医学
危险分层
急诊医学
外科
肝细胞癌
急诊外科
回顾性队列研究
失血
接收机工作特性
切除术
急诊科
风险评估
普通外科
风险模型
选择(遗传算法)
作者
Boram Lee,Jai Young Cho,Ho‐Seong Han,Yoo‐Seok Yoon,Hae Won Lee,Yeshong Park,Seung Yeon Lim,Hyelim Joo
摘要
BACKGROUND: Emergency conversion from minimally invasive liver resection (MILR) to open surgery, typically due to unexpected bleeding, is associated with poorer outcomes than elective conversion or completed MILR. However, prior studies rarely distinguished emergency from elective conversion and lacked integrated preoperative prediction tools. This study aimed to develop and validate an artificial intelligence (AI)-based preoperative model to predict emergency conversion. METHODS: We retrospectively analyzed patients undergoing MILR for hepatocellular carcinoma between 2010 and 2023, including only those with emergency conversion due to unexpected bleeding. Thirteen preoperative variables were used to develop a Random Forest model with five-fold cross-validation. Model performance was assessed by the area under the ROC curve (AUC) and calibration. Patients were stratified into four risk groups, and operative and postoperative outcomes were compared. RESULTS: Among 994 patients, 128 (12.9%) underwent emergency conversion. The model demonstrated good discrimination (AUC 0.79; 95% CI, 0.72-0.86). Observed conversion rates increased stepwise across risk groups (7.3%, 9.5%, 16.0%, and 37.7%). Higher risk groups were associated with longer operative times, greater blood loss, and prolonged hospital stays (all p < 0.001). CONCLUSION: This AI-based model enables preoperative risk stratification for emergency conversion during MILR and may support optimized patient selection and surgical planning.
科研通智能强力驱动
Strongly Powered by AbleSci AI