术前预测肝细胞癌伴微血管侵犯风险列线图模型的建立与验证  

Development and validation of a nomogram model for preoperative prediction of hepatocellular carcinoma with microvascular invasion

作　　者：万康康 潘树波[1] 倪良平 熊奇如 谢胜学[1] 王龙胜[2] 刘涛 孙浩楠 马菊 王惠敏 喻宗繁[1] Wan Kangkang;Pan Shubo;Ni Liangping;Xiong Qiru;Xie Shengxue;Wang Longsheng;Liu Tao;Sun Haonan;Ma Ju;Wang Huimin;Yu Zongfan(Department of General Surgery,the Second Affiliated Hospital of Anhui Medical University,Hefei 230000,China;Department of Radiology,the Second Affiliated Hospital of Anhui Medical University,Hefei 230000,China;Anhui Medical University,Hefei 230000,China)

机构地区：[1]安徽医科大学第二附属医院普外科,合肥230000 [2]安徽医科大学第二附属医院放射科,合肥230000 [3]安徽医科大学,合肥230000

出　　处：《中华肝胆外科杂志》2023年第8期561-566,共6页Chinese Journal of Hepatobiliary Surgery

基　　金：安徽省高校自然科学基金(KJ2021A0325);安徽医科大学第二附属医院临床孵育计划(2022LCYB15)。

摘　　要：目的基于术前增强CT影像学特征和临床资料建立预测肝细胞癌(HCC)微血管侵犯(MVI)的列线图模型并验证其效能。方法回顾性分析安徽医科大学第二附属医院2018年5月至2022年5月手术治疗的210例HCC患者的临床资料,其中男性172例,女性38例,年龄(59±10)岁。按7∶3比例通过系统抽样法随机分为训练组(n=147)和验证组(n=63)。收集患者的术前增强CT影像学特征和临床资料,logistic回归分析HCC合并MVI的危险因素,建立并验证包含危险因素的列线图。通过绘制训练组和验证组的受试者工作特征(ROC)曲线、校准曲线、决策曲线分析(DCA)、临床影响曲线(CIC)评估列线图预测HCC患者MVI状态的诊断效能。结果多因素分析结果显示,甲胎蛋白≥400μg/ml、肿瘤内坏死、肿瘤长径≥3 cm、肿瘤边界不清楚、肿瘤周围有子灶是预测HCC发生MVI的独立危险因素。基于以上因素建立列线图预测模型,该列线图模型在训练组和验证组中预测HCC合并MVI的ROC曲线下面积分别为0.866(95%CI:0.807~0.924)和0.834(95%CI:0.729~0.939)。DCA结果表明,当净收益率>0时预测模型阈值在训练组和验证组中分别为7%~93%和12%~87%。CIC结果显示,本预测模型判断的发生MVI的群体与实际发生MVI群体高度匹配。结论本研究建立的基于影像学特征和临床资料的列线图预测模型可在术前较为准确地预测HCC患者的MVI状态,经验证该模型具有良好的临床应用前景。Objective To develop and validate a nomogram model for predicting microvascular invasion(MVI)in hepatocellular carcinoma(HCC)based on preoperative enhanced computed tomography imaging features and clinical data.Methods The clinical data of 210 patients with HCC undergoing surgery in the Second Affiliated Hospital of Anhui Medical University from May 2018 to May 2022 were retrospectively analyzed,including 172 males and 38 females,aged(59±10)years old.Patients were randomly divided into the training group(n=147)and validation group(n=63)by systematic sampling at a ratio of 7∶3.Preoperative enhanced computed tomography imaging features and clinical data of the patients were collected.Logistic regression was conducted to analyze the risk factors for HCC with MVI,and a nomogram model containing the risk factors was established and validated.The diagnostic efficacy of predicting MVI status in patients with HCC was assessed by receiver operating characteristic(ROC)curve,calibration curves,decision curve analysis(DCA),and clinical impact curve(CIC)of the subjects in the training and validation groups.Results The results of multifactorial analysis showed that alpha fetoprotein≥400μg/ml,intra-tumor necrosis,tumor length diameter≥3 cm,unclear tumor border,and subfoci around the tumor were independent risk factors predicting MVI in HCC.A nomogram model was established based on the above factors,in which the area under the curve(AUC)of ROC were 0.866(95%CI:0.807-0.924)and 0.834(95%CI:0.729-0.939)in the training and validation groups,respectively.The DCA results showed that the predictive model thresholds when the net return is>0 ranging from 7%to 93%and 12%to 87%in the training and validation groups,respectively.The CIC results showed that the group of patients with predictive MVI by the nomogram model are highly matched with the group of patients with confirmed MVI.Conclusion The nomogram model based on the imaging features and clinical data could predict the MVI in HCC patients prior to surgery.

关 键 词：癌 肝细胞 计算机断层扫描 微血管侵犯 列线图 

分 类 号：R735.7[医药卫生—肿瘤]