基于临床指标和MRI特征的列线图预测肝细胞癌患者微血管转移模式的研究
A nomogram based on clinical factors and MRI features for prediction of microvascular metastasis patterns in patients with hepatocellular carcinoma
瞿琦 1鲁梦恬 1刘子鑫 1张继云 1顾春燕 2姜吉锋 1张学琴 1张涛1
作者信息
- 1. 南通大学附属南通第三医院(南通市第三人民医院)放射科,南通 226006
- 2. 南通大学附属南通第三医院(南通市第三人民医院)病理科,南通 226006
- 折叠
摘要
目的 基于临床指标和钆塞酸二钠(Gd-EOB-DTPA)增强MR1特征构建列线图,探讨该列线图对肝细胞癌(HCC)患者血管包绕肿瘤细胞簇(VETC)阳性和/或微血管侵犯(MVI)阳性的预测价值.方法 回顾性分析2014年8月至2021年12月在南通大学附属南通第三医院行肝癌根治性切除术的240例HCC患者的临床资料,其中男性169例,女性71例,年龄(58.4±9.1)岁,年龄范围35～82岁.术后病理结果显示,VETC和/或MVI阳性的HCC患者有121例,VETC和MVI均为阴性的HCC患者有119例.收集患者的性别、年龄、术前维生素K缺乏或拮抗剂Ⅱ诱导的蛋白(PIVKA-Ⅱ)、甲胎蛋白(AFP)水平等临床资料.多因素logistic回归分析确定VETC和/或MVI阳性的风险因素,并构建列线图模型.采用C指数和受试者工作特征(ROC)曲线评价模型的区分度,绘制校准图评价模型的校准度.结果 多因素logistic回归结果显示,血清PIVKA-Ⅱ≥40 mAU/ml(OR=3.686,95％CI:1.883～7.217,P＜0.001)、瘤周强化(OR=2.513,95％CI:1.154～5.476,P=0.020)、≥50％动脉期低强化成分(OR=3.042,95％CI:1.535～6.028,P=0.001)、肝胆期瘤周低信号(OR=2.682,95％CI:1.250～5.754,P=0.011)及低肝胆期肿瘤-肝实质信号强度比(OR=0.024,95％CI:0.004～0.133,P＜0.001)的HCC患者,VETC和/或MVI阳性的风险高.整合上述5个独立危险因素构建列线图模型,其C指数为0.828,ROC曲线下面积为0.83(95％CI:O.77～0.87),且校准图显示模型的预测概率和实际概率之间具有良好的一致性.结论 基于Gd-EOB-DTPA增强MRI特征和临床指标建立的列线图对HCC患者VETC和/或MVI阳性的预测效能较好.
Abstract
Objective To investigate the predictive value of a nomogram based on clinical factors and gadoxetic acid disodium(Gd-EOB-DTPA)-enhanced MRI features for predicting positive of vessels encapsulating tumor clusters(VETC)and/or positive of microvascular invasion(MVI)in patients with hep-atocellular carcinoma(HCC)(VM-HCC).Methods Clinical data of 240 patients with HCC who under-went radical hepatectomy in Affiliated Nantong Hospital 3 of Nantong University from August 2014 to Decem-ber 2021 were retrospectively analyzed.There were 169 males and 71 females,aged(58.4±9.1)years old,ranging from 35 to 82 years old.According to the status of MVI and VETC detected by postoperative pathology:121 cases HCC patients with VETC and/or MVI positive,and 119 cases HCC patients with VETC and MVI negative.Clinical data such as gender,age,protein induced by vitamin K absence or antagonist-Ⅱ(PIVKA-Ⅱ),and alpha-fetoprotein(AFP)level were collected.Multivariate logistic regression analysis was used to determine the independent predictors related to VM-HCC,and a nomogram model was construc-ted.C-index and receiver operating characteristic(ROC)curves were used to evaluate the model's discrimi-nation,and calibration curves were used to evaluate the model's consistency.Results Multivariate logistic regression showed that patients with PIVKA-Ⅱ ≥40 mAU/ml(OR=3.686,95％CI:1.883-7.217,P＜0.001),arterial peritumoral enhancement(OR=2.513,95％CI:1.154-5.476,P=0.020),≥50％arterial phase hypovascular component(OR=3.042,95％CI:1.535-6.028,P=0.001),peritumoral hypointensity on hepatobiliary phase(OR=2.682,95％CI:1.250-5.754,P=0.011),and lower tumor to liver contrast ratio on hepatobiliary phase(OR=0.024,95％CI:0.004-0.133,P＜0.001)had a high risk of VM-HCC.The above five independent risk factors were integrated to construct a nomogram model,the C-index was 0.828,and area under the curve was 0.83(95％CI:0.77-0.87),and the calibration curve had a good consistency.Conclusion The nomogram model based on indicators of clinical and Gd-EOB-DTPA-enhanced MRI has an effective performance in predicting microvascular metastasis patterns in HCC patients.
关键词
肝肿瘤/钆塞酸二钠/磁共振成像/微血管侵犯/血管包绕肿瘤细胞簇Key words
Liver neoplasms/Gadolinium ethoxybenzyl diethylenetriamine penta acetic acid/Magnetic resonance imaging/Microvascular invasion/Vessels encapsulating tumor clusters引用本文复制引用
基金项目
江苏省研究生科研与实践创新计划(SJCX22-1633)
江苏省研究型医院学会感染影像科研专项科技项目(GY202203)
南通市科技计划(MS22022056)
出版年
2024
NETL
NSTL
万方数据