Predictive Model of Oxaliplatin-induced Liver Injury Based on Artificial Neural Network and Logistic Regression
Rui Huang1, Yuanxuan Cai1, Yisheng He2
1School of Pharmacy, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China.
Journal of Clinical and Translational Hepatology
|January 1, 2024
Summary
Artificial neural network (ANN) models show improved prediction of oxaliplatin-induced liver injury (OILI) compared to logistic regression. Key predictors include patient age, chemotherapy details, and co-administered drugs.
Area of Science:
- Hepatology
- Medical Informatics
- Oncology
Background:
- Oxaliplatin-induced liver injury (OILI) poses a risk to patients undergoing chemotherapy.
- Predictive tools for identifying high-risk individuals are currently lacking.
- Early identification of OILI risk is crucial for patient management.
Purpose of the Study:
- To develop and evaluate artificial neural network (ANN) and logistic regression (LR) models for predicting OILI risk.
- To compare the performance of ANN and LR models in identifying patients at risk of OILI.
Main Methods:
- Prospective data collection from 10 hospitals on patients treated with oxaliplatin.
- Utilized the updated Roussel Uclaf causality assessment method (RUCAM) to diagnose OILI.
- Developed and evaluated ANN and LR models using patient and medication characteristics.
Main Results:
- The incidence of OILI was 3.65%.
- The ANN model demonstrated superior discrimination (AUC 0.920 vs. 0.833) and calibration compared to the LR model.
- Important predictors identified included age, chemotherapy regimens/cycles, oxaliplatin dosage, and use of glucocorticoids/antihistamines.
Conclusions:
- The ANN model offers improved discriminative and calibration abilities for predicting OILI risk over the LR model.
- Co-administered chemotherapy drugs may influence OILI, suggesting potential idiosyncratic reactions.
- Further research is needed to validate prophylactic medication strategies for OILI prevention.
