胃癌病人术后发生静脉血栓栓塞症危险因素分析以及预测模型的建立

Analysis of risk factors of postoperative venous thromboembolism in patients with gastric cancer and establishment of prediction model

目的 探讨影响胃癌病人术后发生静脉血栓栓塞症(VTE)的相关危险因素,并建立预测模型,验证该模型的预测价值。方法 纳入2019年1月~2021年6月我院接受根治性手术的胃癌病人160例作为建模组,167例作为验证组。收集病人临床病理资料,根据建模组胃癌病人术后6个月内VTE的发生情况将其分为VTE组和N-VTE组,对两组胃癌病人的临床病理因素进行单因素分析,再将单因素分析中有统计学意义的指标代入多因素Logistic回归模型进行多因素分析,得到影响胃癌病人术后发生VTE的独立危险因素,基于多因素分析结果得到的独立危险因素结合β值,根据列线图原理对独立危险因素进行赋分,构建列线图模型,并应用R软件绘制列线图,以Bootstrap法和校准曲线进行列线图模型的内外部验证,计算区分度评价指标C指数,并通过拟合优度(H-L)检验评价预测模型的校准能力。结果 建模组160例胃癌病人均接受胃癌根治手术,按术后6个月内是否发生VTE分为VTE组23例和N-VTE组137例,多因素分析结果显示,年龄≥60岁、病灶直径≥5 cm、糖尿病、TNM/T分期为T3~T4期、淋巴结转移5个指标为影响胃癌病人术后发生VTE的独立危险因素(P<0.05),构建列线图:P=1/(1+e^(-X)),X=1.885×年龄(≥60岁=1,<60岁=0)+2.051×糖尿病(是=1,否=0)+2.646×病灶直径(≥5 cm=1,<5 cm=0)+2.952×TNM/T分期(1~2期=0,3~4期=1)+0.694×淋巴结转移(是=1,否=0)-0.436,列线图模型的C指数分别为0.847(95%CI:0.784~0.932)和0.832(95%CI:0.772~0.910),H-L检验提示胃癌病人术后发生VTE的预测值与实际值符合度良好(P>0.05)。结论 本研究建立了关于预测胃癌病人术后发生VTE风险的列线图模型,该模型能够较为准确地预测胃癌病人术后的VTE发生风险。

Objective To explore the related risk factors of postoperative venous thromboembolism(VTE) in patients with gastric cancer,establish a prediction model and verify the predictive value of the model.Methods 160 gastric cancer patients who underwent radical surgery at the First Affiliated Hospital of Hainan Medical College from January 2019 to June 2021 were included as the modeling group,167 cases as validation group.Their clinicopathological data were collected.All modeling group patients were divided into VTE group and N-VTE group according to the occurrence of VTE within 6 months after operation.The clinicopathological factors of the two groups were analyzed by univariate analysis.Then,the statistically significant indexes in the univariate analysis were substituted into the multivariate logistic regression model for multivariate analysis to obtain the independent risk factors affecting the postoperative VTE of patients with gastric cancer.The independent risk factors obtained based on the results of multivariate analysis were combined β Value,assign scores to independent risk factors according to the principle of nomogram,construct the nomogram model,draw the nomogram with R software,internal and external validation of nomogram model with Bootstrap method and calibration curve,calculate the discrimination evaluation Index C index,and evaluate the calibration ability of the prediction model through goodness of fit(H-L).Results 160 modeling group patients with gastric cancer underwent radical gastrectomy.According to the occurrence of VTE within 6 months after operation,they were divided into VTE group(23 cases)(14.38%) and N-VTE group(137 cases)(85.62%).Multivariate analysis showed that the age of 60 years old,the diameter of the lesion was more than 5 cm,the stage of diabetes,the TNM/T stage was 3-4,and the lymph node metastasis was the independent risk factors affecting the postoperative VTE of patients with gastric cancer(P<0.05).Construct nomogram:P=1/(1+e~(-X)),X=1.885×Age(≥ 60 years=1,< 60 years=0)+2.051×Diabetes mellitus(=1,no =0) +2.646×Lesion diameter(≥ 5 cm=1,<5 cm=0) + 2.952 × TNM/T stage(stage 1-2 = 0,stage 3-4 = 1) + 0.694 × Lymph node metastasis(yes = 1,no = 0)-0.436.The C index of nomogram model was 847(95%CI:0.784-0.932) and 0.832(95%CI:0.772-0.910).H-L test showed that the predicted value of postoperative VTE in patients with gastric cancer was in good agreement with the actual value(P>0.05).Conclusion A nomogram model for predicting the risk of postoperative VTE in patients with gastric cancer was established.It was verified that the model can accurately predict the risk of postoperative VTE in patients with gastric cancer.