalize treatment strategies for better patient outcomes.
Area of Science:
- Neurosurgery
- Medical Informatics
- Oncology
Background:
- Vestibular schwannomas (VS) are benign tumors impacting hearing, tinnitus, and balance.
- Current surgical and radiosurgery treatments face challenges in predicting post-operative hearing preservation.
- Machine learning (ML) shows promise for predicting hearing outcomes in VS patients.
Purpose of the Study:
- To systematically evaluate the diagnostic accuracy of ML models for predicting hearing preservation in VS patients undergoing surgery.
- To assess the effectiveness of ML in a clinical setting for VS management.
Main Methods:
- A systematic review and meta-analysis adhering to PRISMA-DTA guidelines.
- Inclusion of studies applying ML to predict hearing preservation in VS patients.
- Extraction of diagnostic accuracy metrics (sensitivity, specificity, accuracy).
Main Results:
- Analysis of 15 ML models from 3 studies.
- Pooled sensitivity: 0.856, specificity: 0.853, accuracy: 0.839.
- Area under the ROC curve of 0.883 indicates high diagnostic effectiveness, despite observed heterogeneity.
Conclusions:
- ML models demonstrate high accuracy in predicting hearing preservation for VS patients.
- Significant heterogeneity across studies necessitates further research for model optimization and generalizability.
- ML holds potential for personalized treatment strategies and improved patient outcomes in VS management.
