Translational cancer research · 2026
Predictive modelling of duodenal stump leakage after gastric cancer and long-term oncological outcomes.
СтатьяНазвание 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.
Shu Xufeng, Wu Ahao, Wang Xiaoqiang, Yuan Chendong, Zhang Guoyang, Xiong Jianbo, Chen Wenzheng, Wang Guosen, Li Zhengrong, Jie Zhigang
Аннотация
BACKGROUND: Duodenal stump leakage (DSL) is a significant postoperative complication of gastric cancer that has a detrimental impact on patient prognosis. However, there is currently no established DSL prediction model. Furthermore, the long-term prognostic implications of DSL in gastric cancer patients remain unclear. The aim of this study was to construct a risk prediction model for DSL to guide clinical decision-making and assess the long-term prognosis of DSL.
METHODS: A retrospective analysis of clinical characteristics from 2,511 patients was conducted to develop and evaluate a DSL prediction model using multiple machine learning techniques. These included least absolute shrinkage and selection operator, support vector machine (SVM), decision tree (DT), random forest (RF), k-nearest neighbor, and logistic regression (LR) for clinical feature selection. Kaplan-Meier survival curves were generated to assess the long-term prognostic significance of DSL in gastric cancer.
RESULTS: In the training cohort, the RF and SVM models achieved areas under the curve (AUCs) of 1.000 and 1.000, respectively, while in the test cohort, their AUCs were 0.944 and 0.895. The LR model showed greater stability in the test cohort (AUC 0.954). The three most important risk factors were preoperative albumin, distal margin distance, and postoperative day 4 C-reactive protein (CRP). DSL did not significantly affect overall survival (OS) (P=0.13) but was associated with reduced recurrence-free survival (RFS) (P=0.03).
CONCLUSIONS: The construction of DSL risk prediction models with RF can facilitate more effective clinical management. Furthermore, our findings indicate that DSL affects RFS after gastric cancer surgery.
METHODS: A retrospective analysis of clinical characteristics from 2,511 patients was conducted to develop and evaluate a DSL prediction model using multiple machine learning techniques. These included least absolute shrinkage and selection operator, support vector machine (SVM), decision tree (DT), random forest (RF), k-nearest neighbor, and logistic regression (LR) for clinical feature selection. Kaplan-Meier survival curves were generated to assess the long-term prognostic significance of DSL in gastric cancer.
RESULTS: In the training cohort, the RF and SVM models achieved areas under the curve (AUCs) of 1.000 and 1.000, respectively, while in the test cohort, their AUCs were 0.944 and 0.895. The LR model showed greater stability in the test cohort (AUC 0.954). The three most important risk factors were preoperative albumin, distal margin distance, and postoperative day 4 C-reactive protein (CRP). DSL did not significantly affect overall survival (OS) (P=0.13) but was associated with reduced recurrence-free survival (RFS) (P=0.03).
CONCLUSIONS: The construction of DSL risk prediction models with RF can facilitate more effective clinical management. Furthermore, our findings indicate that DSL affects RFS after gastric cancer surgery.