基于外周血细胞的肾癌患者术后预后风险评估与精准化预测

Risk assessment and precision prediction of postoperative prognosis in patients with renal cell carcinoma based on peripheral blood cells

目的 分析肾癌患者术后总生存(OS)的影响因素,建立Cox回归模型并初步评估其对肾癌预后预测的价值。方法 回顾性分析2011年1月至2016年12月于山西医科大学第一医院接受手术治疗的251例肾癌患者的临床资料。采用Kaplan-Meier方法分析并绘制生存曲线、应用Log-rank检验进行单因素分析,结合Cox回归模型分析筛选肾癌预后影响因素;通过绘制森林图和线段式列线图完成对预测模型的可视化,利用时依受试者工作特征(ROC)曲线检验模型的稳定性。结果 多因素分析显示,Fuhrman分级(HR=1.76,95%CI:1.11~2.80)、中性粒细胞(HR=3.18,95%CI:1.57~6.44)、单核细胞(HR=1.58,95%CI:1.01~2.49)、血小板与淋巴细胞比值(HR=1.42,95%CI:1.03~1.97)是肾癌预后的危险因素,中性粒细胞与淋巴细胞比值(HR=0.57,95%CI:0.35~0.94)是其保护因素。在研究对象中随机抽取一个样本(12号个体),列线图可视化结果显示,该个体对应的术后3年、5年、8年死亡率分别为0.39%、1.18%和13.10%,术后3年、5年、8年生存率的ROC曲线下面积分别为0.692、0.679和0.714。结论 本研究基于血小板与淋巴细胞比值、中性粒细胞与淋巴细胞比值、单核细胞、中性粒细胞、Fuhrman分级构建了一种可以预测肾癌患者预后的模型,并验证该模型对肾癌患者OS具有良好的预测效果。

Objective To analyze the influencing factors of postoperative overall survival(OS)in patients with renal cell carcinoma(RCC),establish a Cox regression model and preliminarily evaluate its value,and explore a personalized prediction model for the prognosis of RCC.Methods The clinical data of 251 patients with RCC who received surgical treatment in the First Hospital of Shanxi Medical University from January 2011 to December 2016 were retrospectively analyzed.Kaplan-Meier method was used to draw the survival curve,Log-rank test was used for single factor analysis,and Cox regression model was used to analyze and screen the factors influencing the prognosis of RCC.The visualization of the prediction model was completed by drawing the forest map and the line segment diagram,and the stability of the model was tested by receiver operating characteristic curve(ROC).Results Multivariate analysis showed that Fuhrman grade(HR=1.76,95%CI:1.11-2.80),neutrophils(HR=3.18,95%CI:1.57-6.44),monocytes(HR=1.58,95%CI:1.01-2.49),PLR(HR=1.42,95%CI:1.03-1.97)were risk factors for the prognosis of RCC.NLR(HR=0.57,95%CI:0.35-0.94)was a protective factor.A sample(No.12 individual)was randomly selected from the study subjects,and the visualization results of the nomogram showed that the corresponding 3-,5-,and 8-year mortality rates of this individual were 0.39%,1.18%,and 13.10%,respectively.The area under the ROC curve of the model at 3,5 and 8 years were 0.692,0.679 and 0.714,respectively.Conclusions Based on PLR,NLR,monocyte,neutrophil and Fuhrman grades,a model was constructed to predict the prognosis of patients with RCC,and the model was verified to have a good predictive effect on OS in patients with RCC.