基于中性粒细胞-淋巴细胞比值与血清C反应蛋白-清蛋白比值的累积评分在胶质母细胞瘤患者中的预后意义

Prognostic significance of a cumulative scoring system based on the neutrophil-to-lymphocyte ratio and serum C-reactive protein-to-albumin ratio for glioblastoma patients

目的探讨联合中性粒细胞-淋巴细胞比值(NLR)与血清C反应蛋白-清蛋白比值(CAR)在多形性胶质母细胞瘤(GBM)患者中的预后潜力。方法回顾性分析2015年1月至2020年1月于该院诊断并接受系统治疗的123例成年GBM患者的临床病理资料与实验室检查数据,通过受试者工作特征曲线确定NLR与CAR的最佳截断值,曲线下面积(AUC)评价两种炎症-免疫标志物的对患者总生存率的预测价值。Kaplan-Meier生存分析评估NLR、CAR及NLR-CAR评分与GBM患者预后的关系。采用Cox比例风险回归分析确定具有独立预后意义的临床变量并构建预测GBM患者总生存率的列线图模型。通过一致性指数和校准曲线评价模型的预测能力。结果CAR的最佳截断值为0.06,AUC为0.710(95%CI:0.619~0.802),而NLR的最佳截断值为2.23,AUC值为0.696(95%CI:0.596~0.796)。生存分析表明治疗前高CAR(χ^(2)=7.964,P=0.005)与高NLR(χ^(2)=15.625,P<0.001)均与GBM患者不良预后有关。基于CAR和NLR的最佳截断值构建累积评分系统,0、1和2分患者的中位总生存期分别为43.6个月(95%CI:26.5~60.9)、24.4个月(95%CI:11.2~37.6)和10.0个月(95%CI:7.4~12.6),评分越高,患者预后越差(2分vs.0分:χ^(2)=22.557,P<0.001;2分vs.1分:χ^(2)=6.816,P=0.009;1分vs.0分:χ^(2)=9.964,P=0.002)。单、多变量Cox回归分析证实,全切除与高CAR-NLR评分是GBM患者死亡的独立风险因子,而辅助放疗和IDH1突变是患者的保护因素。基于上述4个独立预后变量构建预测GBM患者总生存率的列线图模型,结果展示该模型的一致性指数为0.775(95%CI:0.746~0.803),校准曲线表明预测生存概率与患者实际生存概率具有高度一致性,提示该列线图具有较好的预测能力。结论基于CAR与NLR的累积评分为GBM患者的预后提供了更为详尽的风险分层,可作为一项客观有效的预后评估工具。

Objective To investigate the prognostic potential of neutrophil-to-lymphocyte ratio(NLR)and serum C-reactive protein-to-albumin ratio(CAR)for glioblastoma multiforme(GBM)patients.Methods The clinicopathological data and laboratory test of 123 adult GBM patients who underwent systematical treatment in a hospital from January 2015 to January 2020 were retrospectively analyzed.The best cut-off values of NLR and CAR for the survival prediction were identified using receiver operation characteristics curves,and the area under curve(AUC)value was used to evaluate the predictive value of immune-inflammatory markers.Kaplan-Meier survival analysis evaluated the relationship between NLR,CAR,and CAR-NLR score and prognosis in patients with GBM.Cox proportional hazard regression analysis was used to identify clinical variables with independent prognostic significance and to construct a nomogram model for predicting overall survival rate of GBM patients.The prediction ability of the nomogram model was evaluated by consistency index and calibration curve.Results The best cut-off value of CAR was 0.06,with an AUC value of 0.710(95%CI:0.619-0.802).The best cut-off value of NLR was 2.23 and its AUC value was 0.696(95%CI:0.596-0.796).Survival analysis indicated that pretreatment high CAR(χ^(2)=7.964,P=0.005)and high NLR(χ^(2)=15.625,P<0.001)were associated with poor prognosis in GBM patients.The cumulative scoring system were constructed by combining these hematological parameters.The Kaplan-Meier curves showed that the median overall survival of patients with 0,1 and 2 point were 43.6 months(95%CI:26.5-60.9),24.4(95%CI:11.2-37.6)months and 10.0 months(95%CI:7.4-12.6),respectively.The higher point for the cumulative scoring system,the worse the prognosis for GBM patients(2 point vs.0 point:χ^(2)=22.557,P<0.001;2 point vs.1 point:χ^(2)=6.816,P=0.009;1 point vs.0 point:χ^(2)=9.964,P=0.002).The univariate and multivariate Cox regression analysis demonstrated that gross total resection and high point of CAR-NLR were independent risk factors for death in GBM patients,while adjuvant radiotherapy and IDH1 mutation were significant protective factors.A nomogram model for predicting overall survival rate of GBM patients was developed based on the above four independent prognostic variables.The data revealed that the consistency index of the nomogram for survival prediction was 0.775(95%CI:0.746—0.803),suggesting that the nomogram model had a strong predictive ability.In addition,the calibration plots showed an excellent consistency between predicted survival probabilities and actual observations.Conclusion The cumulative scoring system based on CAR and NLR provides a good prognostic stratification for GBM patients,and it could be a reliable tool of prognostic assessment.