结直肠癌癌前病变内镜下治疗后复发风险列线图预测模型的建立及评价

Establishment and Evaluation of a Nomogram Prediction Model for Recurrence Risk of Precancerous Lesions of Colorectal Cancer after Endoscopic Treatment

目的探讨结直肠癌癌前病变内镜下治疗后复发的独立危险因素,构建个体化列线图预测模型并进行验证和评价。方法收集2018年5月至2019年9月于哈尔滨医科大学附属第四医院消化内科住院并行内镜下治疗的209例结直肠癌癌前病变患者的临床资料,应用R软件Caret包将患者随机分为训练组(128例)和验证组(81例),采用单因素和多因素Logistic回归分析复发的独立危险因素,应用R软件构建列线图预测模型,应用受试者工作特征曲线下面积、校准曲线和决策曲线对模型进行评价,并用验证组进行验证。结果209例患者中复发132例,复发率为63.16%;训练组中复发81例,未复发47例;验证组复发51例,未复发30例。多因素Logistic回归结果显示,年龄、饮酒史、病变部位、体质量指数是结直肠癌癌前病变内镜下治疗后复发的独立影响因素(OR=1.088,95%CI 1.035~1.144;OR=3.619,95%CI 1.260~10.394;OR=3.316,95%CI 1.355~8.116;OR=1.213,95%CI 1.017~1.447)(P<0.05或P<0.01)。将多因素Logistic回归分析筛选出结直肠癌癌前病变内镜下治疗后复发的独立危险因素和有临床意义的变量息肉个数、组织分级纳入列线图,列线图的曲线下面积在训练组和验证组中分别为0.781(95%CI 0.700~0.863)和0.680(95%CI 0.561~0.800)。校准曲线和决策曲线显示,该预测模型在训练组和验证组中均具有良好的校准度和临床有效性。结论年龄、饮酒史、病变部位、体质量指数是预测结直肠癌癌前病变内镜下治疗后复发的影响因素,本研究构建的列线图模型可直观、个体化分析结直肠癌癌前病变内镜下治疗后复发的风险,为制订个体化复查方案提供参考依据。

Objective To explore the independent risk factors for recurrence after endoscopic treatment of precancerous lesions of colorectal cancer and construct an individualized nomogram prediction model,and validate and evaluate it.Methods The clinical data and follow-up information of 209 patients who underwent endoscopic treatment in the Department of Gastroenterology of the Fourth Affiliated Hospital of Harbin Medical University from May 2018 to Sep.2019 were collected.They were randomly divided into a training group(128 cases)and a validation group(81 cases)with Caret package of R software.Univariate and multivariate Logistic regression analyses were used to identify independent risk factors.R software was used to construct the nomogram prediction model,which was evaluated using area under the receiver operating characteristic curves,calibration curve,and decision curve analysis.And validation was performed using the validation group.Results 132 relapsed in the 209 patients,the recurrence rate was 63.16%;81 relapsed in the training group,47 without recurrence;51 relapsed in the validation group,30 without recurrence.Multivariate Logistic regression analysis showed that age,history of alcohol consumption,lesion location and body mass index were independent risk factors for recurrence after endoscopic treatment of precancerous lesions of colorectal cancer(OR=1.088,95%CI 1.035-1.144;OR=3.619,95%CI 1.260-10.394;OR=3.316,95%CI 1.355-8.116;OR=1.213,95%CI 1.017-1.447)(P<0.05 or P<0.01).The nomogram included these independent risk factors as well as clinically significant variables such as number of polyps and histological grade,and the area under curve of the nomogram in the training group and validation group were 0.781(95%CI 0.700-0.863)and 0.680(95%CI 0.561-0.800),respectively.The calibration curve and decision curve demonstrated good calibration accuracy and clinical effectiveness in both groups.Conclusion Age,history of alcohol consumption,lesion location and body mass index are influencing factors for predicting recurrence after endoscopic treatment of precancerous lesions of colorectal cancer.The nomogram model constructed in this study can realize direct and individualized analysis of the risk of recurrence of precancerous lesions of colorectal cancer after endoscopic treatment,providing a reference for developing individualized follow-up plans.