Surgery today · 2022
Using machine learning models to predict acute pancreatitis in children with pancreaticobiliary maljunction.
СтатьяРандомизированное исследованиеНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Han Xiao, Geng Jia, Zhang Xin-Xian, Zhao Lian, Wang Jian, Guo Wan-Liang
Аннотация
PURPOSE: To develop a model to identify risk factors and predictors of acute pancreatitis in children with pancreaticobiliary maljunction (PBM).
METHODS: We screened consecutive PBM patients treated at two centers between January, 2015 and July, 2021. For machine learning, the cohort was divided randomly at a 6:4 ratio to a training dataset and a validation dataset. Three parallel models were developed using logistic regression (LR), a support vector machine (SVM), and extreme gradient boosting (XGBoost), respectively. Model performance was judged primarily based on the area under the receiver operating curves (AUC).
RESULTS: A total of 99 patients were included in the analysis, 17 of whom suffered acute pancreatitis and 82 did not. The XGBoost (AUC = 0.814) and SVM (AUC = 0.813) models produced similar performance in the validation dataset; both outperformed the LR model (AUC = 0.805). Based on the SHapley Additive exPlanation values, the most important variable in both the XGBoost and SVM models were age, protein plugs, and white blood cell count.
CONCLUSIONS: Machine learning models, especially XGBoost and SVM, could be used to predict acute pancreatitis in children with PBM. The most important contributing factor to the models were age, protein plugs, and white blood cell count.
METHODS: We screened consecutive PBM patients treated at two centers between January, 2015 and July, 2021. For machine learning, the cohort was divided randomly at a 6:4 ratio to a training dataset and a validation dataset. Three parallel models were developed using logistic regression (LR), a support vector machine (SVM), and extreme gradient boosting (XGBoost), respectively. Model performance was judged primarily based on the area under the receiver operating curves (AUC).
RESULTS: A total of 99 patients were included in the analysis, 17 of whom suffered acute pancreatitis and 82 did not. The XGBoost (AUC = 0.814) and SVM (AUC = 0.813) models produced similar performance in the validation dataset; both outperformed the LR model (AUC = 0.805). Based on the SHapley Additive exPlanation values, the most important variable in both the XGBoost and SVM models were age, protein plugs, and white blood cell count.
CONCLUSIONS: Machine learning models, especially XGBoost and SVM, could be used to predict acute pancreatitis in children with PBM. The most important contributing factor to the models were age, protein plugs, and white blood cell count.