e16016 Background: Esophageal squamous cell carcinoma (ESCC) is a common malignancy without effective therapy. Immune checkpoint–oriented immunotherapies have shown considerable promise and the advent of esophageal microbiome provides researchers with new ideas. Methods: DNA was extracted from blood, oral mucosal, saliva, urine, fecal samples from 20 ESCC patients before and after immunotherapy. Total microbial genomic DNA samples were extracted using an OMEGA Soil DNA Kit (D5625-01). The V3–V4 regions of bacterial 16S rRNA genes were amplified by PCR using the forward primer and the reverse primer and were sequenced with Illumina MiSeq platform. In order to comprehensively evaluate the α diversity of microbial communities, we used Chao1 and Observed Species indices to characterize the richness, Shannon and Simpson indices to characterize the diversity. PCoA were used to analyze differences in β diversity. Functions of 16S rRNA sequences were predicted using the PICRUSt2 and KEGG databases. Results: A comparison of blood, oral mucosal, saliva, urine, fecal samples of ESCC patients before and after immunotherapy showed that α diversity was not statistically significant. In terms of β diversity, no statistically significant differences were detected within blood, oral mucosal, saliva, urine, fecal samples of ESCC patients before and after immunotherapy. In ESCC patients treated before immunotherapy, the α diversity and β diversity of blood, oral mucosal, saliva, urine, fecal samples were different, and in ESCC patients treated after immunotherapy had the same rule. At the phylum level, the top 5 microbes in ESCC patients before and after immunotherapy were Proteobacteria, Firmicutes, Bacteroidetes, Actinobacteria, Fusobacteria. At the genus level, the top 5 microbes in ESCC patients before immunotherapy were Aquabacterium, Streptococcus, Prevotella, Veillonella, Bacteroides, and in ESCC patients after immunotherapy were Aquabacterium, Streptococcus, Prevotella, Faecalibacterium, Veillonella. In terms of the microbial functions in ESCC patients before and after immunotherapy, the metabolic pathways accounted for the most. Conclusions: This study is conducive to exploring new mechanisms for tumor cells to evade host immune surveillance, providing new ideas and new strategies for the microecology-based immunotherapy of ESCC.