A systematic review of artificial intelligence and machine learning for gut microbiome-based CRC screening

Area of Science:

• Microbiome Research
• Artificial Intelligence in Oncology
• Bioinformatics

Background:

• Gut microbiome dysbiosis is strongly linked to colorectal cancer (CRC) development.
• Emerging research utilizes artificial intelligence (AI) and machine learning (ML) with microbiome data for CRC detection.

Purpose of the Study:

• To systematically evaluate the diagnostic performance of AI/ML models using gut microbiome data for CRC detection.
• To compare different AI/ML-based microbiome screening approaches.
• To identify microbial genera consistently associated with CRC and assess study quality.

Main Methods:

• A systematic review of studies from January 2023 to November 2025 was conducted across major databases.
• Included studies applied AI/ML models to human gut microbiome data for CRC screening, reporting diagnostic metrics.
• Risk of bias was assessed using QUADAS-2, and certainty was evaluated with the GRADE approach; data were narratively synthesized due to heterogeneity.

Main Results:

• Twelve studies met inclusion criteria, showing moderate diagnostic performance for AI/ML models.
• Internal validation sets had Area Under the Curve (AUC) values from 0.61-0.98, and external validation sets ranged from 0.70-0.87.
• Specific microbial genera, including *Porphyromonas*, *Fusobacterium*, and *Peptostreptococcus*, were significantly enriched in CRC patients.

Conclusions:

• AI/ML gut microbiome models offer moderate AUC for CRC detection, potentially enhancing noninvasive screening.
• Prospective validation is needed, and limitations include heterogeneity in models, methodologies, and validation strategies.