摘要
目的基于Meta分析评估影像组学模型在肝细胞癌术前微血管侵犯(microvascular invasion,MVI)预测中的效能。材料与方法从PubMed、EMBASE、Cochrane Library、知网和万方数据库中全面检索关于影像组学模型术前预测肝癌MVI的研究,检索时间截至2022年8月20日。根据纳入与排除标准筛选文献并对纳入文献进行数据提取和质量评估。采用Stata 16软件进行Meta分析,绘制森林图和总受试者工作特征(summary receiver operating characteristic,SROC)曲线。此外,我们采用单变量分析、亚组分析和敏感性分析的方法以寻找异质性来源,并绘制漏斗图以评估是否存在发表偏倚。结果32项研究符合我们的标准并被纳入。共3059例患者,其中存在MVI者1339例,不存在MVI者1720例。汇总显示敏感度、特异度、曲线下面积(area under the curve,AUC)值分别为81%[95%置信区间(confidence interval,CI):78%~84%],82%(95%CI:79%~85%)和0.89(95%CI:0.85~0.91)。结论影像组学模型在预测肝细胞癌MVI方面显示出良好的预测性能;其可行性验证和临床转化仍然需要更多高质量研究。
Objective:To evaluate the predictive power of radiomics models for preoperative evaluation of microvascular invasion(MVI)in hepatocellular carcinoma(HCC)by Meta analysis.Materials and Methods:Comprehensive search of studies up to August 20,2022,on preoperative prediction of liver cancer MVI by radiomics models were screened in PubMed,EMBASE,Cochrane Library,CNKI,and Wanfang Database.Data extraction and quality assessment of the retrieved studies were performed according to inclusion and exclusion criteria.We used the STATA version 16 to analyze the raw data,drew forest plots and summary receiver operating characteristic(SROC)curves,and conducted the subgroup analysis and sensitivity analysis to find the heterogeneity.Moreover,Deek’s funnel plots were performed to assess publication bias.Results:Thirty-two studies met our criteria and were included.There were 3059 patients,including 1339 with MVI and 1720 without MVI.For the predictive performance of radiomics models,the pooled sensitivity,specificity,and area under the curve(AUC)values were 81%(95%CI:78%-84%),82%(95%CI:79%-85%),and 0.89(95%CI:0.85-0.91),respectively.Conclusions:Radiomics models showed a promising prediction performance for predicting MVI in HCC.However,more high-quality studies are still needed for further feasibility validation and clinical translation.
作者
张瞳
吴慧
胡鹤
高凯华
杨姣
ZHANG Tong;WU Hui;HU He;GAO Kaihua;YANG Jiao(Department of Radiology,Affiliated Hospital of Inner Mongolia Medical University,Hohhot 010050,China)
出处
《磁共振成像》
CAS
CSCD
北大核心
2023年第1期82-88,共7页
Chinese Journal of Magnetic Resonance Imaging
基金
内蒙古自然科学基金(编号:2021MS08026)
内蒙古医科大学面上项目(编号:YKD2021MS045)。
关键词
肝细胞癌
微血管侵犯
影像组学
META分析
磁共振成像
hepatocellular carcinoma
microvascular invasion
radiomics
Meta-analysis
magnetic resonance imaging
