Molecular Signature of Peri-Implantitis: An AI-Driven Transcriptomic Analysis Reveals a Foreign-Body Inflammatory Phenotype; A Pilot Study

Peri-implantitis (PI) is a highly destructive inflammatory disease characterized by progressive bone loss surrounding dental implants. Although it shares clinical features with chronic periodontitis (CP), its aggressive and often refractory nature suggests a distinct pathophysiology. This study employs an AI-driven transcriptomic approach using publicly available human gingival gene expression data (GEO accession: GSE106090¹) to delineate the molecular mechanisms that distinguish PI from CP. Differential expression analysis identified approximately 240 significantly deregulated genes (|log₂FC| > 1, p < 0.05) between PI and CP. The PI transcriptome was dominated by genes associated with inflammation and tissue destruction — notably IL1B, TNF, MMP9, CXCL8, and SPP1 — while structural and reparative genes such as COL1A1 and BMP2 were markedly downregulated. Pathway enrichment analysis revealed strong activation of Cytokine–cytokine receptor interaction, NF κB signaling, and Osteoclast differentiation pathways, indicating a hyper-inflammatory, bone resorptive phenotype. Unsupervised Uniform Manifold Approximation and Projection (UMAP) analysis confirmed two distinct molecular clusters separating PI from CP, validating their transcriptomic divergence. Furthermore, an L1-regularized logistic regression (LASSO) model identified a compact seven-gene biomarker panel that classified PI and CP with high accuracy (AUC ≈ 0.90). Collectively, these findings provide the first AI-based molecular evidence that peri-implantitis represents a foreign-body–driven inflammatory disorder distinct from classic periodontitis. This work establishes a reproducible molecular map that may serve as a foundation for precision diagnostics and targeted therapeutics in implant dentistry.