A Machine-Learning Method of Predicting Vital Capacity Plateau Value for Ventilatory Pump Failure Based on Data...

Area of Science:

• Biomedical Engineering
• Data Science in Medicine
• Pediatric Pulmonology

Background:

• Ventilatory pump failure is a significant cause of mortality in patients with neuromuscular diseases.
• Vital capacity plateau value (VCPLAT) is a critical indicator for assessing ventilatory pump failure in conditions such as congenital myopathy, Duchenne muscular dystrophy, and spinal muscular atrophy.
• Predicting VCPLAT in pediatric patients is challenging due to the intricate relationship between the value and individual patient conditions.

Purpose of the Study:

• To establish a predictive model for VCPLAT in pediatric patients with neuromuscular diseases using data mining and machine learning.
• To improve the accuracy of VCPLAT prediction, aiding in the assessment of disease severity and clinical decision-making.
• To develop a robust and validated model that outperforms existing prediction methods.

Main Methods:

• Correlation analysis and recursive feature elimination with cross-validation (RFECV) were employed for feature selection.
• A Light Gradient Boosting Machine (LightGBM) algorithm was utilized to build the VCPLAT prediction model.
• The model's performance was rigorously evaluated using 10-fold cross-validation and comparison with other prediction models.

Main Results:

• The proposed LightGBM model demonstrated superior performance with an explained variance score (EVS) of 0.949 and R-squared (R²) of 0.948.
• Key performance metrics included mean absolute error (MAE) of 0.028, mean squared error (MSE) of 0.002, root mean square error (RMSE) of 0.045, and median absolute error (MedAE) of 0.015.
• The model exhibited strong performance on independent test datasets, confirming its predictive accuracy and effectiveness.

Conclusions:

• The developed machine learning model accurately and effectively predicts VCPLAT in pediatric patients with neuromuscular diseases.
• Accurate VCPLAT prediction can assist clinicians in determining disease severity and making informed decisions for patient diagnosis and treatment.
• This data-driven approach offers a valuable tool for augmenting clinical judgment in managing pediatric neuromuscular respiratory conditions.