A diagnostic model based on DNA methylation haplotype block characteristics for identifying papillary thyroid carcinoma from thyroid adenoma

Papillary thyroid carcinoma (PTC) is the most prevalent form of thyroid cancer. Methylation of some genes plays a crucial role in the tendency to malignancy as well as poor prognosis of thyroid cancer, suggesting that methylation features can serve as complementary markers for molecular diagnosis. In this study, we aimed to develop and validate a diagnostic model for PTC based on DNA methylation markers. A total of 142 thyroid nodule tissue samples containing 84 cases of PTC and 58 cases of thyroid adenoma (TA) were collected for reduced representation bisulfite sequencing (RRBS) and subsequent analysis. The diagnostic model was constructed by the logistic regression (LR) method followed by 5-cross validation and based on 94 tissue methylation haplotype block (MHB) markers. The model achieved an area under the receiver operating characteristic curve (AUROC) of 0.974 (95% CI, 0.964-0.981) on 108 training samples and 0.917 (95% CI, 0.864-0.973) on 27 independent testing samples. The diagnostic model scores showed significantly high in males (P = 0.0016), age <= 45 years (P = 0.026), high body mass index (BMI) (P = 0.040), lymph node metastasis (P = 0.00052) and larger nodules (P = 0.0017) in the PTC group, and the risk score of this diagnostic model showed significantly high in recurrent PTC group (P = 0.0005). These results suggest that the diagnostic model can be expected to be a powerful tool for PTC diagnosis and there are more potential clinical applications of methylation markers to be excavated.