磁共振成像2025，Vol.16Issue(4)：60-69,10.DOI:10.12015/issn.1674-8034.2025.04.010
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
摘要
Abstract
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 included in our analysis.Conclusions:MRI radiomics shows strong potential for non-invasive KRAS status prediction in rectal cancer,though study heterogeneity exists.Future research should focus on improving research quality and validating models with multicenter datasets to boost accuracy and reliability.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Key words
colorectal cancer/Kirsten rat sarcoma viral oncogene homolog/magnetic resonance imaging/radiomics/Meta-analysis分类
临床医学引用本文复制引用
马小梅,和建伟,贾应梅,王莉莉..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[J].磁共振成像,2025,16(4):60-69,10.基金项目
Natural Science and Technology Foundation of Gansu Province(No.23JRRA1764) （No.23JRRA1764）
Research Fund Project of Gansu Provincial People's Hospital(No.23GSSYF-4). 甘肃省科技计划项目(编号:23JRRA1764) （No.23GSSYF-4）
甘肃省人民医院院内科研基金项目(编号:23GSSYF-4) （编号:23GSSYF-4）
