肝细胞癌术后早期复发的预测模型构建与验证  

Construction and Validation of a Predictive Model for Early Recurrence after Hepatocellular Carcinoma Surgery

作　　者：周文考 陈嘉佳 黄何森 黄灵炎 王明山 牛文霞 王越[2] 唐慧敏 Zhou Wenkao;Chen Jiajia;Huang Hesen;Huang Lingyan;Wang Mingshan;Niu Wenxia;Wang Yue;Tang Huimin(Xiang’An Hospital of XiaMen University,Fujian Xiamen 361000,China;Liaoning Cancer Hospital&Institute,Liaoning Shenyang 110000,China)

机构地区：[1]厦门大学附属翔安医院,福建厦门361000 [2]辽宁省肿瘤医院,辽宁沈阳110000

出　　处：《青岛医药卫生》2024年第3期181-185,共5页Qingdao Medical Journal

基　　金：厦门市青年项目(3502Z202372092);厦门市医疗卫生指导性项目(3502Z20224ZD1155);厦门大学附属翔安医院青年基金资助项目(XAH23003)。

摘　　要：目的 通过对本中心HCC数据的研究,构建一个能够精准评估HCC术后早期复发风险的预测模型。方法 纳入2014年3月-2017年3月在辽宁省肿瘤医院进行手术的401原发性HCC患者,根据术后2年是否复发分为早期复发组(n=176人)和早期未复发组(n=225人)。卡方检验、独立样本t检验和单因素Mann-Whitney U检验对组间数据行差异性分析。“glmnet”程辑包对影响早期复发的危险因素进行单因素、多因素Logistic回归。“rms”程辑包行列线图的绘制,“pROC”绘制验证ROC曲线,评估模型效能。结果 差异分析和单因素逻辑回归显示腹水、AFP、肿瘤数目、肿瘤大小、巴塞罗那分级、微血管侵犯、分化级别、肿瘤包膜侵犯和切除方式是影响HCC术后早期复发的危险因素(P<0.05),而腹水、巴塞罗那分级、微血管侵犯和分化级别是影响其早期复发的独立危险因素(P<0.05)。以单因素逻辑回归中危险因素构建的预测模型具有较高的诊断效能,其训练集AUC:0.710(0.650—0.771),验证集AUC:0.714(0.620—0.808)。结论 基于本中心数据构建的早期复发预测模型具有较高的预测价值,但还需进一步验证和优化。Objective To develop a predictive model that can accurately assess the risk of early re-currence of HCC post-surgery,based on data from our center.Methods A total of 401 primary HCC patients who underwent surgery at Liaoning Cancer Hospital&Institute between March 2014 and March 2017 were included.They were divided into an early recurrence group(n=176)and an early non-recurrence group(n=225)based on whether recurrence occurred within 2 years post-surgery.Chi-square test,independent samples t-test,and uni-variate Mann-Whitney U test were used to analyze the differences between the groups.The"glmnet"package was used for uni-variate and multi-variate logistic regression analysis of risk factors for early recurrence.The"rms"package was used to create a nomogram,and"pROC"was used to plot and validate the ROC curve to evaluate the model's performance.Results The differential analysis and uni-variate logistic regression showed that ascites,AFP,number of tumors,tumor size,Barcelona Clinic Liver Cancer(BCLC)stage,micro-vascular invasion,differentiation grade,tumor capsule inva-sion,and resection method were risk factors for early recurrence of HCC post-surgery(P<0.05).Ascites,BCLC stage,microvascular invasion,and differentiation grade were independent risk factors for early recurrence(P<0.05).The predictive model constructed with the risk fac-tors identified in uni-variate logistic regression showed high diagnostic performance,with an AUC of 0.710(0.650-0.771)for the training set and 0.714(0.620-0.808)for the validation set.Conclusion The early recurrence prediction model constructed based on data from our center has high predictive value,but further validation and optimization are needed.

关 键 词：肝细胞癌 复发风险 预测模型 内部验证 

分 类 号：R737.9[医药卫生—肿瘤]