To investigate values of serum midkine (MK) as a diagnostic biomarker in DTC before surgery, and as a prognostic biomarker before ¹³¹I ablation therapy.

A total of 162 patients (70 patients with DTC and 92 patients with benign thyroid nodules) participated in the surgical cohort, 75 healthy subjects served as controls. Diagnostic values of pre-surgical MK and Tg for DTC were conducted by ROC curves. A total of 214 DTC patients participated in the ¹³¹I treatment cohort. Prognostic values of pre-¹³¹I-ablative MK and Tg to predict ¹³¹I-avid metastases were performed by ROC curves. Independent two-sample t test or Mann-Whitney u test was used to analyze the data. The relationship between MK and Tg was analyzed by Pearson correlation analysis. Metastasis-free survival was analyzed by Kaplan-Meier method.

MK and Tg were positively correlated (r=0.917, P<0.05). Pre-surgical MK and Tg levels were significantly higher in DTC patients than those in benign thyroid nodule patients (z=-7.283 and-3.191, both P<0.05) and those in controls (z=-7.328 and-4.384, both P<0.05). The best cut-off value of MK for differentiating DTC from benign thyroid nodules was 323.12 ng/L and the diagnostic accuracy was 75.31% (122/162), which was better than the diagnostic accuracy of Tg (60.49%, 98/162). Pre-¹³¹I-ablative Tg demonstrated perfect ability to predict metastases, with cut-off value of 19.50 μg/L and diagnostic accuracy of 96.73%(207/214). MK also performed well with cut-off value of 504.71 ng/L and diagnostic accuracy of 89.25%(191/214). DTC patients with MK or Tg levels higher than thresholds (500 ng/L, 20 μg/L) showed a significantly worse ¹³¹I-avid metastasis-free survival by Kaplan-Meier analysis (χ²=118.539 and 209.823, both P<0.05).

MK can not only be used to screen DTC patients, but also be used to predict metastases before ¹³¹I ablative therapy. It is suitable to serve as a serum biomarker for DTC.

Thyroid neoplasms; Cytokines; Thyroglobulin; Diagnosis, differential