腹腔镜结肠癌根治术后肝转移风险列线图预测模型的建立及其临床意义

Establishment and clinical significance of nomograph model for liver metastasis risk prediction after laparoscopic radical resection of colon cancer

摘要:

目的 构建腹腔镜结肠癌根治术后发生肝转移风险的列线图预测模型，并探讨该模型的临床意义。方法 选取2016年9月1日－2018年12月31日于该院行腹腔镜结肠癌根治术的155例结肠癌患者作为研究对象，根据肝部转移的具体情况，分为肝转移组（n = 36）和无肝转移组（n = 119）。收集两组患者的性别、年龄、肿瘤部位、组织类型、TNM分期、原发灶浸润深度、淋巴转移、肿瘤沉积、肿瘤大小、术后癌胚抗原（CEA）、术后化疗、输血和梗阻等临床资料，分析结肠癌术后肝部转移的影响因素，采用多因素Cox回归分析确定独立危险因素，建立腹腔镜结肠癌根治术后肝转移风险的列线图预测模型，并对该模型进行验证。结果 单因素分析表明，两组患者TNM分期、术后化疗、肿瘤沉积、肿瘤大小、淋巴转移和梗阻比较，差异均有统计学意义（P < 0.05）；多因素Cox回归分析结果显示，TNM分期 ≥ Ⅲ期、肿瘤沉积、淋巴转移和梗阻是腹腔镜结肠癌根治术后发生肝转移的独立危险因素（P < 0.05），术后化疗为保护性因素，肿瘤大小（P = 0.824）为排除指标。使用森林图对5项临床因素的多因素Cox回归结果进行展示。根据5项临床指标建立列线图模型，1、3和5年的肝转移率校准曲线与列线图结果相贴合，受试者操作特征曲线（ROC curve）的曲线下面积（AUC）为0.843。结论 腹腔镜结肠癌根治术后肝转移风险的列线图预测模型，具有较高的精准度和区分度，能够为腹腔镜结肠癌根治术后的患者提供更为精准的肝转移风险预测。

Abstract:

Objective To construct a nomogram model for predicting the risk of liver metastasis in patients after laparoscopic radical resection of colon cancer, and explore its clinical significance.Methods This study included a total of 155 patients who underwent laparoscopic radical resection of colon cancer from September 1, 2016, to December 31, 2018. Follow-up was conducted after surgery, and based on the presence or absence of liver metastasis, the patients were divided into the liver metastasis group (n = 36) and the non-liver metastasis group (n = 119). Clinical data such as gender, age, tumor site, tissue type, TNM stage, infiltration depth of primary lesion, lymphatic metastasis, tumor deposition, tumor size, postoperative carcinoembryonic antigen (CEA), postoperative chemotherapy, blood transfusion and obstruction were collected in the two groups. The influencing factors of postoperative liver metastasis of colon cancer were analyzed, and independent risk factors were determined by multivariate Cox regression analysis. To establish a correlation nomogram prediction model for the risk of liver metastasis after laparoscopic radical resection of colon cancer and verify the model.Results The univariate analysis results of clinical data of two groups showed that there were statistically significant differences in six indicators: TNM stage, postoperative chemotherapy, tumor deposition, tumor size, lymphatic metastasis and obstruction (P < 0.05). The subsequent multivariate Cox regression analysis showed that stage ≥ III, tumor deposition, lymphatic metastasis and obstruction were independent risk factors for liver metastasis after laparoscopic radical resection of colon cancer (P < 0.05), while chemotherapy after operation was a protective factor, and tumor size (P = 0.824) was an exclusion index. Cox regression results of five clinical factors were displayed by forest diagram. The nomogram model was established according to five clinical indicators. The calibration curve of liver metastasis rate at 1, 3 and 5 years was consistent with the nomogram results, and the area under the ROC curve (AUC) was 0.843.Conclusion A new reliable nomograph model for predicting the risk of liver metastasis in patients after laparoscopic radical resection of colon cancer has been established. The model has good accuracy and differentiation, and can provide more accurate risk prediction of liver metastasis for patients after laparoscopic radical resection of colon cancer.