Gut microbiome identifies risk for colorectal polyps

Area of Science:

• Microbiome research
• Gastroenterology
• Biomarker discovery

Background:

• Colorectal cancer (CRC) screening is crucial for early detection.
• Current screening methods like colonoscopy can be invasive.
• Identifying non-invasive biomarkers for polyp detection is a significant unmet need.

Purpose of the Study:

• To characterize the gut microbiome in individuals with and without polyps.
• To evaluate the potential of the gut microbiome as a non-invasive biomarker for colorectal cancer (CRC) risk.
• To develop machine learning models for polyp prediction using microbiome data.

Main Methods:

• Collected rectal swab, stool, and sigmoid biopsy samples from 231 subjects undergoing colonoscopy.
• Performed 16S rRNA sequencing on 552 samples to identify operational taxonomic units (OTUs).
• Utilized non-parametric statistics and machine learning to build classifiers for polyp prediction based on distinct OTUs.

Main Results:

• Colonoscopy detected polyps in 56% of the 218 subjects analyzed.
• Machine learning models using informative OTUs from home-collected stool samples achieved >75% classification accuracy.
• Combined Naïve Bayes and Neural Network models reduced the false negative rate to 5%.

Conclusions:

• Gut microbiome analysis combined with machine learning offers a promising non-invasive approach for polyp screening.
• This strategy has the potential to optimize colonoscopy use and reduce CRC-related morbidity and mortality.
• The findings suggest a significant reduction in healthcare costs associated with CRC screening and management.