Recent advancements in artificial intelligence-powered cancer prediction from oral microbiome.

Abstract

Oral cancer is a major global health burden, ranking sixth in prevalence, with oral squamous cell carcinoma (OSCC) being the most common type. Importantly, OSCC is often diagnosed at late stages, underscoring the need for innovative methods for early detection. The oral microbiome, an active microbial community within the oral cavity, holds promise as a biomarker for the prediction and progression of cancer. Emerging computational techniques in the artificial intelligence (AI) field have enabled the analysis of complex microbiome data sets to unravel the association between oral microbiome composition and oral cancer. This review provides a comprehensive overview of learning-based algorithms applied to oral microbiome data for cancer prediction. In particular, this work discusses how typical machine learning (ML) algorithms, such as logistic regression, random forests, and artificial neural networks, identify the unique microbial patterns associated with oral cancer and other malignancies. A search was conducted in Pubmed covering a 10-year period. The goal was to identify previous studies focused on the role of the oral microbiome in oral cancer prediction using AI-powered tools. The search strategy identified 3382 records in total, of which 44 studies met the inclusion criteria. While AI has shown a transformative power in understanding and revealing the oral microbiome's role in cancer studies, its application in clinical settings requires further efforts on standardization of protocols, curation of diverse cohorts, and validation through large-scale multi-centric and longitudinal studies. The integration of AI with oral microbiome analysis holds significant promise for improving early detection, risk stratification, and personalized treatment strategies for OSCC. By identifying unique microbial patterns associated with cancer, AI-driven models offer a noninvasive, cost-effective tool to predict disease progression and guide clinical decision-making. However, translating these advancements into routine clinical practice requires standardized protocols, diverse patient cohorts, and validation through large-scale, longitudinal studies. Once implemented, this approach could transform oral cancer management, enabling timely interventions and improving patient outcomes.