已发表论文

利用机器学习探索老年胃肠道肿瘤手术患者术后住院时间与炎症相关因素之间的关联

 

Authors Liu J, Hu Y, Zhan X, Wang H, Li C

Received 24 May 2025

Accepted for publication 26 August 2025

Published 12 September 2025 Volume 2025:18 Pages 12589—12597

DOI https://doi.org/10.2147/JIR.S542616

Checked for plagiarism Yes

Review by Single anonymous peer review

Peer reviewer comments 2

Editor who approved publication: Professor Ning Quan

Jing Liu,1,2,* Yubang Hu,2,* Xiaoxuan Zhan,2,* Huanwei Wang,2 Cai Li1 

1Department of Anesthesiology, Nanfang Hospital, Southern Medical University, Guangzhou, People’s Republic of China; 2Department of Anesthesiology, the Eighth Affiliated Hospital, Sun Yat-Sen University, Shenzhen, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Cai Li, Department of Anesthesiology, Nanfang Hospital, Southern Medical University, No. 1838 Guangzhou Dadao Bei, Baiyun District, Guangzhou, Guangdong, 510515, People’s Republic of China, Email licaisysu@163.com

Background: The influencing factors of postoperative recovery in elderly patients undergoing gastrointestinal tumor surgery are complex, and inflammatory responses have been confirmed to play a critical role. However, the specific impact of inflammatory cells on postoperative length of hospital stay (LOS) in elderly patients following general anesthesia for gastrointestinal surgery remains unclear.
Methods: This study collected perioperative data from elderly patients who underwent gastrointestinal tumor surgery at a tertiary hospital between 2019 and 2023. By analyzing the relationship between perioperative factors and postoperative LOS, we first used the LASSO algorithm for variable selection and then employed XGBoost modeling to identify factors associated with rapid (LOS < 7 days) and delayed (LOS > 12 days) discharge. We further explored how inflammatory cells affect postoperative recovery in elderly patients, with results visualized using SHAP (SHapley Additive exPlanations) plots.
Results: A total of 300 elderly patients who underwent gastrointestinal tumor resection were included, with 58.7% male and a mean age of 72 ± 6.4 years. Key factors associated with rapid discharge included operative time, age, blood loss, total intraoperative fluid input, urine output, preoperative neutrophil percentage, hemoglobin, and preoperative creatinine. Specifically, a preoperative neutrophil percentage in the 60%– 70% range was linked to faster discharge. For delayed discharge, significant factors included age, operative time, preoperative creatinine, preoperative hemoglobin, preoperative neutrophil percentage, intraoperative urine output, the difference in neutrophil counts between pre- and postoperative periods, total intraoperative fluid input, and blood loss. Patients with preoperative neutrophil percentages of 70%– 80% and neutrophil count differences of 5%– 25% were more likely to experience delayed discharge.
Conclusion: In elderly patients undergoing gastrointestinal tumor resection, preoperative neutrophil percentage and the change in neutrophil percentages from pre- to postoperative periods significantly influence postoperative LOS.

Keywords: elderly patients, gastrointestinal tumor surgery, inflammatory cells, machine learning