肝细胞癌微血管侵犯危险因素分析及术前预测列线图模型构建

doi:10.7659/j.issn.1005-6947.2021.07.003

首页 > 按期查看>2021年第7期 >772-779. DOI:10.7659/j.issn.1005-6947.2021.07.003

肝细胞癌微血管侵犯危险因素分析及术前预测列线图模型构建
DOI:
                        10.7659/j.issn.1005-6947.2021.07.003
                    
作者:
                        
通讯作者:
                        
作者单位:安徽医科大学附属省立医院 肝脏外科，安徽 合肥 230001
作者简介:邓家仲，安徽医科大学附属省立医院硕士研究生，主要从事肝脏外科临床研究。
通信作者:荚卫东，Email: jwd1968@sina.com
基金项目:2019年中国科学院科技促进发展局STS计划重点基金资助项目（KFJ-STS-ZDTP-080）；安徽省重点研究与开发计划基金资助项目（1704a0802150）。

Analysis of risk factors for microvascular invasion in hepatocellular carcinoma and construction of preoperative predictive nomogram

Author:

Affiliation:

Department of Hepatic Surgery, Provincial Hospital Affiliated to Anhui Medical University, Hefei 230001, China

Fund Project:

摘要

图/表

访问统计

参考文献

相似文献

引证文献

资源附件

音频文件

视频文件

摘要:

背景与目的 术前有效预测微血管侵犯（MVI）对肝细胞癌（HCC）患者的临床决策、术后辅助治疗和全面的预后评估具有重要的临床价值。因此，本研究探讨HCC合MVI的危险因素并建立术前风险预测列线图模型，以期为临床提供参考。方法回顾分析2017年1月—2020年11月安徽省立医院收治的535例HCC患者临床资料，将患者按入院时间分为模型组（433例），验证组（102例）。进行单因素和多因素分析，以确定MVI的独立危险因素，应用R软件建立预测术前HCC的MVI风险的列线图模型，用Bootstrap法进行模型的内部验证，用验证组进行模型的外部验证，采用一致性指数、较正曲线及受试者工作特征（ROC）曲线来评估列线图的预测价值。结果模型组多因素分析显示，NLR>2.282（OR=1.864，95% CI=1.184~2.933）、GGT>60 IU/L（OR=2.554，95% CI=1.631~4.001）、lgAFP（OR=1.455，95% CI=1.21~1.75）、肿瘤大小（OR=1.177，95% CI=1.084~1.277）、无完整假包膜（OR=2.019，95% CI=1.286~3.171）是术前预测HCC患者MVI的独立危险因素，并以此建立的列线图模型一致性指数在模型组和验证组为分别为0.785（95% CI=0.742~0.828）、0.824（95% CI=0.737~0.91）。模型与校准预测曲线贴合良好，通过Youden指数计算出列线图的最佳临界值为103分，临界值下的敏感度、特异度、阳性预测值和阴性预测值在模型组分别为86%、61%、67%和82%，在验证组中分别为82%、56%、53%和83%。结论 NLR>2.282、GGT>60 IU/L、lgAFP、肿瘤大小、无完整假包膜是HCC发生MVI的独立影响因素，以此建立的列线图模型术前预测MVI效能良好，可直观的分析术前合并微血管侵犯的发生风险，甄别出高风险人群。

Abstract:

Background and Aims For patients with hepatocellular carcinoma (HCC), the effective prediction of presence or absence of microvascular invasion (MVI) is of great importance in clinical decision making, postoperative adjuvant therapy and systematic prognostic evaluation. Therefore, this study was conducted to investigate the risk factors for MVI in HCC and to establish a preoperative predictive nomogram, so as to provide a clinical reference.Methods The clinical data of 535 patients with HCC treated in Anhui Provincial Hospital from January 2017 to November 2020 were retrospectively analyzed. According to admission time, they were divided into model group (433 cases) and validation group (102 cases). Univariate and multivariate analyses were carried out to determine the independent risk factors for MVI. R software was used to establish a nomogram model to predict the preoperative MVI risk of hepatocellular carcinoma. Bootstrap analysis was used for internal validation of the model, and validation group was used for external validation of the model. C-index, calibration and receiver operating characteristic (ROC) curves were used to evaluate the predictive value of the nomogram.Results In the model group, multivariate analysis showed that NLR>2.282 (OR=1.864, 95% CI=1.184-2.933), GGT>60 IU/L (OR=2.554, 95% CI=1.631-4.001), lgAFP (OR=1.455, 95% CI=1.21-1.75), tumor size (OR=1.177, 95% CI=1.084-1.277) and absence of complete capsule (OR=2.019, 95% CI=1.286-3.171) were independent risk factors for preoperative prediction of MVI in patients with HCC. The C-index of the nomogram model established based no above factors were 0.785 (95% CI=0.742-0.828) and 0.824 (95% CI=0.737-0.91) in model group and validation group, respectively. The model fitted well with the calibration prediction curve. Based on the Youden index, the optimal critical value of the nomogram was 103. The sensitivity, specificity, positive predictive value and negative predictive value under the critical value were 86%, 61%, 67% and 82% in the model group, and 82%, 56%, 53% and 83% in the validation group, respectively.Conclusion NLR>2.282, GGT>60 IU/L, lgAFP, tumor size and absence of complete capsule were the independent risk factors for MVI in HCC. The established nomogram has a good preoperative performance in predicting MVI, which can directly analyze the preoperative risk of MVI and identify the high-risk population.