食管癌患者术后静脉血栓栓塞症的危险因素分析及预测模型建立

Risk factor analysis and estoblishment of predictive modeling of postoperative venous thromboembolism in patients with esophageal cancer

目的 探究食管癌患者术后静脉血栓栓塞症(VTE)的危险因素,构建个性化预测患者术后VTE发生率的统计学模型。方法 回顾性分析2020年1月至2021年6月于河北医科大学第四医院胸外科行标准食管癌根治术的127例患者的临床资料,分为VTE组(32例)和非VTE组(95例)。采用logistic回归分析探究食管癌患者术后VTE的危险因素,构建预测食管癌患者术后VTE风险的列线图。结果 VTE组年龄、高血压、糖尿病、体重指数、术前D-二聚体、手术时长、开左胸高于非VTE组,差异有统计学意义(P <0.05)。多因素logistic分析显示,糖尿病是VTE发生的独立危险因素(OR> 1,P <0.05);而手术时长及手术方式是影响VTE发生的独立保护因素(OR <1,P <0.05)。构建了预测食管癌患者术后VTE风险的列线图,校准曲线与理想曲线较为贴近,ROC曲线的曲线下面积为0.769,决策曲线显示阈值概率在0~73.1%时对患者进行干预可能会使患者获得临床受益。结论 糖尿病、手术时长、手术方式是食管癌患者术后VTE的独立危险因素,本研究中所构建的列线图模型对术后VTE发生有较高的预测价值。

Objective To investigate the risk factors of postoperative venous thromboembolism(VTE) in patients with esophageal cancer and to construct a statistical model for personalized prediction of the incidence of postoperative VTE in patients. Methods The clinical data of 127 patients who underwent standard radical esophageal cancer surgery in the Department of Thoracic Surgery of the Fourth Hospital of Hebei Medical University from January 2020 to June 2021 were retrospectively analyzed and divided into VTE group(32 patients) and non-VTE group(95 patients). Logistic regression analysis was used to explore the risk factors for postoperative VTE in patients with esophageal cancer and to construct a column line graph predicting the risk of postoperative VTE in patients with esophageal cancer. Results Age, hypertension,diabetes mellitus, body mass index, preoperative D-dimer, duration of surgery, and open left chest were higher in the VTE group than in the non-VTE group, with statistically significant differences(P < 0.05). Multi-factor logistic analysis showed that diabetes was an independent risk factor for the development of VTE(OR > 1, P < 0.05);while the length of surgery and the mode of surgery were independent protective factors for the development of VTE(OR < 1, P < 0.05). Column plots were constructed to predict the risk of postoperative VTE in patients with esophageal cancer, and the calibration curve was more closely aligned with the ideal curve, with an area under the curve of 0.769 for the ROC curve, and the decision curve showed that interventions in patients with a threshold probability of 0 to 73.1% may result in clinical benefit. Conclusion Diabetes mellitus, duration of surgery, and mode of surgery are independent risk factors for postoperative VTE in patients with esophageal cancer, and the column line graph model constructed in this study has a high predictive value for the occurrence of postoperative VTE.