Related Experiment Video
Updated: Sep 4, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.9K
Prediction of Lumbar Drainage-Related Meningitis Based on Supervised Machine Learning Algorithms
Peng Wang1, Shuwen Cheng1, Yaxin Li2
1Department of Neurosurgery, Cancer Prevention and Treatment Institute of Chengdu, Chengdu Fifth People's Hospital (The Second Clinical Medical College, Affiliated Fifth People's Hospital of Chengdu University of Traditional Chinese Medicine), Chengdu, China.
Frontiers in Public Health
|July 15, 2022
Summary
Machine learning models can predict lumbar drainage-related meningitis (LDRM). The random forest model demonstrated the best performance in forecasting LDRM risk, with site leakage as a key indicator.
Area of Science:
- Medical Informatics
- Machine Learning in Healthcare
- Infectious Disease Epidemiology
Background:
- Lumbar drainage is a common clinical procedure.
- Forecasting meningitis risk associated with lumbar drainage (LDRM) remains challenging.
- This study aimed to develop predictive models for LDRM using machine learning.
Purpose of the Study:
- To establish and evaluate supervised machine learning models for predicting lumbar drainage-related meningitis (LDRM).
- To identify key predictors for LDRM.
- To compare the performance of different machine learning algorithms in LDRM prediction.
Main Methods:
- Utilized a cohort of 273 lumbar drainage cases.
- Employed supervised machine learning algorithms: Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN).
- Evaluated models using metrics like AUROC, AUPRC, TPR, TNR, specificity, sensitivity, accuracy, and kappa coefficient, with internal validation.
Main Results:
- All models achieved an AUROC exceeding 0.8 in the training set.
- The RF and SVM models showed better performance in the testing set, with RF demonstrating superior predictive accuracy and efficiency.
- Site leakage was identified as the most significant predictor influencing accuracy.
Conclusions:
- Random Forest (RF) and Support Vector Machine (SVM) models are capable of predicting LDRM.
- The RF model exhibited the best overall performance in predicting LDRM.
- Site leakage is the most impactful predictor for LDRM.
