FABP3 methylation as a novel biomarker for the differentiation and classification of benign and malignant thyroid nodules

Background: Differentiation between benign and malignant thyroid nodules has been a challenge in clinical practice. We aim to explore a novel biomarker to determine the malignancy of thyroid nodules. Materials and methods: In the discovery study, 32 tissue samples from benign thyroid nodule (BTN) and thyroid cancer (TC) patients were analyzed by Methylation 850K array and RNA-Sequencing. TC associated FABP3 methylation was further verified by mass spectrometry in two independent studies (221 BTN vs. 222 TC in Validation I and 191 BTN vs. 256 TC in Validation II). Logistic regression analysis and non-parametric tests were used for the analysis between groups. Results: Altered and inversely correlated methylation and expression in FABP3 gene in TC was found in the discovery study (P = 2.90E-05 for the methylation and P = 0.040 for the expression), and verified in the two validation studies (P values range from 0.012 to 6.30E-10-12). FABP3 methylation could sufficiently differentiate TC from BTN (AUC = 0.77), and could be further improved when combined with the BRAFV600E mutations (AUC = 0.87). The association between FABP3 hypomethylation and TC was enhanced in women, in patients with younger age, with larger tumor size and with lower FT3. FABP3 methylation was varied in BTN and TC subtypes, with the highest level in adenoma and the lowest in anaplastic thyroid cancer. Conclusion: Our study suggested that altered FABP3 methylation in tissue samples as a potential biomarker to distinguish malignant and benign thyroid nodules, and might be helpful for the pathological classification of TC.