Associations between postoperative serum omics-derived proteins and aggressive clinicopathological features of papillary thyroid cancer

Papillary thyroid cancer (PTC) exhibits marked biological heterogeneity, ranging from indolent tumors to more advanced forms characterized by angioinvasion, multifocality, and metastatic spread. Despite extensive molecular characterization at the tissue level, the clinical relevance of circulating omics-derived proteins in reflecting disease aggressiveness remains insufficiently explored. This study evaluated postoperative serum levels of selected thyroid cancer tissue omics-derived proteins, such as Malectin (MLEC), Thioredoxin-Interacting Protein (TXNIP), Nephronectin (NPNT), Fibronectin-1 (FN1), and Alpha-1-Acid Glycoprotein (ORM1), in relation to clinicopathological features of aggressive PTC. Serum concentrations of the analyzed proteins were measured using ELISA. The study group comprised 79 patients with histopathologically confirmed angioinvasive, multifocal, or metastatic PTC, while the reference group included 25 patients with very low-risk, non-invasive PTC. Subgroup analyses included angioinvasive PTC, multifocal PTC, and an advanced PTC category comprising patients with coexisting aggressive features, including angioinvasion, multifocality, and metastatic disease. Group comparisons were performed using the Mann-Whitney U test with false discovery rate (FDR) correction and effect size estimation. Exploratory univariable logistic regression and ROC curve analyses were subsequently performed to quantify the direction and strength of the associations and to assess discriminatory performance. Among the evaluated serum markers, only postoperative NPNT concentrations were significantly lower in patients with angioinvasive, multifocal, and advanced PTC compared with the reference group (angioinvasive: median 2.21 ng/mL vs. 2.78 ng/mL, p < 0.01, FDR = 0.01; multifocal: median 2.13 ng/mL, p = 0.01, FDR = 0.04; advanced: median 2.01 ng/mL, p = 0.01, FDR = 0.02). In contrast, serum levels of MLEC, TXNIP, FN1, and ORM1 did not differ significantly between aggressive and non-invasive PTC subgroups after correction for multiple testing. Exploratory univariable logistic regression analyses confirmed an inverse association between NPNT levels and aggressive clinicopathological features (OR range: 0.33-0.42), while ROC analyses indicated moderate discrimination (AUC approximately 0.72-0.73). An additional exploratory LASSO model incorporating NPNT and lipid parameters yielded an apparent AUC of 0.81; however, this result requires cautious interpretation because the model was not externally validated. Reduced postoperative serum NPNT levels are associated with histopathological features of more advanced PTC. Although these findings do not establish NPNT as a clinically applicable or validated biomarker, they suggest an association between postoperative NPNT concentrations and aggressive clinicopathological PTC phenotypes. The proposed links with extracellular matrix-related pathways and lipid metabolism remain hypothetical. Further prospective and preoperative studies are warranted to clarify the biological and clinical significance of NPNT in thyroid cancer progression.