Surgical endoscopy · 2026
Machine learning-based models for intraoperative blood loss of retroperitoneal laparoscopic adrenalectomy.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
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.
He Jing, Zhao Jiahui, Luo Yong, Li Mingchuan, Han Yili, Wei Dechao, Wang Yongxing, Hou Zhu
Аннотация
OBJECTIVE: Intraoperative blood loss (IBL) is a critical precipitating factor of intraoperative and postoperative complications. This study aims to identify risk factors and employ machine learning (ML) techniques to develop models for predicting IBL in patients undergo retroperitoneal laparoscopic adrenalectomy (RLA).
METHODS: A retrospective study was performed with data from patients who underwent unilateral RLA at Beijing Anzhen Hospital from 2014 to 2021. The data were randomly divided into training and validation sets. A volume ≥ 100 ml was defined as high IBL. Model training incorporated feature selection with least absolute shrinkage and selection operator (LASSO) and algorithms including logistic regression (LR), random forest (RF), neural network (NN), K-nearest neighbor (KNN), support vector machine (SVM) and AdaBoost. Various indicators were utilized to evaluate model performance.
RESULTS: A total of 530 RLA cases were included. LASSO regression identified gender, surgeons' experience, disease type, lesion diameter and lesion location as predictive factors. Six ML models were derived and evaluated. Among which, LR, RF and NN models exhibited exceptional discriminatory power. The area under the curve (AUC) of the three models were 0.775, 0.769, 0.729 in the validation set, respectively. A nomogram and a online calculator were created. The calibration curve, decision curve analysis and clinical impact curve revealed its excellent clinical utility.
CONCLUSION: This study presented and compared the use of various ML models for predicting IBL in RLA, which could assist in assessing surgical risks and tailored treatments without additional examination, and providing insights for the clinical application of ML.
METHODS: A retrospective study was performed with data from patients who underwent unilateral RLA at Beijing Anzhen Hospital from 2014 to 2021. The data were randomly divided into training and validation sets. A volume ≥ 100 ml was defined as high IBL. Model training incorporated feature selection with least absolute shrinkage and selection operator (LASSO) and algorithms including logistic regression (LR), random forest (RF), neural network (NN), K-nearest neighbor (KNN), support vector machine (SVM) and AdaBoost. Various indicators were utilized to evaluate model performance.
RESULTS: A total of 530 RLA cases were included. LASSO regression identified gender, surgeons' experience, disease type, lesion diameter and lesion location as predictive factors. Six ML models were derived and evaluated. Among which, LR, RF and NN models exhibited exceptional discriminatory power. The area under the curve (AUC) of the three models were 0.775, 0.769, 0.729 in the validation set, respectively. A nomogram and a online calculator were created. The calibration curve, decision curve analysis and clinical impact curve revealed its excellent clinical utility.
CONCLUSION: This study presented and compared the use of various ML models for predicting IBL in RLA, which could assist in assessing surgical risks and tailored treatments without additional examination, and providing insights for the clinical application of ML.