Related Experiment Video
Updated: Jan 16, 2026
Intraventricular Drug Delivery and Sampling for Pharmacokinetics and Pharmacodynamics Study
Published on: March 31, 2022
Precision dosing of voriconazole in immunocompromised children under 2 years: integrated machine learning and
Li Shen1,2, Mengdi Hu1, Xiaoyong Xu1
1Department of Clinical Pharmacy, Children's Hospital of Fudan University, National Children's Medical Center, Shanghai, China.
Insights
Machine learning and population pharmacokinetic modeling optimized voriconazole dosing for young children. This personalized approach improves drug efficacy and safety in pediatric patients under two years old.
Area of Science:
- Pharmacology
- Pediatric Medicine
- Computational Biology
Background:
- Voriconazole (VRZ) dosing in children under two years requires optimization for improved therapeutic outcomes.
- Individual variability in voriconazole pharmacokinetics necessitates personalized dosing strategies.
Purpose of the Study:
- To develop an individualized dosing strategy for voriconazole in pediatric patients under two years of age.
- Integrate machine learning (ML) and population pharmacokinetic (PopPK) modeling for precise VRZ dosage recommendations.
Main Methods:
- Retrospective analysis of 76 pediatric patients' data, including baseline characteristics and therapeutic drug monitoring samples.
- Developed a PopPK model using NONMEM and applied six ML algorithms, including Boruta feature selection.
- Evaluated models using MSE, RMSE, MAE, and R2, followed by external validation and SHAP analysis for interpretability.
Main Results:
- An XGBoost model achieved high accuracy in predicting voriconazole concentrations (R2 = 0.81, RMSE = 0.53) and performed well in external validation (R2 = 0.75).
- Population pharmacokinetic parameters for apparent clearance (CL/F) and volume of distribution (V/F) were established.
- SHAP analysis identified clearance, weight, and laboratory values as key predictors for VRZ concentration.
Conclusions:
- Personalized voriconazole treatment is crucial for children under 24 months.
- The developed XGBoost model shows significant potential for guiding initial voriconazole dose recommendations in this vulnerable pediatric population.
Objective:
This study aimed to develop an individualized dosing strategy for voriconazole (VRZ) in children under 2 years of age by integrating machine learning (ML) and population pharmacokinetic (PopPK) modeling.
Methods:
This retrospective observational study included 76 eligible pediatric patients for model development, analyzing their baseline characteristics and laboratory parameters. A population pharmacokinetic (PopPK) model using NONMEM® software was performed to assess the clearance (CL) and volume of distribution (V) of VRZ. The individual CL and V were included as input variables. The Boruta algorithm was employed for feature selection, after which six machine learning algorithms were applied. The models were evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R2) to identify the optimal algorithm, which then underwent independent external validation. The selected final model was analyzed for interpretability using Shapley Additive Explanations (SHAP).
Results:
A total of 76 pediatric patients were enrolled for model development, consisting of 58 males (76.3%) and 18 females (23.7%), with a median age of 11 months and a median weight of 8.05 kg. We analyzed 110 therapeutic drug monitoring (TDM) samples of VRZ from these participants. A one-compartment model with first-order absorption and elimination described the population pharmacokinetics of VRZ. Population estimates for apparent clearance (CL/F) and volume of distribution (V/F) were 17.9 L/h/70kg (RSE, 10.8%) and 788 L/70kg (RSE, 15.4%), respectively. An XGBoost model accurately predicted voriconazole concentrations (R2 = 0.81, RMSE = 0.53) with a relative error of ±20% for most observations. In the external validation, the XGBoost model demonstrated an R2 of 0.75, RMSE of 0.14. SHAP analysis identified clearance, weight, and laboratory values as significant predictors.
Conclusion:
This study emphasized the importance of personalized treatment in utilizing VRZ for children under 24 months. The XGBoost model demonstrated potential in identifying an initial dose recommendation for VRZ.
More Related Videos
12:29Live Imaging of Antifungal Activity by Human Primary Neutrophils and Monocytes in Response to A. fumigatus
Published on: April 19, 2017
06:14Optimized LC-MS/MS Method for the High-throughput Analysis of Clinical Samples of Ivacaftor, Its Major Metabolites, and Lumacaftor in Biological Fluids of Cystic Fibrosis Patients
Published on: October 15, 2017
Related Concept Videos
01:29Drug Dosing: Infants and Children
01:26Pharmacokinetics in Pediatric Patients: Drug Excretion
01:25Determination of Multiple Dosing Parameters: Loading and Maintenance Doses
01:17Pharmacokinetics in Pediatric Patients: Drug Distribution
01:24Pharmacokinetics in Pediatric Patients: Drug Metabolism
01:01Dosage Regimens: Partial Pharmacokinetic Parameters