预测结直肠癌远处转移的临床预测模型:结合术前CEA、CA19-9及临床病理因素

Clinical predictive model for predicting distant metastasis of colorectal cancer:combined with preoperative CEA,CA19-9 and clinicopathological factors

目的探讨术前肿瘤标志物CEA、CA19-9联合检测相关临床病理因素在结直肠癌远处转移中的预测价值。方法回顾性分析了接受R0根治术的2 758例结直肠癌患者临床资料,包括年龄、性别、肿瘤分化程度、肿瘤直径、肿瘤位置、CEA、CA19-9、肿瘤组织学类型、远处转移等临床资料,通过对可能反映结直肠癌患者出现远处转移的因素进行单因素及多因素分析,进而选择合适的变量构建列线图,并通过校准曲线来验证列线图的识别性。绘制受试者工作特征曲线(ROC)检验列线图对于伴随远处转移的结直肠癌患者的预测性能。结果单因素及多因素分析结果表示,术前CEA、CA19-9、肿瘤部位、肿瘤分化程度是反映结直肠癌患者出现远处转移的重要因素(P<0.05)。上述变量构建列线图,该预测模型的ROC值为0.771(95%置信区间:0.730~0.813)。结论术前CEA、CA19-9联合临床病理信息构建的预测模型是无创预测结直肠癌患者发生远处转移的有效方法。

Objective To explore the predictive value of preoperative tumor markers CEA and CA19-9 in predicting distant metastasis of colorectal cancer.Methods The clinical data of 2758 patients with colorectal cancer who underwent radical resection of R0 were analyzed retrospectively,including age,sex,degree of tumor differentiation,tumor diameter,tumor site,CEA,CA19-9,tumor histological type and distant metastasis.The factors that may reflect distant metastasis in patients with colorectal cancer were analyzed by univariate and multivariate analysis,and then appropriate variables were selected to construct the Nomogram.The recognition of the Nomogram is verified by the calibration curve.Subject operating characteristic analysis(ROC)was used to compare the ability of the Nomogram to predict distant metastasis in patients with colorectal cancer.Results Univariate and multivariate results showed that preoperative CEA,CA19-9,tumor site and tumor differentiation were important factors reflecting distant metastasis in patients with colorectal cancer(P<0.05).The above variables are used to construct the Nomogram,and the ROC value of the prediction model is 0.771(95%confidence interval:0.730~0.813).Conclusion The predictive model constructed by preoperative CEA and CA19-9 combined with basic pathological information is an effective method for non-invasive prediction of distant metastasis in patients with colorectal cancer.