Ensemble learning based functional independence ability estimator for pediatric brain tumor survivors

Area of Science:

• Pediatric Oncology
• Rehabilitation Medicine
• Artificial Intelligence in Healthcare

Background:

• Childhood brain tumors significantly impact cognitive function, daily activities, and quality of life.
• Comprehensive assessments are crucial for identifying support needs in school, home, and community settings.
• Predicting functional recovery aids in tailoring treatment plans and monitoring progress.

Purpose of the Study:

• To evaluate the efficacy of machine learning models in predicting functional independence scores in pediatric brain tumor survivors.
• To compare the predictive performance of ensemble learning models against individual machine learning models for functional outcomes.

Main Methods:

• Selected key features relevant to functional recovery.
• Employed various common machine learning models for prediction.
• Utilized the Functional Independence Measure for Children (WeeFIM) scores as the outcome measure.

Main Results:

• Ensemble learning models demonstrated superior predictive performance compared to individual models.
• The developed machine learning models accurately predicted WeeFIM scores.
• Ensemble models proved more suitable for predicting functional assessment scores in this population.

Conclusions:

• Machine learning, particularly ensemble methods, offers a valuable tool for clinicians to predict functional outcomes in childhood brain tumor survivors.
• Accurate prediction facilitates personalized treatment strategies and enhances patient support.
• This approach can optimize recovery trajectories and improve long-term functional independence.