Jin Donghui, Yuan Ligong, Li Feng, Wang Shuaibo, Mao Yousheng
Department of Thoracic Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, China.
Surg Endosc. 2022 Nov;36(11):8144-8153. doi: 10.1007/s00464-022-09249-z. Epub 2022 Apr 20.
BACKGROUND: Pneumonia is a common complication after minimally invasive esophagectomy (MIE), which prolongs hospital stay, adding to the cost and increasing the risk to patients' lives. This study aimed to identify risk factors and establish a predictive nomogram for postoperative pneumonia (PP). METHODS: This case control study included 609 patients with esophageal cancer who underwent MIE between March 2015 and August 2019 in Cancer Hospital, Chinese Academy of Medical Sciences. We randomly divided the data into training and validation sets in the ratio of 7:3 and performed univariate and multivariate logistic regression analyses to acquire independent risk factors of the training set. We constructed a nomogram based on the independent risk factors. The concordance index (C-index), receiver operating characteristic (ROC) curve, calibration plot, and decision curve analysis (DCA) plots were used to evaluate the discrimination of the nomogram. Validation set was applied to confirm the predictive value of the nomogram. RESULTS: In the univariate analysis, age, gender, abdominal procedure method, thoracic operative time, duration of chest tube placement, anastomotic leakage, and recurrent laryngeal nerve palsy were found to be correlated with the incidence of PP. In multivariate analysis, all variables except thoracic operative time were found to be independent risk factors for PP. A nomogram was constructed based on these independent risk factors. The C-index of the training and validation sets was 0.769 and 0.734, respectively, and the areas under the curve (AUC) of ROC curves of the training and validation sets were 0.769 and 0.686, respectively. The calibration plots and DCA plots of the training and validation sets showed the accuracy and predictive value of the nomogram. CONCLUSION: The nomogram could accurately identify the risk factors for PP. We could predict the occurrence of PP based on this nomogram and take corresponding measures to reduce the incidence of PP.
背景:肺炎是微创食管切除术后常见的并发症,会延长住院时间,增加费用并提高患者生命风险。本研究旨在确定术后肺炎(PP)的危险因素并建立预测列线图。 方法:本病例对照研究纳入了2015年3月至2019年8月在中国医学科学院肿瘤医院接受微创食管切除术的609例食管癌患者。我们以7:3的比例将数据随机分为训练集和验证集,并进行单因素和多因素逻辑回归分析以获取训练集的独立危险因素。我们基于独立危险因素构建了列线图。采用一致性指数(C指数)、受试者工作特征(ROC)曲线、校准图和决策曲线分析(DCA)图来评估列线图的辨别力。应用验证集来确认列线图的预测价值。 结果:在单因素分析中,发现年龄、性别、腹部手术方式、胸腔手术时间、胸管放置时间、吻合口漏和喉返神经麻痹与PP的发生率相关。在多因素分析中,发现除胸腔手术时间外的所有变量都是PP的独立危险因素。基于这些独立危险因素构建了列线图。训练集和验证集的C指数分别为0.769和0.734,训练集和验证集的ROC曲线下面积(AUC)分别为0.769和0.686。训练集和验证集的校准图和DCA图显示了列线图的准确性和预测价值。 结论:该列线图可以准确识别PP的危险因素。我们可以基于此列线图预测PP的发生并采取相应措施降低PP的发生率。
Zhonghua Zhong Liu Za Zhi. 2023-12-23
Front Med (Lausanne). 2025-6-11
Diagnostics (Basel). 2024-2-17
Anticancer Res. 2019-5
World J Gastroenterol. 2019-1-21
Ann Surg. 2019-2