Gut microbiota signatures and predictive model of KPS in advanced colorectal cancer patients

The gut microbiota is associated with host health and disease, but its relationship with functional performance status (KPS) in colorectal cancer (CRC) patients remains unexplored. We performed 16 S rRNA gene sequencing on fecal samples from 73 advanced CRC patients (47 high-KPS, 26 low-KPS) and used subsequent microbiota analysis alongside random forest modeling to identify KPS-specific signatures. β-Diversity analysis revealed distinct microbial communities between groups. The high-KPS group (Group A) was enriched in beneficial taxa such as Bifidobacterium and Prevotella_2, while the low-KPS group (Group B) showed an increase in Enterococcus. Metabolic pathway inference indicated enrichment of pathways linked to tumor progression (e.g., cytochrome P450 metabolism) in the low-KPS group. A random forest model constructed with 10 differential genera achieved high predictive accuracy (AUC = 0.992). This study describes for the first time an association between gut microbiota composition and performance status in CRC patients, identifying specific microbial patterns associated with KPS in advanced disease. These findings provide a rationale for future research into microbiota-targeted interventions and a basis for using microbial biomarkers to assess patient status and guide therapeutic strategies.