低位直肠癌病人肛全直肠系膜切除术后中期排便功能障碍发生风险Nomogram模型构建

Construction of a nomogram model of the risk of mid-term defecation dysfunction after TaTME in patients with low rectal cancer

目的:分析影响低位直肠癌病人行肛全直肠系膜切除术(transanal total mesorectal excision,TaTME)后中期排便功能障碍发生的危险因素,并建立Nomogram模型。方法:选择2016年1月至2020年1月172例低位直肠癌并接受腹腔镜辅助TaTME手术病人作为建模组,选择2020年2月至2021年2月间接受相同手术病人32例作为验证组;建模组病人根据术后6个月随访时是否发生低位前切除综合征(low anterior resection syndrome,LARS)分为LARS组和非LARS组,比较2组病人病历资料,采用单因素和多因素logistic回归模型对影响LARS发生的独立危险因素进行分析,并采用R软件建立预测低位直肠癌病人TaTME术后LARS发生的列线图模型,并进行验证。结果:病人的LARS评分为(14.78±4.54)分,其中LARS病人84例(48.84%),非LARS病人LARS 88例(51.16%)。单因素、多因素logistic回归分析显示术前放疗、肛门括约肌损伤、吻合口距离肛缘距离、肿瘤直径、坐骨棘间径长度均是影响LARS的独立危险因素(P<0.05~P<0.01)。列线图模型预测建模组、验证组的C-index分别为0.836和0.827。ROC曲线显示,列线图模型预测建模组、验证组发生LARS的AUC分别为0.829、0.808。结论:术前放疗、肛门括约肌损伤、吻合口距离肛缘距离、肿瘤直径、坐骨棘间径长度是影响是否发生LARS的独立危险因素,以上述指标建立的列线图模型具有良好预测效能。

Objective:To analyze the independent risk factors of mid-term defecation dysfunction after transanal total mesorectal excision(TaTME)in patients with low rectal cancer,and establish a nomogram model.Methods:A total of 172 patients with low rectal cancer who received laparoscopic assisted TaTME from January 2016 to January 2020 were selected as the training group,and 32 patients who received the same operation from February 2020 to February 2021 were selected as the validation group;the patients in training group were divided into LARS and non-LARS group according to the occurrence of low anterior resection syndrome(LARS)at the follow-up of 6 months after surgery.The medical records of the patients in two groups were compared.The independent risk factors affecting the occurrence of LARS were analyzed by using univariate and multivariate logistic regression models.The nomogram model for predicting the LARS in low rectal cancer patients after TaTME was established by using R software and validated.Results:The LARS score of patients was(14.78±4.54)points,including 84 LARS patients(48.84%)and 88 non-LARS patients(51.16%).Univariate and multivariate logistic regression analysis showed that preoperative radiotherapy,anal sphincter injury,distance between anastomotic stoma and anal margin,tumor diameter,and length of ischial spine were independent risk factors for LARS(P<0.05 to P<0.01).The C-indexes of the nomogram model in both training group and validation group were are 0.836 and 0.827,respectively.ROC curve showed that the AUCs of the nomogram model in both training group and validation group were 0.829 and 0.808,respectively.Conclusions:Preoperative radiotherapy,anal sphincter injury,anastomotic distance from anal margin,tumor diameter,and ischial interspinous diameter are the independent risk factors for LARS.The nomogram model established with the above indicators has good predictive efficacy.