Intratumoral and fecal microbiota to reveal microbial markers associated with gastric carcinogenesis.

e16014 Background: The relationship between dysbiosis of the gastrointestinal microbiota and gastric cancer (GC) has been extensively studied. However, microbiota alterations in GC patients vary widely across studies, and there are still no diagnostic biomarkers for early GC that can be replicated in multiple populations. Therefore, this study aimed to characterize the gastrointestinal microbial communities associated with gastric carcinogenesis. The diversity results showed significant differences in GC patients' intratumoral microbiota compared with the non-GC population. Methods: 1,642 gastric tissue samples and 394 fecal samples were tested using 16s rRNA amplicon sequencing. After identifying taxa with differential biomarkers, we built a random forest model to distinguish gastric cancer from normal controls. We validated the generalizability and specificity of the model using an external cohort. In addition, we used the KEGG database for predictive analysis of colony correlation functions. Results: The positive correlation between GC-enriched bacteria increased, and the positive correlation between GC-depleted bacteria decreased compared to healthy individuals. Functional analysis revealed that carbohydrate and iron metabolism pathways significantly enriched in the GC phase. In addition, we explored non- Helicobacter pylori ( HP) commensals, which play an essential role in HP-negative and HP-positive GC. Finally, random forest models were constructed using the specific taxa associated with GC obtained from the Lefse analysis, and it was found that the combination of Lactobacillus and Streptococcus included alone could well discriminate between GC patients and healthy individuals in fecal samples (area under the curve (AUC) = 0.7949). We also validated it in an independent cohort (AUC = 0.7712). Conclusions: This study demonstrates the validity of GC-specific microbial markers across multiple populations, which may contribute to early non-invasive diagnosis of GC.