e15583 Background: The peritoneal metastasis (PM) in patients with colorectal cancer (CRC) is associated with poor survival, especially in those with Ras mutations. However, knowledge on the mechanism of molecular biology in CRC peritoneal metastasis (CRCPM) is limited, and the impact of tumor immune microenvironment (TME) on PM pathogenesis and the prognosis of CRCPM remain unclear. Therefore, we characterized the TME of CRCPM and evaluated its potential diagnostic and prognostic values. Methods: This study involved 21 patients with metastatic CRC (mCRC), of whom 11 with CRCPM were classified into the experimental group and 10 with mCRC without PM (noCRCPM) were classified into the control group. Formalin-fixed and paraffin-embedded tissue of primary tumors from all patients was examined using the NanoString RNA sequencing system. The clinicopathological variables and TME biomarkers were compared using the chi-square test and Cox regression analysis. According to the results of multivariate analysis, a prognostic nomogram was generated, and its prediction ability was measured using the concordance index (C-index). Survival curves were generated using the Kaplan–Meier method, and survival comparison between groups was conducted using the log-rank test. Results: The TME in the CRCPM and noCRCPM groups was different, especially the expression of NOS2, TNSF9, KIR3DL2, MAGEA12, MRC1, KIR3DL1, and CD244 and the content of DC cells, macrophages, exhausted CD8+T cells, NK CD56dim cells, and M2 macrophages. Univariate analysis revealed that the expression of PTPN11, TIE1, MAGEA4, PDGFB, and PMS2 was significantly correlated with progression-free survival (PFS) time. These genes were subjected to the least absolute shrinkage and selection operator regression analysis and established a risk score model. Multiple Cox regression analysis showed that the risk score was a significant independent prognostic factor. However, considering the particularity of serum carcinoembryonic antigen (CEA), we built a combined model that included the CEA level and risk score. It could perform well in predicting the PFS time in patients with CRCPM (C-index 0.77). Conclusions: The TME of the primary lesion was significantly different between the CRCPM and noCRCPM groups. The model that combined the risk score and CEA level was a better outcome predictor among patients with CRCPM.