Welcome to Visited Lingnan Modern Clinics In Surgery, Today is

Lingnan Modern Clinics In Surgery ›› 2023, Vol. 23 ›› Issue (04): 309-314.DOI: 10.3969/j.issn.1009-976X.2023.04.004

• Original Articles and Clinical Research • Previous Articles     Next Articles

Development and validation of a clinical prediction model for predicting postoperative complications after radical resection of colon cancer

LIU Xiao-qing1, ZHANG Yu-chao2, WANG Cheng-li1   

  1. 1. Department of Anesthesiology, Sun Yat-sen Memorial Hospital of Sun Yat-sen University, Guangzhou 510120, China;
    2. Department of Gastroenterology, Sun Yat-sen Memorial Hospital of Sun Yat-sen University, Guangzhou 510120, China
  • Contact: WANG Cheng-li, wangchli3@mail.sysu.edu.cn

结肠癌根治术后并发症的临床预测模型的构建与验证

刘晓清1, 张育超2, 王成立1,*   

  1. 1.中山大学孙逸仙纪念医院麻醉科, 广州 510120;
    2.中山大学孙逸仙纪念医院胃肠外科, 广州 510120
  • 通讯作者: *王成立, Email : wangchli3@mail.sysu.edu.cn

Abstract: Objective The aim of this study was to develop and validate a nomogram for predicting postoperative complications after radical resection of colon cancer. Methods Retrospective analysis was conducted based on the relevant data of patients who underwent radical resection of colon cancer at Sun Yat-sen Memorial Hospital of Sun Yat-sen University from January 2019 to June 2021 which was extracted from the electronic medical record system. The main outcome variables of the study were postoperative complications. Cases that meet the inclusion criteria were randomly divided into a model training cohort and an internal validation cohort in a ratio of 7∶3. Firstly, the LASSO regression was used to screen potential influencing factors with non-zero regression coefficients. Then, multiple factor stepwise regression was performed on the selected potential risk factors ad select variables with P<0.05 to further construct a prediction model and visualize it using nomogram. The discriminability, calibration, and clinical application value of the model were evaluated through receiver operating characteristic curve (ROC curve), calibration curve, and decision curve analysis (DCA). Results The final prediction model included four variables: age, duration of surgery, postoperative admission to ICU, and preoperative D-dimer level. The area under the operating characteristic curve (AUC) of the training set and the validation set were 0.791 and 0.889, respectively. The calibration curve and DCA curve indicate that the model has good calibration and clinical practicality. Conclusion We have developed and validated a new predictive model for postoperative complications after radical resection of colon cancer, which only has 4 predictive factors but has good predictive performance. This new tool may help reduce postoperative complications and accelerate patient recovery.

Key words: colon cancer, postoperative complications, nomogram

摘要: 目的 构建和验证预测结肠癌根治术后并发症的列线图。方法 从电子病历系统提取2019年1月~2021年6月在中山大学孙逸仙纪念医院行结肠癌根治术的患者相关资料进行回顾性分析。研究的主要结局指标是术后是否发生并发症。将符合入选标准的病例按照7:3的比例随机分为模型训练集和内部验证集,首先采用最小绝对收缩和选择算子回归即LASSO回归筛选出回归系数不为零的潜在影响因素,然后对筛选出的潜在影响因素进行多因素逐步回归,选择P<0.05的变量进一步构建预测模型并用列线图进行可视化展示。通过ROC曲线、校准曲线和决策曲线分析(DCA)分别评价模型的区分度、校准度和临床应用价值。结果 最终的列线图模型包括年龄、手术时间、术后是否入ICU和术前D-二聚体水平4个变量。训练集与内部验证集的受试者操作特征曲线下面积(AUC)分别为0.791(95%CI:0.771~0.801)和0.889(95%CI:0.873~0.899)。校准曲线和DCA曲线表明该模型具有良好的校准度和临床适用性。结论 本研究开发并验证了一种新的结肠癌根治术后并发症的预测模型,该模型只有4个预测因子,但预测性能较好。这种新型工具应用可能有助于减少术后并发症并加速患者康复。

关键词: 结肠癌, 术后并发症, 列线图

CLC Number: