Machine learning prediction model for post- hepatectomy liver failure in hepatocellular carcinoma: A multicenter
Jitao Wang1,2, Tianlei Zheng3,4, Yong Liao1
1Xingtai Key Laboratory of Precision Medicine for Liver Cirrhosis and Portal Hypertension, Xingtai People's Hospital, Xingtai, Hebei, China.
Frontiers in Oncology
|November 21, 2022
Summary
A new machine learning (ML) model accurately predicts post-hepatectomy liver failure (PHLF) in hepatocellular carcinoma (HCC) patients. This novel approach offers improved prediction accuracy compared to existing non-invasive methods.
Area of Science:
- Hepatocellular Carcinoma (HCC) Research
- Machine Learning in Medicine
- Surgical Complication Prediction
Background:
- Post-hepatectomy liver failure (PHLF) is a critical complication following liver surgery for hepatocellular carcinoma (HCC).
- Accurate prediction of PHLF is essential for improving patient outcomes and surgical decision-making.
Purpose of the Study:
- To develop and validate a novel machine learning (ML) model for predicting PHLF.
- To utilize the Light Gradient Boosting Machines (LightGBM) algorithm for enhanced predictive accuracy.
Main Methods:
- A cohort of 875 HCC patients undergoing hepatectomy was divided into training, validation, and testing groups.
- Shapley additive explanation (SHAP) was employed to identify key predictive variables.
- The ML model's performance was evaluated using AUC, sensitivity, specificity, and decision curve analysis (DCA).
Main Results:
- The ML model achieved high AUC values: 0.944 (training), 0.870 (validation), and 0.822 (testing).
- The developed ML model demonstrated superior predictive performance compared to existing non-invasive models.
Conclusions:
- A novel, validated ML model effectively predicts PHLF using common clinical parameters.
- This advanced ML model offers a more valuable tool for PHLF prediction than current non-invasive methods.
