A bioinformatics pipeline for identifying clinically relevant immune and adhesion biomarkers in thyroid risk...

The substantial biological heterogeneity of thyroid cancer, particularly within follicular-patterned lesions, underscores the need for improved molecular tools for risk stratification. Genetic variability in the pathways regulating cell adhesion, immune interactions, and extracellular matrix remodeling may influence tumor behavior and the complexity of diagnosis. In this study, we conducted integrated in silico and genetic analyses to evaluate the role of polymorphisms in genes encoding immunoglobulin superfamily adhesion molecules, integrins, junctional proteins, matrix metalloproteinases, and extracellular matrix-associated proteins. Using multiple bioinformatics platforms, we screened 407,812 polymorphisms across 22 candidate genes and prioritized 133 variants with high predicted functional impact. Eight selected single-nucleotide polymorphisms were genotyped in a cohort of 648 individuals, including patients with benign thyroid nodules (n = 152), malignant thyroid nodules (n = 171), and healthy controls (n = 325). Clinical validation revealed that MADCAM1 rs3745925 significantly distinguished follicular adenoma from controls (p = 0.017, OR: 3.15; 95% CI: 1.63-6.05), goiter (p = 0.019, OR: 3.18; 95% CI: 1.49-6.85), and papillary thyroid carcinoma (p = 0.009, OR: 3.49; 95% CI: 1.73-7.09). This association remained robust after Bonferroni correction, underscoring its potential as a priority candidate. Additionally, ITGAM rs1143683, ITGAL rs2230433, and ICAM1 rs5498 were associated with tumor multifocality (p = 0.0428, p = 0.0008, and p < 0.0001, respectively). These findings demonstrate the feasibility of integrating bioinformatics-driven variant prioritization with clinical validation methods. Among the evaluated polymorphisms, MADCAM1 rs3745925 emerged as a promising auxiliary biomarker warranting further evaluation for the characterization of follicular-patterned thyroid lesions.