摘要
背景:同一分期的结肠腺癌患者预后不尽相同,因此,需要寻找可靠的分子标记物来指导个体化治疗决策。本研究旨在鉴定出可用于结肠腺癌预后预测的基因特征。方法:分别从GEO数据库和TCGA数据库中获取332例和431例结肠腺癌患者的基因表达数据集。在训练集(93例)中分析基因特征与无复发生存之间的关系,并分别在内部验证集(94例)和两个外部验证集(145例和431例)中进行验证。结果:通过单变量和LASSO Cox回归分析,共鉴定出11个基因(NDRG1、FLT1、LBP、FABP4、ADIPOQ、AGT、ACVRL1、CCL11、CDC42、TRAV9_2和POMC)。根据风险评分模型,以中位数为临界值,将患者分为高危组和低危组。训练集1、3、5年复发的曲线下面积分别为0.970、0.849和0.859;训练集高危组患者无复发生存率明显低于低危组。此模型的预测准确性在各个验证集中均得到了验证。与其他四种模型相比,本模型对1、3、5年复发显示了更优的预测能力。结论:此基因模型对结肠腺癌复发有良好的预测作用,其预后分级的准确性有待进一步研究证实。
Background Prognosis varies among patients within the same colon adenocarcinoma(COAD)stage,indicating the need for reliable molecular markers to enable individualized treatment.This study aimed to investigate gene signatures that can be used for better prognostic prediction of COAD.Methods Gene-expression profiles of COAD patients were obtained from the Gene Expression Omnibus database(n=332)and The Cancer Genome Atlas database(n=431).The relationship between gene signature and relapse-free survival was analysed in the training set(n=93)and validated in the internal validation set(n=94)and external validation sets(n=145 and 431).Results Overall,11 genes(N-myc downstream regulated gene 1[NDRG1],fms-like tyrosine kinase 1[FLT1],lipopolysaccharide binding protein[LBP],fatty acid binding protein 4[FABP4],adiponectin gene[ADIPOQ],angiotensinogen gene[AGT],activin A receptor,type II-like kinase 1[ACVRL1],CC chemokine ligand 11[CCL11],cell division cycle 42[CDC42],T-cell receptor alpha variable 9_2[TRAV9_2],and proopiomelanocortin[POMC])were identified by univariable and least absolute shrinkage and selection operator(LASSO)Cox regression analyses.Based on the risk-score model,the patients were grouped into the high-risk or low-risk groups using the median risk score as the cut-off.The area under the curve(AUC)values for 1-,3-,and 5-year recurrence were 0.970,0.849,and 0.859,respectively.Patients in the high-risk group had significantly poorer relapsefree survival than did those in the low-risk group.The predictive accuracy of the 11-gene signature was proven in the validation sets.Our gene signature showed better predictive performance for 1-,3-,and 5-year recurrence than did the other four models.Conclusions The 11-gene signature showed good performance in predicting recurrence in COAD.The accuracy of the signature for prognostic classification requires further confirmation.
基金
supported by National Key Clinical Discipline,the Fundamental Research Funds for the young teacher training program of Sun Yat-sen University[grant number 18ykpy02]
the“5010 Clinical Research Program”of Sun Yat-sen University[grant number 2010012]
the Natural Science Foundation of Guangdong Province,China[grant number 2020A1515010428]
the Medical Science Research Grant from the Health Department of Guangdong Province[grant number A2018007].
