本文已被：浏览 33次 下载 11次
投稿时间：2023-10-15 网络发布日期：2025-01-20
投稿时间：2023-10-15 网络发布日期：2025-01-20
中文摘要: 目的 联合使用logistic回归和关联规则分析法对食管癌根治术后急性呼吸衰竭（ARF）的危险因素进行分析。
方法 选取2019年6月至2022年6月河西学院附属张掖人民医院食管癌根治术后患者146例的临床资料进行回顾性研究。将患者分为ARF组（ n =49）和无ARF组（ n =97）。对ARF的危险因素进行单因素分析，并用logistic回归进行多因素分析；采用Python语言编制FP-Growth算法程序,对ARF组行关联规则分析，计算临床特征间有效强关联规则。
结果 Logistic回归分析显示，影响ARF发生的危险因素为吸烟史（ OR =3.039 , P =0.018）、吻合口瘘（ OR =5.041 , P ＜0.01）、胸腔粘连（ OR =7.993 , P ＜0.01）、低蛋白血症（ OR =3.831 , P ＜0.01）。关联规则分析显示，临床特征间有效强关联规则共11条：（1） 二项规则分别为年龄（60～69岁）∩ARF、吸烟史∩ARF、肺部手术史∩ARF、手术时长（≥3 h）∩ARF、吻合口瘘∩ARF、胸腔粘连∩ARF、低蛋白血症∩ARF，其置信度 C 值（即ARF发生的概率）为0.73～0.85；（2） 三项规则分别为年龄（60～69岁）∩肺部手术史∩ARF、肺部手术史∩胸腔粘连∩ARF、手术时长（≥3 h）∩低蛋白血症∩ARF、吸烟史∩低蛋白血症∩ARF，其 C 值提高至0.94及以上。
结论 相比logistic回归分析，关联规则分析中二项规则更为丰富，三项规则结果进一步缩小了ARF高危范围。两种方法联用有利于联合筛选食管癌根治术后ARF的危险因素，三项规则在指导临床针对性干预方面更有应用价值。
中文关键词: 食管癌根治术 急性呼吸衰竭 危险因素 Logistic回归分析 关联规则分析
Abstract:Objective To analyze the risk factors of acute respiratory failure (ARF) after radical resection for esophageal cancer by combining logistic regression analysis and association rule analysis.
Methods The clinical data of 146 patients after radical resection for esophageal cancer in Zhangye People's Hospital Affiliated to Hexi University from June 2019 to June 2022 were retrospectively studied. Patients were divided into the ARF group ( n =49) and the non-ARF group ( n =97). Univariate analysis of risk factors for ARF was performed, and multivariate analysis was performed by logistic regression in two groups. The FP-Growth algorithm program was compiled in Python, and the association rule analysis was performed to calculate the effective strong association rules between the clinical features in the ARF group.
Results Logistic regression analysis showed that the risk factors of ARF were smoking history ( OR =3.039,P =0.018), anastomotic fistula( OR =5.041,P ＜0.01), thoracic adhesion ( OR =7.993,P ＜0.01) and hypoproteinemia ( OR =3.831,P ＜0.01). The analysis of association rules showed that there were 11 effective strong association rules between clinical features: (1) the two rules were age (60~69 years old) ∩ ARF, smoking history ∩ARF, lung surgery history ∩ARF, operation duration(≥3 h) ∩ ARF, anastomotic fistula ∩ARF, thoracic adhesion ∩ARF, hypoproteinemia ∩ARF, and their confidenceC value (i.e., the probability of ARF occurrence) was 0.73~0.85; (2) the three rules were age (60~69 years old)∩ lung surgery history ∩ARF, lung surgery history ∩ thoracic adhesion ∩ARF, operation duration (≥3 h) ∩ hypoproteinemia ∩ARF, smoking history ∩ hypoproteinemia ∩ARF, and theirC value increased to 0.94 and above.
Conclusion Compared with logistic regression analysis, the two rules in the association rule analysis are more abundant, and the results of the three rules further narrow the high-risk range of ARF. The combination of the two methods is conducive to the joint screening of risk factors for ARF after radical resection for esophageal cancer, and the three rules are more valuable in guiding clinical intervention.
keywords: Radical resection for esophageal cancer Acute respiratory failure Risk factor Logistic regression analysis Association rule analysis
文章编号： 中图分类号：R735.1 文献标志码：A
基金项目:甘肃省高等学校创新基金项目（2021B-234)
附件
| Author Name | Affiliation |
| LEI Xiuwen, ZHU Xiaolei, TIAN Long | epartment of Cardiothoracic Oncology, Zhangye People's Hospital Affiliated to Hexi University, Zhangye, Gansu 734000, China |
引用文本：
雷秀文，朱晓磊，田龙.两种方法对食管癌根治术后急性呼吸衰竭危险因素的分析[J].中国临床研究,2025,38(1):62-66.
雷秀文，朱晓磊，田龙.两种方法对食管癌根治术后急性呼吸衰竭危险因素的分析[J].中国临床研究,2025,38(1):62-66.