Salivary transcriptomic signatures reveal early molecular reprogramming in oral potentially malignant disorders: A non-invasive precision oncology approach in indian patients

Background: Oral potentially malignant disorders (OPMDs) such as oral lichen planus, leukoplakia, and erythroplakia exhibit variable malignant transformation risk, often eluding early clinical recognition. Saliva offers a non-invasive, molecularly rich biofluid for detecting transcriptomic alterations reflective of oncogenic field changes. Methods: In this cross-sectional study, unstimulated saliva samples from 96 participants (48 OPMD cases, 48 age- and sex-matched controls) were subjected to RNA sequencing using the Illumina NovaSeq 6000 platform at CSIR-IGIB, India. Quality control (FastQC), alignment (STAR), and differential expression analyses (DESeq2 and edgeR) were performed. Functional enrichment employed GSEA and Reactome pathways. Diagnostic modeling used LASSO-regularized logistic regression and ROC analysis. Results: Transcriptomic profiling revealed 1,126 significantly dysregulated genes (|log₂FC| > 1.5, FDR < 0.05), including upregulation of MMP9, IL6, CXCL8, TP63, and downregulation of KRT13, LOR, SPRR2A. Pathway enrichment implicated PI3K/AKT, p53, and epithelial-mesenchymal transition signaling. The 15-gene salivary signature yielded an AUC = 0.94 (95% CI 0.89-0.97) for discriminating OPMD from controls, with sensitivity = 91% and specificity = 88%. Sub-analysis identified TP63-CXCL8 co-activation as a predictor of transformation potential (p = 3.2 × 10⁻⁵, adjusted p < 0.01). Conclusion: High-throughput salivary transcriptomics demonstrates robust discriminatory power for identifying molecular dysregulation preceding oral carcinogenesis. This 15-gene signature may serve as a non-invasive biomarker panel for early risk stratification of OPMD, with potential application in population-level screening programs in resource-limited settings.