Oral Microbiota Composition Predicts Early Childhood Caries Onset.

Area of Science:

• Oral microbiology
• Pediatric dentistry
• Biomarker discovery

Background:

• Early childhood caries (ECC) is a prevalent chronic disease impacting children's quality of life and imposing significant societal costs.
• Current caries risk assessment (CRA) methods lack accuracy, consistency, and longitudinal validation.
• Molecular and microbial biomarkers offer a promising avenue for precise ECC risk and onset prediction.

Purpose of the Study:

• To investigate the predictive potential of oral microbiota composition for early childhood caries (ECC) onset.
• To identify specific microbial biomarkers indicative of ECC risk.
• To evaluate the efficacy of 16S rRNA gene sequencing and machine learning in ECC risk assessment.

Main Methods:

• 16S ribosomal RNA (rRNA) gene sequencing was performed on saliva samples collected every 6 months for 24 months from initially caries-free children aged 1-3 years.
• Machine learning models were developed to analyze microbiota composition and distinguish between children who developed ECC and those who remained caries-free.
• Nested cross-validation was employed to assess model performance in predicting ECC onset.

Main Results:

• Machine learning models accurately distinguished between caries-affected and non-affected groups at initial visits (AUC=0.71).
• Models demonstrated high discrimination between ECC-converted and healthy children just before diagnosis (AUC=0.89).
• Key discriminatory bacterial features identified include *Rothia mucilaginosa*, *Streptococcus* sp., and *Veillonella parvula*.

Conclusions:

• Oral microbiota profiling using high-throughput 16S rRNA gene sequencing is a predictive tool for ECC onset.
• Identified bacterial species serve as potential biomarkers for ECC risk.
• These findings support the development of novel, microbiota-based strategies for ECC prevention and management.