食管癌患者三切口食管切除术后发生颈部吻合口瘘的列线图预测模型构建

Construction of nomogram prediction model for post-operative jugular anastomotic fistula of tri-incisional esophagectomy in patients with esophageal cancer

目的分析食管癌患者三切口食管切除术(McKeown术)后颈部吻合口瘘(AF)发生的影响因素,绘制列线图构建颈部AF预测模型。方法选取2017年3月至2021年3月收治的503例拟行McKeown术的食管癌患者作为研究对象,记录基线资料,根据术后颈部AF发生情况分为颈部AF组(发生颈部AF)和无颈部AF组(未发生颈部AF),比较2组基线资料,经Logistic回归分析检验食管癌患者McKeown术后颈部AF发生的影响因素,绘制列线图构建颈部AF预测模型,并绘制决策曲线检验颈部AF预测模型对食管癌患者McKeown术后颈部AF发生风险的预测价值。结果503例食管癌患者中,20例在McKeown术后发生AF,发生率为3.97%;颈部AF组肿瘤位置在颈段、合并心脑血管疾病、合并胃供血动脉钙化的比例高于无颈部AF组,白蛋白、FEV1%低于无颈部AF组,手术时间长于无颈部AF组,差异有统计学意义(P<0.05)。Logistic回归,结果显示,肿瘤位置在颈段、合并心脑血管疾病、手术时间长、合并胃供血动脉钙化是食管癌患者McKeown术后颈部AF发生的影响因素(OR>1,P<0.05);白蛋白、FEV1%水平高是食管癌患者McKeown术后颈部AF发生的保护因素(OR<1,P<0.05);绘制列线图构建颈部AF风险预测模型,验证模型显示C-index值为0.873,模型区分度良好,校准曲线Y与X直线相近,模型准确度良好;绘制决策曲线,结果显示,在阈值0~1.0内,颈部AF风险预测模型预测食管癌患者McKeown术后颈部AF发生的净受益率基本>0,有临床意义。结论食管癌患者McKeown术后发生颈部AF与肿瘤位置、心脑血管疾病、白蛋白、FEV1%、手术时间、胃供血动脉钙化有关,根据这些因素建立的颈部AF风险预测模型具有一定的预测价值,可获得较高的净受益率。

Objective To analyze the factors of post-operative jugular anastomotic fistula(AF)of tri-incisional esophagectomy(McKeown)in patients with esophageal cancer(EC),and to draw the nomogram to construct jugular AF prediction model.Methods Totally 503 EC patients who were proposed to take McKeown procedure from March 2017 to March 2021 were selected as the research subjects,the baseline data of the patients were recorded,and the patients were divided into jugular AF group(with jugular AF)and non-jugular AF group(without jugular AF)based on the incidences of post-operative jugular AF.The baseline data of two groups were compared.Logistic regression analysis was applied to test the factors of post-operative jugular AF following McKeown procedure.The nomogram was drawn to asset up the jugular AF prediction model,the decision curve was drawn to assess the predictive value.Results After operation,there were 20 patients with post-operative AF with an incidence rate for 3.97%among 503 EC patients;the proportion of cervical tumor,cardiovascular/cerebrovascular diseases,and gastric feeding artery calcification of patients in the jugular AF group was higher than those in the non jugular AF group,the incidence of albumin and FEV1%was lower than those in the non-jugular AF group,the operation duration was longer than that in the non jugular AF group,with statistically significant differences(P<0.05).Logistic regression results showed that the cervical tumor,cardiovascular/cerebrovascular diseases,longer operation duration and gastric feeding artery calcification were factors for post-operative jugular AF following McKeown procedure(OR>1,P<0.05);while high level of albumin and FEV1%was the protective factor(OR<1,P<0.05).The nomogram was drawn to set up the jugular AF risk prediction model.The validation model showed that the C index value was 0.873,the model discrimination was satisfactory,the calibration curve y was close to Line X,and the model accuracy was fit.The decision curve was drawn,and the results showed that the net benefit rate of post-operative jugular AF risk prediction model was basically greater than 0within the threshold range of 0~1.0,being worthy of clinical significance.Conclusion After McKeown procedure,the incidence of post-operative jugular AF of EC patients is related to tumor location,cardiovascular/cerebrovascular diseases,albumin,FEV1%,operation duration and gastric feeding artery calcification;the risk prediction model of jugular AF established according to the factors delivers certain prediction value and may acquire a high net benefit rate.