Construction of a novel microbial abundance prognostic risk (MAPR) model for predicting prognosis in CRC - based on bioinformatics

Abstract

Abstract Background Previous studies have demonstrated the significant role of the microbiome in the prognosis of colorectal cancer (CRC) patients. However, few studies have utilized bioinformatics to analyze the prognostic value of the microbiome in CRC. In this study, we constructed a CRC microbial abundance prognostic risk (MAPR) model and evaluated its prognostic value. Methods The TCGA CRC microbiome data (TCGA-CRC-microbiome) was downloaded from the cBioPortal website. Univariate, LASSO, and multivariate Cox regression analyses were performed to investigate the relationship between CRC microbial abundance and survival. The MAPR model was constructed based on the above analyses. The predictive ability of the MAPR model was evaluated using Kaplan-Meier (KM) survival curves, receiver operating characteristic (ROC) curves, independent prognostic analysis, and nomogram models. Results Using 11 microbial genera which exhibited adverse overall survival (OS) in CRC patients from overall 1406 microbes in the TCGA-CRC microbiome dataset to construct a MAPR model. This model was constructed and assessed for prognostic value using different survival endpoints. The results indicated that the high-risk group had shorter OS, progression-free interval (PFI), disease-specific survival (DSS), and disease-free interval (DFI). High-risk status served as an independent adverse prognostic factor, with greater prognostic value than other clinical indicators. Compared to the MAPR-unincorporated CRC nomogram, the four nomograms incorporating MAPR significantly improved the predictive ability. Conclusion The successful establishment of CRC's MAPR and its unique prognostic value provide a novel perspective for further investigations into the prognostic mechanisms of CRC patients.