Microbiome-Based Colon Cancer Patient Stratification and Survival Analysis

Area of Science:

• Microbiome research
• Cancer biology
• Genomics and bioinformatics

Background:

• Colorectal cancer (CRC) is a major global health concern, ranking third in diagnosis and second in mortality.
• CRC's severity is linked to immune evasion, asymptomatic early stages, aging populations, and lifestyle factors like diet and obesity.
• The gut microbiome significantly influences CRC development, offering potential targets for early detection and treatment outcome prediction.

Purpose of the Study:

• To identify distinct clusters of gut microbiomes within colorectal cancer patients.
• To analyze survival outcomes based on specific microbiome features within identified clusters.

Main Methods:

• Applied consensus clustering, Kruskal-Wallis test, random forest, and LASSO for feature selection.
• Utilized K-means clustering, identifying three distinct patient clusters with varied microbiome diversity.
• Performed survival analysis using Kaplan-Meier curves and Cox regression on selected microbiome features.

Main Results:

• Selected 140 significant microbial features from an initial 1311 candidates.
• Identified eight specific microbial species (N4likevirus, Ambidensovirus, Synechococcus, Thermithiobacillus, Hydrocarboniphaga, Rhodovibrio, Gloeobacter, Candidatus Nitrosotenuis) associated with CRC clustering and survival.
• Demonstrated clear differences in alpha and beta diversity and clinical variables among the three identified clusters.

Conclusions:

• The study highlights significant heterogeneity within the colorectal cancer microbiome.
• Specific microbial species identified influence CRC disease prognosis.
• Further research is needed to elucidate the biological mechanisms of these microbes and the applicability of this approach to other cancers.