目的:研究大肠癌患者肠道菌群结构变化及其与病理学分期关系以及基于菌群构建大肠癌的诊断模型。方法:收集155例大肠癌患者(大肠癌组)和442例健康对照者(对照组)粪便标本，进行16S RNA测序。将所有手术患者进行术后病理学分期并采用线性判别分析效应量分析不同分期之间的菌群特征。随机森林选择特征性菌群，利用受试者工作曲线对大肠癌进行诊断模型建模。结果:大肠癌组和对照组肠道菌群的丰度和多样性均无明显差异。大肠癌组厚壁菌门丰度降低，拟杆菌门、梭杆菌门显著高于对照组。其中消化链球菌属、梭杆菌属和芽孢杆菌属细菌在大肠癌组中显著富集，而产丁酸细菌n Faecalibacterium菌、瘤胃球菌属和罗氏菌属在大肠癌组中丰度显著降低。Ⅳ期大肠癌患者肠道菌群丰度显著降低于Ⅰ～Ⅲ期。Ⅰ期大肠癌患者肠杆菌-志贺菌属显著富集，Ⅲ期大肠癌患者拟杆菌科和拟杆菌属的丰度显著增加，Ⅳ期大肠癌患者毛罗菌科富集而芽孢杆菌目减少。基于18个属的细菌模型可以区分大肠癌组和对照组，曲线下面积0.858，联合粪便免疫化学试验后诊断模型的敏感性和特异性均超过98%，曲线下面积0.992。n 结论:大肠癌组肠道菌群结构发生显著改变，不同病理学分期的大肠癌患者肠道菌群也出现特征性改变。其中Ⅳ期大肠癌患者菌群多样性显著降低。基于肠道细菌和粪便免疫化学试验构建的模型可以有效区别大肠癌患者和正常人群。“,”Objective:To study the alteration of gut microbiota structure in colorectal cancer (CRC) and also the relationships between gut microbiota and pathological stages. To establish the diagnostic model for colorectal cancer based on gut microbiota.Methods:Fecal specimens from CRC patients and healthy controls were collected for 16S RNA sequencing. For all patients who received surgeries, tumor pathological stages were identified and gut microbiota of patients in different stages were characterized with LEfSe. Random forests were used for feature selection of gut microbiota and the receiver working curve was applied for the diagnostic model of CRC.Results:There were no significant differences in the richness and diversity of gut microbiota between CRC patients and healthy controls. CRC patients had reduced abundance of Firmicutes, while the relative abundances of Bacteroides and Fusobacteria were significantly higher than that of healthy controls. Among them, n Peptostreptococcus, Fusobacterium and n Bacillus were significantly enriched in CRC patients. In contrast, the butyrate-producing bacteria such as n Faecalibacterium, Ruminococcus, and n Roseburia were with significantly reduced abundance in CRC patients. The compositions of gut microbiota of cancer patients with different pathological stages were significantly different. The microbial diversity of stage Ⅳ CRC was significantly lower than that of stage Ⅰ-Ⅲ. n Escherichia-n Shigella was significantly enriched in patients with stage I CRC, the abundance of n Bacteroides and n Bacteroides were significantly increased in patients with stage Ⅲ CRC, and patients with stage Ⅳ CRC had increased relative abundance of Lachnospiraceae and decreased abundance of n Bacillus. The diagnostic model based on bacteria of 18 genera was effective in distinguishing CRC patients from healthy controls (AUROC=0.858). The model showed both sensibility and specificity of over 98% with the AUROC of 0.992 while combined with fecal immunochemical test.n Conclusions:The structure of gut microbiota in CRC patients changed significantly, and the microbial diversity of stage Ⅳ CRC patients significantly decreased. Diagnostic model based on gut microbiota was effective for distinguishing CRC patients from healthy controls.