基于MRI影像组学预测结直肠癌KRAS基因突变的系统回顾和Meta分析  

Assessing the efficacy of MRI radiomics for KRAS mutation prediction in colorectal cancer:insights from a systematic review and Meta-analysis

作　　者：马小梅[1] 和建伟[1] 贾应梅 王莉莉[1] MA Xiaomei;HE Jianwei;JIA Yingmei;WANG Lili(Department of Radiology,Gansu Provincial Hospital,Lanzhou 730013,China)

机构地区：[1]甘肃省人民医院放射科,兰州730013

出　　处：《磁共振成像》2025年第4期60-69,共10页Chinese Journal of Magnetic Resonance Imaging

基　　金：甘肃省科技计划项目(编号:23JRRA1764);甘肃省人民医院院内科研基金项目(编号:23GSSYF-4)。

摘　　要：目的 本研究旨在评价利用MRI影像组学在无创地预测结直肠癌患者Kirsten大鼠肉瘤病毒致癌基因同源物(kirsten rat sarcoma viral oncogene homolog,KRAS)突变方面的研究质量,并全面评估相关预测模型的诊断精确度。材料与方法 使用PubMed、 Embase、 The Cochrane Library、 Web of Science、 Scopus、中国知网和万方数据等数据库进行文献检索,收集自2015年1月至2024年10月所有符合纳入和排除标准的涉及利用MRI预测结直肠癌KRAS基因突变的文献。对纳入研究使用修订版诊断准确性研究质量评估(quality assessment of diagnostic accuracy studies-2,QUADAS-2)和放射组学质量评分(radiomics quality score,RQS)进行方法学质量评估;并对纳入研究的异质性进行检验,使用Stata 18软件计算合并敏感度、特异度和诊断优势比(diagnostic odds ratio,DOR),同时进行集成受试者工作特征(summary receiver operating characteristics,SROC)曲线分析。结果 共纳入17项研究,涉及2684个病例,基于MRI影像组学术前预测直肠癌KRAS基因状态的合并敏感度、合并特异度和曲线下面积(area under the curve,AUC)分别为79%[95%置信区间(confidence interval,CI):75%~83%]、74%(95%CI:68%~80%)和0.85 (95%CI:0.81~0.88);合并敏感度、合并特异度均存在中异质性,异质性统计量I2值分别为:56.80%(95%CI:34.08%~79.53%)、77.35%(95%CI:67.22%~87.48%);Q值分别为39.35(P<0.001)、75.05(P<0.001);亚组分析和单变量Meta分析结果表明所有变量都对异质性有一定影响(P<0.05);Deek's漏斗图基本对称,斜率系数差异无统计学意义(P=0.11),表明纳入的研究没有统计学上显著性的发表偏倚。结论 MRI影像组学可以无创地预测直肠癌KRAS基因状态,具有较高的预测效能,然而,影像组学研究是异质的,建议未来的研究者更加重视研究方法的质量,并利用多中心数据集对预测模型进行外部验证,以增强模型的精确度和可信度。Objective:To assess the research quality of utilizing MRI radiomics for non-invasive prediction of Kirsten rat sarcoma viral oncogene homolog(KRAS) mutations in colorectal cancer and evaluate the diagnostic accuracy of associated prediction models.Materials and Methods:A comprehensive literature search was conducted utilizing databases including PubMed,Embase,The Cochrane Library,Web of Science,Scopus,CNKI,and WanFang Data.This search aimed to identify all relevant studies that satisfied the established inclusion and exclusion criteria regarding the use of MRI for predicting KRAS gene mutations in colorectal cancer,covering the period from January 2015 to October 2024.The methodological quality of the selected studies was evaluated using the quality assessment of diagnostic accuracy studies 2(QUADAS-2) and radiomics quality score(RQS) tools.Furthermore,heterogeneity among the included studies was assessed,and pooled weighted sensitivity,specificity,and diagnostic odds ratio(DOR) were calculated using Stata 18 software,along with a summary receiver operating characteristic(SROC) analysis.Results:A total of 17 studies,encompassing 2684 cases,were included in the analysis.The sensitivity,specificity area under the curve(AUC) values of preoperative prediction of KRAS gene status in rectal cancer utilizing MRI radiomics were 79% [95% confidence interval(CI):75% to 83%],74%(95% CI:68%to 80%),and 0.85(95% CI:0.81 to 0.88),respectively.Both sensitivity and specificity combined results showed moderate heterogeneity,with I2 heterogeneity statistics values of 56.80%(95% CI:34.08% to 79.53%) and 77.35%(95% CI:67.22% to 87.48%),respectively;Q values were 39.35(P < 0.001) and 75.05(P < 0.001),respectively.The results of subgroup analysis and univariate Meta-analysis indicated that all variables had a certain impact on heterogeneity(P < 0.05).Deek's funnel plot was basically symmetrical,and the slope coefficient was not statistically significant(P = 0.11),suggesting that there was no significant publication bias in the studies

关 键 词：结直肠癌 Kirsten大鼠肉瘤病毒致癌基因同源物 磁共振成像 影像组学 META分析 

分 类 号：R445.2[医药卫生—影像医学与核医学] R735.3[医药卫生—诊断学]