基于增强MRI的预测模型在小肝癌微血管侵犯术前判断中的应用  

Application of enhanced MRI-based prediction model on preoperative diagnosis of microvascular invasion in small hepatocellular carcinoma

作　　者：王浩东 高娟 Wang Haodong;Gao Juan(Department of Diagnostic Radiology,First Affiliated Hospital of Naval Medical University,Shanghai 200433,China;Third Ward Of Special Needs Diagnosis and Treatment,Third Hospital Affiliated to Naval Medical University,Shanghai 200438,China)

机构地区：[1]海军军医大学第一附属医院放射诊断科,上海200433 [2]海军军医大学第三附属医院特需诊疗三病区,上海200438

出　　处：《中华转移性肿瘤杂志》2024年第4期326-332,共7页Chinese Journal of Metastatic Cancer

摘　　要：目的本研究旨在构建增强MRI影像学特征的预测模型,并探讨其对预测小肝癌患者微血管侵犯(MVI)的术前应用价值。方法选取2019—2022年海军军医大学附属医院行手术治疗的134例小肝癌患者资料为研究对象。采用Logistic回归模型多因素分析小肝癌发生MVI的危险因素,采用受试者工作特性(ROC)曲线评价预测模型临床价值。结果全组患者分为MVI(+)组45例和MVI(-)组89例。肿瘤长径、肿瘤包膜不完整、瘤内动脉、动脉期强化、影像学中观察到晕征及毛刺征与小肝癌患者发生MVI有关(P<0.05),且均是小肝癌发生MVI的危险因素(P<0.05)。预测模型为P=1/[1+e^((2.782-0.956×肿瘤长径-1.580×肿瘤包膜不完整-0.743×瘤内动脉-1.429×动脉期强化-1.358×晕征-0.758×毛刺征))],Hosmer-Lemshow检验模型拟合度良好(P=0.607),ROC曲线下面积为0.847,敏感性为97.14%、特异性为81.25%。结论术前MRI增强扫描及弥散加权成像特征可提示小肝癌存在MVI,基于上述特征建立的回归预测模型具有良好的区分度和校准度。ObjectiveThis study aimed to investigate the value of constructing a regression prediction model of enhanced MRI imaging features for predicting microvascular invasion(MVI)in patients with small hepatocellular carcinoma(HCC)for preoperative application.MethodsSelected 134 patients for research objects with small HCC who underwent surgery in Affiliated Hospital of Naval Medical University from July 2019 to March 2022.Multivariate Logistic regression model was used to analyze the risk factors of MVI in small HCC.The subject operating characteristic(ROC)curve was used to evaluate the clinical value of predictive model.ResultsThe whole group of patients was divided into 45 patients in the MVI(+)group and 89 patients in the MVI(-)group.Tumor length diameter,incomplete tumor envelope,intra-tumor artery,arterial phase enhancement,halo sign and burr sign observed in imaging were related to MVI in patients with small HCC(P<0.05),and were all risk factors for MVI in small HCC(P<0.05).Predictive model:P=1/(1+exp(2.782-0.956×tumor length diameter-1.580×incomplete tumor envelope-0.743×intra-tumor artery-1.429×arterial phase enhancement-1.358×halo sign-0.758×burr sign).The predictive model was subjected to the Hosmer-Lemshow test,which was a good fit(P=0.607),the area under the ROC curve was 0.847,and the sensitivity was 97.14%and specificity was 81.25%.ConclusionsPreoperative MRI enhancement scans and diffusion-weighted imaging can suggest the presence of MVI in small HCC.The regression prediction model based on the above features has good discrimination and calibration.

关 键 词：小肝癌 微血管侵犯 增强磁共振 回归预测模型 

分 类 号：R445.2[医药卫生—影像医学与核医学] R735.7[医药卫生—诊断学]