肝癌电子杂志 ›› 2021, Vol. 8 ›› Issue (2): 62-68.

• 护理园地 • 上一篇    下一篇

NRS-2002和CONUT评分与中晚期肝细胞癌患者预后分析及Nomogram模型构建

黄桂荣1, 吴德平1, 陈方鹏1, 韩山山2,*   

  1. 1.皖西卫生职业学院附属医院肝胆外科,安徽 六安 237000;
    2.北京朝阳急诊抢救中心普外科,北京 100021
  • 收稿日期:2020-12-09 发布日期:2021-07-16
  • 通讯作者: *韩山山,E-mail: HanSS_1992@163.com
  • 作者简介:黄桂荣, 主管护师,皖西卫生职业学院附属医院,肝胆外科

NRS-2002 and CONUT scores and prognostic analysis of patients with advanced hepatocellular carcinoma and Nomogram model construction

Huang Guirong1, Wu Deping1, Chen Fangpeng1, Han Shanshan2,*   

  1. 1. Hepatobiliary surgery, Affiliated Hospital of West Anhui Health Vocational College, Lu'an 237000, Anhui, China;
    2. General Surgery,Beijing Chaoyang Emergency Rescue Center,Beijing 10021,China
  • Received:2020-12-09 Published:2021-07-16

摘要: 目的: 采用营养风险筛查-2002(nutritional risk screening-2002,NRS-2002)与控制营养状态(controlling nutritional status,CONUT)评分评估中晚期肝细胞癌(hepatocellular carcinoma,HCC)患者经导管动脉栓塞化疗(transcatheter arterial chemoembolization,TACE)术前营养状态,并分析术后死亡影响因素,构建Nomogram模型。方法: 选取2016年1月至2020年1月皖西卫生职业学院附属医院接受TACE治疗的113例中晚期HCC患者。采用单因素及多因素Cox比例风险模型分析中晚期HCC患者死亡影响因素。将死亡影响因素作为构建Nomogram模型的预测指标。采用内部数据集验证、一致性指数(consistency index, C-index)、决策曲线分析(decision curve analysis,DCA)及时间-ROC曲线评估Nomogram模型预测性能。结果: NRS-2002筛查出76例(67.3%)患者营养不良;CONUT评分筛查出58例(51.3%)患者营养不良。Kappa值为0.238(P<0.05),筛查结果一致性一般。NRS-2002及CONUT评分筛查出来的营养不良患者生存时间低于营养正常患者(P<0.05)。多因素Cox风险比例模型分析结果显示,腹水(是)、BCLC分级(C级)、PVTT(是)、TACE次数(> 2次)、TBIL(>27μmol/L)、NRS-2002(营养不良)是中晚期HCC患者死亡的独立危险因素(P<0.05)。内部数据集验证结果显示,6个月生存率、12个月生存率、24个月生存率的C-index分别为0.701(95%CI:0.668~0.801)、0.697(95%CI:0.611~0.794)、0.651(95%CI:0.601~0.754)。预测中晚期HCC患者TACE术后死亡Nomogram模型的风险阈值>0.16而<0.89时,提供显著附加临床收益。Nomogram模型的AUC在各个时间点均高于腹水、BCLC分级、PVTT、TACE次数、TBIL及NRS-2002。结论: 该Nomograms模型具有较高临床实践性及准确性,可指导中晚期HCC患者治疗管理。

关键词: 肝细胞癌, 经导管动脉栓塞化疗, 生存率, 模型

Abstract: Objective: Nutritional risk screening-2002 (NRS-2002) and controlling nutritional status (CONUT) scores were used to evaluate the nutritional status of patients with advanced hepatocellular carcinoma (HCC) before transcatheter arterial chemoembolization (TACE).The influencing factors of postoperative death were analyzed and Nomogram model was constructed. Methods: 113 patients with advanced HCC who received TACE therapy in Affiliated Hospital of West Anhui Health Vocational College from January 2016 to January 2020 were selected.Univariate and multivariate Cox proportional hazards models were used to analyze the factors affecting the death of patients with advanced HCC.The influential factors of death were used as predictors of the Nomogram model.Internal dataset validation, consistency index(C-index), decision curve analysis (DCA), and time-ROC curve were used to evaluate the predictive performance of Nomogram model. Results: NRS-2002 screened 76 (67.3%) patients with malnutrition, and 758 (51.3%) patients with malnutrition were screened by CONUT score. Kappa value was 0.238 (P < 0.05). The survival time of malnutrition patients screened by NRS-2002 and CONUT score was lower than that of patients with normal nutrition (P < 0.05). Multivariate Cox proportional hazards model analysis showed that ascites (yes), BCLC grade (grade C), PVTT (yes), TACE frequency (>2), TBIL (> 27 μ mol/L) and NRS-2002 (malnutrition) were the independent risk factors of death in patients with advanced HCC (P < 0.05). The results of internal validation showed that the C-index of 6-month survival rate, 12-month survival rate and 24-month survival rate were 0.701 (95%CI: 0.668—0.801), 0.697 (95%CI: 0.611—0.794) and 0.651 (95%CI: 0.601—0.754), respectively. When the risk threshold of nomogram model was higher than 0.16 and lower than 0.89 for predicting the death of patients with advanced HCC after TACE, it provides significant additional clinical benefits. The AUC of Nomogram model was higher than that of BCLC, PVTT, ascites, TACE times, TBIL and NRS-2002. Conclusion: The Nomogram model has high clinical practice and accuracy, and can guide the treatment and management of patients with advanced HCC.

Key words: Hepatocellular carcinoma, Transcatheter arterial chemoembolization, Survival rate, Model